HANNA NURMI RHEUMATOID ARTHRITIS-ASSOCIATED INTERSTITIAL LUNG DISEASE ASSESSMENT OF THE FACTORS ASSOCIATED WITH THE COURSE OF THE DISEASE

Transkriptio

1 PUBLICATIONS OF THE UNIVERSITY OF EASTERN FINLAND Dissertations in Health Sciences HANNA NURMI RHEUMATOID ARTHRITIS-ASSOCIATED INTERSTITIAL LUNG DISEASE ASSESSMENT OF THE FACTORS ASSOCIATED WITH THE COURSE OF THE DISEASE

2

3 Rheumatoid Arthritis-associated Interstitial Lung Disease Assessment of the factors associated with the course of the disease

4

5 HANNA NURMI Rheumatoid Arthritis-associated Interstitial Lung Disease Assessment of the factors associated with the course of the disease To be presented by permission of the Faculty of Health Sciences, University of Eastern Finland for public examination in Kuopio, on Friday, June 8 th, 2018, at 12 noon. Publications of the University of Eastern Finland Dissertations in Health Sciences Number 458 Department of Respiratory Medicine, Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland Kuopio 2018

6 Grano Oy Jyväskylä, 2018 Series Editors: Professor Tomi Laitinen, M.D., Ph.D. Institute of Clinical Medicine, Clinical Physiology and Nuclear Medicine Faculty of Health Sciences Professor Hannele Turunen, Ph.D. Department of Nursing Science Faculty of Health Sciences Professor Kai Kaarniranta, M.D., Ph.D. Institute of Clinical Medicine, Ophthalmology Faculty of Health Sciences Associate Professor (Tenure Track) Tarja Malm, Ph.D. A.I. Virtanen Institute for Molecular Sciences Faculty of Health Sciences Lecturer Veli-Pekka Ranta, Ph.D. (pharmacy) School of Pharmacy Faculty of Health Sciences Distributor: University of Eastern Finland Kuopio Campus Library P.O.Box 1627 FI Kuopio, Finland ISBN (print): ISBN (pdf): ISSN (print): ISSN (pdf): ISSN-L:

7 III Author s address: Center of Medicine and Clinical Research, Division of Respiratory Medicine, Kuopio University Hospital and Department of Respiratory Medicine, Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences University of Eastern Finland KUOPIO FINLAND Supervisors: Professor Riitta Kaarteenaho, M.D., Ph.D. Research Unit of Internal Medicine Medical Research Center Oulu Department of Internal Medicine and Respiratory Medicine University of Oulu and Oulu University Hospital OULU FINLAND Docent Minna Purokivi, M.D., Ph.D. Center of Medicine and Clinical Research Division of Respiratory Medicine Kuopio University Hospital KUOPIO FINLAND Reviewers: Professor Hannu Puolijoki, M.D., Ph.D. University of Tampere Central Hospital of Southern Ostrobothnia SEINÄJOKI FINLAND Docent Paula Rytilä, M.D., Ph.D., Adj. Prof. University of Helsinki Chief Medical Officer, Vice President Global Medical Affairs and Pharmacovigilance, R&D Orion Corporation, Orion Pharma ESPOO FINLAND Opponent: Docent Maija Halme, M.D., Ph.D. Department of Pulmonary Diseases University of Helsinki Helsinki University Central Hospital HELSINKI FINLAND

8 IV

9 V Nurmi, Hanna Rheumatoid Arthritis-Associated Interstitial Lung Disease Assessment of the factors associated with the course of the disease University of Eastern Finland, Faculty of Health Sciences Publications of the University of Eastern Finland. Dissertations in Health Sciences p. ISBN (print): ISBN (pdf): ISSN (print): ISSN (pdf): ISSN-L: ABSTRACT Interstitial lung disease (ILD) is one of the most common lung manifestations in patients with rheumatoid arthritis (RA), occurring in approximately 10% of patients with RA and increasing both their morbidity and mortality. RA-ILD is not considered as one single disease entity; instead it includes several different subtypes, each of which seems to have a distinct disease course. Moreover, the disease course even within the same subtype can be highly variable from patient to patient, which complicates the estimation of prognosis. The categorization into different subtypes is often performed by high-resolution computed tomography (HRCT). Different kinds of scoring systems for ILDs have been developed over the years, trying to help in the evaluation of an individual s prognosis. These scoring models have, however, mainly been developed for idiopathic pulmonary fibrosis (IPF), and their suitability for RA- ILD is largely unknown. Our aim was to evaluate the course of the disease of RA-ILD patients in Kuopio University Hospital (KUH) health care district. The study material consisted of retrospectively gathered data of 60 RA-ILD patients treated between the years in the KUH pulmonology clinic. Clinical, pulmonary function tests and death certificate data were gathered using a specially designed form. The HRCTs of the patients were reevaluated and the radiological re-categorization was conducted according to the current criteria. Firstly, we evaluated comorbidities and causes of death, as well as investigated the course of the disease in different subtypes. Secondly, we tested the applicability of three different prediction models, previously mostly applied in patients with IPF, and searched for other factors that could be useful for evaluating the prognosis of RA-ILD patients. Finally, we compared the presence and extent of various HRCT observations in different subtypes and compared radiological findings with clinical data. Most of the patients (36/60%) showed a radiological pattern of usual interstitial pneumonia (UIP). These patients had higher numbers of hospitalizations for respiratory reasons and deaths as well as greater use of oxygen than patients with other subtypes. RA- ILD was the most common primary cause of death, even though several comorbidities coexisted. We observed that the risk predicting models, such as the gender-age-physiologic variables model (GAP), were applicable for evaluating the risk of death of patients with RA-ILD in a similar manner as in those with IPF. The baseline diffusion capacity to carbon monoxide (DLCO), the composite physiologic index (CPI) and several radiological findings, such as the extents of reticulation and traction bronchiectasis also predicted survival. Moreover, the extents of honeycombing, traction bronchiectasis and architectural distortion correlated with hospitalizations due to respiratory reasons. This thesis clarified the numbers and subtypes of RA-ILD patients in KUH region as well as revealing the variable course of the disease. Our study may help clinicians to identify those patients at the highest risk of death which could lead to more individualized followup and treatment protocols in the future.

10 VI National Library of Medicine Classification: WE 346, WF 600, WN 206 Medical Subject Headings: Lung Diseases, Interstitial; Arthritis, Rheumatoid; Risk Factors; Prognosis; Tomography; Respiratory Function Tests; Death Certificates; Cause of Death; Comorbidity; Hospitalization; Retrospective Studies; Humans; Finland

11 VII Nurmi, Hanna Nivelreumaan liittyvän interstitiaalisen keuhkosairauden luokittelu ja taudinkulku Itä-Suomen yliopisto, terveystieteiden tiedekunta Publications of the University of Eastern Finland. Dissertations in Health Sciences s. ISBN (print): ISBN (pdf): ISSN (print): ISSN (pdf): ISSN-L: TIIVISTELMÄ Interstitiaalinen keuhkosairaus on yksi nivelreuman tärkeimpiä keuhkoilmentymiä, jota esiintyy n. 10 %:lla nivelreumaa sairastavista potilaista. Nivelreumaan liittyvä interstitiaalinen keuhkosairaus (RA-ILD) lisää merkittävästi näiden potilaiden sairastavuutta ja kuolleisuutta. Se ei kuitenkaan ole yksi yhtenäinen sairaus, vaan ryhmä monia eri alatyyppejä, joilla on erilainen taudinkulku ja ennuste ja joista osa johtaa keuhkojen fibrotisoitumiseen. Yksittäisen sairastuneen kohdalla taudin kulun ja oletetun eliniän ennustaminen on erittäin haastavaa. Potilaiden luokittelu eri alatyyppeihin tehdään pääasiallisesti ohutleike-tietokonetomografian (HRTT) perusteella. Aiemmin on kehitetty useita riskinarviointimenetelmiä, joilla on pyritty arvioimaan ILDpotilaita suuremman ja pienemmän kuoleman riskin luokkiin. Nämä menetelmät on kuitenkin valtaosin kehitetty idiopaattista keuhkofibroosia (IPF) sairastaville, eikä niiden soveltuvuudesta RA-ILD:ssä ole aikaisempaa tutkittua tietoa. Tavoitteenamme oli selvittää RA-ILD:n taudinkulkua Pohjois-Savon sairaanhoitopiirin alueelta kerätyssä kohortissa. 60 potilaan aineisto kerättiin retrospektiivisesti vuosina Kuopion yliopistollisen sairaalan keuhkoklinikassa hoidetuista potilaista, joiden kliiniset sekä kuolintodistusten tiedot ja keuhkojen toimintakokeiden tulokset kerättiin yksityiskohtaista tiedonkeruukaavaketta käyttäen ja joiden HRTT-kuvat arvioitiin uudelleen. HRTT-kuvien perusteella potilaat luokiteltiin nykysuositusten mukaisesti eri ILD alatyyppeihin. Ensimmäisessä osatyössä kartoitimme liitännäissairauksia ja kuolinsyitä, sekä vertasimme taudinkulkua eri alatyypeissä. Toisessa osatyössä testasimme miten IPF: iin kehitetyt ennustemallit toimivat RA-ILD potilaiden kohdalla ja selvitimme muita tekijöitä, joita mahdollisesti voitaisiin hyödyntää yksittäisen potilaan kuoleman vaaraa arvioidessa. Kolmannessa tutkimuksessa tarkastelimme radiologisia löydöksiä RA- ILD:n eri alatyypeissä, sekä niiden korrelointia kliinisiin tekijöihin. Aineistostamme valtaosa (36/60%) kuului ns. tavallisen interstitiaalisen pneumonian (UIP) alaryhmään, jossa esiintyi muihin alaryhmiin verrattuna enemmän keuhkoperäisistä syistä johtuvia sairaalahoitojaksoja, happihoidon tarvetta ja kuolemantapauksia. Yleisin peruskuolinsyy oli RA-ILD, vaikka erilaiset liitännäissairaudet olivat yleisiä. Toisessa tutkimuksessa osoitimme, että erilaiset kuolemanriskin arviointimallit, kuten gender-agephysiologic variables - malli (GAP), ovat käyttökelpoisia myös RA-ILD-potilaiden kuolemanriskin arviossa samaan tapaan kuin IPF-potilailla. Lisäksi havaitsimme lähtötason kokonaisdiffuusiokapasiteetin ja ns. composite physiologic index (CPI)- pistemäärän ennustavan kuolleisuutta. Kolmannessa osatyössä todettiin useiden radiologisten löydösten, kuten retikulaation ja traktiobronkiektasioiden laajuuden, olevan yhteydessä lyhentyneeseen elinikään sekä hunajakennojen, traktiobronkiektasioiden ja arkkitehtuurin vääristymän laajuuksien korreloivan keuhkoperäisten osastohoitojaksojen määrään. Tutkimuksen myötä tiedämme, minkä verran ja minkä typpisiä RA-ILD potilaita alueellamme on. Saimme lisätietoa taudinkulun eroista eri alatyypeissä ja toivottavasti

12 VIII pystymme jatkossa tunnistamaan paremmin suuressa riskissä olevat potilaat. Tämä voisi mahdollistaa seurantakäytäntöjen ja hoitojen yksilöllisemmän suunnittelun. Luokitus: WE 346, WF 600, WN 206 Yleinen Suomalainen asiasanasto: keuhkosairaudet; keuhkofibroosi; nivelreuma; riskitekijät; riskinarviointi; tietokonetomografia; ennusteet; kuolintodistukset; kuolemansyyt; kuolleisuus; liitännäistaudit; elinikä; sairaalahoito; happihoito; Pohjois-Savo; Suomi

13 IX Maailma on kaunis ja hyvä elää sille, jolla on aikaa ja tilaa unelmille. Ja mielen vapaus, ja mielen vapaus Vexi Salmi

14 X

15 XI Acknowledgements This study was carried out in the Department of Respiratory Medicine, University of Eastern Finland and in the Center of Medicine and Clinical Research, Division of Respiratory Medicine, Kuopio University Hospital during the years First of all, I would like to express my deepest gratitude to my supervisor Professor Riitta Kaarteenaho. You have provided both guidance and support in the first steps of my research career and advised me carefully throughout the process. You have always found the time to answer my questions and your dedication made possible the completion of this study. Your devotion to science is truly inspirational. I am also deeply grateful to my other supervisor Docent Minna Purokivi. You have encouraged and supported me in many ways. I am sincerely thankful for your empathy and advice when life delivered a number of misfortunes. You often understand my temper. Moreover, you have arranged my leaves of absence; these allowed me to complete this study at full speed. I wish to thank all my co-authors and study group members for their collaboration and support. This study would not have been possible without the excellent radiologists Hannu-Pekka Kettunen and Sanna Suoranta, who performed the enormous job of screening and re-categorizing the HRCTs. Many thanks to Tuomas Selander for his patience when guiding me through the basics of SPSS. The cheerful peer support of Miia Kärkkäinen has helped me carry on and the assistance of research nurse Satu Nenonen saved a lot of my time and energy at the beginning of this project. I sincerely thank Ewen MacDonald for reviewing English language in all the original publications as well as this thesis. I express my gratitude to the official reviewers of this dissertation, Professor Hannu Puolijoki and Docent Paula Rytilä, who gave me professional, constructive and helpful comments about this manuscript. I am also grateful to the anonymous reviewers of the original publications for their comments, which helped me improve the manuscripts. I warmly thank all funders of my research: the Foundation of the Finnish Anti- Tuberculosis Association, the Jalmari and Rauha Ahokas Foundation, the Väinö and Laina Kivi Foundation, the Research Foundation of the Pulmonary Diseases, the Kuopio region Respiratory Foundation, the North Savo Regional Fund of the Finnish Cultural Foundation and a state subsidy to the Kuopio University Hospital. I am grateful for being a part of the Department of Respiratory Medicine personnel. I am surrounded by skillful clinicians, enthusiastic researchers and warm and intelligent people. You have believed in me, supported me and offered me much useful advice. Your company in every-day work as well as in numerous parties has made me laugh and relax, which has been very important during this process. Special thanks to Professor Heikki Koskela, who acted as a mentor when I first started my resident s training in respiratory medicine; his guidance has continued since that time and he is partially responsible for planting the seed of a scientific way of thinking. I also wish to thank my parents-in-law Orvokki and Kari for their support and love, not to mention taking care of our children which has enabled me and Samipetteri to enjoy some time together as well as coping with the long working hours. I thank all my dear friends and relatives, especially Jenni, Juha, Annukka, Anne, Anniina, Laura, Tiina and Jussi, for their friendship, support, listening ears, delicacies, sparkling wine, enjoyable company, sharing all the precious moments in life and giving me other things to think outside my scientific work. A big thank you to the Sawotta girls for their companionship in music, the power of which is astonishing.

16 XII I am extremely grateful to my parents, Eeva and Kalervo, for their love and support during my whole life. You have taught me the importance of both hard work and relaxing in the summer cottage. I was fortunate to grow up in a stable and loving home. Most importantly, I thank my spouse Samipetteri for sharing his life with me in both the good times and the bad ones, for your unconditional love and patience. When facing a hill, you push me forward, when plunging downhill, you slow down my speed. You make me laugh and boost my spirits. Your worlds best cinnamon buns have comforted me on so many occasions. You have been the best father to our precious children, Iita and Paavo, for whom I m more grateful than anything else. You three are the loves of my life. Hanna Nurmi Kuopio, March 2018

17 XIII List of the original publications This dissertation is based on the following original publications: I Nurmi H, Purokivi M, Kärkkäinen M, Kettunen H-P, Selander T, Kaarteenaho R. Variable course of disease of rheumatoid arthritis-associated usual interstitial pneumonia compared to other subtypes. BMC Pulm Med 16: , II Nurmi H, Purokivi M, Kärkkäinen M, Kettunen H-P, Selander T, Kaarteenaho R. Are risk predicting models useful for estimating survival of patients with rheumatoid arthritis-associated interstitial lung disease? BMC Pulm Med 17: , III Nurmi H, Kettunen H-P, Suoranta S-K, Purokivi M, Kärkkäinen M, Selander T, Kaarteenaho R. Several high-resolution computed tomography findings associate with survival and clinical features in rheumatoid arthritis-associated interstitial lung disease. Resp Med 2018;134:24-30 The publications were adapted with the permission of the copyright owners.

18 XIV

19 XV Contents 1 INTRODUCTION REVIEW OF THE LITERATURE RHEUMATOID ARTHRITIS INTERSTITIAL LUNG DISEASES EXTRA-ARTICULAR MANIFESTATIONS IN RA OVERVIEW OF RA-ILD History Definition EPIDEMIOLOGY PATHOPHYSIOLOGY AND RISK FACTORS Genetics Citrullination and autoimmune response Smoking and other patient-dependent risks Factors relating severity of RA Other potential biomarkers for RA-ILD CLINICAL FEATURES Symptoms and clinical findings PFT and chest radiography Bronchoalveolar lavage CLASSIFICATION OF RA-ILD DIAGNOSTICS HRCT The radiological features of the RA-ILD subtypes The histological features of most common RA-ILD subtypes The role of surgical lung biopsy THE COURSE OF THE DISEASE Disease progression RA-ILD and prognosis Acute exacerbations ASSESSMENT OF PROGNOSIS Radiological predictors of mortality Histopathological predictors of mortality Pulmonary function tests, 6MWT and prognosis Patient- and RA-related predictors of mortality THE RISK PREDICTION MODELS IN ILDS COMORBIDITIES Comorbidities in RA-ILD Comorbidities in RA CAUSES OF DEATH TREATMENT Whom and how to treat? Immunosuppressive agents... 27

20 XVI Synthetic disease modifying antirheumatic drugs (DMARDs) Biologic agents Pulmonary rehabilitation Lung transplantation Antifibrotic drugs Treatment of RA-ILD exacerbation Other treatments Palliative care AIMS OF THE STUDY MATERIAL AND METHODS DATA SOURCES AND PATIENT SELECTION GATHERING OF DEMOGRAPHIC INFORMATION (I, II, III) RADIOLOGICAL EVALUATION Re-classification of HRCTs (I, II, III) Further interpretation of the CTs and the scoring system (III) STAGING SYSTEMS (II) STATISTICAL ANALYSIS ETHICAL CONSIDERATIONS RESULTS PATIENT CHARACTERISTICS Demographics Medication for RA PFT Radiological subtypes GAP and ILD-GAP (II) Comparison of the demographics in UIP and non-uip patients (I) Comparisons within RA-UIP subgroup RADIOLOGICAL FINDINGS Disease progression Inter-observer agreement (III) The HRCT findings in different subtypes (III) Original radiological reports HISTOLOGICAL DATA AND BAL COMORBIDITIES (I) CAUSES OF DEATHS (I) CORRELATIONS BETWEEN CLINICAL DATA, PFT AND RADIOLOGY (III) THE COURSE OF THE DISEASE Differences between RA-UIP and non-uip patients (I) Survival (I, II) Predictors of mortality (II, III) VALIDATION OF THE GAP AND ILD-GAP MODELS (II) DISCUSSION GENERAL DISCUSSION OF THE STUDY DESIGN Search for the patients and sample size Data gathering and missing data Implication of the RA medication... 53

21 XVII Diagnostics Reliability of the radiological re-categorization CLINICAL FEATURES OF THE COHORT Subject characteristics and PFT Radiological features and their correlation to RA duration (III) BAL results Original radiological reports Disease course in UIP and non-uip patients (I) COMORBIDITIES AND CAUSES OF DEATHS (I) SURVIVAL (I) PREDICTORS OF MORTALITY (II, III) Pulmonary function tests and CPI Clinical factors Radiological factors associating with decreased survival VALIDATION OF THE GAP AND ILD-GAP MODELS (II) FUTURE PERSPECTIVES CONCLUSIONS REFERENCES... 62

22 XVIII

23 XIX Abbreviations 6MWT Six-minute Walk Test DMARD Disease modifying ACPA Anticitrullinated protein antirheumatic drug antibodies ERS European Respiratory Society AE Acute exacerbation ESR Erythrocyte sedimentation ALAT Latin American Thoracic rate Association ExRA Extra-articular manifestations ANA Antinuclear antibodies in rheumatoid arthritis ARDS ATS BAL Acute respiratory distress syndrome American Thoracic Society Bronchoalveolar lavage FEV1 Forced expiratory volume in 1 second FIN-RACo Finnish rheumatoid arthritis combination therapy CAD Coronary artery disease fnsip Fibrotic nonspecific CI Confidence interval interstitial pneumonia COPD Chronic obstructive FPF Familial pulmonary fibrosis pulmonary disease FVC Forced vital capacity cnsip Cellular nonspecific GAP Gender, age, and interstitial pneumonia physiological variables CPI Composite physiologic index GER Gastro-esophageal reflux CRP Clinical-radiologic- GGO Ground-glass opacity physiologic scoring system HAQ Health-assessment CT Computed tomography questionnaire CTD Connective tissue diseases HAQ-DI HAQ Disability Index score DAD Diffuse alveolar damage HLA Human leukocyte antigen DAS-28 Disease activity score in 28 joints HR HRCT Hazard ratio High-resolution computed DIP Desquamative interstitial tomography pneumonia ICD International Classification of DLCO Diffusion capacity to carbon monoxide Diseases

24 XX IIP Idiopathic interstitial RTX Rituximab pneumonias SD Standard deviation IL Interleukin SE Shared epitope ILD Interstitial lung disease SLB Surgical lung biopsy insip Idiopathic nonspecific SSc-ILD Systemic sclerosis-associated interstitial pneumonia interstitial lung disease IPF Idiopathic pulmonary fibrosis TBB Transbronchial biopsy JRS Japanese Respiratory Society TBCx Transbronchial cryobiopsy KL-6 Krebs von den Lungen TERT Telomerase reverse KUH Kuopio University Hospital transcriptase LDH Lactate dehydrogenase TLC Total lung capacity LEF Leflunomide TNF Tumor necrosis factor LIP Lymphocytic interstitial UIP Usual interstitial pneumonia pneumonia VAS Visual analogue pain scale LTx lung transplantation VATS Video-assisted thoracoscopic MDD Multidisciplinary discussion surgery MMF Mycophenolate mofetil VEGF Vascular endothelial growth MMP Matrix metalloproteinase factor MTX Methotrexate NSIP Nonspecific interstitial pneumonia OP Organizing pneumonia PDGF Platelet derived growth factor PFT Pulmonary function test RA Rheumatoid arthritis RA-ILD Rheumatoid arthritisassociated interstitial lung disease RB Respiratory bronchiolitis RF Rheumatoid factor ROSE Risk stratification score

25 XXI

26

27 1 1 Introduction Rheumatoid arthritis (RA) is a systemic inflammatory disease that affects approximately 1% of the global population (1) and about 0.8% of the Finnish population (2). Patients with RA have greater mortality than the healthy population and their average life expectancy is shortened by approximately 10 years (3). The majority of deaths are due to extra-articular manifestations (ExRA) of the disease, of which interstitial lung disease (ILD) is one of the most important (4). Approximately every tenth patient with RA develops clinically evident ILD with respiratory symptoms and/or a decline in pulmonary function tests (PFT) during the course of the rheumatoid disease (5). In a substantial percentage i.e % of asymptomatic RA patients, high-resolution computed tomography (HRCT) scans have revealed evidence of interstitial lung involvement and a large proportion of those patients with subclinical disease deteriorate with time (6,7). Similarly, as in HRCT-based studies, autopsy studies have detected high prevalences of up to 35% for rheumatoid arthritis-associated interstitial lung disease (RA-ILD) (8). RA-ILD greatly affects the lives of RA patients, increasing both morbidity and mortality (9). Recent studies have shown that despite the decline in overall mortality in RA, deaths attributable to RA-ILD have substantially increased (10). However, the course of the disease is highly heterogenic, as some patients remain stable for years or even decades, while others develop an insidious progressive disease (11). Predicting the survival of an individual patient with ILD is challenging (9). Several factors i.e. physiological, radiological and histopathological characteristics, as well as demographic variables have been proposed to predict disease progression and survival (12). Several indexes combining single factors into multifaceted scoring systems have been developed over the past years to help in the risk prediction (13), but these models have primarily been developed for idiopathic pulmonary fibrosis (IPF) and have not been previously investigated or validated in RA-ILD patients. There has been very little research conducted concerning RA-ILD in Finland after some studies that were conducted in the 1980s (14). Subsequently, radiological technology has developed, significantly improving the diagnostic accuracy and enabling a modern classification of the ILDs. Moreover, the first randomized controlled studies of combination therapy on RA were performed in the late 1990s, the long-term effects of which have been documented in a Finnish study (15). Since then, the recommended treatment of newly diagnosed RA has involved a combination of methotrexate (MTX), sulfasalazine, hydroxychloroquine plus prednisolone; this recommendation has stabilized treatment protocols and thus may have an impact on the course of disease on RA and RA-ILD as well. In addition, the repertoire of drugs has expanded with the arrival of biological drugs, the first of which was taken into use in 1999 (16) and therefore results from the studies from the 1980s are no longer completely applicable. The purpose of the present study was to evaluate a cohort of RA-ILD- patients treated in Kuopio University Hospital (KUH) health care district, re-classify the cases with ILD according to the current criteria, evaluate the course of the disease, comorbidities, causes of death, prognostic factors for survival in different subtypes as well as testing the suitability of the prediction models previously developed for IPF. We wanted to evaluate how many and which subtype of patients with RA-ILD exist in the KUH region, information which was formerly unknown due to the non-standardized diagnosis coding. We wanted to explore the course of the disease and examine how the lung disease has affected the patients lives and lifespans. A secondary aim was to seek means to help clinicians in the difficult estimation of an individual s prognosis and to test the prediction models of IPF in

28 RA-ILD patients. The identification of the high-risk patients could help to plan individualized monitoring and in particular, to consider when to proceed to lung transplantation (LTx) or perhaps to a treatment trial. Hopefully, this thesis will improve the recognition of RA-ILD and provide clinicians with tools for identifying those patients who are at the highest risk of death. 2

29 3 2 Review of the literature 2.1 RHEUMATOID ARTHRITIS RA is a systemic, chronic progressive inflammatory disease that is characterized by destructive joint disease, systemic inflammation, and in most of the patients, the presence of autoantibodies to either rheumatoid factor (RF) or citrullinated proteins or to both (17). The prevalence of clinically significant RA is about 0.8% and the incidence of RA is about 40 / of the adult Finnish population (2). 2.2 INTERSTITIAL LUNG DISEASES ILDs are a heterogeneous group of differently behaving rare diseases, characterized by varying degrees pulmonary inflammation and fibrosis formation. Most of the cases are idiopathic, but ILDs can also be attributable to exogenous factors, such as connective tissue disorders (CTD) (e.g. RA), exposure to organic dusts (e.g. asbestos), or exposure to certain drugs. ILDs are commonly categorized into four categories: idiopathic interstitial pneumonias (IIP), ILDs of known causes, granulomatous diseases and a remnant group of other ILDs (Figure 1) (18,19). Figure 1. Classification of ILDs. Modified from the 2002 consensus classification of the IIPs (18) and the 2013 update (20). IIP = idiopathic interstitial pneumonias; ILD = interstitial lung disease; CTD = connective tissue diseases; LAM = lymphangioleiomyomatosis; HX = Langerhans histiocytosis; IPF = idiopathic pulmonary fibrosis; NSIP = nonspecific interstitial pneumonia; RB-ILD = respiratory bronchiolitis interstitial lung disease; DIP = desquamative interstitial pneumonia; COP = cryptogenic organizing pneumonia; AIP = acute interstitial pneumonia; LIP = lymphocytic interstitial pneumonia; PPFE = pleuroparenchymal fibroelastosis.

30 4 2.3 EXTRA-ARTICULAR MANIFESTATIONS IN RA Since RA is a systemic inflammatory disease, it is recognized that there can be extensive variability between different ExRAs (21). There is a wide spectrum of pulmonary (Table 1), as well as cardiac and other organ manifestations (Table 2). RA-ILD is one of the most important ExRAs significantly impacting on morbidity and mortality of the patients with RA (9). Table 1. Frequency and impact of pulmonary manifestations in patients with rheumatoid arthritis (RA). Adapted and modified from Lake et al (22). Frequency Impact Pleural abnormalities Pleuritis Effusion Pleural thickening Other (unexpandable lung, empyema, chyliform effusion, pneumothorax, hemothorax, pyopneumothorax, bronchopleural fistula) Upper airway Crico-arytenoid immobility with vocal cord abnormality, cord nodules, recurrent laryngeal or vagus nerve vasculitis, cord paralysis Lower airway Airflow obstruction ++ + Obliterative bronchiolitis Bronchiectasis + + Interstitial lung disease Apical fibrosis and Caplan syndrome + + Nodules Pulmonary hypertension Vasculitis Related to RA + + Related to treatment Pneumonitis Pleuritis / effusion + + Lung cancer Pulmonary thromboembolism + ++ Parenchymal Vascular Musculoskeletal related Chest wall immobility and respiratory failure Infection Treatment related Increased risk RA = rheumatoid arthritis.

31 5 Table 2. Other organ manifestations in addition to those classified as pulmonary extra-articular manifestations (ExRA) in patients with rheumatoid arthritis. Adapted and modified from Prete et al (23). Affected tissue or organ Skin Heart Nervous system Eyes Hematological system Kidneys 2.4 ExRA Not severe Nodules Raynaud s phenomenon Valvular heart disease Myocarditis Arrhythmias Secondary Sjögren syndrome Sicca syndrome - ExRA Severe Petechiae, purpura Ulcers, gangrene Pericarditis Coronary vasculitis and aortitis Mono/polyneuritis multiplex Central nervous system vasculitis Episcleritis or scleritis Retinal vasculitides Felty s syndrome Glomerulonephritis Interstitial nephritis Amyloid deposition OVERVIEW OF RA-ILD History The first descriptions of three RA patients with rapidly progressive fibrosing pneumonitis were published in 1948 by Ellman and Ball (24). The first review article concerning a few sporadic cases was published in 1965 (25). In the 1960s, there were some doubts about whether there was any association between RA and pulmonary fibrosis, even though several research groups had investigated the relationship between these two disorders and moreover, risk factors for the RA-related pulmonary fibrosis could already be detected (26). By the end of the 1970s, typical symptoms and clinical signs, PFT findings and typical radiological and histological findings were defined. In that era, the common appearances in the chest X-rays were described as non-specific diffuse bilateral shadows in the lower zones (27) which was likely a mixture of the currently known different entities. A pivotal change in the field of pulmonary fibrosis research occurred in the early 1990s with the development of the HRCT technique (28). Gradually, the computed tomography (CT) findings of RA-ILD were described (29,30) resulting in a more uniform terminology and enabling the identification of different subtypes (31) Definition Currently, there is no official international accepted definition or criteria for RA-ILD or its different subtypes. The definitions are commonly adopted from the American Thoracic Society/European Respiratory Society (ATS/ERS) statement on IPF (32) and the IIPs (18), the latter being updated in 2013 (20). These criteria include an underlying RA diagnosis and ILD on HRCT scan or lung biopsy or both, without any identifiable etiology to account for the lung changes. With this definition, respiratory infections, treatment-related ILDs and e.g. rheumatoid nodules can be distinguished from RA-ILD (33).

32 6 2.5 EPIDEMIOLOGY The reported prevalence and incidence of RA-ILD have varied in different studies. The differences derive from different study populations as some studies have included only asymptomatic patients or those with recently diagnosed RA, whereas in other reports, the cohorts have consisted of longstanding and/or symptomatic RA patients. Naturally, the development of modern and more sensitive diagnostic technology has increased the estimations of RA-ILD prevalence. Moreover, the variable RA-ILD definitions used in the past make any comparison of the reported prevalences difficult. The prevalence has been estimated as low as 4.5% (25) or 1.6% in those studies that used chest radiographs in ILD diagnostics (26). In one investigation using CT-based diagnostics, only one fifth of the RA-patients with abnormal CT scans had visible ILD changes in their chest X-rays (7); nowadays the prevalence has been shown to be much higher with the evolution of more sensitive HRCT imaging (34,35). One study which applied the diffusion capacity to carbon monoxide (DLCO), estimated the prevalence of RA-ILD to be as high as 41% (36), whereas another investigation using autopsy material of 81 RA patients found ILD in every third RA patient with advanced disease (8). In a cohort exploring RA patients diagnosed less than two years earlier, 58% of the patients had changes suggestive of ILD in either chest X-ray, HRCT, PFT, bronchoalveolar lavage (BAL) and/or 99Tc-DTPA (technetium-99-m-labelled diethylenetriamine pentaacetate) scan. Of these, 76% were asymptomatic (7). Another study with recent onset RA patients detected HRCT abnormalities and/or abnormal PFTs in 45% of the patients, of which 10% were symptomatic (34). The prevalence of clinically relevant ILD was estimated between 4 and 7.9%, with a 30- year cumulative incidence of 7.7% in a large population-based cohort of RA patients (4). Similar results were reported in another study in which clinically significant ILD was observed in approximately 6.8% of women and 9.5% of men with RA (10). A subsequent report from Turesson et al. described comparable results estimating the 30-year cumulative incidence as 7% (5). In the cohort study of Koduri et al, the annual incidence rate for the development of RA-ILD was 4.1/1000 (95%CI: ), with a 15-year cumulative incidence of 62.9/1000 (95%CI: ) (37). Overall, the lifetime risk of a clinically significant RA-ILD is nowadays shown to be approximately 10% (9), whereas in unselected populations, subclinical ILD has been detected in 20-30% (10,38,39) although in some reports it has been estimated to be present in two thirds of RA patients (40,41). Selected studies reporting the prevalence and/or incidence of RA-ILD are shown in table 3.

33 126 Koduri, 2010 (37) 65 <1 year, 61 >3 years not reported 54 Bilgici, 2005 (41) Mori, 2008 (43) Bongartz, 2010 (4) ± Turesson, 2003 (5) <2 years Not reported 8 ± Not reported 336 Dawson, 2001 (39) 13.7 ± < 2 years Not reported Not reported Duration of RA Number of the patients 516 Gabbay, 1997 (7) Walker, 1969 (26) Frank, 1973 (36) Suzuki, 1994 (8) Saag, 1996 (42) Study, year Probable ILD = X-ray + treating physician s diagnosis of ILD. Definite ILD = Diagnosis of ILD by a pulmonologist + 2/3 of following: ILD on CT / x-ray, restrictive PFT, biopsy confirmation HRCT HRCT Clinical judgement and decreased VC or DLCO by 15% from normal HRCT HRCT Chest X-ray, HRCT, PFT, BAL and/or 99Tc-DTPA scan Chest X-ray and PFT Autopsy DLCO Chest X-ray Method 2.9% 4.0% definite, 7.9% definite + probable 11.9% 67.3% - Chest X-ray abnormality 12%. FVC <80% % pred. 12.5%, DLCO <80% %pred. 19.0%, any of the above 32.4% Abnormalities in 58% (one or more investigations), 22% PFT, 6% X-ray, 33% HRCT. 18.7% 34.6% 41.4% 1.6% Prevalence of RA-ILD Annual incidence 4.1/1000, 15year cumulative incidence 6.3% 10-, 20- and 30- year cumulative incidences 3.5%, 6.3% and 7.7% - 30-year cumulative incidence 6.8% Incidence All symptomatic 42.6% symptomatic 23.8% symptomatic not reported 71.4% dyspnoea, 46.4% productive cough Not reported 14% Not reported Not clearly reported Usually not symptomatic Not reported Symptoms Table 3. Summary of studies investigating the prevalence and/or incidence of RA-ILD in different study populations using different diagnostic methods. 7

34 Habib, 2011 (34) Olson, 2011 (10) Restrepo, 2015 (44) Zhang, 2017 (35) 40 <2 years HRCT and/or PFT Abnormal HRCT 27.5%, abnormal PFT 32.5%, abnormal PFT and HRCT 20%, abnormal PFT and HRCT with symptoms 10% - 90% asymptomatic > Not reported ICD-9 and ICD-10 codes 6.8 % in women, 9.5% in men - Not reported ± 10.8 Chest X-ray, CT, HRCT or lung biopsy ±9 (range 2 weeks 40 years) 8.8% - All symptomatic HRCT 43.1% - 41% symptomatic of patients with HRCT changes DLCO = diffusion capacity to carbon monoxide; HRCT = high-resolution computed tomography; PFT = pulmonary function tests; BAL = bronchoalveolar lavage; VC = vital capacity; FVC = forced vital capacity; RA = rheumatoid arthritis; ILD = interstitial lung disease; 99Tc-DTPA scan = technetium-99-m-labelled diethylenetriamine pentaacetate scan; ICD = international classification of diseases. 8

35 9 2.6 PATHOPHYSIOLOGY AND RISK FACTORS Genetics The mechanism of pulmonary fibrosis in ILD is poorly understood. Available data suggest a role for both genetic and environmental factors. It has been speculated that there is some underlying genetic vulnerability with some form of injury to the lung triggering the fibrosis formation (45). Specific human leukocyte antigen (HLA) variants, such as HLA-B40 and HLA-DR4, have been associated with RA-ILD (46,47). Some polymorphisms of the HLA-DRB shared epitope (SE) have been associated with an increased risk of ILD, while others seem to protect from ILD (48). In a Japanese study, HLA-DQB1*06, HLA-DRB1*15 and *16 alleles were associated with an increased risk of ILD, and HLA-DRB1*04 and HLA-DQB1*04 appeared to be protective against the development of RA-ILD, although the majority of the HLA-DRB1 subtype alleles had no significant, either negative or positive, associations (49). An association between HLA-DRB1*1502 and ILD was also observed in the study of Mori et al. (50). In the study of Restrepo et al., the association between ILD and smoking was seen only in those patients with an HLA-DRB1 SE, which was speculated to reflect a geneenvironment interaction (44). The risk of RA-ILD was shown to be increased in patients with the non-m1m1 alpha one antitrypsin phenotype (51). One study has investigated the MUC5B polymorphism, which is associated with IPF, but found no association between it and RA-ILD (52). Other potential genetic factors believed to be associated with the development of ILDs include surfactant protein abnormalities (53), telomerase reverse transcriptase (TERT) mutations and telomere length (54), but not all of them have been investigated in cohorts containing of RA patients. RA-ILD and IPF have many similarities in terms of their histopathology and epidemiology, thus raising questions about whether these two fibrotic lung diseases could have a similar genetic background. The recent study of Juge et al. performed whole exome sequencing on 101 patients with RA-ILD (55). Restricting their analysis to nine genes linked to familial pulmonary fibrosis (FPF), they found mutations in the TERT, RTEL1, PARN or SFTPC coding regions in 11.9% of patients with RA-ILD. Patients with mutations in the TERT, RTEL1 or PARN genes were also found to have short telomeres in their peripheral blood leukocytes, suggesting that these mutations were biologically relevant, although the findings will still need to be confirmed in the future. These results suggest shared genetic risk factors in RA-ILD and FPF (55) Citrullination and autoimmune response Citrullination is a post-translational modification of proteins in which arginine is converted to citrulline, resulting in a change in the structure of the protein and an increase in its immunogenicity. Several diseases including IPF have been associated with abnormal citrullination of peptides (22). Protein citrullination leads to the production of anticitrullinated protein antibodies (ACPA), which are commonly present in patients with RA and can be detected in the serum for several years before clinical disease onset (56). High titers of ACPAs in patients with RA have been shown to be associated with an increased risk of ILD in several different studies (50,57,58), but this association was not detected in one publication (59). ACPAs are thought to cause synovial inflammation through the deposition of immune complexes and targeting of citrullinated synovial proteins such as vimentin, filaggrin and fibronectin (22). As is the case in other ILDs, protein citrullination promotes autoimmune responses that further contribute to tissue damage through inflammatory responses characterized by cellular infiltration and the release of selected cytokines, chemokines and

36 10 growth factors, such as tumor necrosis factor (TNF), vascular endothelial growth factor (VEGF), platelet derived growth factor (PDGF) and interleukins (IL). These influence the differentiation and proliferation of fibroblasts, increased synthesis and deposition of extracellular matrix, as well as increased activity of matrix metalloproteinases resulting in ILD (60,61) (Figure 2). In a recent study, tissue samples were obtained from lung and synovial biopsies of RA patients and identical citrullinated protein contents were found in both specimens (62). Some investigators have suggested that RA begins in the lungs, by descripting cohorts of ACPA positive patients with lung disease without RA, some of whom developed articular disease afterwards (63,64). However, the initiation point of the autoimmunity process in RA has also been suggested as being in the oral mucosa or gastrointestinal system (65). Figure 2. Schematic illustration of the concepts in the pathogenesis of RA-ILD. Original figure reproduced from the review article of Shaw et al (60), with permission of the ERS 2015 (European Respiratory Review Mar 2015, 24 (135) 1-16; DOI: / ). RA-ILD = rheumatoid arthritis-associated interstitial lung disease; HLA = human leukocyte antigen; ECM = extracellular matrix; MMP = metalloproteinases; TNF = tumor necrosis factor; VEGF = vascular endothelial growth factor; PDGF = platelet derived growth factor; IL = interleukins; CCP = anti-cyclic citrullinated peptide.

37 Smoking and other patient-dependent risks Smoking has been shown to promote citrullination of lung proteins, thus leading to the appearance of ACPAs (66,67). Moreover, the risk of developing ACPAs is increased in heavy smokers carrying at least one copy of the HLA-DRB1 SE alleles (68). Thus, it seems that smoking, the HLA-DRB1 SE and ACPAs interact to increase the risk of RA (44,68). The exact relationship between tobacco smoke and the development of RA-ILD is unknown but a dual effect has been proposed, with one explanation being the above-mentioned smokinginduced protein citrullination in the lungs and the subsequent ACPA-promoted lung injury and another pathway being the independent elevated risk of lung injury and fibrosis caused by smoking (69) (Figure 3). Smoking has associated with RA (70), its severity (71), RA-ILD (42,71-73) as well as other ExRAs (21), in many studies. However, the development of ILD does not require smoke exposure but can also be encountered in lifelong non-smokers (74). Higher risks for the ILD development associating with male sex (7,73,75) and aging (37,73,76) have been confirmed in many studies (Table 4). In the study of Koduri et al. the likelihood of having ILD increased by 64% for every 10-year increase in age (37). In another study, more than a four-fold risk of ILD was detected in patients over 65 years (50). Figure 3. Schematic illustration of the dual role of smoking in the pathogenesis of RA-ILD. Original figure adapted and modified from review article of Johnson et al (69). ACPA = anticitrullinated protein antibodies; HLA-DRB = human leukocyte antigen-d-related beta alleles; SE = shared epitope; RA-ILD = rheumatoid arthritis-associated interstitial lung disease.

38 Factors relating severity of RA Different kinds of measurements have revealed that the RA severity is related to the development of RA-ILD (45). High disease activity and extensive disability within the first 2 years after RA diagnosis were shown to predict subsequent development of severe ExRA (77). Similarly, the presences of erosive joint disease and/or rheumatoid nodules, as well as high levels of erythrocyte sedimentation rate (ESR) have been identified as risk factors for the development of ILD (4,37). Commonly used scores, such as disease activity score in 28 joints (DAS28) and Health Assessment Questionnaire Disability Index score (HAQ-DI) are also associated with RA-ILD (37,42,44). There is also some evidence that high levels of circulating RF increase the risk of ILD development in RA, as well as the risk of other ExRAs (78-80). The explanation or exact mechanism for this association is unclear but might be due to the formation of circulating immune complexes (60). However, the interpretation of the RF titer is challenging, since e.g. tobacco smoking can increase RF serum titer in the general population as well as in patients with RA (81,82) Other potential biomarkers for RA-ILD A study by Harlow et al. described citrullinated versions of Heat Shock Protein 90 isoforms as potential biomarkers for RA-ILD (83). The levels of Krebs von den Lungen (KL-6) have been shown to correlate with the severity of CT findings in RA-ILD (84). Previously Oyama et al reported that an increase in serum KL-6 levels in RA associated with the presence of active interstitial pneumonitis (85). Moreover, elevated KL-6 levels have been claimed to associate with the presence of pulmonary fibrosis in patients with systemic sclerosis (SSc) (86), although they can also be present in IIPs, hypersensitivity pneumonitis, radiation pneumonitis and Pneumocystis jirovecii pneumonia (86). In a quite recent study, multiplex enzyme-linked immunosorbent assays (ELISAs) and Luminex xmap technology were used to assess 36 cytokines/chemokines, matrix metalloproteinases (MMPs) and acute-phase proteins in two different cohorts. In this study, levels of MMP-7 and interferon-γ-inducible protein 10 (IP-10)/CXCL10 were elevated in the serum of RA patients with ILD, versus RA patients without ILD (87).

39 13 Table 4. Different risk factors associated with interstitial lung disease related to rheumatoid arthritis (RA). Table partly adapted and extended from the review article of C. Johnson (69). Clinical factors Study (first n Age Male Later RA RF ACPAs DAS- HLA- Cigarette HAQ-DI High Tumour author) sex onset RA duration titer 28 DRB1 SE smoking score ESR markers Kelly (73) Restrepo (44) Bongartz (4) Koduri (37) Wang (88) Wang (89) 41 + Song (90) Mori (50) Akiyama (91) Doyle (72) Rocha-Munoz (92) Wang (93) Saag (42) RA = rheumatoid arthritis; RF = rheumatoid factor; ACPAs = anticitrullinated protein antibodies; DAS-28 = disease activity score in 28 joints; HLA DRB1 SE = human leucocyte antigen DRB1 shared epitope; HAQ-DI = Health Assessment Questionnaire Disability Index score; ESR =erythrocyte sedimentation rate.

40 CLINICAL FEATURES RA-ILD most commonly occurs in patients in their late 50s to 60s (94,95). The onset of ILD has been shown to appear at an average of 10 years after RA diagnosis (94,96). ILD may, however, precede the development of arthritis in RA, as was the case in 10 out of 111 patients (9.0%) initially diagnosed as IPF (97). Others have reported ILD predating the articular disease in approximately 10-17% of the patients (73,96), and the time interval from ILD to RA diagnosis has varied from 6 months to 8 years (96,97) Symptoms and clinical findings Many patients with RA display no respiratory symptoms despite the presence of radiographic or PFT abnormalities. In a study of 52 RA patients, HRCT abnormalities were observed in almost 70% of the patients of whom only 40% experienced respiratory symptoms (41). In another study, approximately one-third of asymptomatic patients with RA had interstitial lung changes in their CT scan, and of those, more than half progressed within 2 years (6). Even though some of the patients are asymptomatic, many patients suffer from shortness of breath during exercise or cough, which is commonly dry and insidious (48,98). Once present, the symptoms usually progress over time. Less common respiratory symptoms include chest tightness/pain, wheezing and sputum secretion. In addition, fatigue or generalized weakness can also be seen (98,99). In particular, dyspnoea on exertion can be difficult to notice in patients whose joint disease limits exercising. Common physical examination findings include bi-basilar crackles, which are present in most of the patients (17). Finger clubbing can also be seen, but not as frequently as in IPF patients (100). Hypoxemia can be seen either during exercise, for example using the 6minute walking test (6MWT), or also at rest in the advanced cases (9,22,101) PFT and chest radiography In the majority of the patients, spirometry usually demonstrates a restrictive defect with a lowered forced vital capacity (FVC), although the spirometry can also be completely normal, especially in the early stage of the disease. Decreased DLCO has been reported to be the most sensitive test for predicting the presence of ILD on HRCT (39) as well as predicting the extent of disease (102); this has been detected in up to 25-60% of asymptomatic RA patients (7,39). The chest X-ray is often normal in patients with early RA-ILD and has a low sensitivity for detection of ILD since more than 50% of patients with a normal chest X-ray will have abnormalities visible in their HRCT (7,39). When present, the common X- ray changes are reticular, reticulonodular or honeycomb changes, mainly in the basal area of the lungs (17) Bronchoalveolar lavage BAL is not routinely performed in RA-ILD patients as BAL patterns are nonspecific and abnormalities in BAL can be present also in the absence of ILD (103). Increased concentrations of neutrophils, eosinophils and lymphocytes, as well as a decreased CD4/CD8 ratio have been described in RA-ILD patients (7,96,104,105). RA patients with advanced interstitial lung disease on HRCT have been shown to have significantly higher BAL fluid cellular concentrations as their counterparts with mild or no ILD changes (105). Some investigators have found an increased cellularity in BAL compared to healthy controls but not when compared to RA patients without ILD (106). Another report investigated 24 RA patients; nine with significant ILD with symptoms showed increased

41 15 percentages of neutrophilia, five asymptomatic patients with evidence of ILD showed elevated BAL lymphocyte numbers, whereas those without any evidence of ILD had normal BAL (104). In the study of Gabbay et al 29 out of 36 recent onset RA-patients underwent a BAL and of those 52% showed elevations in the numbers of neutrophils, eosinophils and/or lymphocytes (7). Moreover, an abnormal cell profile in BAL was present in all clinically significant RA-ILD patients, whereas 10 out of 13 subclinical cases and none of 11 RA patients without evidence of ILD displayed any signs of BAL pathology (7). The results of BAL analysis may have a modest ability to distinguish between the different subtypes of ILD, as a slight neutrophilia may be more frequent in usual interstitial pneumonia (UIP) while lymphocytosis is more common in non-specific interstitial pneumonia (NSIP) and organizing pneumonia (OP) (75), but convincing studies are lacking which would have correlated BAL findings with HRCT patterns of disease (107). As most of the RA patients are immunosuppressed because of the RA medication, they experience an increased risk of developing various infections. Thus, the most useful role of BAL in clinical practice is in patients with an acute/subacute respiratory worsening when the clinician needs to rule out the appearance of an opportunistic infection (45,75). Moreover, a significant eosinophilia in BAL could suggest a drug reaction rather than ILD (108) 2.8 CLASSIFICATION OF RA-ILD The subtypes of RA-ILD have been categorized according to the subdivisions of the IIPs. Since 2002, the IIPs have been classified into seven different entities according to the ATS/ERS International Multidisciplinary Consensus Classification of the IIPs (18) and the guidelines were updated in 2013 (20). In the updated IIP Consensus Classification, the previously used term cryptogenic fibrosing alveolitis was removed and the different entities are categorized as the major, rare and unclassifiable IIPs. The major IIPs can further be categorized as chronic fibrosing IIPs, smoking-related IIPs and acute/subacute IIPs (Table 5). Of these IIP patterns, UIP, NSIP, OP and diffuse alveolar damage (DAD) are the most commonly observed patterns in RA-ILD, but respiratory bronchiolitis (RB), lymphocytic interstitial pneumonia (LIP) and desquamative interstitial pneumonia (DIP) are also possible, although rare (109). Unlike the situation in other CTDs, the most common radiologic and histopathologic pattern of RA-ILD is UIP (31,96,101,110,111). Table 5. Categorization of major idiopathic interstitial pneumonias according to the ATS/ERS updated International Multidisciplinary Classification (20). Category Radiologic and/or histologic pattern Multidisciplinary diagnosis Chronic fibrosing Usual interstitial pneumonia (UIP) Nonspecific interstitial pneumonia (NSIP) Idiopathic pulmonary fibrosis (IPF) Idiopathic nonspecific interstitial pneumonia (insip) Smoking-related Respiratory bronchiolitis (RB) Desquamative interstitial pneumonia (DIP) Respiratory bronchiolitis-interstitial lung disease (RB-ILD) Desquamative interstitial pneumonia (DIP) Acute/subacute Organizing pneumonia (OP) Diffuse alveolar damage (DAD) Cryptogenic organizing pneumonia (COP) Acute interstitial pneumonia (AIP)

42 DIAGNOSTICS HRCT The development of HRCT techniques has reduced the need for lung biopsies and the diagnostics of RA-ILD is nowadays mainly performed radiologically. According to the official diagnostic criteria, IPF can be diagnosed without biopsy in the presence of a typical HRCT pattern, i.e. definite UIP, the characteristics of which are shown in table 6. Similar criteria are used in RA-UIP diagnosis due to the lack of its own criteria (32). A good correlation between histopathological and radiological features has been previously observed and a definite UIP pattern in HRCT has been demonstrated to be a sensitive and specific technique for detecting the histopathologic UIP pattern in both IPF and RA-ILD ( ). For example, Hozumi et al showed that the specificity and sensitivity of HRCT for the histological UIP pattern were 100% and 41.7%, respectively (116). In another study conducted by Assayag et al. the definite UIP pattern in HRCT was highly predictive of the histological UIP pattern, having a 96% specificity and a 45% sensitivity (113). Thus, HRCT provides a useful non-invasive tool for diagnosing RA-ILD patients, especially those with a definite UIP pattern. A multidisciplinary discussion (MDD) between pulmonologists, radiologists and pathologists increases the accuracy of the diagnosis and is recommended as a part of the diagnostic work in ILDs (117). The group of experts is recommended to include also a rheumatologist when handling the CTD-ILD cases. Table 6. The radiological criteria of UIP pattern according to the ATS/ERS/Japanese Respiratory Society (JRS)/Latin American Thoracic Association (ALAT) guidelines 2011 (32). UIP = usual interstitial pneumonia. Definite UIP Subpleural basal predominance Possible UIP Subpleural basal predominance Not UIP Upper- or mid lung predominance Reticular abnormalities Reticular abnormalities Peribronchovascular predominance Honeycombing with or without No inconsistent findings Severe ground glass opacities (ground traction bronchiectasis (column 3) glass pattern > reticulation) No inconsistent findings (column 3) Profuse micronodules Cysts (other than honeycombing) Mosaic-attenuation pattern/air trapping Consolidation The radiological features of the RA-ILD subtypes The typical radiological features of the most common RA-ILD subtypes are shown in Table 7 and Figure 4. The UIP pattern occurs in 40-62% of the patients with RA-ILD (48). As described above, the distribution of UIP on HRCT is typically basal and peripheral, often also patchy. UIP is characterized by the presence of reticular opacities, which often are associated with traction bronchiectasis. Honeycombing is common and essential for making a definite diagnosis.

43 17 Ground-glass opacities (GGO) are also common, but usually less extensive than the reticulation (Figure 4a) (32). Table 7. Prevalences and radiological features of the four most common RA-ILD subtypes. Adapted and modified from review articles of Cavagna et al (75) and Lake & Proudman 2014 (22). Prevalence Radiological pattern Usual interstitial pneumonia (UIP) ++++ Subpleural, peripheral, basal predominance, Nonspecific interstitial pneumonia (NSIP) reticulation, honeycombing with or without traction bronchiectasis, architectural distortion, less diffuse GGO, absence of inconsistent features +++ Bilateral GGO, possible reticulation, may have traction bronchiectasis, little or no honeycombing, often subpleural sparing Organizing pneumonia (OP) ++ Patchy multiple peripheral consolidations, subpleural and peribronchial, often migratory, air bronchograms can be seen Diffuse alveolar damage (DAD) + Patchy or diffuse ground glass changes with basal consolidation, rapid progression GGO = ground-glass opacities. The NSIP pattern, comprising 11-38% of the RA-ILD cases (48), is characterized by basilar bilateral predominant GGO, reticulation with or without traction bronchiectasis, little or no architectural distortion or honeycombing and often a subpleural sparing distribution (Figure 4b) (31,118,119). Sometimes, the features are less typical and can overlap with the UIP pattern (120). The characteristic radiological findings of OP pattern, which is clearly less common than UIP and NSIP, covering approximately 0-11% of RA-ILDs (22), include patchy, peripheral and often migratory consolidations, nodules and sometimes a central GGO surrounded by a ring shaped denser airspace consolidation, also called as the atoll sign/reversed halo sign (Figure 4c) (22, ). DAD pattern is the fourth commonest, although rare, RA-ILD subtype and its typical radiological appearance in the early phases includes patchy or diffuse ground glass changes with basal consolidation and a rapid progression (Figure 4d) (22,124). Later, the DAD appearance can change to the organizing stage associated with distortion of bronchovascular bundles and traction bronchiectasis (20). The rest of the subtypes are rare in RA-ILD patients (37,96,101,111). The CT findings of smoking-related RB include centrilobular nodules, patchy ground glass attenuation, and thickening of the walls of central and peripheral airways (125). These findings can be reversible if the patient stops smoking or is treated with corticosteroids (18). In all cases of DIP, the lower zone distributed GGO is present on CT (126). Irregular linear opacities and a reticular pattern are frequent but not extensive and usually present only in the basal parts. Honeycombing can also be seen in less than one-third of cases, and usually is quite limited in extent (18,126). Finally, the dominant CT finding of LIP is usually GGO, but perivascular cysts or perivascular honeycombing can also be observed and a reticular abnormality is seen in about half of the patients (127).

44 18 Figure 4. Representative high-resolution computed tomography (HRCT) images from the subjects with rheumatoid arthritis- associated interstitial lung disease. A) 64-year-old female with fibrotic changes of usual interstitial pattern (UIP) pattern in basal and subpleural predominance, traction bronchiectasis and honeycombing. B) 69-year-old male revealing peripheral ground glass opacity and typical subpleural sparing representing the nonspecific interstitial pneumonia (NSIP) pattern. C) 55-year-old female showing evidence of bilateral peribronchovascular and peripheral patchy consolidations typical for organizing pneumonia (OP). D) 59-year-old male patient with diffuse alveolar damage (DAD) shows diffuse ground glass opacity, reticulation and traction bronchiectasis The histological features of most common RA-ILD subtypes In patients who undergo a lung biopsy, the criteria of IPF, i.e. a definite UIP pattern, are applied in RA-ILD patients. They include marked fibrosis / architectural distortion with or without honeycombing, predominantly subpleural / paraseptal distribution, patchy involvement of lung parenchyma by fibrosis and presence of fibroblast foci in the absence of features suggesting alternative diagnosis (32). The histopathological diagnostics of other RA-ILD subtypes are made using the consensus classification of IIPs (18). Accordingly, histological features of NSIP can show either cellular (cnsip) or fibrosing (fnsip) patterns the first showing primarily chronic inflammation and the latter usually homogenous interstitial fibrosis lacking the UIP features. The typical histological features of OP are patches involving alveolar ducts and alveoli with the preservation of lung architecture, mild interstitial inflammation and a uniform temporal appearance. The histopathological appearance of DAD contains alveolar

45 19 septal thickening due to organizing fibrosis, airspace consolidation and hyaline membranes with the features being uniform temporal appearance and diffuse (18) The role of lung surgical biopsy As the traditional bronchoscopic transbronchial forceps lung biopsy (TBB) has only sensitivity for UIP detection of 30% (128), the golden standard for the classification or diagnosis of an IIP subtype has historically been a surgical lung biopsy (SLB) (9), although nowadays HRCT has reduced its necessity. Even though it is recommended in non-definite UIP cases, its utilization has significantly declined and most patients do not undergo SLB due to the risks associated with it, as well as due to the adoption of MDD and clinicalradiologic criteria for definitive diagnosis of IPF (129). The most feared feature of SLB is the mortality which is reported to be 1.5% for elective and 16% for non-elective procedures (130). There is evidence that the high mortality is particularly present in patients with IPF i.e. with the UIP pattern (131). Acute exacerbations (AE) are partly responsible for the mortality. Kondoh et al. observed AEs in 2.1% of the 236 ILD patients who underwent a surgical biopsy (132). In a retrospective analysis of 10 patients with AE (6 idiopathic NSIP (insip), 3 RA-UIP, 1 SSc-ILD), eight of them had undergone SLB prior to AE and in two of them the AE occurred shortly after the SLB and thus was thought to be induced by the SLB (133). It is suggested that a higher risk of AE is associated with a lower baseline DLCO and lower vital capacity and that the risk of AE is increased in patients who exhibited AE signs and symptoms prior to biopsy (134). If a biopsy is to be performed, the preferred method is video-assisted thoracoscopic surgery (VATS), rather than open-lung biopsy (75). Nowadays, in some centers, a new diagnostic method has become available, namely transbronchial cryobiopsy (TBCx), in which the sample size is much larger than in standard TBBs (135). One study compared TBCx and SLB in a series of 117 IPF-patients (58 TBCx, 59 SLB) with fibrotic and not definite UIP on HRCT and observed a major increase in diagnostic confidence after TBCx, similar to SLB, highlighting the usefulness of TBCx in the diagnostics (136). The mortality from cryobiopsy is significantly less (0.1%) (137) than that for SLB (130) and moreover, a decreased length of hospital stay and less adverse events such as persistent fever, prolonged air leak and acute exacerbation of ILD have been reported in TBCx compared to SLB (137). Some risks are, however, attached also to TBCx, the most common of which are pneumothorax (10-20%) and mild-to-moderate bleeding (412%) (135,137) THE COURSE OF THE DISEASE Disease progression Retrospective studies on the natural course of RA-ILD have revealed a variable clinical course, with some patients having a stable or slowly progressive course for a decade, and others having a fulminant course with death less than six months after the onset of respiratory symptoms (11). For example, in a small study cohort (28 RA-ILD patients) of Fujii et al, 57% of patients had no change in HRCT after a 1 year of follow-up, whereas nine patients had progressed and three improved (138). In a study of initially asymptomatic early RA-ILD patients, over 50% had a radiological disease progression on HRCT during a mean follow-up of 1.5 years (6). In another study, 34% of patients progressed radiographically over the 2-year follow-up and the progressive disease was more commonly associated with reticular abnormalities than with ground glass changes, suggesting that UIP patients have a greater risk of progression (11). In a retrospective study of 84 RA-UIP patients monitored for 33 months, Song et al. observed 44% of the patients

46 20 being stable, 33% progressing, 17% deteriorating with an AE and 6% improving during the follow-up (139). Reduced DLCO, presence of bibasilar crackles and the extent and distribution of HRCT changes have been reported to associate with the disease progression (defined as >15% fall in DLCO) with evidence of increasing fibrosis changes on HRCT or ILD-related death (11). In a very recent retrospective study of 167 RA-ILD patients, the UIP pattern, unlike NSIP, was a risk factor for DLCO worsening. Moreover, a lower baseline DLCO, lower baseline FVC, and higher changes in PFT during the first 6 months increased the risk for progression (140) RA-ILD and prognosis The available studies demonstrate that patients with RA-ILD have a 3-fold risk of death compared to RA patients without ILD (98). While the overall mortality in RA has declined, the numbers of deaths due to RA-ILD have increased, especially in women and in older aged patients (10). Hakala et al. showed that hospitalization was not that common among RA-ILD patients, but those who were hospitalized because of ILD had a median survival of only 3.5 years (141). The results of studies investigating survival have, however, been variable and mostly conducted with several types of CTD, and not solely with RA. Several studies have previously observed better survival outcomes in CTD-ILD when compared to the idiopathic ILD subtypes (110,142,143), which might be a consequence of lesser fibroblastic foci observed in CTD-UIP compared to IPF (142). Often studies comparing the survival in CTD-ILD vs. IIP have included patients with SSc, polymyositis and dermatomyositis, in which the most common ILD subtype is NSIP, unlike in RA-ILD, which may influence the results. Some investigators have proposed the prognosis of CTD-ILD to be even worse than that of IPF/IIP after adjusting for age, gender, smoking habit and exposure to oral corticosteroids (144), or adjusting just for age (145). RA-ILD may be a possible exception to the putative hypothesis of better prognosis of CTD-ILDs compared to IIPs having revealed similar survival rates than IPF, i.e. 2-3 years (4,37), especially in RA-UIP cases (95,110,146). The study of Song et al., however, calculated a median survival of 53 months in RA-UIP versus 41 months in IPF (139). Similarly, in the study of Rajasekaran et al. for example, RA-ILD had a higher median survival (60 vs. 27 months) when compared to identical patterns of IIPs (147) and some studies have described an even better prognosis, with median survival of approximately 6-8 years in patients with RA-ILD (143,148). Overall, the question concerning the prognosis of RA-ILD is controversial and therefore predicting of lifetime of an individual patient is challenging. On average it can be concluded that the median survival ranges between 2.6 and 8 years from the time of RA-ILD diagnosis (4,45,95,146,148) Acute exacerbations AE has been previously defined as an acute, clinically significant, respiratory deterioration of unidentifiable cause, and has been mostly linked to IPF (32). Recently, an updated definition of AE in IPF was proposed and the exclusion of specific triggers of AE, such as infection, aspiration or drug toxicity, is no longer required (149). AE is characterized by newly developed bilateral GGO and/or consolidations on chest X-ray or CT scans and worsening / development of dyspnoea (150). AE has also been detected in other ILDs than IPF, including insip, chronic hypersensitivity pneumonitis and CTD-ILD, among which AE is most commonly present in RA-ILD (133,150,151). The one-year cumulative incidence of AE in IPF has been reported to be 9-16% depending on whether infections were included or excluded (152), whereas that of RA-ILD was 2.8% in a study by Hozumi et al (116). In the same study, AE was associated with a high mortality up to 64%, similarly as in IPF, and an older age at ILD diagnosis, the

47 21 UIP pattern on HRCT and the use of MTX correlated with the development of AE (116). Suda et al. investigated 83 CTD-ILD patients (including 25 RA-ILD); five RA-ILD patients developed AE with a one-year incidence of 2.6% and 75% mortality (151) ASSESSMENT OF PROGNOSIS Most studies investigating the prognostic factors in ILD have focused on IPF, but some studies on RA-ILD have also been published and reviewed quite recently (12). Several factors have been proposed to predict disease progression and survival which can be categorized as patient-specific, RA-specific and ILD-specific variables (Figure 5) Radiological predictors of mortality A retrospective study of 144 RA-ILD, whose diagnosis was based on either HRCT (n=120) or lung biopsy (n=24), revealed that patients with DAD had the worst prognosis following UIP-patients, whose survival was significantly shorter than that of NSIP-patients (148). Several other investigators have demonstrated that the UIP-pattern on HRCT is associated with worse survival than the NSIP- or non-uip-patterns (73,95,153). The extent of the disease has also been shown to be related with increased mortality in RA-ILD with extensive disease (>20% of lung affected) associating with a two-fold relative risk of dying compared with limited disease (73,154,155). Similar results have been observed in SSc-ILD (156), IPF (157,158) and fnsip (158) as well. In a study conducted by Walsh et al., 168 patients with CTD-ILD, including 39 patients with RA-ILD were examined, and the extents of honeycombing and severity of traction bronchiectasis were associated with increased mortality (159). The presence and extent of traction bronchiectasis and honeycombing, as well as the presence of reticulation were all significantly associated with worse survival time on bivariate survival analysis of Kim et al. investigating 82 patients with RA-ILD (95). Figure 5. Predictors of mortality in RA-ILD in the unadjusted analysis. Original figure adapted and modified from review article of Assayag et al (12). RA = rheumatoid arthritis; ILD = interstitial lung disease; ESR = erythrocyte sedimentation rate; LDH = lactate dehydrogenase; HAQ DI = Health-assessment questionnaire disability index score; DLCO % = diffusion capacity to carbon monoxide, percent of the predicted value; FVC = forced vital capacity; UIP = usual interstitial pneumonia.

48 Histopathological predictors of mortality Yousem and colleagues were the first to report that among RA-ILD patients, those with a histologic pattern UIP in surgical lung biopsy specimens had the worst prognosis (160). Subsequently, several studies have demonstrated that patients with a histological UIP pattern have a worse prognosis as compared to those with other subtypes (96, ). Contradictory results have, however, also been reported. Yoshinouchi et al. (164) studied 16 patients with RA-ILD (7 UIP, 7 NSIP, 2 UIP/NSIP hybrid pattern) and in contrast to the other studies, survival in the NSIP group seemed worse as more deaths occurred in patients with NSIP, although 2/3 NSIP deaths were from non-respiratory causes. Solomon et al. investigated 48 biopsy-proven RA-ILD cases and found no difference in survival between patients with UIP and NSIP patients (146). In the same study, the NSIP group was further divided into fnsip and cnsip. After pooling unclassified ILD, UIP and fnsip together, these so-called fibrotic ILDs (N=23) had clearly shorter survival times than the nonfibrotic (bronchiolitis, DAD, DIP, LIP, cnsip, OP) patients (N=25) and thus the conclusion was that the presence of any fibrosis on biopsy predicted mortality (146). A subgroup analysis of the 168 CTD-ILD-patients was performed by Walsh et al on 51 biopsy-proven patients, whose histopathological diagnoses were compared to HRCT data (159). The 51 patients were divided into four groups based on the concordance of the histopathological and radiological diagnoses. It revealed that the patients with discordant UIP had a better prognosis than concordant UIP-patients but worse than that of those with fnsips (159) Pulmonary function tests, 6MWT and prognosis When the fibrotic subtypes of IIPs were examined, it has been demonstrated that pulmonary physiology could be an even stronger predictor of mortality than the histopathologic pattern (165) and that in IPF patients, changes in FVC % predicted and DLCO % predicted have associated with mortality (166,167) especially in patients with a desaturation in 6MWT (166). In RA-ILD studies, the baseline DLCO has been considered as an independent risk factor for death in some univariate/bivariate and multivariate models (95,153,159). Moreover, the change in DLCO has been also identified as a prognostic factor (139,153). Baseline FVC and the change in this parameter have also been shown to be prognostic factors (95,139,153). Reduced walking distance and oxygen desaturation below 88% during a 6MWT are associated with mortality in IPF (32,168,169), but this has not been examined in patients with RA-ILD Patient- and RA-related predictors of mortality Older age (37,146,170) and male sex (95,171) have been detected as risk factors for mortality in several studies. In addition, in one cohort the risk of death was almost doubled in patients with low socio-economic status, although this finding did not quite reach statistical significance (37). Some factors reflecting the RA severity have also been proposed to associate with worse survival. The severity can be assessed in several different ways, but for example high baseline visual analogue pain scale (VAS), increased ESR, DAS and HAQ-DI have been detected to correlate with worse survival (37,170,171). Furthermore, the presence of RF and high lactate dehydrogenase (LDH) pointed towards a poorer prognosis (154).

49 THE RISK PREDICTION MODELS IN ILDS Since the course of the disease is highly variable among ILD patients and no specific biomarker has been found that would predict patient s survival, several models based on combinations of the above-mentioned separate factors have been developed recently, most of which were developed for IPF, but some of them have been expanded to all ILDs, even though not yet validated in RA-ILD patients (Table 8). In the Clinical-Radiologic-Physiologic scoring system (CRP) age, smoking history, finger clubbing, arterial blood oxygen during exercise, the percent predicted total lung capacity (TLC), and the extent of profusion of interstitial opacities and evidence of pulmonary hypertension on chest X-ray were used to construct a score (maximum score 100) (172). The composite physiologic index (CPI), published 2 years after the previous system, displayed some important advantages, since it contained only PFT and gas transfer values but omitted radiological scoring or exercise testing. The formula (CPI= 91 - (0.65 x DLCO% predicted) (0.53 x FVC % predicted) + (0.34 x forced expiratory volume in 1 second (FEV1) % predicted) was derived by fitting PFT results against disease extent on CT in a regression model and this was shown to correlate better with disease extent than the individual PFT test (173). The model developed by Goh et al was designed for survival prediction of SSc-ILD patients (174). The model integrated PFT results and HRCT findings, grading the disease extent on HRCT as minimal (defined as clearly <20%) or severe (defined as clearly >20%) and using a FVC threshold of 70% in indeterminate cases to categorize the patients either as having a limited or extensive disease. This kind of integrated staging had a greater prognostic value than either HRCT or FVC thresholds on their own (174). In the model by du Bois et al. age, respiratory hospitalisation, FVC % predicted, and 24week change in FVC were combined (175), whereas the Risk Stratification score (ROSE) by Mura et al. was based on combining the Medical Research Council Dyspnoea Score (MRCDS), 6MWT and the CPI (176). Probably the model that has obtained widest clinical utilisation is the GAP index. This was introduced by Ley et al. in 2012 and it combines gender (G), age (A) and two lung physiology variables (P), i.e. FVC and DLCO, into a multidimensional index and staging system with three stages (I-III) proposing 1-year mortality of 6, 16 and 39%, respectively (177) (Table 9). The GAP model has also been utilized in the prognosis of other chronic ILDs in addition to IPF. The modified model was named as ILD-GAP, with the assumption that patients with chronic hypersensitivity pneumonitis, insip and CTD-ILD enjoyed a better survival than those suffering from IPF (178). The cohort, in which the expanded ILDGAP model was applied, included the following patients: 307 IPF, 206 chronic hypersensitivity pneumonitis, 45 insip, 173 unclassifiable ILD and 281 CTD-ILD. The number of RA-ILD patients in the CTD-ILD group was not reported and it is unclear which of the prognostic models - GAP or ILD-GAP - would be better suited for RA-ILD, given the high proportion of UIP patients in RA-ILD (96).

50 24 Table 8. Prediction models in ILD. Adapted and modified from the PhD Thesis of Charlotte Hyldgaard 2015 (19). Study (first Name of Disease The factors that are included No. of author, year) the model in the model patients King 2001 CRP IPF Age, smoking, finger clubbing, 238 No arterial blood oxygen during exercise, TLC, chest X-ray Wells 2003 CPI IPF FVC, FEV1, DLCO 212 Yes Goh SSc-ILD FVC, HRCT 215 Yes Du Bois IPF Age, resp. hosp., FVC, 24-week 1099 No change in FVC Mura 2012 ROSE IPF MRCDS, 6MWT, CPI 70 No Ley 2012 GAP IPF Age, FVC, DLCO 228 Yes Validated Ryerson 2013 ILD-GAP ILD Age, FVC, DLCO 1012 In IPF CRP = Clinical-Radiologic-Physiologic scoring system; CPI = composite physiologic index; ROSE = Risk Stratification score; GAP = gender, age, physiologic variables; IPF = idiopathic pulmonary fibrosis; ILD = interstitial lung disease; SSc = systemic sclerosis; TLC = total lung capacity, FVC = forced vital capacity; FEV1 forced expiratory volume in 1 second, DLCO = diffusion capacity to carbon monoxide; HRCT = highresolution computed tomography; Resp.hosp = respiratory hospitalization; MRCDS = Medical Research Council Dyspnoea Score; 6MWT = Six-minute walk test. Table 9. The GAP (gender, age and physiology) index and staging system. Adapted and modified from the original GAP publication of Ley et al (177). The maximum possible point score is 8 and the patients are further divided to three stages according to their total scores. The model-predicted 1-, 2-, and 3-year mortality is shown by stage. Predictor Details Points 0 1 G Gender Female Male A Age >65 P Physiology FVC % predicted > < DLCO % predicted > Cannot perform Total possible points 8 Stage I II III Points Mortality years FVC = forced vital capacity; DLCO = diffusion capacity to carbon monoxide.

51 COMORBIDITIES Comorbidities in RA-ILD Comorbidities in RA-ILD have not been previously studied in detail. A Danish cohort of 679 RA-ILD patients reported that the burden of comorbidity assessed by the Charlson Comorbidity Index was higher in the RA-ILD group compared to their RA counterparts without ILD. They also reported that ischemic heart disease, congestive heart failure and diabetes were more frequent in the RA-ILD group, although the difference in ischemic heart disease frequency was rather small and the authors considered that it did not account for the increased mortality observed in the RA-ILD group (179) Comorbidities in RA Even though studies concerning RA-ILD comorbidities are infrequent, many studies have investigated the co-existing diseases in RA population. An elevated risk has been reported for atherosclerosis, and cardiovascular disease, of which the prevalence has been estimated to be times higher in RA patients compared to the general population. The risk for myocardial infarction and for stroke is approximately doubled in RA. The prevalence of diabetes seems to be increased, whereas the data are somewhat inconsistent as to whether the prevalence of hypertension is higher in RA than in the general population. Patients with RA are also at an increased risk for cardiac heart failure. Moreover, thyroid dysfunction and depressive symptoms may be as many as 2 to 3 times more common in RA as in the general population (180) CAUSES OF DEATH ILD is the most common pulmonary cause of death, and the second commonest overall cause of death in RA (8,181,182). The causes of deaths of the patients with RA-ILD have not been previously systematically studied, although some data have been published. In the study of Nakamura et al., nine of the 54 biopsy-proven cases had died; in 7 out of the 9 cases, the cause of death was respiratory failure due to disease progression (161). In the study of Koduri et al. in 28 of the 39 deceased RA-ILD patients, the cause of death was attributed to RA-ILD, while the remainder recorded causes of death were bronchopneumonia (n=4), ischemic heart disease (n=3), heart failure (n=2), pulmonary embolism (n=2), cerebrovascular disease (n=2) and miscellaneous reasons (n=5) (37) TREATMENT Whom and how to treat? The management of ILD in patients of RA is challenging due to the lack of randomized controlled trials. Treatment is typically initiated in patients who suffer from respiratory symptoms and show a progressive course of the disease, whereas asymptomatic patients are often simply monitored (119). A 10% decline in FVC in a RA-ILD patient suggests a higher risk of mortality and may help in decisions to treat (153). On the other hand, the treatment is generally focused on controlling the systemic disease even in the cases with a stable lung disease (102). An adaptation of the unclassifiable ILDs categorization by the disease behavior (20) may be useful in deciding the monitoring strategy and treatment choices (Table 10) (22). Several therapeutic agents have been suggested, but no randomized controlled trials have been performed to guide the clinical practice (103). There is rather limited evidence of

52 26 specific therapeutic agents; this is mainly based on case reports, case series and retrospective cohorts. Another concern is the possible role of some agents in the progression and/or exacerbations of ILD (17,183). It remains unclear whether the recent onset ILD or the exacerbation of pre-existing ILD observed during a specific therapy reflects a pulmonary side effect of the drug or whether these events would have occurred regardless of the medication (183). Usually it is advised to perform functional and the radiological evaluation of the lung before starting a new RA treatment ( ). In Finland, the recommended treatment of newly diagnosed RA is the so-called FIN-RACo-strategy (Finnish rheumatoid arthritis combination therapy), which includes MTX, sulfasalazine, hydroxychloroquine plus prednisolone and aims at rapid remission (15). Traditionally, in patients with the development of ILD, the MTX has been discontinued, but recent reviews have recommended that the decision on whether or not to discontinue drugs should be considered on a case-by-case basis and no strict guidelines are given (22,183). There is evidence that the response to therapy varies in different RA-ILD subtypes, with OP pattern typically showing rapid responses to corticosteroid therapy often with a complete recovery (187) and NSIP pattern having more favourable responses to therapy than the UIP pattern (22,48,119). For example, in the retrospective study of Song et al., 41% of RA-UIP patients were treated with high-dose corticosteroids combined with azathioprine, cyclophosphamide or cyclosporine due to disease progression or poor initial lung function. Fifty percent of the patients improved or had stable lung function but no difference was observed in outcome between the treated and untreated groups (139). Table 10. Different monitoring and treatment consideration strategies according to disease behaviour. Original table adapted from the review article of Lake et al (22). Clinical behaviour Treatment and goal Monitoring strategy Potentially reversible with risk of irreversible disease (e.g. drugrelated lung disease in RA) Remove cause, treat to obtain a response to reverse changes. Short-term (3-6 months) observation to confirm disease regression, or occasionally need for palliation. Reversible disease with risk of progression (e.g. some RA-NSIP, RA-OP) Stable with residual disease (e.g. some RA-NSIP, some RA-UIP) Progressive, irreversible disease with potential for stabilization (e.g. some RA-NSIP, some RA- UIP) Treat to initially achieve response and then rationalize longer term therapy. No treatment if stable, aiming to maintain status. Consider treatment trial to stabilize. Short-term observation to confirm treatment response. Long-term observation to ensure that gains are preserved. Long-term observation to assess disease course. Long-term observation to assess disease course. Progressive, irreversible disease despite therapy (e.g. RA-DAD, most RA-UIP, some RA-NSIP) In the absence of contraindications, consider treatment trial in selected patients to slow progression. Short (DAD) or long-term observation to assess disease course, and need for transplant or effective palliation. RA = rheumatoid arthritis; NSIP = nonspecific interstitial pneumonia; OP = organizing pneumonia; UIP = usual interstitial pneumonia; DAD = diffuse alveolar damage.

53 Immunosuppressive agents The treatment of RA-ILD has been typically empirical with corticosteroids being used as first-line agents. Patients who responded to steroids often had immunosuppressive drugs such as azathioprine added to steroids (103). Prolonged durations of even moderate doses of corticosteroids can, however, cause variable adverse effects, for example impairment of muscle strength (188). Moreover, the risk for serious infection was recently determined in 181 RA-ILD patients. The risk was highest in the first year after ILD diagnosis and among patients receiving 10mg or more daily prednisone, with an overall infection rate of 7.4 per 100 person-years which is similar to that of RA patients without ILD (189). The cases with OP and NSIP patterns respond better to corticosteroids but the effect on RA-UIP patients is unclear (48). Cyclophosphamide treatment was thought to be beneficial in SSc-ILD (190), but a metaanalysis concluded that it did not induce clinically significant improvements of PFT in SScILD patients (191). No randomized controlled trial has been performed to clarify its benefits in RA-ILD (103). Mycophenolate mofetil (MMF) has been shown to be safe and effective in some patients with CTD-ILDs. In one study of 125 MMF- treated patients with CTD-ILD, including 18 RA-ILD-patients, a slight improvement in FVC was observed in 18 of the total group, with less improvement seen in UIP patterned patients than in the others (192). A prospective study of 10 CTD-ILD-patients, including 3 RA-ILD cases, treated with MMF revealed that all ten patients experienced an improvement in their respiratory symptoms, quality of life and activity levels. Two out of 8 subjects with repeated HRCT exhibited a radiological improvement, 6 stabilized and none worsened, whereas one out of 9 patients showed worsening in PFT, 3 with improvement and 5 stabilized (193) Synthetic disease modifying antirheumatic drugs (DMARDs) The addition of DMARD (MTX, leflunomide (LEF) or azathioprine in this study) to lowered dose of prednisone was associated with an improvement in baseline FVC in a study of 40 RA-ILD-patients (194). In a subgroup analysis of the above-mentioned study, those with a lower fibrosis score on CT improved and those with a UIP pattern had a higher mean fibrosis score (194). The treatment of RA-ILD with DMARDs is, however, complicated due to the possibilities of the development of drug-related pneumonitis or worsening of the existing ILD linked in several DMARDs (103). The recommended first-line DMARD for RA is MTX, which has been shown to be associated with the progression of preclinical ILD and an increased risk of pneumonitis in some studies (22,103). Still, some researchers think that it is not conclusively evidenced, that MTX induces / exacerbates the underlying RA-ILD or leads to a greater risk of pulmonary death (103) and a decision on whether or not to avoid the drug needs consideration of both the joint and lung diseases (22). Leflunomide has also been associated with rapid onset pneumonitis, ILD and nodule development with an almost two-fold risk of ILD compared to those not receiving LEF (195). A case report from Finland described 5 patients with newly developed severe ILDs after combination therapy with LEF and MTX (196). In some case-reports, positive responses have been described with the use of cyclosporine in RA-ILD patients (197), but in the study of Tokano et al., no persisting response to cyclosporine was found in four steroid-resistant RA-ILD cases, even though positive lasting responses were observed in other CTD-ILDs (198) Biologic agents A list of biologic therapies for treatment of RA is shown in table 11. Concern about the respiratory safety of anti-tnf agents first appeared after the publication of a few rapid,

54 28 fatal exacerbations of previously asymptomatic RA-ILD after the introduction of infliximab (199) and subsequently cases of induction or exacerbation of RA-ILD have been repeatedly reported (200,201). However, in contrast, some case-reports have claimed to have detected stabilization or improvement of RA-ILD after the administration of infliximab (202,203). Etanercept has also been linked to exacerbation of pre-existing lung disease in patients with RA (204), but conversely, in a study of 367 patients with RA-ILD treated with either anti- TNF agents (n=299) or traditional RA treatments (n=68), no difference was found in mortality (170). Detorakis et al. examined prospectively 82 RA patients, 42 with and 40 without RA-ILD, treated with anti-tnf agents (68 infliximab, 10 etanercept, 4 adalimumab). Control groups consisted of 44 patients with pre-existing RA-ILD and 44 patients without RA-ILD, treated with non-biologic DMARDs (68 MTX alone, 20 MTX + hydroxychloroquine). There were no episodes of newly emerged ILDs or ILD exacerbation in the anti-tnf- treatment group and moreover, this group displayed an improvement in bronchial wall thickening and air trapping over time (205). Little information is available regarding the safety and benefits of biological DMARDs other than the TNF-alpha inhibitors in RA-ILD patients, including anakinra, tocilizumab, abatacept and rituximab. Tocilizumab improved RA-ILD in one single case report (206), whereas other reports have addressed ILD occurrence or exacerbation with tocilizumab therapy (207,208). One case series with abatacept claimed that there was stabilization of lung function in RA patients who had developed ILD on anti-tnf agents (209), but another study reported worsening of HRCT findings in a single RA-ILD patient (210). Weinblatt et al. examined the pooled data from eight clinical trials of intravenous abatacept therapy for RA, and discovered a low incidence rate of 0.09 for ILD development. However, the study did not answer the question of abatacept s safety in RA patients with pre-existing ILD (211). There is no data regarding to non-infective pulmonary complications, nor potential therapeutical / beneficial role in ILD of anakinra. Rituximab (RTX) is a monoclonal antibody against B-cell marker CD20 (119), often considered as a rescue therapy in RA-ILD, although based on limited evidence (185). Some reports are available regarding the use of RTX in RA-ILD patients. Keir et al. claimed that some of the studied 33 CTD-ILD patients (two with RA-ILD) experienced an improvement in FVC and stability of DLCO in the 6-12 months following RTX treatment, in contrast to the clear decline in both parameters evident prior to rituximab (212). In a 48-week pilotstudy of RTX in RA-ILD patients, only one out of 10 patients enjoyed a respiratory improvement, five remained stable, one deteriorated and 3 were unable to complete the study with one of these dying because of pneumonia / acute respiratory distress syndrome (ARDS) (213). Quite recently, experiences over 10 years from a single centre were published demonstrating a progression in 32%, an improvement in 16% and a stable disease in 52% of the 56 RTX-treated RA-ILD patients (214). Of those who deteriorated/died, 79% had severe ILD before RTX treatment, suggesting that the drug had not been contributory to the deterioration (214). No prospective studies are available identifying predictors of progression of RA-ILD in patients treated with biologic agents and thus, there are no evidence-based guidelines to help clinicians to decide which patients to treat or not to treat with biological therapy. The ongoing clinical trials (uploaded on ) on biologic drugs are listed in table 12. It has been hypothesized that those patients with severe of progressive RA-ILD would probably have a higher risk of developing drug-induced AEs (200). It is crucial to gain a better understanding of factors that predict poor outcomes of biologic therapy. One suggested treatment strategy is shown in figure 6.

55 29 Table 11. Available biologic therapies for rheumatoid arthritis. Substance Brand name Main effect Application form Infliximab Remicade, Remsima, Inflectra TNF inhibition iv Etanercept Enbrel TNF inhibition sc Adalimumab Humira TNF inhibition sc Certolizumab Cimzia TNF inhibition sc Golimumab Simponi TNF inhibition sc Tocilizumab Roactemra IL-6 inhibition iv Anakinra Kineret IL-1 inhibition sc Rituximab Mabthera, Ritemvia B-cell depletion iv Abatacept Orencia Inhibition of the iv T-cell activation TNF = Tumor necrosis factor; IL = interleukin; iv = intravenous; sc = subcutaneous. Figure 6. Suggested treatment options for RA-ILD. Lung transplantation is recommended to be considered for non-responding progressive RA-ILD patients. Original figure adapted and modified from the review article of Roblez-Perez and Molina-Molina (185). ILD = interstitial lung disease; RA = rheumatoid arthritis; RA-ILD = rheumatoid arthritisassociated interstitial lung disease; UIP = usual interstitial pneumonia; OP = organizing pneumonia; NSIP = nonspecific interstitial pneumonia Pulmonary rehabilitation In IPF patients, pulmonary rehabilitation has been shown to improve dyspnea and increase exercise capacity as well as the quality of life at least in the short term (9). In a recent randomized controlled study, the patients with IPF or asbestosis showed more improvement in the 6MWT, in symptoms and in quality of life measured by different

56 30 questionnaires, than those with patients with CTD-ILD. However, the CTD-ILD group was rather small (23 patients out of the total 142 ILD patients) and included only 8 RA-ILD patients (215). Thus, the potential of pulmonary rehabilitation in RA-ILD is still undefined and can be limited by the functional impairment due to the joint disease, which may require RA-specific rehabilitation protocols (103) Lung transplant Patients with progressive disease should be evaluated for LTx (186). There is limited data concerning LTx in CTD-ILD with most of them being studied in patients with SSc-ILD. In a retrospective study of Yazdani et al., ten patients with RA-ILD, 53 with IPF and 17 with SSc-ILD underwent LTx. Groups were matched for age and transplant year and no statistically significant differences were observed between the cumulative survival rates of different disease entities (216). Moreover, the quality of life of RA-ILD patients increased significantly measured by all three different scores in use (216). Another study compared the survival and outcomes after LTx between IPF and non-scleroderma CTD-ILD, including 68 (24.7%) RA-ILD patients, and found no significant differences in survival, acute or chronic rejection, or extrapulmonary organ dysfunction (217). A study with a wide spectrum of different CTD patients, including 36 RA-ILD patients, found that the cumulative survival in patients with CTD-ILD was like that of IPF patients (218). Thus, it seems that LTx would be beneficial in selected patients, but it may be contraindicated because of age, comorbidities or poor functional ability (186). In many organ transplant centers, the presence of CTD is considered as a contraindication for lung transplant. In Finland, however, these patients have been accepted for transplantation if the CTD is well managed, has been stable for years and there are no other organ manifestations of the CTD that would rule out the surgery (219). The common indications and contraindications for transplantation are applied regardless of the presence or absence of CTD. Of the 282 LTx:s performed in Finland, the reason for transplant was IIP in 97 patients of whom 12 suffered from CTD-ILD and two from RA-ILD (219) Antifibrotic drugs There is no data concerning antifibrotic agents, such as pirfenidone and nintedanib, for the treatment of RA-ILD. Ongoing clinical trials (uploaded on ) on antifibrotics for RA-ILD are listed in Table Treatment of RA-ILD exacerbation Clinical data on treatment for non-ipf exacerbation is lacking (133). There are no controlled trials investigating the treatment of AE in IPF or in other ILDs. Thus, the treatment is empirical. Patients with AE-IPF are often treated with corticosteroids, either a pulse dose or lower doses of oral prednisone. Antibiotics are commonly used, since bacterial infection is difficult to exclude (220). In a retrospective analysis of 10 patients (3 RA-ILD, 1 SSc-ILD, 6 insip), nine patients received broad-spectrum antibiotics and high-dose systemic corticosteroid therapy (6 i.v. pulse therapy and three 1mg/kg/d dosage). Six patients needed mechanical ventilation and all of them died. Four patients with idiopathic NSIP survived, when all the CTD-ILD patients died (133) Other treatments Smoking is a potential risk factor for developing RA-ILD and COPD with or without emphysematous lung damage. It can also affect the severity of joint disease and therefore all patients with RA should be encouraged to stop smoking and provided with the appropriate advice and assistance about cessation of smoking. In the patients in whom the

57 31 ILD has already developed, smoking cessation is crucial. Annual vaccinations for influenza, as well as pneumococcal vaccination are recommended for RA patients on immunosuppressive therapy as well as in those with a chronic lung disease (9). Some physicians also recommend prophylaxis against Pneumocystis jirovecii for all patients receiving immunosuppressive therapy (98) Palliative care Cough and dyspnea are common symptoms of ILDs and can cause tremendous suffering and reduce the patient s quality of life. Opioids may be beneficial and the treatment of possible concomitant gastro-esophageal reflux should be considered (220). Supplemental oxygen therapy may be required in advanced disease, although it is not known, whether oxygen alters the course of the disease (9). Psychological / psychosocial issues need to be addressed and different kinds of support e.g. psychotherapy, counselling or pharmacological therapy, may be beneficial (220).

58 32 Table 12. Ongoing trials of biologic and antifibrotic drugs on RA-ILD patients according to ClinicalTrials.gov (uploaded on ). Name of the study ClinicalTrials Gov identifier Phase II study of pirfenidone in patients with RA-ILD Phase Disease Status Duration (weeks) Estimated enrolment Intervention Primary outcome NCT RA-ILD Recruiting Pirfenidone, 2403mg/day, three times /d variable Disease progression FVC>10% Efficacy and Safety of Nintedanib in Patients with Progressive Fibrosing Interstitial Lung Disease (PF-ILD) * NCT PF-ILD Recruiting Exp. Nintedanib Comp.: placebo Annual rate of decline in FVC AbatacePt in RA-ILD (APRIL) NCT RA-ILD Not yet recruiting I.v. Abatacept approximately 10mg/kg fortnightly for the first 4 weeks, then every 4 weeks for a total of 20 weeks Disease progression FVC>10% Evaluation of Efficacy and Safety of Rituximab With Mycophenolate Mofetil in Patients With Interstitial Lung Diseases (EvER- ILD) NCT CTD-NSIP, IPAF-NSIP, insip Recruiting Exp. MMF + RTX Comp.: MMF+placebo Change in FVC in % of predicted * Includes CTD-ILD patients who have medically indicated need for change in CTD medication and have not yet been initiated of new therapy or withdrawal of therapy for CTD within 6 weeks prior to Visit 1). PF = progressive fibrosis; MMF = mycophenolate mofetil; RTX = rituximab; IPAF = interstitial pneumonitis with autoimmune features; NSIP = nonspecific interstitial pneumonia; Exp = experiment, Comp. = comparative group; FVC = forced vital capacity; RA-ILD = rheumatoid arthritis-associated interstitial lung disease; CTD = connective tissue diseases; insip = idiopathic NSIP.

59 33 3 Aims of the Study The aim of this thesis was to investigate a cohort of RA-ILD in patients treated between the year 2000 and the end of 2014 in the Kuopio University Hospital (KUH) pulmonology clinic and to evaluate the course of the disease. The specific aims were: 1. To investigate the numbers and subtypes of the patients with RA-ILD treated in KUH during the years , to evaluate the course of the disease, medication, pulmonary function test results, survival, comorbidities and causes of death and to compare these parameters between UIP and non-uip cases. 2. To determine the applicability of CPI, GAP and ILD-GAP scores for predicting the prognosis of the patients with RA-ILD and to examine the association between individual PFT and demographic factors with the survival of the patients. 3. To evaluate the HRCT findings in patients with RA-ILD and to compare the presence and extent of different radiologic findings in different RA-ILD subtypes, as well as to identify associations between radiological findings and clinical factors, survival and pulmonary function tests.

60 34 4 Material and Methods 4.1 DATA SOURCES AND PATIENT SELECTION The study cohort consisted of patients treated in the KUH pulmonology in-patient or outpatient clinic between and The patients were identified from the database of KUH using two International Classification of Diseases (ICD-10) codes, namely J84.X and M05.X/M06.X. From these patients, we only included those subjects that had been examined or treated in the pulmonology in-patient or out-patient clinic between and for any respiratory symptoms or any suspected pulmonary disease, thus omitting those RA patients with no symptoms or chest X-ray abnormalities. For the third study, another search was performed using the code J99.0*M05.1, but only one extra UIP patient fulfilling the study inclusion criteria was detected. The two first searches resulted in the identification of 1047 patients, and their patient records were evaluated to identify those patients suffering from clinically relevant RA-ILD (Figure 7). At baseline, the patients with ILD but without RA (i.e. patients with IIP, other CTDs or allergic alveolitis) and those with RA, whose visits to pulmonology clinic were because of some other lung diseases (such as asthma, COPD, obstructive sleep apnea) were excluded. We also excluded suspected but not confirmed RA-ILD patients, for whom HRCT, or some other comparable radiological examination capable of achieving a reliable analysis of the lung parenchyma were not available, as were those patients whose RA diagnosis was not certain according to the 1987 classification criteria (221), or who developed later mixed CTD- like symptoms. Another 38 patients were excluded subsequently after the evaluation by the radiologist and/or after a multidisciplinary discussion due to the very minor signs or nonspecific features for ILD, leaving a total of 59 RA-ILD (60 in study no. III) patients to be studied in detail and classified. Figure 7. The study protocol and the final re-categorization of the patients with RA-ILD in the first two studies. In the third study, one additional UIP patient was found. *additional 2 DAD findings included in OP group (n=1) and UIP group (n=1). RA = rheumatoid arthritis, ILD = interstitial lung disease, HRCT = high-resolution computed tomography, MDD = multidisciplinary discussion, UIP = usual interstitial pneumonia, NSIP = nonspecific interstitial pneumonia, OP = organizing pneumonia, DAD = diffuse alveolar damage.

61 GATHERING OF DEMOGRAPHIC INFORMATION (I, II, III) Clinical information was inclusively gathered from the patient records of KUH, primary health care centers and other hospitals using a specially designed form (Table 13). The gathered laboratory test results included RF and anti-nuclear antibody (ANA) titer, as well as ACPAs and arterial blood examples. ACPAs were not available for half of the patients. The results of PFT were gathered at baseline and, when available, during the follow-up at 6 months, 1 year, 2 year and subsequently annually, including also the most recent available results. The reference values of Viljanen were used when assessing PFT results (222). Any medication in use prior to ILD diagnosis was recorded, as was the lifelong medication used for RA. In addition, possible palliative therapy, e.g. opioids, for RA-ILD was recorded. Histological data (BAL, TBB, SLB, autopsy samples) also was collated, when available. The numbers of hospitalizations due to either respiratory problems (including infections, suspected drug reactions and suspected acute exacerbations) or cardiac problems like unstable angina pectoris, myocardial infarctions, arrhythmias and cardiac failures were collected. Table 13. Detailed list of demographic and other data that was gathered from the patient records of KUH, primary health care centers and other hospitals. Gathered data Details Date of birth Sex Occupation Family history of pulmonary fibrosis Smoking Duration, amount, pack-years, passive exposure Exposure to asbestos Radiation therapy of the thorax region Duration of RA Date of RA diagnosis RA-related surgery Date of the first visit due to ILD Comorbidities Use of oxygen Rehabilitation due to ILD Symptoms at baseline Cough, dyspnea at rest/exercise, hemoptysis, pain, fever Respiratory status findings at baseline Inspiratory crackles, finger clubbing, dyspnea at rest Baseline laboratory test RF, ANA titer, ACPAs, arterial blood examples Death certificate data Primary and immediate causes of death, place of death Pulmonary function test results DLCO and spirometry (FVC, FEV1, FEV1/FVC) Medication Any medication prior to ILD diagnosis, lifelong RA medication, opioids in palliative purposes BAL, TBB, SLB, autopsy samples Histological data Hospitalizations due to cardiac and respiratory reasons RA = rheumatoid arthritis; ILD = interstitial lung disease; RF = rheumatoid factor; ANA antinuclear antibodies; ACPAs = anticitrullinated protein antibodies; DLCO = diffusion capacity to carbon monoxide; FVC = forced vital capacity; FEV1 = forced expiratory volume in 1 second; BAL = bronchoalveolar lavage; TBB = transbronchial biopsy; SLB = surgical lung biopsy.

62 RADIOLOGICAL EVALUATION Re-classification of HRCTs (I, II, III) An experienced radiologist evaluated baseline HRCTs blinded to the demographic data and without consideration of the reports accompanying the original CT results. In study III the re-classification was performed independently by two radiologists. Radiological ILD recategorization was conducted according to the 2013 IIP classification (20) as UIP, NSIP, OP, DAD and unclassified subgroups. The radiological RA-UIP criteria were applied from those of IPF (32). Mainly patients with a definite UIP pattern were included in the UIP group, but three patients who displayed a slightly upper (n=2) or mid-lung (n=1) predominated distribution, were included after a multidisciplinary discussion. Patients with possible UIP, i.e. a subpleural and basal predominated reticular abnormality without honeycombing, were not included in the UIP group. RA-NSIP was defined as the predominance of GGO, possible visible subpleural sparing and possible fine reticulation with minor or no honeycombing. RA-OP was defined as single or multiple patchy consolidations. The unclassified subgroup consisted of those patients that did not fit the definition of any specific subtype. In addition to the baseline CT, the most recent HRCT was also evaluated in a similar manner from 33 patients who had a follow-up CT available. In those patients, the final subgroup was determined based on the analysis of both CTs Further interpretation of the CTs and the scoring system (study III) In addition to the radiological subgrouping performed by the two radiologists, the radiological findings were further assessed in detail in study III by the first radiologist, using a form designed for the study. The presence and the extent of the following findings were evaluated separately: GGO, reticulation, honeycombing, emphysema, consolidation, crazy-paving appearance, bronchiectasis, traction bronchiectasis, nodules, thickening of the bronchovascular bundle, cysts, mosaic attenuation, air trapping (when applicable), rounded atelectasis, architectural distortion, pleural plaques, pleural effusion and tumours. The definitions of these findings used in this study are those issued by the Fleischner Society (223). The most prominent observation was designated in each HRCT. Both lungs were divided into three zones. The upper zones were at or superior to the aortic arch, the middle zones were between the aortic arch and pulmonary veins and the lower zones were at or below the pulmonary veins. The extents of GGO, reticulation and honeycombing were semi-quantitatively graded on a scale from 0 to 4 as follows: 0 = finding absent, 1 = minor peripheral scattered changes, 2 = uniform peripheral or minor central changes, 3 = substantial peripheral changes that penetrated deeply into the lung parenchyma, 4 = very abundant peripheral and central changes. The total score of these three findings was obtained by summing the grades for all six zones i.e. the maximum score was 24. Emphysema, traction bronchiectasis, architectural distortion and pleural plaques were scored similarly, adding up the given grades in six zones, but now scored with a threepoint scale (0-3; 0 = absent, 1 = single scattered changes, 2 = larger single changes or several minor changes, 3 = uniform or substantial changes) resulting in a score ranging from STAGING SYSTEMS (II) CPI was calculated using the formula (173): CPI= 91 (0.65 x DLCO % predicted) (0.53 x FVC % predicted) + (0.34 x FEV1 % predicted). Requisite data was available from 51 patients, from whom the GAP / ILD-GAP scores were calculated by gender, age, FVC % predicted and DLCO % predicted following the division of patients into GAP / ILD-GAP

63 37 stages I and II as previously described (177,178). There were no stage III (or IV in ILD-GAP) patients in our study material (Figure 8). Figure 8. Flowchart of a patient s enrollment into the study showing the subdivision into the different GAP/ILD-GAP groups (Study II). RA-ILD = rheumatoid arthritis-associated interstitial lung disease; GAP = gender, age, physiologic variables. 4.5 STATISTICAL ANALYSIS The different statistical tests that were used in the study are listed in Table 14. In studies I and II, the distribution of the continuous variables was verified with Shapiro-Wilk test. If the distribution was normally divided, the comparison was made using an independent Ttest; otherwise Mann-Whitney U-test was applied. The chi-squared test or Fisher test, when appropriate, was used for comparison of the categorical variables. Gender, amounts of different RA-ILD subtypes, smoking habits, laboratory results, use of medications, symptoms, inspiratory crackles, comorbidities, use of oxygen, numbers of observed deaths and the presence of different radiological findings were calculated as percentages. Age at the time of RA-ILD, at diagnosis or at death, RA duration, PFT results, CPI score, hospitalizations and the extents of different radiological findings were expressed as mean ± SD. Survival analysis was done using the Kaplan-Meier method and survival curves were compared using the log-rank test. Survival time was calculated from the first visit to the pulmonology clinic due to ILD to the date of death or November 4, 2015 when the vital status was ascertained. Survival results are expressed as median (95% confidence interval). In the second study, the observed 1-, 2-, and 3-year mortality rates were calculated and these were supplemented with an estimate of the confidence interval by using the Wilson score. Next, the observed mortality and the risk of death predicted by the GAP / ILD-GAP models were compared using Hosmer-Lemeshow goodness-of-fit test. Finally, Cox regression analysis was used to identify factors that predicted mortality. In the third study, the associations between different radiological findings and survival were estimated using the Kaplan-Meier method and the univariate Cox regression analysis. The correlations between radiological finding scores and PFT as well as with other clinical factors were estimated using the Spearman rank correlation coefficient. Agreement between the radiologist s re-categorization was expressed as a kappa value (κ). Values of κ were considered as moderate and κ values as good agreement.

64 38 We considered a p-value <0.05 as statistically significant. All data was analyzed using IBM Statistics SPSS software, version Table 14. Summary of the statistical methods used in studies I, II and III Statistical test Use of the test Study in which used Chi-Square test Fisher test Shapiro-Wilk test Independent T-test Comparison of categorical variables Comparison of categorical variables The distribution of the continuous variables Comparison of normally distributed continuous variables Comparison of continuous variables, not normally distributed Mean survivals, survival comparison I, I, I, I, Prognostic factors for mortality Confidence interval estimation for the observed mortality Correlations between the extents of different radiological findings and PFT/clinical data Suitability of GAP/ILD-GAP models for predicting RA-ILD mortality Agreement between the radiologists recategorization II, III II Mann-Whitney U-test Kaplan-Meier method, log rank test Cox s regression analysis Wilson score Spearman rank correlation coefficient Hosmer-Lemeshow goodness-of-fit test Cohen s kappa 4.6 II, III II, III II II, III I, II I, II, III III II III ETHICAL CONSIDERATIONS In this retrospective study, most of the patients were deceased and no consents to participate were gathered due to the register-based nature of research in accordance with the Finnish legislation. The study protocol was approved by the Ethical Committee of Kuopio University Hospital (statement 17/2013). Organizational permission of Kuopio University Hospital was obtained as well as permissions from the Finnish National Institute for Health and Welfare (THL/1052/ /2013) and Statistics Finland (TK ), which enabled data collection from other hospitals, primary health care centers and death certificates. This study was conducted in compliance with the Declaration of Helsinki.

65 39 5 Results 5.1 PATIENT CHARACTERISTICS Demographics Fourteen patients were diagnosed by the end of year 2000, 19 between the years 2001 and 2007, and 27 between the years The highest amounts of new diagnoses were seen in the years 2012 and 2014, with 6 and 7 new RA-ILD diagnoses, respectively. Of the 60 HRCT-confirmed RA-ILD patients, 34 (56.7%) were male and 35 (59.3%) were current or former smokers. The mean age at diagnosis was 66 ± 11.2 years, ranging from 32 to 87 years. The mean follow-up time was 4.2 ± 5.2 years and in eight patients (13.6%) ILD diagnosis was made before RA diagnosis. In the cases where ILD followed the RA diagnosis, the time interval between the two diagnoses was 1 year in 13%, 3 years in 27% and 5 years in 40%, when the longest time interval between RA and ILD was 52 years. RF was positive in 84.5%, ANA in 19.6% and ACPAs were present in 68% of the patients from whom the data was available. The majority, i.e. 61.5% of the patients suffered from cough and almost as many (60.3%) from dyspnea Medication for RA Seventy-five percent of patients were receiving some medication for RA at the time of RAILD diagnosis and in 11 of them, the medication was changed or discontinued after the ILD diagnosis. In all of them, however, the ILD continued to progress. At any time, almost all (54/90%) had received glucocorticoids, over half (35/58.3%) MTX and 14/23.3% had received biologic drugs PFT Baseline spirometry was missing from five, and baseline DLCO from eight patients. Thirtyone patients (56.4%) had normal FVC at baseline. Twenty-five (48.1%) had a normal baseline DLCO. In 18 patients (34.6%), both the baseline DLCO and FVC were normal Radiological subtypes The re-categorization was based on both the baseline and, if available, on the follow-up HRCT, as well as on the rarely available histological samples. From the 60 individuals with RA-ILD, 36 (60%) revealed the UIP pattern, 8 (13%) NSIP, 7 (12%) OP and 8 (13%) unclassified pattern in their HRCTs (Figure 9). One case with DAD was observed, with a previously normal HRCT and then a rapidly progressing dyspnoea, severe hypoxemia as well as newly developed bilateral GGO changes in the HRCT assessment. In the follow-up, additional two DAD patterns were observed in patients with OP and UIP diagnoses prior to DAD GAP and ILD-GAP (II) Thirty-nine (76.5%) patients, from whom GAP/ILD-GAP scores could be calculated, belonged to stage I with the remaining categorized into stage II group. There were no stage III patients. Stage I patients were younger (p=0.024) and more likely to have never smoked (p=0.033) than the stage II patients. Baseline FVC, FEV 1 and DLCO were better preserved in stage I patients. RA-UIP patients were divided equally between both stages.

66 40 Figure 9. Final diagnoses after the re-classification of the HRCTs, with the consideration of the histological and clinical data. UIP = usual interstitial pneumonia; NSIP = nonspecific interstitial pneumonia; OP = organizing pneumonia; DAD =diffuse alveolar damage Comparison of the demographics in UIP and non-uip patients (I) Dyspnea (p= 0.022) and inspiratory crackles (p=0.007) were more often present with the RAUIP patients compared to non-uip individuals. No differences were observed between subgroups with respect to age, smoking, baseline PFT or RA, serology. No statistically significant differences were found either in the use of MTX or biologic drugs Comparisons within RA-UIP subgroup Among the RA-UIP patients, more males were former or current smokers, when only two (11%) were lifelong non-smokers, compared to 13 (81.3%) female non-smokers (p<0.001) (Table 9). All RA-UIP male were RF positive compared to 67% of the females (p=0.009). The male RA-UIP individuals had poorer baseline DLCO (p<0.001), FEV1 (p=0.006), CPI-points (p=0.005) and baseline GAP score (p=0.018) than the female counterparts (Table 15). When the UIP group of 36 patients was divided into subgroups of slowly (patient alive > 5 years after the diagnosis) and rapidly (patient dying <5 years after the diagnosis) progressing cases, the baseline CPI score was higher in the rapid group (p=0.037) and the mean number of hospitalizations due to cardiac illness was higher in the slow group (p= 0.043). The median survival was shorter (16 months vs. 152 months, p<0.001) and the number of deaths higher in the rapid group (p=0.024) (Table 15).

67 41 Table 15. The differences of clinical characteristics between genders and rapidly / slowly progressed RA-UIP cases. RA-UIP (n=36) MALE (n=20/55.6%) FEMALE (n=16/44.4%) P-value RAPID (=0-5Y) (n=12/40%) SLOW (>5Y) (n=18/60%) Age at dg 66.0 ± ± ± ± ± Age at death 74.1 ± ± ± ± ± Number of deaths 24 (66.7) 13 (65.0) 11 (68.8) F 12 (100.0) 11 (61.1) F P-value Smoking Non-smokers Ex-smokers Current smokers Serology RF positivity ANA positivity 15 (42.9) 14 (40.0) 6 (17.1) 30 (85.7) 6 (23.1) 2 (10.5) 12 (63.2) 5 (26.3) 20 (100.0) 4 (28.6) 13 (81.3) 2 (12.5) 1 (6.3) 10 (66.7) 2 (16.7) <0.001 F F F 5 (41.7) 4 (33.3) 3 (25.0) 11 (91.7) 2 (25.0) 8 (47.1) 8 (47.1) 1 (5.9) 14 (77.8) 2 (14.3) GAP points 2.2 ± ± ± ± ± CPI points 27.5 ± ± ± ± ± RA duration, years 15.9 ± ± ± ± ± Lung functions DLCO % pred FVC % pred FEV1 % pred 71.7 ± ± ± ± ± ± ± ± ± 16.3 < ± ± ± ± ± ± 16.3 MTX ever 18 (50.0) 9 (47.4) 9 (56.3) F 6 (50.0) 8 (44.4) Biological ever 7 (19.4) 3 (15.8) 4 (25.0) F 3 (25.0) 2 (11.1) F Prednisolone ever 32 (88.9) 17 (89.5) 15 (93.8) F 12 (100.0) 15 (83.3) F Dyspnoea 25 (71.4) 16 (84.2) 9 (56.3) (54.5) 14 (77.8) F Cough 18 (62.1) 10 (66.7) 8 (57.1) (72.7) 6 (46.2) F Crackles 30 (83.3) 15 (75.0) 15 (93.8) F 12 (100.0) 15 (83.3) F Hospitalization due to respiratory illness 1.9 ± ± ± ± ± Hospitalization due to cardiac illness 0.7 ± ± ± ± ± Median survival 88.0 ( ) 64.0 ( ) 92.0 ( ) ( ) ( ) F F F <0.001 F =Fisher test. In 6 patients, the follow-up time was too short to be able to categorize either as slow or rapid. The following data was missing: smoking status, RF, and information of dyspnoea from 1 patient, data about ANA from 10 patients, information of cough from 7 patients. RF = rheumatoid factor; ANA = antinuclear antibodies; RA = rheumatoid arthritis; DLCO = diffusion capacity to carbon monoxide; FVC = forced vital capacity; FEV1 Forced expiratory volume in one second; GAP = gender, age, physiology; CPI = composite physiologic index; RA = rheumatoid arthritis; MTX = methotrexate; UIP = usual interstitial pneumonia.

68 RADIOLOGICAL FINDINGS Disease progression In addition to baseline imaging, an additional HRCT was available for 33 patients (17 UIP, 4 OP, 7 NSIP, 4 unclassified, 1 DAD) (unpublished data). The second HRCT of 17 RA-UIPpatients showed clear signs of a progression in 11 cases, a mild progression in two cases and unchanged imaging in two cases (time between scans ranging from 19 to 77 months). In two cases, the possible progression could not be evaluated due to the poor quality of the second scan. In the 7 RA-NSIP patients, two clear progressions and one milder progression were observed, whereas the changes had diminished in three cases. One RA-NSIP exhibited new changes in the second imaging as the first scan had revealed no signs of any kind of ILD. In the RA-OP subgroup, two patients healed completely, one remained unchanged between scans and one developed DAD-like changes. Two of the unclassified ILD cases displayed a mild progression, remaining still as unclassifiable, and two remained unchanged. Nine patients, who could not be specifically categorized based on the first scan, developed more specific ILD pattern in their second scan, resulting in the appearance of an RA-UIP pattern in 7 cases and an RA-NSIP pattern in two cases Inter-observer agreement (III) When categorized into five subgroups (UIP, NSIP, OP, DAD, unclassified), the overall agreement between the radiologists was moderate (κ = 0.492). When categorized only into two groups i.e. definite UIPs vs. others, the inter-agreement rose slightly (κ = 0.592). When definite UIPs and the unclassified subgroup likely representing possible UIPs were pooled as one category, the agreement improved to good (κ= 0.629) The HRCT findings in different subtypes (III) Reticulation (93.1%) and GGO (72.4%) were the most common findings and observed in every subgroup to some extent. Reticulation was more abundant in UIP vs. OP (p<0.001), UIP vs. unclassified (p=0.041), NSIP vs. OP (p=0.020) and unclassified vs. OP (p=0.001), but similar in extent in UIP and NSIP patients. GGO was more extensive in the NSIP group vs. the unclassified subgroup (p=0.017). Honeycombing and architectural distortion were most often seen in patients with UIP and the extent of those findings was also significantly higher compared to all other subgroups (p<0.001). Emphysema (29.3%), bronchiectasis (24.5%) and pleural plaques (32.6%) were observed in approximately every third patient, without any statistically significant differences between subgroups with respect to the prevalence of the findings, although emphysema was more extensive in the UIP subgroup vs. RA-OP patients Original radiological reports Original reports were available from 59 patients, with 1 report missing from a patient that was re-classified as OP. A large proportion of the reports were non-specific descriptive reports, with the modern IIP classification being most often used in patients that were reclassified as NSIP or OP, whereas the UIP pattern was less mentioned (Table 17). More recent reports, i.e. those after the year 2011, were more specific than their older counterparts. The raw data of the original reports is shown in Table 16.

69 43 Table 16. Raw data of individual patients original radiological reports with selected details. Recategorized RA-ILD subtypes Years of the baseline / latest CTs Original radiological report Selected details of the original reports UIP 2002/- UIP lung fibrosis with basal predominance, interlobular septa thickening and comb formation. Consistent with IPF UIP 1999/2003 UIP peripheral fibrotic changes with basal predominance, reticulation and honeycombing, could be IPF, inconsistent with allergic alveolitis or asbestosis UIP 2003/- descriptive subpleural, peripheral fibrosis, basal and posterior predominance, few traction bronchiectasis, few honeycombs. Consistent with rheumatoid lung UIP 1997/1998 descriptive pleural plaques, peripheral basal fibrosis, impression of honeycombing, some traction bronchiectasis. Etiology could be asbestos but the role of MTX cannot be excluded UIP 2000/2000 descriptive extremely extensive fibrosis with honeycombing UIP 2006/2015 descriptive advanced fibrosis with honeycombing UIP 2004/- descriptive fibrotic changes with basal predominance, traction bronchiectasis and slight honeycombing. Could be idiopathic or RA-related fibrosis UIP 2002/- descriptive basal and peripheral reticulation and honeycombing, consistent with lung fibrosis UIP 2012/- UIP most probably RA-related UIP-changes UIP 1996/2004 descriptive progression. In both sides, clear fibrotic changes, which are most extensive in right middle lobe and less so in lower lobes, which speaks against RArelated fibrosis UIP 2005/2013 UIP most probably RA-related UIP-changes UIP 2008/- descriptive fibrosis advanced to honeycombing stage UIP 2004/2012 descriptive abundant fibrosis, clear progression UIP 2008/2011 descriptive apparently RA-related fibrosis, advanced to honeycombing stage UIP 2004/2013 UIP radiological UIP pattern, could be RA-related UIP 2009/- descriptive symmetrical peripheral fibrosis, advanced to honeycombing stage. Methotrexate reaction usually has more acute nature with GGO UIP 2006/2007 UIP rapidly progressive fibrosis, most likely IPF or a drug reaction UIP 2011/- UIP/NSIP distribution of the changes and the existing honeycombing are consistent with UIP, but in places there s an impression of subpleural sparing and GGO, could be UIP or NSIP, possibly related to RA or drugs UIP 2009/- descriptive old fibrosis, partly in honeycombing stage UIP 2012/2013 UIP subpleural, basal, honeycombing, consistent with UIP. Could be related to RA. NSIP also possible the pattern of which mostly suggests a drug reaction UIP 2010/- descriptive extensive honeycombing as a sign of mature fibrosis, no parts of the lungs spared UIP 2005/- descriptive very extensive honeycombing referring to fibrosis UIP 2006/- descriptive honeycombing-staged fibrosis in the basal parts. Quite extensive fibrotic changes that could be RArelated

70 UIP 2000/2007 descriptive bilateral peripheral extensive fibrotic changes, basal predominance, honeycombing, GGO both sides as a sign of disease activity UIP 2010/- UIP very severe lung fibrosis, reticulation and honeycombing, probably related to RA. Not quite typical UIP, even though these naturally cannot be differentiated radiologically UIP 2007/- descriptive basal predominance, fibrotic changes with honeycombing, these non-specific changes can be related to use of sulfasalazine or RA UIP 2007/2013 UIP consistent with UIP pattern, even though unchanged compared to older scans UIP 2014/- descriptive bilateral, basal honeycombing UIP 2010/2014 UIP / fnsip upper-lobe predominated UIP-like fibrosis with reticulation and scarce honeycombing. Possible UIP or fnsip UIP 2000/- descriptive bilateral basal peripheral honeycombing and traction bronchiectasis UIP 1995/- descriptive extensive honeycomb fibrosis in left lower lobe UIP 2014/- UIP radiological UIP pattern most likely associated with RA UIP 1993/- descriptive basal, peripheral fibrosis, partly in honeycombing stage. Could be RA-related UIP 2009/2012 UIP bilateral basal extensive honeycombing, UIP-like fibrosis UIP 2006/2015 descriptive lung fibrosis, with subpleural honeycombing and scarring UIP 2004/- descriptive fibrotic honeycomb formation NSIP 2012/- descriptive left: mild fibrotic changes. In the right side slightly more extensive fibrosis, with minor traction bronchiectasis, reticulation and GGO NSIP 2008/2013 descriptive bilateral middle and lower-lobe predominated interlobular septa thickening and bronchiectasis consistent with fibrosis, which might be RA-related NSIP 2008/2015 NSIP fibrotic, primarily NSIP-consistent changes, could be RA-ILD or a drug reaction NSIP 2005/2012 NSIP basal reticular abnormalities, no honeycombing. Could be e.g. RA-associated NSIP NSIP 2007/2009 UIP/fNSIP Not IPF-like UIP. Rather RA-related UIP or fnsip based on slow progression no honeycombing NSIP 2006/2012 NSIP the distribution of the changes and the lack of honeycombing speaks for NSIP NSIP 2006/2007 NSIP interlobular septa thickening, GGO, no honeycombing. NSIP pattern, probably RA-related NSIP 2014/2014 NSIP GGO, slight reticulation, a few posterior honeycomb cysts. Not typical for asbestosis, rather RA-NSIP OP 2005/2009 Original report not found OP 2012/- COP several patchy subpleural consolidations, primarily consistent with COP OP 2007/- descriptive two separate pneumonis-like consolidations OP 2011/2012 NSIP/COP/IPF extensive GGO, consolidation in the left side, small area with local honeycombing. Differential diagnosis contains NSIP, COP and apparently also IPF OP 2013/2013 COP multifocal consolidations consistent with COP OP 2014/- COP multifocal consolidations, primarily COP, although vasculitis also possible OP 2014/2015 COP single peripheral consolidation, malignancy possible, or COP 44

71 45 Unclass. 2003/- UIP interstitial intralobular fibrosis in basal parts. Could be related to rheumatoid lung, but the lack of noduli makes it slightly unlikely. Overall, the findings are very close to those of IPF Unclass. 1998/- descriptive subpleural fibrotic changes especially in basal parts, reticulation, interlobular septa thickening, honeycombing and traction bronchiectasis. Asbestos cannot be excluded, but also RA and drug reaction is possible Unclass. 2002/- descriptive clear fibrosis in basal areas, peripheral, minor subpleural fibrotic changes also in upper lobes, traction bronchiectasis, incipient honeycomb formation. Can be caused by CTD Unclass. 2012/2013 possible UIP / RB-ILD early fibrosis, peripheral reticulation, small centrilobular nodules, local few honeycomb cysts, could be UIP or RB-ILD Unclass. 2007/2009 descriptive small basal and peripheral fibrotic changes that could represent idiopathic pulmonary fibrosis due to RA or drug hypersensitivity Unclass. 1997/2005 descriptive TB-related scarring in right upper lobe, bronchiectasis both sides, not proper lung fibrosis nor GGO as a sign of an active process Unclass. 2008/- descriptive minor interlobular septa thickening, might be incipient fibrosis, very unspecific finding Unclass. 2013/2015 possible UIP reticular abnormalities with basal predominance, traction bronchiectasis, no clear honeycombing RA-ILD = rheumatoid arthritis-associated interstitial lung disease; UIP = usual interstitial pneumonia; NSIP = nonspecific interstitial pneumonia; OP = organizing pneumonia; Unclass. = unclassified; fnsip = fibrotic nonspecific interstitial pneumonia; COP = cryptogenic organizing pneumonia; RB-ILD =respiratory bronchiolitis- interstitial lung disease; RA = rheumatoid arthritis; MTX =methotrexate; GGO = groundglass opacities; CTD = connective tissue diseases; IPF = idiopathic pulmonary fibrosis; TB = tuberculosis. Table 17. Summary of the contents of the original radiological reports in different subtypes. RA-ILD subtype in recategorization Original radiological reports available n (%) Contents of the original reports n (%) IIP classification used Descriptive report UIP n=36 36 (100) 13 (25.0) UIP (n=11) UIP/fNSIP (n=2) NSIP n=8 8 (100.0) 6 (75.0) NSIP (n=5) UIP/NSIP (n=1) OP n=7 6 (85.7) 5 (71.4) COP (n=4) COP/NSIP/UIP (n=1) Unclassified n=8 8 (100.0) 3 (37.5) UIP (n=1) possible UIP (n=1) possible UIP / RB-ILD (n=1) 23 (63.9) 2 (25.0) 1 (14.3) 5 (62.5) RA-ILD = rheumatoid arthritis-associated interstitial lung disease; IIP = idiopathic interstitial pneumonias; UIP = usual interstitial pneumonia; NSIP = nonspecific interstitial pneumonia; OP = organizing pneumonia; fnsip = fibrotic nonspecific interstitial pneumonia; COP = cryptogenic organizing pneumonia; RB-ILD =respiratory bronchiolitis- interstitial lung disease.

72 HISTOLOGICAL DATA AND BAL Seven out of nine available TBB samples were non-diagnostic, while two showed a suspected OP of which both were suitable with OP also radiologically. One CT-guided transthoracal needle biopsy was performed in which the histopathologic diagnosis was OP. One patient had undergone a VATS biopsy. This patient died within a week from the biopsy which revealed a DAD diagnosis. Altogether 10 autopsies were performed, of which six were clinical autopsies and four forensic autopsies. Histological analyses of the autopsies were available in four cases, two of which revealed histopathological UIP- patterns, one OP-pattern and one nonspecific description of lung fibrosis. One RA-UIP patient underwent LTx but did not survive. The pathological examination of the original lungs revealed a severe and extensive UIP- pattern. BAL samples were gathered from 22 patients with the total instilled volume of saline mentioned in 5 cases, ranging from 50 to 300 ml. The differential cell count was available in 16 cases. The mean value of macrophages was 63.7 ± 26.6, that of lymphocytes 16.6 ± 13.4, of neutrophils 16.5 ± 18.4 and of eosinophils 3.5 ± 7.9 (Table 14). The mean percentage of macrophages was significantly higher in UIP-patients compared to BAL samples of OP individuals (p=0.002). The mean percentages of neutrophils differed significantly between UIP vs. OP subtypes and NSIP vs. OP subtypes (Table 18). Table 18. The mean percentages of cell counts in the 16 BAL samples from which the differential cell count was available. Cell type (mean ± SD) Macrophages RA-ILD (n=16) 63.7 ± 26.6 RA-UIP (n=8) 79.9 ± 18.5* RA-NSIP (n=3) 52.7 ± 33.6 RA-OP (n=4) ± 13.0 Unclassified (n=1) 75.0 Lymphocytes 16.3 ± ± ± ± Neutrophils 16.5 ± ± 8.6* 17.0 ± ± Eosinophils 3.5 ± ± ± ± p<0.05: * UIP vs. OP, NSIP vs. OP BAL = bronchoalveolar lavage; RA-ILD = rheumatoid arthritis-associated interstitial lung disease; UIP = usual interstitial pneumonia; NSIP = nonspecific interstitial pneumonia; OP = organizing pneumonia; SD = standard deviation. 5.4 COMORBIDITIES (I) The most common comorbidities were hypertension (30/50.8%), coronary artery disease (CAD) (21/35.6%), chronic obstructive pulmonary disease (COPD) (17/28.8%), cardiac insufficiency (16/27.1%), diabetes (13/22.0%) and asthma (12/20.3%). Gastroesophageal reflux (GER) was reported in six (10.2%) and hypothyroidism in three (5.1%) patients. There were two lung cancers (3.4%) and nine (15.3%) other cancers, including basal cell carcinoma (n=2), diffuse large B cell lymphoma (n=2), urinary bladder carcinoma (n=1), colon adenocarcinoma (n=1), squamous cell carcinoma in upper lip (n=1) and in tongue (n=1) and one ventricle carcinoma. Six (10.2%) patients suffered from tuberculosis (five in lungs and one intestinal) and 7 (11.9%) from depression. Obstructive sleep apnea (2/3.4%) was infrequent in our cohort. No statistically significant differences in the comorbidities were found between the UIP and non-uip groups, although COPD was more common in the UIP group (p= 0.088). Asthma was more common in women (p= 0.016) and COPD in men (p<

73 ). Comorbidities were divided equally in non-smokers and ever-smokers, except for COPD (p<0.001). 5.5 CAUSES OF DEATHS (I) According to the death certificates of the 33 deceased patients in study I, RA-ILD was the most common primary cause of death in 13 cases and reported as commonly in men and women (8 vs. 5, p=1.000), in non-smokers and ever-smokers (6 vs. 7, p= 0.522) and in UIP and non-uip individuals (10 vs. 3, p=0.701) (Figure 10). CAD was the second most common primary cause of death in seven cases, equally in UIP and non-uip patients (p=0.161). RA was the primary cause of death in five cases. The other reported primary causes of death were Alzheimer s disease, universal atherosclerotic disease with acute lower limb ischemia, acute pancreatitis, intestinal tuberculosis, COPD, massive bleeding due to pelvic fracture, lung cancer and suspected viral infection in the central nervous system - one case of each disease. Pneumonia and CAD were equally common as the immediate cause of death (both 10/30.3%; 6 UIP, 4 non-uip) and RA-ILD (5/15.2%; 4 UIP, 1 non-uip) was also prevalent (Figure 10). The following immediate causes of deaths were reported each in single cases: lung cancer, RA, diabetes, RA associated secondary amyloidosis with renal failure, acute pancreatitis, intestinal tuberculosis and gastroenteritis. Figure 10. The most common primary and immediate causes of deaths in 59 patients with RAILD. RA = rheumatoid arthritis; ILD = interstitial lung disease; CAD = coronary artery disease. 5.6 CORRELATIONS BETWEEN CLINICAL DATA, PFT AND RADIOLOGY (III) The strongest negative correlation (r= , p=0.001) was seen between the extent of emphysema and DLCO, which also correlated with the extent of architectural distortion (r= , p=0.033). A negative correlation between GGO extent and the duration of RA was observed (r= , p=0.023). The extents of honeycombing (r=0.266, p=0.046), traction

74 48 bronchiectasis (r=0.333, p=0.012) and architectural distortion (r=0.353, p=0.007) correlated with hospitalizations due to respiratory reasons. 5.7 THE COURSE OF THE DISEASE Differences between RA-UIP and non-uip patients (I) The patients with the definite UIP pattern needed more long-term oxygen therapy than their non-uip counterparts (p=0.016). In addition, their DLCO result declined more during the disease course (p=0.021), and their main number of hospitalizations due to respiratory causes was higher than in the non-uip group (p=0.004) a phenomenon which was not present with hospitalizations due to cardiac reasons (Table 19). Table 19. Factors associating with the differential course of the disease in the patients with rheumatoid arthritis associated usual interstitial pneumonia (RA-UIP) and non-uip patterns (RA-non-UIP). RA-ILD (n=59) RA-UIP (n=35/59.3%) RA-non-UIP (n=24/40.7%) Pvalue Oxygen therapy, n (%) 8 (13.6) 8 (22.9) 0 (0) Hospitalization due to respiratory 1.29 ± ± ± Hospitalization due to cardiac illness 0.6 ± ± ± Latest available FVC % pred 82 ± ± ± Latest available DLCO % pred 61 ± ± ± illness (mean ± SD) RA = rheumatoid arthritis, ILD = interstitial lung disease; UIP = usual interstitial pneumonia; FVC =forced vital capacity; DLCO = diffusion capacity to carbon monoxide; SD = standard deviation Survival (I, II) Overall, 34 (56.7%) of the 60 patients had died with the clear majority (87.9%) of the patients dying in hospital or in other health care centers, whereas only 4 patients died at home or outdoors. The average age of death was 75.0 ± 9.1 years, ranging from 54.8 to 91.7 years. The RA-UIP patients died slightly younger than the non-uip- patients (73.6 ± 9.8 vs ± 6.5, p=0.187). The survival analyses in the first two studies were performed after the exclusion of the RA-DAD patient, thus the total number of patients was 58, and the number of the deceased was 33. More patients had died in the UIP group versus non-uip patients (p= 0.046). The median survival in the whole group was months with no statistically significant differences between UIP vs. non-uip patients, male vs. female or between non-smokers vs. the combined group of current and former smokers (Table 20). The median survivals were 152 months and 61 months (p=0.017) in the GAP / ILD GAP stages I and II, respectively (Figure 11). In the non-uip group, there was a trend towards a longer survival in female vs. male (p=0.093) patients and non-smokers vs. ever-smokers (p=0.218), whereas no such trend was observed in the UIP-group.

75 49 Table 20. Survival of the patients (months) in different subgroups. Median survival, months 95% confidence interval Overall UIP non-uip Male Female Non-smokers Ever-smokers GAP/ILD-GAP stage I GAP/ILD GAP stage II P-value UIP = usual interstitial pneumonia; GAP = gender, age, physiologic variables- model. Figure 11. Comparison of the survival curves of the patients with rheumatoid arthritisassociated ILD categorized into either GAP / ILD-GAP stage I or II. The survival was significantly worse in GAP / ILD-GAP stage II (p=0.017, Log Rank). GAP = gender, age, physiologic variables- model; ILD = interstitial lung disease.

76 Predictors of mortality (II, III) All tested risk predicting models, i.e. GAP, ILD-GAP and CPI, were significant predictors of mortality in the Cox univariate model, as were the age at diagnosis, baseline DLCO and hospitalization due to respiratory reasons. In addition, several radiological features, i.e. the extents of reticulation, traction bronchiectasis and architectural distortion, were associated with decreased survival. For every increased GGO score point the mortality risk increased by 8 %, nearly reaching statistical significance (p=0.051) (Table 21). After adjusting for age, CPI score and baseline DLCO remained as significant predictors of mortality, whereas respiratory hospitalization and GAP/ILD-GAP lost their statistical significance. The median survival of four patients with pleural fluid was 10 months compared to 107 months in those without pleural effusions (p<0.001). Neither the presence of GGO, honeycombing nor reticulation associated statistically significantly with survival when assessed by either the presence or absence of these features. Smoking, FVC, male sex, UIP pattern or the use of methotrexate did not affect survival in a statistically significantly manner. Table 21. Prognostic factors for survival in patients with RA-ILD using a univariate Cox model. Hazard ratio 95% CI P-value Age at diagnosis DLCO % pred Resp.hospitalization Card. hosp NS CPI-points GAP score ILD-GAP score Extent of GGO Extent of reticulation Extent of traction bronchiectasis Extent of architectural distortion DLCO = diffusion capacity to carbon monoxide; Resp.hospitalization = hospitalization for respiratory reasons; Card. hosp. = hospitalization for cardiac reasons; CPI = composite physiologic index; GAP =gender, age, physiologic variables; GGO = ground-glass opacity; CI = confidence interval. 5.8 VALIDATION OF THE GAP AND ILD-GAP MODELS (II) Both GAP and ILD-GAP models predicted the mortality accurately, since both prediction models fitted the Wilson score confidence interval and no apparent differences were seen between the observed cumulative mortality and the predicted risk of mortality (Table 22). The observed mortality and the risk of death predicted by these models were also compared using the Hosmer-Lemeshow goodness-of-fit test. All the p-values were >0.05, meaning that there were no statistically significantly differences between estimated and observed mortality (Figure 12). ILD-GAP was more accurate in its prediction of 1-year cumulative mortality in both stages, whereas the GAP model was slightly more accurate at 2- and 3-year mortality prediction.

77 51 Table 22. Predicted and observed cumulative mortality of the patients with RA-ILD. GAP / ILD-GAP stage Observed (95% CI calculated by Wilson score) Predicted by GAP index and staging system Predicted by ILD-GAP Stage I 0.0 ( ) Stage II 8.3 ( ) Stage I 14.3 ( ) Stage II 9.1 ( ) Stage I 17.6 ( ) Stage II 27.3 ( ) Y mortality 2-Y mortality 3-Y mortality GAP = gender, age, physiologic variables. Figure 12. The Hosmer-Lemeshow statistic test shows that predicted and observed risks do not differ significantly (p>0.05). The x-axis shows the 1-y, 2-y and 3-y risk of mortality as predicted by the GAP and ILD-GAP staging system and the y-axis shows the observed risk. In every figure, stage I is on the left side and stage II on the right side. The vertical lines represent the confidence interval of the observed mortality rate. GAP = gender, age, physiologic variables.

78 52 6 Discussion The development of modern RA therapies has resulted in a decline in RA-related death, but the prevalence, burden and mortality of RA-ILD are all increasing (10,224). The increased prevalence could partly reflect an increased detection of this ILD (60) and another explanation is the increased survival of patients with RA, since advanced age is risk factor for the development of RA-ILD (4,37,73). In Denmark, the number of prevalent RA patients more than doubled during the years while the incidence remained stable, which reflects the increased survival among patients with RA (179). At the same time with the increased survival of RA, the prognosis of RA-ILD has been reported to be poor (37) with RA-ILD being responsible for 7% of the deaths in RA patients (10). The prevalence or incidence of RA-ILD in Finland has not been studied. Hakala et al. reported the RA-ILD incidence in Finland to be roughly one case per 3500 patient years, but only included those that were hospitalized and moreover included those with a suspected drug-induced lung disease (141). We detected 60 clinically relevant RA-ILD patients, which can be considered as representative sample size since KUH is responsible for the specialist medical care of the citizens in its catchment area and the prevalence of RA in Finland is approximately 0.8% (2). A slight trend towards an increased incidence might be present in our cohort, since in years 2012 and new patients were found compared to 0-4 new patients detected in each of the previous years, and in the last 7 years of data collection, the total number of new patients amounted to 27, compared to 19 new patients between years One should bear in mind the variable ICD-coding used and the fact that patients without HRCT were excluded from the cohort. Thus, this can only be considered as a directional result. This study represents real-life data from clinically significant RA-ILD in Finnish patients in whom we conducted a retrospective study by gathering and re-evaluating the demographic and clinical data extremely thoroughly. We have shown that there is variability in the course of the disease depending on the presence or absence of definite UIP pattern in HRCT. The patients with the UIP pattern more often 1) needed oxygen therapy, 2) were admitted to hospital care due to respiratory reasons and 3) suffered from an accelerated PFT decline in comparison with those with the non-uip subtype. We applied the GAP and ILD/GAP scores in RA-ILD patients and, as far as we are aware, we are the first group to demonstrate the suitability of these risk prediction models in RA-ILD. We have also revealed several radiological and other findings associated with reduced survival. 6.1 GENERAL DISCUSSION OF THE STUDY DESIGN Search for the patients and sample size Identifying the RA-ILD patients from hospital registers was challenging. The coding of the diagnoses has varied over the years and no standardized policy in the use of ICD-10 codes has been applied in clinical practice. Therefore, we decided to use a common pulmonary fibrosis code J84.X, even though it resulted in mainly finding IPF patients. Another search was conducted using common RA diagnoses M05.X and M06.X with the additional criterion that the patient had been examined / treated in the KUH pulmonology clinic. The inclusiveness of the searches was further ascertained with a third search using a code J99.0*M05.1. A similar search strategy with pulmonary fibrosis and RA codes has been used by other study groups also for the identification of suspected ILD patients (10,179). Another frequent method is to gather all RA patients from the studied area/hospital using different

79 53 registers and then evaluate the SLBs or HRCTs in these subjects to identify those with ILD (153,225), but often the exact data of how patients were detected is not provided. Our first search with the above-mentioned ICD-codes encompassed the years , resulting in the identification of approximately 50 RA-ILD patients. At that time, we considered broadening the searches to other Finnish hospitals but that would have delayed the research for several months or even years, as no appropriate registers of RA-ILD patients existed and new investigators would have needed to be recruited. Therefore, we decided to extend the search years instead. In 2015, when the patient search was underway, the current study population of 60 patients was similar in size with the majority of other RAILD publications, except for one or two UK/USA studies, where larger cohorts are possible due to their larger populations. Overall, medical records of over thousand patients had to be reviewed to achieve this study population of 60 RA-ILD patients, which is similar in size as most of the published reports (80,89,161), except for a few multicenter studies (73,225). It was our intention to investigate clinically relevant RA-ILD patients, and thus those patients with only minimal changes in their HRCTs were excluded. It is also probable that all RA-ILD patients have not been diagnosed with ILD Data gathering and missing data One strength in our cohort is that it is very well characterized since we gathered the data highly inclusively. In other investigations the gathered data has varied, especially the symptoms have often not been reported (73,179,225). Some studies lack RA serology (95), some PFT and smoking data (179) and some report only the percentage of current smokers (161). It is rare that all the above-mentioned data as well as comorbidities and causes of deaths have been evaluated from the same cohort and moreover our study is the first where hospitalizations and the use of oxygen have been investigated in RA-ILD patients. In a retrospective study protocol, there are always problems with missing data. To minimize this error, we have gathered the demographic data in a detailed manner using a specially designed form and gathered the data not only from KUH databases, but also from other hospitals and primary health care centres in which the patients had been treated. We believe that this approach has reduced the missing data to an absolute minimum. The smoking data was lacking from one patient (1.7%), RF from two (3.3%), baseline spirometry from five (8.3%) and baseline DLCO results from eight (13.3%) patients. Other retrospective studies have confronted similar problems with missing data. For example, in a study of 77 RA-ILD patients the RF data was missing from 1.3%, smoking data from 11.7% and ACPA data from 44.2% (80). It is also worth remembering that at least in a small country like Finland with a population of 5.5 million, it would take a very long time to gather a similar size RA-ILD cohort using a prospective study protocol Implication of the RA medication In a retrospective study, the patients have received highly variable treatments for RA and/or ILD without being followed using a standardized protocol, making it impossible to evaluate the effects of any specific treatment. We collected and reported, however, the medication used at the time of ILD diagnosis and the lifetime usage of biologic drugs, MTX and corticosteroids. Similar tactics have also been exploited in other reports (37,95,225) with medication history being obtained from a medical record charts similarly as in our study, while some reports have left RA medication unreported (39,73) Diagnostics Since histological data was limited, the re-categorization into different RA-ILD subtypes was performed radiologically, which is a common policy in ILD studies (90,100,194). This re-categorization can be considered as reasonably reliable, since a definite UIP pattern in HRCT has been proven to be a sensitive and specific method to detect the histopathologic

80 54 UIP in both IPF and RA-ILD. Assayag et al investigated 69 biopsy-proven RA-ILD patients from three tertiary care centres who also had undergone a CT scan within 12 months of their SLB (113). They observed that a definite UIP pattern on a CT was highly specific (96%; 95%CI %) with a negative predictive value of 53%, and rather sensitive (45%; 95% CI 30-61%) with a positive predictive value of 95% (113). Similar results have also been reported from IPF studies, with a definite UIP pattern in CT showing specificity of 90-95% and sensitivity of 79 85% in depicting a histopathologic UIP pattern (112,114). Therefore, we are confident that the UIP subgroup in this study reliably represents true RA-UIP patients. It is, however, possible that some patients re-categorized into the NSIP or unclassified subgroups may be suffering from histopathological UIP. Another strength that increases the reliability of our re-categorization is the fact that a large proportion of the cases / HRCT scans were evaluated in a MDD Reliability of the radiological re-categorization The inter-observer agreement of our study (III) was in line with a previous report from Assayag et al (113), in which the agreement of UIP using strict criteria (definite UIP vs. possible UIP + inconsistent with UIP) was good (κ = 0.67) and moderate (κ = 0.52) when using the broader criteria (definite + possible UIP vs. inconsistent). 6.2 CLINICAL FEATURES OF THE COHORT Subject characteristics and PFT The mean age of the patients and the amount of cases in which ILD preceded RA were similar as described in a recent multicentre UK study (73). In our study, there was almost equal numbers of both genders. Since RA is twice as common in females (226), our gender distribution supports previous findings that male sex is a risk factor for ILD development (4,7). The PFTs, especially FVC, were rather well preserved in our cohort, which indicates that the patients had been diagnosed earlier than in many other studies. For example, FVC was normal in 55.6% of our patients, and the mean baseline value was as high as 84.8%, whereas others have reported mean FVC values in a range 69-75% (139,153). This may be partly explained also by our rather large proportion of never-smokers (42.9%) e.g. compared to 36% in the study of Solomon et al (153). In our cohort, dyspnoea was more common and cough as common, as in the biopsy-proven group of 54 RA-ILD cases of Nakamura et al (161). The difference in the numbers of patients suffering from dyspnoea may be explained by the larger proportion of UIP cases in our cohort (59% vs. 28%), although inspiratory crackles, more often present in RA-UIP cases in our cohort, were as common in both studies (161) Radiological features and their correlation to RA duration (III) The proportion of different radiological subtypes after the re-categorization was in line with the previous literature (148,227), with UIP being the most common subtype. Our most commonly observed radiological findings i.e. reticulation (93.1%) and GGO (72.4%) were also in line with previous investigations, since Tanaka et al. reported 90% of 63 RA-ILD patients having GGO and 98% having reticulation (31). In the same study, also the frequencies of honeycombing (60%), emphysema (24%) and traction bronchiectasis (75%) were very similar as encountered in our works (53%, 29% and 60%, respectively), but nodules were much more common in their study (49% vs. 4%), the reason for this discrepancy is unclear (31). Another study with 29 RA-ILD patients reported the most common radiological observations as being reticulation (72%) and GGO (66%) (29). In that study, of the 14/19 patients with likely definite and possible UIPs with reticulation with or without honeycombing underwent a follow-up CT, in which progression of the disease was detected in 79% of the patients (29). In our study, 13/17 RA-UIP patients showed

81 55 progression of the disease. Moreover, nine patients that initially could not be categorized to any specific subgroup later developed more distinct features and 7 of them fulfilled definite UIP criteria based on the follow-up scan. This gives the impression that if RA-ILD patients were followed longer and if HRCT was repeated, then a more accurate diagnosis could probably be made. In the study of Kim et al, the RA duration was longer in the RA-UIP than in the non-uip patients (95). A similar result has been observed by others, with those patients with a predominantly reticular pattern on HRCT having a longer duration of RA than those with predominantly ground-glass pattern (94). In our study, GGO was negatively correlated with the duration of RA, which could indicate that GGO is an early phenomenon in RAILD, thus being in line with the above-mentioned investigations. Conflicting reports have, however, been published. In one study, GGO was as prevalent in early as in longstanding RA (43) and thus, the timeframe of GGO development remains unclear and might vary from patient to patient BAL results In our retrospective study, BAL samples were taken using variable techniques and only 16 out of 22 samples were representative. The total amount of saline was mentioned in 5 cases, varying between 50 to 300 ml given in 2-6 doses. It was impossible to evaluate the impacts of exogenous factors of e.g. smoking, infections, drugs on the BAL results. Ten out of 16 BAL samples with differential cell counts were abnormal, which is in line with several previous studies (7,104,105). Some differences were observed between RA-ILD subtypes, but the number of samples was so limited that it is hard to draw any definitive conclusions. BAL techniques vary in different countries and different hospitals and the indications when BAL is performed are variable. In RA-ILD, BAL should perhaps have a greater role than in other ILDs, since the patients are receiving immunosuppressive medication and thus at a higher risk of developing opportunistic infections and experiencing drug reactions. The differential diagnostics of these illnesses can be challenging and evaluation of BAL can be helpful to some extent. The ATS Guideline of BAL recommends BAL being performed in cases with non-diagnostic HRCT, using a standardized technique and always including a differential cell count (107). As noted in our study, this has not been a common policy in clinical practice Original radiological reports There is no formal recommendation or a guideline about the use of the IIP classification in CTD-ILD patients. The classification has been gradually introduced into use quite recently, especially in the last 5-7 years. In the 2002 ATS/ERS international multidisciplinary consensus classification of IIPs, it was stated that in the case of a suspected diffuse lung disease, the presence or absence of typical UIP pattern should be determined (18). However, as our data shows, the UIP pattern was not mentioned in a routine manner in the radiological reports of RA-ILD patients and actually, the process of adopting the IIP classification into use has been rather slow also in patients with IIPs. Interestingly a high proportion of the NSIP patients had NSIP mentioned already in their original reports which could be due to the fact that previously mainly the NSIP pattern has been associated with CTD-ILDs, whereas UIP mostly was perceived as an IPF-related pattern. Gradually, an increasing number of studies has revealed the high proportion of UIP patients in RA. In our cohort, it seemed that by the year 2011, the UIP/IIP definitions were being better adopted also in the non-idiopathic ILDs and the reports more often contained a specific classification Disease course in UIP and non-uip patients (I) Differences in the course of the disease in distinct RA-ILD subtypes have not been studied extensively, apart from the differences in survival. In our study, the lung disease was more

82 56 progressive in the UIP subgroup based on the number of deaths, greater need for oxygen therapy, hospitalization due to respiratory reasons and the decline of PFT. The last of those findings was confirmed in a recent study, in which the PFTs of 167 RA-ILD patients were evaluated at baseline and then annually, when available, up to 10 years after ILD diagnosis. It was found that the UIP pattern was a risk factor for DLCO progression (140). Moreover, the need for supplemental oxygen was more common to some extent among patients with UIP compared to NSIP patients, similarly as observed in our study, although this difference did not reach statistical significance (140). In longitudinal analyses, significant differences between groups of IPF and CTD-UIP were not detected in changes of FVC% or DLCO%, but nonetheless those subjects with IPF had a poorer survival (228). 6.3 COMORBIDITIES AND CAUSES OF DEATHS (I) The most common comorbidities in our study were hypertension and CAD, which was much more common (36%) in our cohort than in the Danish cohort (13%) of 679 RA-ILD patients (179). We also had a double amount (22% vs. 10%) of diabetics and a three-fold higher amount (27% vs. 9%) of patients with heart failure (179). This could reflect the unfavourable genetic background in eastern Finland. Indeed, an IPF cohort of KUH region revealed very similar comorbidity rates as encountered in our study: CAD 49%, diabetes 27%, heart failure 27%, hypertension 46% (vs. 51% in our study) (229). COPD was observed in almost 30% of the patients, which is in line with a recent study revealing a 48% prevalence of emphysema (230). Interestingly, COPD was more prevalent in UIP individuals, even though smoking was similar in UIP and non-uip groups. There is some evidence for an association between ILD and lung cancer. One study compared 18 RA-ILD patients with 18 matched IPF patients and observed more patients dying from lung cancer in RA-ILD (147). In our study, lung cancer was present in only 2 patients (3.4%), whereas in the IPF cohort from the same district, a higher proportion i.e. 6.8% was observed, even though the amount of non-smokers was almost identical (our 39.7% vs. their 35.2%) in both studies (229). Earlier, some studies have addressed causes of death in RA-ILD patients and reported results similar to ours. Hakala et al. claimed that 80% of the patients died because of progression of the lung disease (141). In another RA-UIP cohort of 10 patients, there were 5 deaths and all of them were attributable to respiratory diseases with three steady progressions of ILD, one pneumonia and one acute exacerbation (96). Moreover, Tsuchiya et al. reported that in their cohort of 144 RA-ILD patients, there were 71 deaths of which most (58/81.7%) were due to respiratory reasons. The reported respiratory complications included 19 UIP exacerbations, 9 pneumonias, 4 other pulmonary infections, 13 ILD progressions, 6 lung cancers, 3 bronchiectasis exacerbations, 2 pneumothoraxes and 2 pneumonitis (drug- or radiation- induced) (148). However, it is somewhat unclear how exacerbations were distinguished from different respiratory infections. Our results concerning causes of deaths are in line with the previous reports, as in our cohort also the patients mostly died because of the ILD. This was especially the case with the RA-UIP patients, with more people in the non-uip group having CAD as the primary cause of death, even though this finding did not reach statistical significance. Other reports have not separated underlying and immediate causes of deaths and thus we are providing some novel data about this topic. It is possible that some of the patients whose immediate cause of death was coded as pneumonia actually suffered from AE, but this is impossible to ascertain in a retrospective study protocol. This hypothesis is perhaps supported by our novel finding of the correlation between the hospitalizations due to respiratory reasons and the extents of honeycombing and architectural distortion, i.e. features most typical in UIP in which the AEs are most common (133) and in which group most deaths occurred.

83 SURVIVAL (I) Previous publications have reported variable results concerning the survival of RA-ILD patients. Some have reported survival as being as poor as in IPF i.e. 2-3 years (4,37,146), while other have demonstrated survival times of up to 6-8 years (143,148) in line with the present study. One possible reason for the longer survival is the fact that in this study we included also patients with the OP pattern. The studies investigating mortality of RA-ILD have often consisted only of UIP and NSIP, but not of other types of ILDs. Another possible reason is the large proportion of patients with well-preserved PFT. Furthermore, almost 24% of the patients had received treatment with biological drugs, which may be speculated to have influenced their survival. In our study, the difference in median survival between UIP and non-uip patients did not reach statistical significance, even though the difference was almost four years (92 vs 137 months). This could be a consequence of our rather small study population. Among the UIP patients, there were eight patients that have stayed alive for over 10 years, three for over 15 years and two with an extremely indolent course of disease who have lived for over 20 years. All the above-mentioned patients had a definite UIP pattern in HRCT and no explanatory reason for their long survival could be detected. In a previous study of biopsyproven 48 RA-ILD patients, no differences in survival for patients with UIP, fnsip or unclassifiable fibrosing ILD were observed, but when these were pooled into one fibrotic category and compared to a non-fibrotic group consisting of cnsip, DAD, DIP, LIP and OP patients, the fibrotic ILDs had significantly worse survival (146). 6.5 PREDICTORS OF MORTALITY (II, III) Pulmonary function tests and CPI Previously, several studies have highlighted the importance of pulmonary physiology when evaluating the risk of death. In fibrotic IIPs, for example, the pulmonary physiology appeared as an even stronger predictor of survival than the histopathologic pattern (165). In our study, the baseline DLCO, but not FVC, was an independent risk factor for mortality in the univariate analysis and retained its significance after adjusting for age. This could be explained by the relatively high mean baseline FVC and/or by the small patient cohort. Not all studies have, however, been unanimous when investigating the role of FVC in predicting mortality. In the study of 84 RA-ILD patients conducted by Kim et al., FVC did not act as an independent predictor of death, despite the lower mean FVC values than in our study (95) whereas others have demonstrated that a lower baseline FVC value and the FVC decline over time were both associated with an increased hazard of death (139,153). Various studies have identified the association of DLCO with survival either in univariate (95,146), or multivariate (95) models, a finding which we confirmed here. Some investigators have also used longitudinal methods and substantiated that a decline of 10% or more in DLCO as being a significant predictor of mortality in RA-ILD (153) or RA-UIP cohorts (139), as well as a predictor of progressive disease (140). Unfortunately, it was not possible to conduct multivariate models with our small cohort and the missing follow-up data prevented longitudinal investigations of PFT. The CPI score has been investigated by other groups. In line with the results of our study, Solomon et al revealed CPI to be a significant predictor of mortality in a univariate model, but not in the multivariate model which controlled for age, gender, smoking, baseline FVC and HRCT pattern (153) Clinical factors Previously male sex has been associated with worse survival. Solomon et al. investigated 137 RA-ILD patients and observed a hazard ratio for mortality of 0.58 in females (153). In

84 58 another retrospective study of 82 patients with RA-ILD, female sex was associated with better survival in both bivariate and multivariate models (95). Male sex, contrary to the above-mentioned reports, was not a statistically significantly associated with mortality in our study. Nonetheless, our survival analyses did reveal a tendency towards a better survival in non-uip females. The mean amount of hospitalizations due to respiratory reasons was an independent predictor of mortality in our univariate analysis, which seems to be a novel finding as we have been unable to find any other report investigating this parameter in RA-ILD patients. However, a similar result has been observed previously in an IPF study, in which a 24week history of respiratory hospitalization was associated with a more than fourfold increase in the risk of death (175). The RF positivity was not a significant prognostic factor in our study nor was it in a Danish population-based cohort study of 679 RA-ILD patients (179), but in contrast, in a very recent study, a high titer of RF was associated with poor survival (80) Radiological factors associating with decreased survival Our study revealed that the extents of reticulation, traction bronchiectasis and architectural distortion were associated with decreased survival, the last of these, as far as we are aware, is a novel finding. We used a semi-quantitative method to evaluate the extents of different findings. A similar method has been previously used in one study of RA-ILD, which we believe is also the only previous RA-ILD study which has investigated the association between different radiological details and mortality (95). In that other study, the extents of GGO, reticulation, traction bronchiectasis and honeycombing were graded as absent, mild, moderate or severe similarly as conducted in our study. They observed that the presence of reticulation, traction bronchiectasis and honeycombing, as well as the extents of honeycombing and traction bronchiectasis were independently associated with worse survival (95). Thus, the results of both studies are rather similar, although they did not investigate the role of architectural distortion as a predictor of mortality. Another study of 168 CTD-ILD patients (including 39 RA-ILD) also reported that the severity of traction bronchiectasis, graded on a scale 0-3, and the extensiveness of honeycombing were associated with an increased death hazard (159). The fact that their study population was mostly constituted of patients with other than RA-related ILDs complicates the comparison of the results. In addition, the method was different, since they estimated extents of all other findings other than traction bronchiectasis to the nearest 5% (159). Some studies have explored the association between CT findings and survival in the patients with IIP. In the study of Edey et al., the extent of traction bronchiectasis was graded on a scale 0-3 when other findings were estimated to the nearest 5% and then the average scores of 6 levels of both lungs were used to calculate a global disease score (231). These investigators observed somewhat similar results as us, since the extents of reticulation and traction bronchiectasis were associated with worse survival. In addition, the overall extent of any lung abnormalities and honeycombing were also related to poor prognosis (231). Another study of 98 biopsy-proofed IPF patients reported that traction bronchiectasis and fibrosis scores were significant predictors of outcome (232). 6.6 VALIDATION OF THE GAP AND ILD-GAP MODELS (II) In our study, both GAP and ILD-GAP could provide relatively accurate estimations of mortality in stage I and II patients, whereas stage III patients were not present in our study for some unknown reason. In a previously published Danish ILD cohort, 32% of the 115 IPF patients belonged to stage I, 48% to stage II and 20% to stage III (157). Other investigators have reported GAP stage III patients to account for approximately 14-22% of the studied IPF cohorts (233,234).

85 59 Interestingly, the ILD-GAP index was more accurate at predicting 1-year mortality, whereas in the 2-year and 3-year prediction, the GAP index was slightly more accurate. Our study was the first one to investigate these risk prediction models in a RA-ILD cohort, although previous studies have been conducted on IPF and SSc-ILD patients. In Korean IPF patients, GAP produced accurate 1-year, but not 3-year, mortality estimates (234). In another cohort of IPF patients, the GAP staging was found to be useful for evaluating the IPF severity, revealing statistically significant differences in survival in different GAP stages (233). The GAP index, however, displayed poor applicability for the predicted 1-year mortality in SSc-ILD patients (235). After the publication of our results (236), two studies have investigated the applicability of the GAP model in RA-ILD patients. The first report investigated 309 RA-ILD patients and stated that the discrimination of the model was similar to that previously reported in IPF patients and that the addition of other variables, for example a definite UIP pattern, did not improve the discrimination of the model (225). They also reported that GAP Index and a staging system had satisfactory capabilities in predicting mortality at 1, 2 and 3 years (225). In another study of 181 RA-ILD patients, the GAP model demonstrated good calibration and discrimination in both sexes and all types of ILD, but the ILD-GAP did not perform as well as the GAP model (237). We observed that the 2-year mortality in stage I patients was much higher than predicted by the ILD-GAP model, which also underestimated the 3-year mortality of stage I patients. It remains unclear which of the risk prediction models is better suited for patients with RA-ILD. The ILD-GAP was originally developed in a study protocol including all kinds of ILDs without considering the variable prognosis and courses of diseases in different CTD- ILDs. It has been pointed out that patients with other CTD-ILDs enjoy a better survival than those with RA-ILD (110,162). At least partly, this could be a consequence of the higher proportion of UIP patients in RA-ILD (96). Therefore, it can be debated whether the ILD- GAP is truly valid for all CTD-ILDs. It is also worth pondering whether the accuracy of GAP/ILD-GAP models would improve with additional variables, such as smoking or symptoms. Originally, when the GAP model was being designed, the researchers considered several commonly available predictor variables, such as body mass index, smoking status and the use of long-term oxygen and the best combination of variables was screened using complex and sophisticated statistical methods (177). Oxygen use was removed due to its different effects in the derivation and validation cohorts. Dyspnoea was omitted because it was not available in the validation cohort (177). Very recently, it was shown that combining position emission tomography data with GAP data was able to improve the models ability to predict mortality (238). Future research might well devise improved versions of the prediction models, but overall for any staging system, it is important to keep it simple and repeatable in order to make possible its use in daily practice. 6.7 FUTURE PERSPECTIVES Historically, when compared with the remarkable advances in clarifying the articular aspects of RA, RA-ILD has remained poorly understood, under-recognized, undertreated, even its very existence doubted. Gradually, RA-ILD research has attracted more interest. With the help of modern imaging possibilities and less risky biopsy procedures, it is now possible to increase our knowledge of this illness. Accurate biomarkers could help to recognize those patients who are at high risk of rapidly progressing disease and those who remain stable, which might allow clinicians to target interventions to patients at the highest risk. Moreover, disease monitoring could be planned more personally if it were possible to identify different courses of diseases already during their early phases. Patients with a potentially lethal disease deserve proper counselling; this is easier to accomplish if we can

86 60 identify those at the highest risk of disease progression or death. Risk prediction models, such as GAP, could offer a framework for discussing prognosis and hopefully their use will increase in clinical practise. Another important and under-used utilization of GAP is in the decision-making of when to refer a patient for LTx. For example, if the 1-, 2- or 3-year risk for mortality predicted by GAP surpasses the published risks for mortality after LTx, the patient should be screened and listed for transplantation, if appropriate. In some countries, risk estimation with GAP is already required in IPF patients that are referred to LTx. Key areas in future RA-ILD research include also other treatment strategies. If we are to discover new treatment options or to ascertain the safety profile of currently available biologic or other drugs, it is crucial to conduct well-designed clinical trials. GAP or other risk prediction models would probably also be useful in various areas of research. If the patients could be categorized into different groups by their predicted risk of death, this would enable the evaluation of identifying which patients would obtain the greatest benefit from the investigated drug. Results of the ongoing clinical trials on antifibrotics are eagerly awaited and perhaps in the future these will become available for RA-ILD or RA-UIP patients. The diagnosis, not to mention treatment, of RA-ILD/CTD-ILD is not straightforward. The disease is so rare, that national/international collaboration, such as national register or biobank, should be considered since this would make possible examining larger study populations and samples. ILD research continues actively in KUH in collaboration with other hospitals. Cohorts of asbestosis and IPF have already been gathered from KUH databases using the same data collection form as used here, making comparative studies possible in the future. This thesis has highlighted the clinical importance of RA-ILD. We found rather many RA-ILD patients in KUH region and our coarse estimations suggested that the incidence is rising also in this region, as other reports have indicated is the case elsewhere. These patients deserve to be well diagnosed and assessed as candidates for modern treatment protocols. The recognition of RA-ILD is important in every day clinical practice as a part of ILD differential diagnostics, since ILD can be the first manifestation of RA and furthermore, since most RA-ILD patients reveal a UIP pattern similarly as IPF.

87 61 7 Conclusions 1. During , a total of 60 clinically relevant RA-ILD patients were treated in the KUH pulmonology clinic. Of those, 60% revealed a UIP pattern, followed by 13.3% NSIP, 11.7% OP, 1 DAD with the rest having an unclassifiable subtype. The RA-UIP patients followed a distinctive and more progressive course of disease based on the higher number of deaths, greater use of oxygen therapy, more extensive decline in PFT and the increased number of hospitalizations. Although different comorbidities were frequent, the RA-ILD patients mostly died because of the ILD itself. 2. Both GAP and ILD-GAP can provide relatively good estimates of mortality in RAILD patients, even though it remains unclear which of these models would be better suited for these patients. In addition, CPI and baseline DLCO are associated with shortened remaining lifetime. 3. Reticulation and GGO are the most common radiological findings in RA-ILD patients, and can be present in every subtype. Rather than the presence of different findings, it is their extent that is clinically important; this can be estimated using a semiquantitative method. The extents of reticulation, traction bronchiectasis and architectural distortion are associated with decreased survival. The two last parameters also correlate with the number of hospitalizations, as does the extent of honeycombing. HRCT findings can be useful when evaluating the risk of death and the course of the disease.

88 62 8 References (1) Gabriel SE. The epidemiology of rheumatoid arthritis. Rheumatic Disease Clinics of North America 2001;27(2): (2) Aho K, Kaipiainen-Seppanen O, Heliovaara M, Klaukka T. Epidemiology of rheumatoid arthritis in Finland. Semin Arthritis Rheum 1998;27(5): (3) Minaur NJ, Jacoby RK, Cosh JA, Taylor G, Rasker JJ. Outcome after 40 Years with Rheumatoid Arthritis: A Prospective Study of Function, Disease Activity, and Mortality. J Rheumatol 2004;31(SUPPL. 69):3-8. (4) Bongartz T, Nannini C, Medina-Velasquez YF, Achenbach SJ, Crowson CS, Ryu JH, et al. Incidence and mortality of interstitial lung disease in rheumatoid arthritis: a populationbased study. Arthritis Rheum 2010 Jun;62(6): (5) Turesson C, O'Fallon WM, Crowson CS, Gabriel SE, Matteson EL. Extra-articular disease manifestations in rheumatoid arthritis: Incidence trends and risk factors over 46 years. Ann Rheum Dis 2003;62(8): (6) Gochuico BR, Avila NA, Chow CK, Novero LJ, Wu H-, Ren P, et al. Progressive preclinical interstitial lung disease in rheumatoid arthritis. Arch Intern Med 2008;168(2): (7) Gabbay E, Tarala R, Will R, Carroll G, Adler B, Cameron D, et al. Interstitial lung disease in recent onset rheumatoid arthritis. Am J Respir Crit Care Med 1997 Aug;156(2 Pt 1): (8) Suzuki A, Ohosone Y, Obana M, Mita S, Matsuoka Y, Irimajiri S, et al. Cause of death in 81 autopsied patients with rheumatoid arthritis. J Rheumatol 1994 Jan;21(1): (9) O'Dwyer DN, Armstrong ME, Cooke G, Dodd JD, Veale DJ, Donnelly SC. Rheumatoid Arthritis (RA) associated interstitial lung disease (ILD). Eur J Intern Med 2013 Oct;24(7): (10) Olson AL, Swigris JJ, Sprunger DB, Fischer A, Fernandez-Perez ER, Solomon J, et al. Rheumatoid arthritis-interstitial lung disease-associated mortality. Am J Respir Crit Care Med 2011 Feb 1;183(3): (11) Dawson JK, Fewins HE, Desmond J, Lynch MP, Graham DR. Predictors of progression of HRCT diagnosed fibrosing alveolitis in patients with rheumatoid arthritis. Ann Rheum Dis 2002 Jun;61(6):

89 63 (12) Assayag D, Lubin M, Lee JS, King TE, Collard HR, Ryerson CJ. Predictors of mortality in rheumatoid arthritis-related interstitial lung disease. Respirology 2014 May;19(4): (13) Kolb M, Collard HR. Staging of idiopathic pulmonary fibrosis: past, present and future. Eur Respir Rev 2014 Jun;23(132): (14) Hakala M. Pulmonary manifestations of rheumatoid arthritis: a clinical, immunogenetic and histological study. Acta Universitatis Ouluensis D (15) Rantalaiho V, Korpela M, Hannonen P, Kautiainen H, Jarvenpaa S, Leirisalo-Repo M, et al. The good initial response to therapy with a combination of traditional disease-modifying antirheumatic drugs is sustained over time: the eleven-year results of the Finnish rheumatoid arthritis combination therapy trial. Arthritis Rheum 2009 May;60(5): (16) Leirisalo-Repo M. Tulehduksellisten reumatautien uudet biologiset lääkkeet. Lääketieteellinen Aikakauskirja Duodecim 2007;123(20):67. (17) Solomon JJ, Brown KK. Rheumatoid arthritis-associated interstitial lung disease. Open Access Rheumatol Res Rev 2012;4: (18) American Thoracic Society, European Respiratory Society. American Thoracic Society/European Respiratory Society International Multidisciplinary Consensus Classification of the Idiopathic Interstitial Pneumonias. This joint statement of the American Thoracic Society (ATS), and the European Respiratory Society (ERS) was adopted by the ATS board of directors, June 2001 and by the ERS Executive Committee, June Am J Respir Crit Care Med 2002 Jan 15;165(2): (19) Charlotte Hyldgaard. A cohort study of Danish patients with interstitial lung diseases: burden, severity, treatment and survival; (20) Travis WD, Costabel U, Hansell DM, King TE,Jr, Lynch DA, Nicholson AG, et al. An official American Thoracic Society/European Respiratory Society statement: Update of the international multidisciplinary classification of the idiopathic interstitial pneumonias. Am J Respir Crit Care Med 2013 Sep 15;188(6): (21) Turesson C. Extra-articular rheumatoid arthritis. Curr Opin Rheumatol 2013 May;25(3): (22) Lake F, Proudman S. Rheumatoid arthritis and lung disease: from mechanisms to a practical approach. Semin Respir Crit Care Med 2014 Apr;35(2): (23) Prete M, Racanelli V, Digiglio L, Vacca A, Dammacco F, Perosa F. Extra-articular manifestations of rheumatoid arthritis: An update. Autoimmun Rev 2011 Dec;11(2): (24) Ellman P, Ball RE. Rheumatoid disease with joint and pulmonary manifestations. Br Med J 1948 Nov 6;2(4583):

90 64 (25) Stack BH, Grant IW. Rheumatoid interstitial lung disease. Br J Dis Chest 1965 Oct;59(4): (26) Walker WC, Wright V. Diffuse interstitial pulmonary fibrosis and rheumatoid arthritis. Ann Rheum Dis 1969 May;28(3): (27) Respiratory complications of rheumatoid disease. Br Med J 1978 Jun 3;1(6125): (28) Murakami DM, Bassett LW, Seeger LL. Advances in imaging of rheumatoid arthritis. Clin Orthop Relat Res 1991 Apr;(265)(265): (29) Akira M, Sakatani M, Hara H. Thin-section CT findings in rheumatoid arthritisassociated lung disease: CT patterns and their courses. J Comput Assist Tomogr 1999;23(6): (30) Remy-Jardin M, Remy J, Cortet B, Mauri F, Delcambre B. Lung changes in rheumatoid arthritis: CT findings. Radiology 1994 Nov;193(2): (31) Tanaka N, Kim JS, Newell JD, Brown KK, Cool CD, Meehan R, et al. Rheumatoid arthritis-related lung diseases: CT findings. Radiology 2004 Jul;232(1): (32) Raghu G, Collard HR, Egan JJ, Martinez FJ, Behr J, Brown KK, et al. An official ATS/ERS/JRS/ALAT statement: idiopathic pulmonary fibrosis: evidence-based guidelines for diagnosis and management. Am J Respir Crit Care Med 2011 Mar 15;183(6): (33) Doyle TJ, Lee JS, Dellaripa PF, Lederer JA, Matteson EL, Fischer A, et al. A roadmap to promote clinical and translational research in rheumatoid arthritis-associated interstitial lung disease. Chest 2014 Mar 1;145(3): (34) Habib HM, Eisa AA, Arafat WR, Marie MA. Pulmonary involvement in early rheumatoid arthritis patients. Clin Rheumatol 2011 Feb;30(2): (35) Zhang Y, Li H, Wu N, Dong X, Zheng Y. Retrospective study of the clinical characteristics and risk factors of rheumatoid arthritis-associated interstitial lung disease. Clin Rheumatol 2017 Apr;36(4): (36) Frank ST, Weg JG, Harkleroad LE, Fitch RF. Pulmonary dysfunction in rheumatoid disease. Chest 1973 Jan;63(1): (37) Koduri G, Norton S, Young A, Cox N, Davies P, Devlin J, et al. Interstitial lung disease has a poor prognosis in rheumatoid arthritis: results from an inception cohort. Rheumatology (Oxford) 2010 Aug;49(8): (38) McDonagh J, Greaves M, Wright AR, Heycock C, Owen JP, Kelly C. High resolution computed tomography of the lungs in patients with rheumatoid arthritis and interstitial lung disease. Br J Rheumatol 1994 Feb;33(2):

91 65 (39) Dawson JK, Fewins HE, Desmond J, Lynch MP, Graham DR. Fibrosing alveolitis in patients with rheumatoid arthritis as assessed by high resolution computed tomography, chest radiography, and pulmonary function tests. Thorax 2001 Aug;56(8): (40) Zrour SH, Touzi M, Bejia I, Golli M, Rouatbi N, Sakly N, et al. Correlations between high-resolution computed tomography of the chest and clinical function in patients with rheumatoid arthritis. Prospective study in 75 patients. Joint Bone Spine 2005 Jan;72(1): (41) Bilgici A, Ulusoy H, Kuru O, Celenk C, Unsal M, Danaci M. Pulmonary involvement in rheumatoid arthritis. Rheumatol Int 2005 Aug;25(6): (42) Saag KG, Kolluri S, Koehnke RK, Georgou TA, Rachow JW, Hunninghake GW, et al. Rheumatoid arthritis lung disease. Determinants of radiographic and physiologic abnormalities. Arthritis Rheum 1996 Oct;39(10): (43) Mori S, Cho I, Koga Y, Sugimoto M. Comparison of pulmonary abnormalities on highresolution computed tomography in patients with early versus longstanding rheumatoid arthritis. J Rheumatol 2008 Aug;35(8): (44) Restrepo JF, Del Rincon I, Battafarano DF, Haas RW, Doria M, Escalante A. Clinical and laboratory factors associated with interstitial lung disease in rheumatoid arthritis. Clin Rheumatol 2015 Sep;34(9): (45) Assayag D, Lee JS, King Jr TE. Rheumatoid arthritis associated interstitial lung disease: a review. Medicina (B Aires) 2014;74(2): (46) Charles PJ, Sweatman MC, Markwick JR, Maini RN. HLA-B40: a marker for susceptibility to lung disease in rheumatoid arthritis. Dis Markers 1991;9(2): (47) Scott TE, Wise RA, Hochberg MC, Wigley FM. HLA-DR4 and pulmonary dysfunction in rheumatoid arthritis. Am J Med 1987;82(4): (48) Hallowell RW, Horton MR. Interstitial lung disease in patients with rheumatoid arthritis: spontaneous and drug induced. Drugs 2014 Mar;74(4): (49) Furukawa H, Oka S, Shimada K, Sugii S, Ohashi J, Matsui T, et al. Association of Human Leukocyte Antigen with Interstitial Lung Disease in Rheumatoid Arthritis: A protective role for shared epitope. PLoS ONE 2012;7(5). (50) Mori S, Koga Y, Sugimoto M. Different risk factors between interstitial lung disease and airway disease in rheumatoid arthritis. Respir Med 2012 Nov;106(11): (51) Michalski JP, McCombs CC, Scopelitis E, Biundo JJ,Jr, Medsger TA,Jr. Alpha 1- antitrypsin phenotypes, including M subtypes, in pulmonary disease associated with rheumatoid arthritis and systemic sclerosis. Arthritis Rheum 1986 May;29(5): (52) Johnson C, Giles JT, Bathon J, Danoff SK. Exploration Of The MUC5B Polymorphism Frequency In Rheumatoid Arthritis Interstitial Lung Disease. A42. INTERSTITIAL LUNG

92 66 DISEASE: EPIDEMIOLOGY, EVALUATION AND PATHOGENESIS: American Thoracic Society; p. A5982. (53) Pantelidis P, Veeraraghavan S, du Bois RM. Surfactant gene polymorphisms and interstitial lung diseases. Respir Res 2002;3:14. (54) Liu T, Ullenbruch M, Young Choi Y, Yu H, Ding L, Xaubet A, et al. Telomerase and telomere length in pulmonary fibrosis. Am J Respir Cell Mol Biol 2013 Aug;49(2): (55) Juge PA, Borie R, Kannengiesser C, Gazal S, Revy P, Wemeau-Stervinou L, et al. Shared genetic predisposition in rheumatoid arthritis-interstitial lung disease and familial pulmonary fibrosis. Eur Respir J 2017 May 11;49(5):2016. Print 2017 May. (56) Avouac J, Gossec L, Dougados M. Diagnostic and predictive value of anti-cyclic citrullinated protein antibodies in rheumatoid arthritis: a systematic literature review. Ann Rheum Dis 2006 Jul;65(7): (57) Bongartz T, Cantaert T, Atkins SR, Harle P, Myers JL, Turesson C, et al. Citrullination in extra-articular manifestations of rheumatoid arthritis. Rheumatology (Oxford) 2007 Jan;46(1): (58) Yin Y, Liang D, Zhao L, Li Y, Liu W, Ren Y, et al. Anti-cyclic citrullinated Peptide antibody is associated with interstitial lung disease in patients with rheumatoid arthritis. PLoS One 2014 Apr 17;9(4):e (59) Inui N, Enomoto N, Suda T, Kageyama Y, Watanabe H, Chida K. Anti-cyclic citrullinated peptide antibodies in lung diseases associated with rheumatoid arthritis. Clin Biochem 2008;41(13): (60) Shaw M, Collins BF, Ho LA, Raghu G. Rheumatoid arthritis-associated lung disease. Eur Respir Rev 2015 Mar;24(135):1-16. (61) Ascherman DP. Interstitial lung disease in rheumatoid arthritis. Curr Rheumatol Rep 2010;12(5): (62) Ytterberg AJ, Joshua V, Reynisdottir G, Tarasova NK, Rutishauser D, Ossipova E, et al. Shared immunological targets in the lungs and joints of patients with rheumatoid arthritis: identification and validation. Ann Rheum Dis 2015 Sep;74(9): (63) Fischer A, Solomon JJ, Du Bois RM, Deane KD, Olson AL, Fernandez-Perez ER, et al. Lung disease with anti-ccp antibodies but not rheumatoid arthritis or connective tissue disease. Respir Med 2012;106(7): (64) Gizinski AM, Mascolo M, Loucks JL, Kervitsky A, Meehan RT, Brown KK, et al. Rheumatoid arthritis (RA)-specific autoantibodies in patients with interstitial lung disease and absence of clinically apparent articular RA. Clin Rheumatol 2009;28(5):

93 67 (65) Demoruelle MK, Deane KD, Holers VM. When and where does inflammation begin in rheumatoid arthritis? Curr Opin Rheumatol 2014 Jan;26(1): (66) Lee DM, Phillips R, Hagan EM, Chibnik LB, Costenbader KH, Schur PH. Quantifying anti-cyclic citrullinated peptide titres: clinical utility and association with tobacco exposure in patients with rheumatoid arthritis. Ann Rheum Dis 2009 Feb;68(2): (67) Makrygiannakis D, Hermansson M, Ulfgren AK, Nicholas AP, Zendman AJ, Eklund A, et al. Smoking increases peptidylarginine deiminase 2 enzyme expression in human lungs and increases citrullination in BAL cells. Ann Rheum Dis 2008 Oct;67(10): (68) Kallberg H, Ding B, Padyukov L, Bengtsson C, Ronnelid J, Klareskog L, et al. Smoking is a major preventable risk factor for rheumatoid arthritis: estimations of risks after various exposures to cigarette smoke. Ann Rheum Dis 2011 Mar;70(3): (69) Johnson C. Recent advances in the pathogenesis, prediction, and management of rheumatoid arthritis-associated interstitial lung disease. Curr Opin Rheumatol 2017 May;29(3): (70) Bergstrom U, Jacobsson LT, Nilsson JA, Berglund G, Turesson C. Pulmonary dysfunction, smoking, socioeconomic status and the risk of developing rheumatoid arthritis. Rheumatology (Oxford) 2011 Nov;50(11): (71) Saag KG, Cerhan JR, Kolluri S, Ohashi K, Hunninghake GW, Schwartz DA. Cigarette smoking and rheumatoid arthritis severity. Ann Rheum Dis 1997;56(8): (72) Doyle TJ, Patel AS, Hatabu H, Nishino M, Wu G, Osorio JC, et al. Detection of Rheumatoid Arthritis-Interstitial Lung Disease Is Enhanced by Serum Biomarkers. Am J Respir Crit Care Med 2015 Jun 15;191(12): (73) Kelly CA, Saravanan V, Nisar M, Arthanari S, Woodhead FA, Price-Forbes AN, et al. Rheumatoid arthritis-related interstitial lung disease: associations, prognostic factors and physiological and radiological characteristics--a large multicentre UK study. Rheumatology (Oxford) 2014 Sep;53(9): (74) Ayhan-Ardic FF, Oken O, Yorgancioglu ZR, Ustun N, Gokharman FD. Pulmonary involvement in lifelong non-smoking patients with rheumatoid arthritis and ankylosing spondylitis without respiratory symptoms. Clin Rheumatol 2006 Mar;25(2): (75) Cavagna L, Monti S, Grosso V, Boffini N, Scorletti E, Crepaldi G, et al. The multifaceted aspects of interstitial lung disease in rheumatoid arthritis. Biomed Res Int 2013;2013: (76) Zou YQ, Li YS, Ding XN, Ying ZH. The clinical significance of HRCT in evaluation of patients with rheumatoid arthritis-associated interstitial lung disease: a report from China. Rheumatol Int 2012 Mar;32(3): (77) Nyhall-Wahlin BM, Petersson IF, Nilsson JA, Jacobsson LT, Turesson C, BARFOT study group. High disease activity disability burden and smoking predict severe extra-

94 68 articular manifestations in early rheumatoid arthritis. Rheumatology (Oxford) 2009 Apr;48(4): (78) Turesson C, Jacobsson LT, Sturfelt G, Matteson EL, Mathsson L, Ronnelid J. Rheumatoid factor and antibodies to cyclic citrullinated peptides are associated with severe extra-articular manifestations in rheumatoid arthritis. Ann Rheum Dis 2007 Jan;66(1): (79) Sakaida H. IgG rheumatoid factor in rheumatoid arthritis with interstitial lung disease. Ryumachi 1995 Aug;35(4): (80) Yang JA, Lee JS, Park JK, Lee EB, Song YW, Lee EY. Clinical characteristics associated with occurrence and poor prognosis of interstitial lung disease in rheumatoid arthritis. Korean J Intern Med 2017 Mar 28. (81) Manfredsdottir VF, Vikingsdottir T, Jonsson T, Geirsson AJ, Kjartansson O, Heimisdottir M, et al. The effects of tobacco smoking and rheumatoid factor seropositivity on disease activity and joint damage in early rheumatoid arthritis. Rheumatology 2006;45(6): (82) Másdóttir B, Jónsson T, Manfreosdóttir V, Víkingsson A, Brekkan Á, Valdimarsson H. Smoking, rheumatoid factor isotypes and severity of rheumatoid arthritis. Rheumatology 2000;39(11): (83) Harlow L, Rosas IO, Gochuico BR, Mikuls TR, Dellaripa PF, Oddis CV, et al. Identification of citrullinated hsp90 isoforms as novel autoantigens in rheumatoid arthritisassociated interstitial lung disease. Arthritis Rheum 2013 Apr;65(4): (84) Kinoshita F, Hamano H, Harada H, Kinoshita T, Igishi T, Hagino H, et al. Role of KL-6 in evaluating the disease severity of rheumatoid lung disease: comparison with HRCT. Respir Med 2004 Nov;98(11): (85) Oyama T, Kohno N, Yokoyama A, Hirasawa Y, Hiwada K, Oyama H, et al. Detection of interstitial pneumonitis in patients with rheumatoid arthritis by measuring circulating levels of KL-6, a human MUC1 mucin. Lung 1997;175(6): (86) Sato S, Nagaoka T, Hasegawa M, Nishijima C, Takehara K. Elevated serum KL-6 levels in patients with systemic sclerosis: association with the severity of pulmonary fibrosis. Dermatology 2000;200(3): (87) Chen J, Doyle TJ, Liu Y, Aggarwal R, Wang X, Shi Y, et al. Biomarkers of rheumatoid arthritis--associated interstitial lung disease. Arthritis Rheumatol 2015 Jan;67(1): (88) Wang JX, Du CG. A retrospective study of clinical characteristics of interstitial lung disease associated with rheumatoid arthritis in Chinese patients. Med Sci Monit 2015 Mar 7;21: (89) Wang T, Zheng XJ, Liang BM, Liang ZA. Clinical features of rheumatoid arthritisassociated interstitial lung disease. Sci Rep 2015 Oct 7;5:14897.

95 69 (90) Song ST, Kim SS, Kim JY, Lee SY, Kim K, Kwon IS, et al. Association of Single Nucleotide Polymorphisms of PADI4 and HLA-DRB1 Alleles with Susceptibility to Rheumatoid Arthritis-Related Lung Diseases. Lung 2016 Oct;194(5): (91) Akiyama M, Kaneko Y, Yamaoka K, Kondo H, Takeuchi T. Association of disease activity with acute exacerbation of interstitial lung disease during tocilizumab treatment in patients with rheumatoid arthritis: a retrospective, case-control study. Rheumatol Int 2016 Jun;36(6): (92) Rocha-Munoz AD, Ponce-Guarneros M, Gamez-Nava JI, Olivas-Flores EM, Mejia M, Juarez-Contreras P, et al. Anti-Cyclic Citrullinated Peptide Antibodies and Severity of Interstitial Lung Disease in Women with Rheumatoid Arthritis. J Immunol Res 2015;2015: (93) Wang T, Zheng XJ, Ji YL, Liang ZA, Liang BM. Tumour markers in rheumatoid arthritis-associated interstitial lung disease. Clin Exp Rheumatol 2016;34(4): (94) Biederer J, Schnabel A, Muhle C, Gross WL, Heller M, Reuter M. Correlation between HRCT findings, pulmonary function tests and bronchoalveolar lavage cytology in interstitial lung disease associated with rheumatoid arthritis. Eur Radiol 2004 Feb;14(2): (95) Kim EJ, Elicker BM, Maldonado F, Webb WR, Ryu JH, Van Uden JH, et al. Usual interstitial pneumonia in rheumatoid arthritis-associated interstitial lung disease. Eur Respir J 2010 Jun;35(6): (96) Lee HK, Kim DS, Yoo B, Seo JB, Rho JY, Colby TV, et al. Histopathologic pattern and clinical features of rheumatoid arthritis-associated interstitial lung disease. Chest 2005 Jun;127(6): (97) Kono M, Nakamura Y, Enomoto N, Hashimoto D, Fujisawa T, Inui N, et al. Usual interstitial pneumonia preceding collagen vascular disease: a retrospective case control study of patients initially diagnosed with idiopathic pulmonary fibrosis. PLoS One 2014 Apr 15;9(4):e (98) Yunt ZX, Solomon JJ. Lung disease in rheumatoid arthritis. Rheum Dis Clin North Am 2015 May;41(2): (99) Pappas DA, Giles JT, Connors G, Lechtzin N, Bathon JM, Danoff SK. Respiratory symptoms and disease characteristics as predictors of pulmonary function abnormalities in patients with rheumatoid arthritis: an observational cohort study. Arthritis Res Ther 2010;12(3):R104. (100) Rajasekaran BA, Shovlin D, Lord P, Kelly CA. Interstitial lung disease in patients with rheumatoid arthritis: A comparison with cryptogenic fibrosing alveolitis. Rheumatology 2001;40(9): (101) Brown KK. Rheumatoid lung disease. Proc Am Thorac Soc 2007 Aug 15;4(5):

96 70 (102) Hamblin MJ, Horton MR. Rheumatoid arthritis-associated interstitial lung disease: diagnostic dilemma. Pulm Med 2011;2011: (103) Iqbal K, Kelly C. Treatment of rheumatoid arthritis-associated interstitial lung disease: a perspective review. Therapeutic Advances in Musculoskeletal Disease 2015;7(6): (104) Garcia JG, Parhami N, Killam D, Garcia PL, Keogh BA. Bronchoalveolar lavage fluid evaluation in rheumatoid arthritis. Am Rev Respir Dis 1986 Mar;133(3): (105) Okuda Y, Takasugi K, Kurata N, Takahara J. [Bronchoalveolar lavage fluid analysis in rheumatoid arthritis]. Ryumachi 1993 Aug;33(4): (106) Perez T, Farre JM, Gosset P, Wallaert B, Duquesnoy B, Voisin C, et al. Subclinical alveolar inflammation in rheumatoid arthritis: superoxide anion, neutrophil chemotactic activity and fibronectin generation by alveolar macrophages. Eur Respir J 1989 Jan;2(1):7-13. (107) Meyer KC, Raghu G, Baughman RP, Brown KK, Costabel U, du Bois RM, et al. An official American Thoracic Society clinical practice guideline: the clinical utility of bronchoalveolar lavage cellular analysis in interstitial lung disease. Am J Respir Crit Care Med 2012 May 01,;185(9): (108) De Lauretis A, Veeraraghavan S, Renzoni E. Connective tissue disease-associated interstitial lung disease: How does it differ from IPF? How should the clinical approach differ? Chronic Respir Dis 2011;8(1): (109) Manjunatha YC, Seith A, Kandpal H, Das CJ. Rheumatoid arthritis: spectrum of computed tomographic findings in pulmonary diseases. Curr Probl Diagn Radiol 2010 Nov-Dec;39(6): (110) Park JH, Kim DS, Park IN, Jang SJ, Kitaichi M, Nicholson AG, et al. Prognosis of fibrotic interstitial pneumonia: idiopathic versus collagen vascular disease-related subtypes. Am J Respir Crit Care Med 2007 Apr 1;175(7): (111) Kim EJ, Collard HR, King TE,Jr. Rheumatoid arthritis-associated interstitial lung disease: the relevance of histopathologic and radiographic pattern. Chest 2009 Nov;136(5): (112) Raghu G, Mageto YN, Lockhart D, Schmidt RA, Wood DE, Godwin JD. The accuracy of the clinical diagnosis of new-onset idiopathic pulmonary fibrosis and other interstitial lung disease: A prospective study. Chest 1999;116(5): (113) Assayag D, Elicker BM, Urbania TH, Colby TV, Kang BH, Ryu JH, et al. Rheumatoid arthritis-associated interstitial lung disease: radiologic identification of usual interstitial pneumonia pattern. Radiology 2014 Feb;270(2): (114) Hunninghake GW, Lynch DA, Galvin JR, Gross BH, Müller N, Schwartz DA, et al. Radiologic Findings Are Strongly Associated with a Pathologic Diagnosis of Usual Interstitial Pneumonia. Chest 2003;124(4):

97 71 (115) Flaherty KR, Thwaite EL, Kazerooni EA, Gross BH, Toews GB, Colby TV, et al. Radiological versus histological diagnosis in UIP and NSIP: survival implications. Thorax 2003 Feb;58(2): (116) Hozumi H, Nakamura Y, Johkoh T, Sumikawa H, Colby TV, Kono M, et al. Acute exacerbation in rheumatoid arthritis-associated interstitial lung disease: a retrospective case control study. BMJ Open 2013 Sep 13;3(9): (117) Flaherty KR, King TE, Raghu G, Lynch JP, Colby TV, Travis WD, et al. Idiopathic interstitial pneumonia: what is the effect of a multidisciplinary approach to diagnosis? Am J Respir Crit Care Med 2004 Oct 15,;170(8): (118) Travis WD, Hunninghake G, King TE, Lynch DA, Colby TV, Galvin JR, et al. Idiopathic nonspecific interstitial pneumonia: report of an American Thoracic Society project. Am J Respir Crit Care Med 2008 Jun 15,;177(12): (119) Suda T. Up-to-Date Information on Rheumatoid Arthritis-Associated Interstitial Lung Disease. Clin Med Insights Circ Respir Pulm Med 2016 May 31;9(Suppl 1): (120) Hartman TE, Swensen SJ, Hansell DM, Colby TV, Myers JL, Tazelaar HD, et al. Nonspecific interstitial pneumonia: Variable appearance at high-resolution chest CT. Radiology 2000;217(3): (121) Kim SJ, Lee KS, Ryu YH, Yoon YC, Choe KO, Kim TS, et al. Reversed halo sign on high-resolution CT of cryptogenic organizing pneumonia: diagnostic implications. AJR Am J Roentgenol 2003 May;180(5): (122) Doyle TJ, Dellaripa PF. Lung Manifestations in the Rheumatic Diseases. Chest 2017 May 25,. (123) Rees JH, Woodhead MA, Sheppard MN, du Bois RM. Rheumatoid arthritis and cryptogenic organising pneumonitis. Respir Med 1991 May;85(3): (124) Johkoh T, Müller NL, Taniguchi H, Kondoh Y, Akira M, Ichikado K, et al. Acute interstitial pneumonia: thin-section CT findings in 36 patients. Radiology 1999 Jun;211(3): (125) Gruden JF, Webb WR. CT findings in a proved case of respiratory bronchiolitis. AJR Am J Roentgenol 1993 Jul;161(1): (126) Hartman TE, Primack SL, Swensen SJ, Hansell D, McGuinness G, Müller NL. Desquamative interstitial pneumonia: thin-section CT findings in 22 patients. Radiology 1993 Jun;187(3): (127) Ichikawa Y, Kinoshita M, Koga T, Oizumi K, Fujimoto K, Hayabuchi N. Lung cyst formation in lymphocytic interstitial pneumonia: CT features. J Comput Assist Tomogr 1994 Sep-Oct;18(5):

98 72 (128) Tomassetti S, Cavazza A, Colby TV, Ryu JH, Nanni O, Scarpi E, et al. Transbronchial biopsy is useful in predicting UIP pattern. Respir Res 2012 Oct 29,;13:96. (129) Kaarteenaho R. The current position of surgical lung biopsy in the diagnosis of idiopathic pulmonary fibrosis. Respir Res 2013 Apr 15;14:43. (130) Hutchinson JP, Fogarty AW, McKeever TM, Hubbard RB. In-Hospital Mortality after Surgical Lung Biopsy for Interstitial Lung Disease in the United States to Am J Respir Crit Care Med 2016 May 15,;193(10): (131) Utz JP, Ryu JH, Douglas WW, Hartman TE, Tazelaar HD, Myers JL, et al. High shortterm mortality following lung biopsy for usual interstitial pneumonia. Eur Respir J 2001 Feb;17(2): (132) Kondoh Y, Taniguchi H, Kitaichi M, Yokoi T, Johkoh T, Oishi T, et al. Acute exacerbation of interstitial pneumonia following surgical lung biopsy. Respiratory medicine 2006 November 1,;100: (133) Park I-, Kim DS, Shim TS, Lim C-, Lee SD, Koh Y, et al. Acute exacerbation of interstitial pneumonia other than idiopathic pulmonary fibrosis. Chest 2007;132(1): (134) Ghatol A, Ruhl AP, Danoff SK. Exacerbations in Idiopathic Pulmonary Fibrosis Triggered by Pulmonary and Nonpulmonary Surgery: A Case Series and Comprehensive Review of the Literature. Lung 2012 /08/01;190(4): (135) Lentz RJ, Argento AC, Colby TV, Rickman OB, Maldonado F. Transbronchial cryobiopsy for diffuse parenchymal lung disease: a state-of-the-art review of procedural techniques, current evidence, and future challenges. J Thorac Dis 2017 Jul;9(7): (136) Tomassetti S, Wells AU, Costabel U, Cavazza A, Colby TV, Rossi G, et al. Bronchoscopic Lung Cryobiopsy Increases Diagnostic Confidence in the Multidisciplinary Diagnosis of Idiopathic Pulmonary Fibrosis. Am J Respir Crit Care Med 2016 Apr 01,;193(7): (137) Ravaglia C, Bonifazi M, Wells AU, Tomassetti S, Gurioli C, Piciucchi S, et al. Safety and Diagnostic Yield of Transbronchial Lung Cryobiopsy in Diffuse Parenchymal Lung Diseases: A Comparative Study versus Video-Assisted Thoracoscopic Lung Biopsy and a Systematic Review of the Literature. Respiration 2016;91(3): (138) Fujii M, Adachi S, Shimizu T, Hirota S, Sako M, Kono M. Interstitial lung disease in rheumatoid arthritis: assessment with high-resolution computed tomography. J Thorac Imaging 1993;8(1): (139) Song JW, Lee HK, Lee Et Al, C K. Clinical course and outcome of rheumatoid arthritisrelated usual interstitial pneumonia. Sarcoidosis Vasc Diffuse Lung Dis 2013 Aug 1;30(2):

99 73 (140) Zamora-Legoff JA, Krause ML, Crowson CS, Ryu JH, Matteson EL. Progressive Decline of Lung Function in Rheumatoid Arthritis-Associated Interstitial Lung Disease. Arthritis Rheumatol 2017 Mar;69(3): (141) Hakala M. Poor prognosis in patients with rheumatoid arthritis hospitalized for interstitial lung fibrosis. Chest 1988 Jan;93(1): (142) Flaherty KR, Colby TV, Travis WD, Toews GB, Mumford J, Murray S, et al. Fibroblastic foci in usual interstitial pneumonia: idiopathic versus collagen vascular disease. Am J Respir Crit Care Med 2003 May 15,;167(10): (143) Navaratnam V, Ali N, Smith CJ, McKeever T, Fogarty A, Hubbard RB. Does the presence of connective tissue disease modify survival in patients with pulmonary fibrosis? Respir Med 2011 Dec;105(12): (144) Hubbard R, Venn A. The impact of coexisting connective tissue disease on survival in patients with fibrosing alveolitis. Rheumatology (Oxford) 2002 Jun;41(6): (145) Kocheril SV, Appleton BE, Somers EC, Kazerooni EA, Flaherty KR, Martinez FJ, et al. Comparison of disease progression and mortality of connective tissue disease-related interstitial lung disease and idiopathic interstitial pneumonia. Arthritis Rheum 2005 Aug 15;53(4): (146) Solomon JJ, Ryu JH, Tazelaar HD, Myers JL, Tuder R, Cool CD, et al. Fibrosing interstitial pneumonia predicts survival in patients with rheumatoid arthritis-associated interstitial lung disease (RA-ILD). Respir Med 2013 Aug;107(8): (147) Rajasekaran A, Shovlin D, Saravanan V, Lord P, Kelly C. Interstitial lung disease in patients with rheumatoid arthritis: Comparison with cryptogenic fibrosing alveolitis over 5 years. J Rheumatol 2006;33(7): (148) Tsuchiya Y, Takayanagi N, Sugiura H, Miyahara Y, Tokunaga D, Kawabata Y, et al. Lung diseases directly associated with rheumatoid arthritis and their relationship to outcome. Eur Respir J 2011;37(6): (149) Collard HR, Ryerson CJ, Corte TJ, Jenkins G, Kondoh Y, Lederer DJ, et al. Acute Exacerbation of Idiopathic Pulmonary Fibrosis. An International Working Group Report. Am J Respir Crit Care Med 2016 Aug 01,;194(3): (150) Churg A, Wright JL, Tazelaar HD. Acute exacerbations of fibrotic interstitial lung disease. Histopathology 2011 Mar;58(4): (151) Suda T, Kaida Y, Nakamura Y, Enomoto N, Fujisawa T, Imokawa S, et al. Acute exacerbation of interstitial pneumonia associated with collagen vascular diseases. Respir Med 2009;103(6):

100 74 (152) Kim DS, Park JH, Park BK, Lee JS, Nicholson AG, Colby T. Acute exacerbation of idiopathic pulmonary fibrosis: frequency and clinical features. Eur Respir J 2006 Jan;27(1): (153) Solomon JJ, Chung JH, Cosgrove GP, Demoruelle MK, Fernandez-Perez ER, Fischer A, et al. Predictors of mortality in rheumatoid arthritis-associated interstitial lung disease. Eur Respir J 2016 Feb;47(2): (154) Lee H, Choe J, Kim S, Park SH, Kim JH, Hyun DS, et al. Important Prognostic Factor in Rheumatoid Arthritis Patients with Interstitial Lung Disease Is Not Usual Interstitial Pneumonia Pattern but Interstitial Lung Disease Extent On Chest High-Resolution Computed Tomography. Arthritis Rheum 2012 OCT;64(10):S528. (155) Sathi N, Urwin T, Desmond S, Dawson JK. Patients with limited rheumatoid arthritisrelated interstitial lung disease have a better prognosis than those with extensive disease. Rheumatology 2011;50(3):620. (156) Moore OA, Goh N, Corte T, Rouse H, Hennessy O, Thakkar V, et al. Extent of disease on high-resolution computed tomography lung is a predictor of decline and mortality in systemic sclerosis-related interstitial lung disease. Rheumatology (United Kingdom) 2013;52(1): (157) Hyldgaard C, Hilberg O, Muller A, Bendstrup E. A cohort study of interstitial lung diseases in central Denmark. Respir Med 2014 May;108(5): (158) Shin KM, Lee KS, Chung MP, Han J, Bae YA, Kim TS, et al. Prognostic determinants among clinical, thin-section CT, and histopathologic findings for fibrotic idiopathic interstitial pneumonias: tertiary hospital study. Radiology 2008 Oct;249(1): (159) Walsh SLF, Sverzellati N, Devaraj A, Keir GJ, Wells AU, Hansell DM. Connective tissue disease related fibrotic lung disease: High resolution computed tomographic and pulmonary function indices as prognostic determinants. Thorax 2014;69(3): (160) Yousem SA, Colby TV, Carrington CB. Lung biopsy in rheumatoid arthritis. Am Rev Respir Dis 1985 May;131(5): (161) Nakamura Y, Suda T, Kaida Y, Kono M, Hozumi H, Hashimoto D, et al. Rheumatoid lung disease: prognostic analysis of 54 biopsy-proven cases. Respir Med 2012 Aug;106(8): (162) Moua T, Zamora Martinez A, Baqir M, Vassallo R, Limper AH, Ryu JH. Predictors of diagnosis and survival in idiopathic pulmonary fibrosis and connective tissue diseaserelated usual interstitial pneumonia. Respir Res 2014 Dec 4;15(1):154. (163) Hakala M, Pääkkö P, Huhti E, Tarkka M, Sutinen S. Open lung biopsy of patients with rheumatoid arthritis. Clin Rheumatol 1990 Dec;9(4):

101 75 (164) Yoshinouchi T, Ohtsuki Y, Fujita J, Yamadori I, Bandoh S, Ishida T, et al. Nonspecific interstitial pneumonia pattern as pulmonary involvement of rheumatoid arthritis. Rheumatol Int 2005;26(2): (165) Jegal Y, Kim DS, Shim TS, Lim CM, Do Lee S, Koh Y, et al. Physiology is a stronger predictor of survival than pathology in fibrotic interstitial pneumonia. Am J Respir Crit Care Med 2005 Mar 15;171(6): (166) Flaherty KR, Andrei AC, Murray S, Fraley C, Colby TV, Travis WD, et al. Idiopathic pulmonary fibrosis: prognostic value of changes in physiology and six-minute-walk test. Am J Respir Crit Care Med 2006 Oct 1;174(7): (167) Collard HR, King TE,Jr, Bartelson BB, Vourlekis JS, Schwarz MI, Brown KK. Changes in clinical and physiologic variables predict survival in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2003 Sep 1;168(5): (168) Caminati A, Bianchi A, Cassandro R, Mirenda MR, Harari S. Walking distance on 6- MWT is a prognostic factor in idiopathic pulmonary fibrosis. Respir Med 2009 Jan;103(1): (169) du Bois RM, Weycker D, Albera C, Bradford WZ, Costabel U, Kartashov A, et al. Sixminute-walk test in idiopathic pulmonary fibrosis: test validation and minimal clinically important difference. Am J Respir Crit Care Med 2011 May 01,;183(9): (170) Dixon WG, Hyrich KL, Watson KD, Lunt M, BSRBR Control Centre Consortium, Symmons DP, et al. Influence of anti-tnf therapy on mortality in patients with rheumatoid arthritis-associated interstitial lung disease: results from the British Society for Rheumatology Biologics Register. Ann Rheum Dis 2010 Jun;69(6): (171) Wolfe F, Caplan L, Michaud K. Rheumatoid arthritis treatment and the risk of severe interstitial lung disease. Scand J Rheumatol 2007;36(3): (172) King TE,Jr, Tooze JA, Schwarz MI, Brown KR, Cherniack RM. Predicting survival in idiopathic pulmonary fibrosis: scoring system and survival model. Am J Respir Crit Care Med 2001 Oct 1;164(7): (173) Wells AU, Desai SR, Rubens MB, Goh NS, Cramer D, Nicholson AG, et al. Idiopathic pulmonary fibrosis: a composite physiologic index derived from disease extent observed by computed tomography. Am J Respir Crit Care Med 2003 Apr 1;167(7): (174) Goh NS, Desai SR, Veeraraghavan S, Hansell DM, Copley SJ, Maher TM, et al. Interstitial lung disease in systemic sclerosis: a simple staging system. Am J Respir Crit Care Med 2008 Jun 1;177(11): (175) du Bois RM, Weycker D, Albera C, Bradford WZ, Costabel U, Kartashov A, et al. Ascertainment of individual risk of mortality for patients with idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2011 Aug 15;184(4):

102 76 (176) Mura M, Porretta MA, Bargagli E, Sergiacomi G, Zompatori M, Sverzellati N, et al. Predicting survival in newly diagnosed idiopathic pulmonary fibrosis: a 3-year prospective study. Eur Respir J 2012 Jul;40(1): (177) Ley B, Ryerson CJ, Vittinghoff E, Ryu JH, Tomassetti S, Lee JS, et al. A multidimensional index and staging system for idiopathic pulmonary fibrosis. Ann Intern Med 2012 May 15;156(10): (178) Ryerson CJ, Vittinghoff E, Ley B, Lee JS, Mooney JJ, Jones KD, et al. Predicting survival across chronic interstitial lung disease: the ILD-GAP model. Chest 2014 Apr;145(4): (179) Hyldgaard C, Hilberg O, Pedersen AB, Ulrichsen SP, Lokke A, Bendstrup E, et al. A population-based cohort study of rheumatoid arthritis-associated interstitial lung disease: comorbidity and mortality. Ann Rheum Dis 2017 Jun 13. (180) Anne Kerola. Epidemiology of Comorbidities in Early Rheumatoid Arthritis University of Helsinki; (181) Kelly C, Hamilton J. What kills patients with rheumatoid arthritis? Rheumatology (Oxford) 2007 Feb;46(2): (182) Young A, Koduri G, Batley M, Kulinskaya E, Gough A, Norton S, et al. Mortality in rheumatoid arthritis. Increased in the early course of disease, in ischaemic heart disease and in pulmonary fibrosis. Rheumatology (UK) 2007;46(2): (183) Mori S. Management of Rheumatoid Arthritis Patients with Interstitial Lung Disease: Safety of Biological Antirheumatic Drugs and Assessment of Pulmonary Fibrosis. Clin Med Insights Circ Respir Pulm Med 2015 Sep 8;9(Suppl 1): (184) Wallace B, Vummidi D, Khanna D. Management of connective tissue diseases associated interstitial lung disease: a review of the published literature. Curr Opin Rheumatol 2016 May;28(3): (185) Robles-Perez A, Molina-Molina M. Treatment Considerations of Lung Involvement in Rheumatologic Disease. Respiration 2015;90(4): (186) Kelly C, Saravanan V. Treatment strategies for a rheumatoid arthritis patient with interstitial lung disease. Expert Opin Pharmacother 2008 Dec;9(18): (187) Mori S, Cho I, Koga Y, Sugimoto M. A simultaneous onset of organizing pneumonia and rheumatoid arthritis, along with a review of the literature. Mod Rheumatol 2008;18(1): (188) Hanada M, Sakamoto N, Ishimatsu Y, Kakugawa T, Obase Y, Kozu R, et al. Effect of long-term treatment with corticosteroids on skeletal muscle strength, functional exercise capacity and health status in patients with interstitial lung disease. Respirology 2016 Aug;21(6):

103 77 (189) Zamora-Legoff JA, Krause ML, Crowson CS, Ryu JH, Matteson EL. Risk of serious infection in patients with rheumatoid arthritis-associated interstitial lung disease. Clin Rheumatol 2016 Oct;35(10): (190) Tashkin DP, Elashoff R, Clements PJ, Roth MD, Furst DE, Silver RM, et al. Effects of 1- year treatment with cyclophosphamide on outcomes at 2 years in scleroderma lung disease. Am J Respir Crit Care Med 2007 Nov 15,;176(10): (191) Nannini C, West CP, Erwin PJ, Matteson EL. Effects of cyclophosphamide on pulmonary function in patients with scleroderma and interstitial lung disease: a systematic review and meta-analysis of randomized controlled trials and observational prospective cohort studies. Arthritis Res Ther 2008;10(5):R124. (192) Fischer A, Brown KK, Du Bois RM, Frankel SK, Cosgrove GP, Fernandez-Perez ER, et al. Mycophenolate mofetil improves lung function in connective tissue disease-associated interstitial lung disease. J Rheumatol 2013;40(5): (193) Saketkoo LA, Espinoza LR. Experience of mycophenolate mofetil in 10 patients with autoimmune-related interstitial lung disease demonstrates promising effects. Am J Med Sci 2009 May;337(5): (194) Rojas-Serrano J, Gonzalez-Velasquez E, Mejia M, Sanchez-Rodriguez A, Carrillo G. Interstitial lung disease related to rheumatoid arthritis: evolution after treatment. Reumatol Clin 2012;8(2): (195) Suissa S, Hudson M, Ernst P. Leflunomide use and the risk of interstitial lung disease in rheumatoid arthritis. Arthritis Rheum 2006;54(5): (196) Rutanen J, Kononoff A, Arstila L, Elfving P, Koskela H, Kaipiainen-Seppänen O. Five cases of interstitial lung disease after leflunomide was combined with methotrexate therapy. Scand J Rheumatol 2014;43(3): (197) Puttick MPE, Klinkhoff AV, Chalmers A, Ostrow DN. Treatment of progressive rheumatoid interstitial lung disease with cyclosporine. J Rheumatol 1995;22(11): (198) Tokano Y, Ogasawara H, Ando S, Fujii T, Kaneko H, Tamura N, et al. Cyclosporin A therapy for interstitial pneumonitis associated with rheumatic disease. Mod Rheumatol 2002 Dec;12(4): (199) Ostor AJK, Crisp AJ, Somerville MF, Scott DGI. Fatal exacerbation of rheumatoid arthritis associated fibrosing alveolitis in patients given infliximab. Br Med J 2004;329(7477):1266. (200) Jani M, Hirani N, Matteson EL, Dixon WG. The safety of biologic therapies in RAassociated interstitial lung disease. Nat Rev Rheumatol 2014 May;10(5):

104 78 (201) Perez-Alvarez R, Perez-de-Lis M, Diaz-Lagares C, Pego-Reigosa JM, Retamozo S, Bove A, et al. Interstitial Lung Disease Induced or Exacerbated by TNF-Targeted Therapies: Analysis of 122 Cases. Semin Arthritis Rheum 2011;41(2): (202) Bargagli E, Galeazzi M, Rottoli P. Infliximab treatment in a patient with rheumatoid arthritis and pulmonary fibrosis [1]. European Respiratory Journal 2004;24(4):708. (203) Vassallo R, Matteson E, Thomas Jr. CF. Clinical response of rheumatoid arthritisassociated pulmonary fibrosis to tumor necrosis factor-a inhibition. Chest 2002;122(3): (204) Hagiwara K, Sato T, Takagi-Kobayashi S, Hasegawa S, Shigihara N, Akiyama O. Acute exacerbation of preexisting interstitial lung disease after administration of etanercept for rheumatoid arthritis. J Rheumatol 2007 May;34(5): (205) Detorakis EE, Magkanas E, Lasithiotaki I, Sidiropoulos P, Boumpas DT, Gourtsoyiannis N, et al. Evolution of imaging findings, laboratory and functional parameters in rheumatoid arthritis patients after one year of treatment with anti-tnf-α agents. Clin Exp Rheumatol 2017 Jan-Feb;35(1): (206) Mohr M, Jacobi AM. Interstitial lung disease in rheumatoid arthritis: response to IL- 6R blockade. Scand J Rheumatol 2011;40(5): (207) Kawashiri SY, Kawakami A, Sakamoto N, Ishimatsu Y, Eguchi K. A fatal case of acute exacerbation of interstitial lung disease in a patient with rheumatoid arthritis during treatment with tocilizumab. Rheumatol Int 2012 Dec;32(12): (208) Hadjinicolaou AV, Nisar MK, Bhagat S, Parfrey H, Chilvers ER, Östör AJK. Noninfectious pulmonary complications of newer biological agents for rheumatic diseases-a systematic literature review. Rheumatology (UK) 2011;50(12): (209) Mera-Varela A, Pérez-Pampín E. Abatacept therapy in rheumatoid arthritis with interstitial lung disease. J Clin Rheumatol 2014 Dec;20(8): (210) Wada T, Akiyama Y, Yokota K, Sato K, Funakubo Y, Mimura T. [A case of rheumatoid arthritis complicated with deteriorated interstitial pneumonia after the administration of abatacept]. Nihon Rinsho Meneki Gakkai Kaishi 2012;35(5): (211) Weinblatt ME, Moreland LW, Westhovens R, Cohen RB, Kelly SM, Khan N, et al. Safety of abatacept administered intravenously in treatment of rheumatoid arthritis: integrated analyses of up to 8 years of treatment from the abatacept clinical trial program. J Rheumatol 2013 Jun;40(6): (212) Keir GJ, Maher TM, Ming D, Abdullah R, De Lauretis A, Wickremasinghe M, et al. Rituximab in severe, treatment-refractory interstitial lung disease. Respirology 2014;19(3):

105 79 (213) Matteson EL, Bongartz T, Ryu JH, Crowson CS, Hartman TE, Dellaripa PF. Open- Label, Pilot Study of the Safety and Clinical Effects of Rituximab in Patients with Rheumatoid Arthritis-Associated Interstitial Pneumonia. Open Journal of Rheumatology and Autoimmune Diseases ;02(03):53. (214) Md Yusof MY, Kabia A, Darby M, Lettieri G, Beirne P, Vital EM, et al. Effect of rituximab on the progression of rheumatoid arthritis-related interstitial lung disease: 10 years' experience at a single centre. Rheumatology (Oxford) 2017 August 01;56(8): (215) Dowman LM, McDonald CF, Hill CJ, Lee AL, Barker K, Boote C, et al. The evidence of benefits of exercise training in interstitial lung disease: a randomised controlled trial. Thorax 2017 Jul;72(7): (216) Yazdani A, Singer LG, Strand V, Gelber AC, Williams L, Mittoo S. Survival and quality of life in rheumatoid arthritis-associated interstitial lung disease after lung transplantation. J Heart Lung Transplant 2014 May;33(5): (217) Courtwright AM, El-Chemaly S, Dellaripa PF, Goldberg HJ. Survival and outcomes after lung transplantation for non-scleroderma connective tissue-related interstitial lung disease. J Heart Lung Transplant 2017 Jul;36(7): (218) Takagishi T, Ostrowski R, Alex C, Rychlik K, Pelletiere K, Tehrani R. Survival and extrapulmonary course of connective tissue disease after lung transplantation. J Clin Rheumatol 2012 Sep;18(6): (219) Halme M. Interstitiaaliset keuhkosairaudet - Keuhkonsiirto interstitiaalisen keuhkosairauden hoitona. Suomen Lääkärilehti 2017;72(38): (220) Skold CM, Bendstrup E, Myllarniemi M, Gudmundsson G, Sjaheim T, Hilberg O, et al. Treatment of idiopathic pulmonary fibrosis: a position paper from a Nordic expert group. J Intern Med 2017 Feb;281(2): (221) Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31(3): (222) Viljanen AA. Reference values for spirometric, pulmonary diffusing capacity and body pletysmographic studies. Scan J Clin Invest 1982;42(suppl. 159):1-50. (223) Hansell DM, Bankier AA, MacMahon H, McLoud TC, Muller NL, Remy J. Fleischner Society: glossary of terms for thoracic imaging. Radiology 2008 Mar;246(3): (224) Bartels CM, Bell CL, Shinki K, Rosenthal A, Bridges AJ. Changing trends in serious extra-articular manifestations of rheumatoid arthritis among United State veterans over 20 years. Rheumatology (Oxford, England) 2010 Sep;49(9):

106 80 (225) Morisset J, Vittinghoff E, Lee BY, Tonelli R, Hu X, Elicker BM, et al. The performance of the GAP model in patients with rheumatoid arthritis associated interstitial lung disease. Respir Med 2017 Jun;127: (226) Crowson CS, Matteson EL, Myasoedova E, Michet CJ, Ernste FC, Warrington KJ, et al. The lifetime risk of adult-onset rheumatoid arthritis and other inflammatory autoimmune rheumatic diseases. Arthritis Rheum 2011 Mar;63(3): (227) Yunt ZX, Chung JH, Hobbs S, Fernandez-Perez ER, Olson AL, Huie TJ, et al. High resolution computed tomography pattern of usual interstitial pneumonia in rheumatoid arthritis-associated interstitial lung disease: Relationship to survival. Respir Med 2017 May;126: (228) Strand MJ, Sprunger D, Cosgrove GP, Fernandez-Perez ER, Frankel SK, Huie TJ, et al. Pulmonary function and survival in idiopathic vs secondary usual interstitial pneumonia. Chest 2014 Sep;146(3): (229) Kärkkäinen M, Kettunen H, Nurmi H, Selander T, Purokivi M, Kaarteenaho R. Effect of smoking and comorbidities on survival in idiopathic pulmonary fibrosis. Respir Res 2017 Aug 22,;18(1):160. (230) Antoniou KM, Walsh SL, Hansell DM, Rubens MR, Marten K, Tennant R, et al. Smoking-related emphysema is associated with idiopathic pulmonary fibrosis and rheumatoid lung. Respirology 2013 Nov;18(8): (231) Edey AJ, Devaraj AA, Barker RP, Nicholson AG, Wells AU, Hansell DM. Fibrotic idiopathic interstitial pneumonias: HRCT findings that predict mortality. Eur Radiol 2011;21(8): (232) Sumikawa H, Johkoh T, Colby TV, Ichikado K, Suga M, Taniguchi H, et al. Computed tomography findings in pathological usual interstitial pneumonia: Relationship to survival. Am J Respir Crit Care Med 2008;177(4): (233) Kishaba T, Shimaoka Y, Fukuyama H, Nagano H, Nei Y, Yamashiro S, et al. Clinical characteristics of idiopathic pulmonary fibrosis patients with gender, age, and physiology staging at Okinawa Chubu Hospital. J Thorac Dis 2015 May;7(5): (234) Kim ES, Choi SM, Lee J, Park YS, Lee CH, Yim JJ, et al. Validation of the GAP Score in Korean Patients With Idiopathic Pulmonary Fibrosis. Chest 2015 Feb 1;147(2): (235) Ryerson CJ, O'Connor D, Dunne JV, Schooley F, Hague CJ, Murphy D, et al. Predicting Mortality in Systemic Sclerosis-Associated Interstitial Lung Disease Using Risk Prediction Models Derived From Idiopathic Pulmonary Fibrosis. Chest 2015 Nov 1;148(5): (236) Nurmi HM, Purokivi MK, Karkkainen MS, Kettunen HP, Selander TA, Kaarteenaho RL. Are risk predicting models useful for estimating survival of patients with rheumatoid arthritis-associated interstitial lung disease? BMC Pulm Med 2017 Jan 13;17(1):2.

107 81 (237) Zamora-Legoff JA, Krause ML, Crowson CS, Ryu JH, Matteson EL. Patterns of interstitial lung disease and mortality in rheumatoid arthritis. Rheumatology (Oxford) 2017 Mar 1;56(3): (238) Win T, Screaton NJ, Porter JC, Ganeshan B, Maher TM, Fraioli F, et al. Pulmonary 18F- FDG uptake helps refine current risk stratification in idiopathic pulmonary fibrosis (IPF). Eur J Nucl Med Mol Imaging 2018 Jan 16,.

108 82

109 ORIGINAL PUBLICATIONS (I-III)

110

111 I Variable course of disease of rheumatoid arthritis-associated usual interstitial pneumonia compared to other subtypes. Nurmi H, Purokivi M, Kärkkäinen M, Kettunen H-P, Selander T, Kaarteenaho R. BMC Pulm Med 16: , Reprinted with the kind permission of BMC Pulmonary Medicine. The original article is an open access article distributed under the terms of Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

112

113 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 DOI /s RESEARCH ARTICLE Open Access Variable course of disease of rheumatoid arthritis-associated usual interstitial pneumonia compared to other subtypes Hanna M. Nurmi 1,2*, Minna K. Purokivi 1, Miia S. Kärkkäinen 2,4, Hannu-Pekka Kettunen 5, Tuomas A. Selander 6 and Riitta L. Kaarteenaho 1,2,3 Abstract Background: In rheumatoid arthritis-associated interstitial lung disease (RA-ILD), occurring in 10 % of patients with patients with RA, usual interstitial pattern (UIP) has shown to associate with poor prognosis but more detailed data about the course of the disease in different subtypes is limited. Our aim was to compare the disease course of patients with RA-ILD categorized into either UIP or other types of ILDs. Methods: Clinical and radiological information of 59 patients with RA-ILD were re-assessed and re-classified into UIP or non-uip groups, followed by a between-group comparison of demographic data, lung function, survival, cause of death and comorbidities. Results: The majority of patients (n = 35/59.3 %) showed a radiological UIP-like pattern in high resolution computed tomography. The median survival was 92 months (95 % CI ) in the UIP-group and 137 months (95 % CI ) in the non-uip-group (p = 0.417). Differences in course of disease were found in the number of hospitalizations for respiratory reasons (mean 1.9 ± 2.6 in UIP vs. 0.5 ± 0.9 in non-uip group, p = 0.004), the use of oxygen therapy (8/22.9 % UIP patients vs. 0 non-uip patients, p = 0.016), number of deaths (23/65.7 % vs. 10/41.7 %, p = 0.046) and decline in diffusion capacity (56 ± 20.6 vs. 69 ± 20.2, p = 0.021). Dyspnea and inspiratory crackles were detected more often in the UIP group. RA-ILD was the most common primary cause of death (39.4 % of cases). Hypertension, coronary artery disease, chronic obstructive pulmonary disease, heart insufficiency, diabetes and asthma were common comorbidities. ILD preceded RA diagnosis in 13.6 % of patients. Conclusions: The course of the disease in RA-UIP patients is different from the other RA-ILD subtypes. Several comorbidities associated commonly with RA-ILD, although ILD was the predominant primary cause of death. Keywords: High-resolution computed tomography, Cause of death, Comorbidity Background Interstitial lung disease (ILD) is a rather common extraarticular manifestation of rheumatoid arthritis (RA) and a major cause of morbidity and mortality in RA patients [1, 2]. Approximately 10 % of patients with RA may develop clinically evident ILD with respiratory symptoms and/or a decline in pulmonary function tests [3]. In * Correspondence: hanna.nurmi@kuh.fi 1 Center of Medicine and Clinical Research, Division of Respiratory Medicine, Kuopio University Hospital, POB 100, KYS Kuopio, Finland 2 Division of Respiratory Medicine, Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland, POB 1627, Kuopio, Finland Full list of author information is available at the end of the article asymptomatic RA patients, high-resolution computed tomography (HRCT) scans commonly reveal evidence of interstitial lung involvement, and a large proportion of those with subclinical disease deteriorate with time [4, 5]. However, the clinical course of RA-ILD is highly heterogenic, as some patients remain stable for years, even decades, while others develop an insidious progressive disease [6]. While the overall mortality in RA has declined, the numbers of deaths due to RA-ILD have increased [7], although the results of studies investigating survival have been variable. Some studies have reported survival of 3 years, similar to that of idiopathic pulmonary fibrosis 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

114 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 2 of 10 (IPF) [8, 9], whereas in others the prognosis of RA-ILD has been significantly better, with median survival of approximately 6 8 years [10, 11]. Since it lacks its own distinctive classification, the subtypes of RA-ILD have been categorized according to the subdivisions of the idiopathic interstitial pneumonias (IIP) [12]. Unlike the situation in other connective tissue diseases (CTD), the most common radiologic and histopathologic pattern of RA-ILD is usual interstitial pneumonia (UIP), whereas nonspecific interstitial pneumonia (NSIP) and other subtypes also exist to a lesser extent [13]. The clinical significance of these different histological and radiological patterns has become nowadays more important since the RA-ILD patient with the UIP pattern (RA-UIP) seems to have a significantly worse prognosis and reduced survival compared to other types such as NSIP and organizing pneumonia (OP) [11, 14 16]. Other differences in the course of the disease in distinct RA- ILD subtypes, in addition to the difference in survival, have not been widely studied so far. Recently the significance of radiologic and histopathological subtyping of RA-ILD was highlighted as one important area for future investigation [17]. Little is known about concomitant diseases or causes of death of RA-ILD patients. The few studies that have addressed cause of death in these patients, have been unanimous that the majority of deaths are due to respiratory disease either after an exacerbation, infection or simply due to the steady progression of the ILD [11, 13, 18]. The aims of this study were to investigate the numbers and subtypes of the patients with RA-ILD treated in Kuopio University Hospital (KUH), in Eastern Finland, during The course of the disease, survival, co-morbidities and cause of death were evaluated and compared between UIP and non-uip cases. Methods Search and evaluation of data The subjects for the study were identified from the database of KUH using two International Classification of Diseases (ICD-10) codes, namely J84.X and M05.X/ M06.X. (Fig. 1). From these patients we only included the subjects that had been examined or treated in the pulmonology in-patient or out-patient clinic between and for any respiratory symptoms or any suspected pulmonary disease, thus omitting those RA patients with no symptoms or chest X-ray abnormalities. A total of 1047 patients were identified and their patient records were evaluated. At baseline, the patients with ILD but without RA (i.e. patients with IIP, other connective tissue disorders (CTD) or allergic alveolitis) and those with RA, whose visits to pulmonology clinic were because of some other lung diseases (such as asthma, chronic obstructive pulmonary disease (COPD), obstructive sleep apnea) were excluded. We also excluded suspected but not confirmed RA-ILD patients, for whom HRCT, or other comparable radiological examination capable of allowing reliable analysis of the lung parenchyma were not available, as were those patients whose RA diagnosis was not certain according to the 1987 classification criteria [19], or who developed later mixed CTD- like symptoms. Another 38 patients were excluded subsequently after the evaluation by the radiologist and/or after a multidisciplinary discussion due to the very minor signs or nonspecific features for ILD, leaving a total of 59 RA-ILD patients to be studied in detail and classified. Clinical information was gathered from the patient records of KUH, primary health care centers and other hospitals using a specially designed form. Demographic data included date of birth, sex, occupation, smoking habits, exposure to asbestos, radiation therapy of the 35 RA-UIP J84.X / M05.X + M06.X + a visit to the pulmonology clinic n=1047 No RA, No ILD, No HRCT -> excluded n=950 Suspected RA-ILD n=97 RA-ILD n=59 MDD/pulmonary radiologist evaluation -> excluded n=38 8 RA-NSIP 7 RA-OP 1 RA-DAD* 8 unclassified Fig. 1 The study protocol and the final categorization of the patients with RA-ILD. *additional 2 DAD findings included in OP group (n = 1) and UIP group (n =1)

115 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 3 of 10 thorax region, date of RA diagnosis, date of the first visit to pulmonology clinic due to ILD, comorbidities, death certificates, use of long term oxygen therapy, symptoms and respiratory status findings at baseline, laboratory test results including rheumatoid factor (RF) and antinuclear antibody (ANA) titer and surgery due to RA. Antibodies against cyclic citrullinated peptide were not available for half of the patients. The results of lung function tests, such as spirometry including forced vital capacity (FVC), forced expiratory volume (FEV1) and diffusion capacity to carbon monoxide (DLCO), were gathered at baseline and, when available, during the follow-up at 6 months, 1 year, 2 year and so on annually, including also the most recent available results. Any medication in use prior to ILD diagnosis and also lifelong medication used for RA were recorded. Histological data also was collated. The numbers of hospitalizations due to either respiratory problems (including infections, suspected drug reactions and suspected acute exacerbations) or cardiac problems like unstable angina pectoris, myocardial infarctions, arrhythmias and cardiac failures were collected. Data from death certificates was also collected. An experienced radiologist evaluated baseline HRCTs from these 59 patients. Radiological ILD categorization was conducted according to the 2013 IIP classification [12]. The radiological RA-UIP criteria were applied from those of IPF [20]. Mainly patients with a definite UIP pattern were included in the UIP group (32 out of 35, 91.4 %). Three patients who displayed a slightly upper (n = 2) or mid-lung (n = 1) predominated distribution, were included after a multidisciplinary discussion. Patients with possible UIP, i.e. a subpleural and basal predominated reticular abnormality without honeycombing, are not included in the UIP group. When available, an additional HRCT during the follow-up was also evaluated to reveal the progression of the lung disease. The study protocol was approved by the Ethical Committee of Kuopio University Hospital (statement 17/ 2013). Statistical analysis The distribution of the continuous variables was verified with Shapiro-Wilk test. If distribution was normally divided, the comparison was made using an independent T-test, otherwise Mann Whitney U-test was applied. The chi-squared test or Fisher test, when appropriate, was used for categorical variables. Sex, smoking habits, laboratory results and the numbers of deaths are calculated as percentages. Age at the time of RA-ILD diagnosis and lung function results are expressed as mean ± SD. The mean values of the first and most recent available FVCs and DLCOs were calculated in both UIP- and non-uip groups to determine whether there had been any change in lung function. The mean values of both groups were compared using the independent T-test to evaluate possible differences in lung function tests at the time of RA-ILD diagnosis and also the difference in lung function development. In survival analyses, we excluded the patient who did not have an underlying ILD preceding acute DAD changes. Survival analysis was done using the Kaplan-Meier method and survival curves were compared using the log-rank test. Survival time was calculated from the first visit to the pulmonology clinic due to ILD to the date of death or November 4, 2015 when the vital status was ascertained. Survival results are expressed as median (95 % confidence interval). We considered a p-value <0.05 as statistically significant. All data was analyzed using IBM Statistics SPSS software, version Results Radiologic findings and demographics Thirty-three (59.5 %) of the patients were male. Most of the patients (n = 35/60.3 %) were current or former smokers (Table 1). Five (15.6 %) male and 18 (69.2 %) female patients were never-smokers (p < 0.001). The mean age at diagnosis was 66 ± 11.1 years (range 32 87) differing non-significantly in subgroups (UIP vs. non-uip, non-smokers vs. ever-smokers, male vs. female). RF was positive in 84.2 %, ANA in 17.8 % and antibodies against cyclic citrullinated peptide (CCP) in 60.8 % of the patients. Themajority(35/59.3%)ofthepatientsshoweda radiological UIP-pattern in HRCT and the remainder werensip(8/13.6%),op(7/11.9%)and8patients whose radiological features remained nonspecific, which we termed as unclassified (13.6 %). A diffuse alveolar damage (DAD) pattern was detected in one patient without an underlying ILD, thus likely representing RA-DAD. Additional two DAD patterns were seen in patients with OP and UIP diagnoses prior to DAD. No statistically significant differences were observed between groups with respect to age, smoking, baseline lung functions or RA serology. Thirty-five (61.4 %) patients suffered from dyspnea and 31 (60.8 %) from cough. Cough was equally common in both groups, but dyspnea occurred more often in the UIP group (p = 0.022). Inspiratory crackles were more commoninuipthaninnon-uippatients(p = 0.007) (Table 1). Medication for RA and RA-ILD Seventy-five percent of patients were receiving some medication for RA at the time of RA-ILD diagnosis (Tables 1 and 2). In 11 cases the RA medication had been markedly changed due to ILD diagnosis. In most cases (9 out of 11), the change was a discontinuation of

116 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 4 of 10 Table 1 Clinical characteristics of the patients with rheumatoid arthritis-associated interstitial lung disease (RA-ILD), which have been classified according to the presence or absence of usual interstitial pneumonia (UIP) pattern in high resolution computed tomography (HRCT) Characteristics RA-ILD RA-UIP RA-non-UIP P-value (n = 59) (n = 35, 59.3 %) (n = 24, 40.7 %) (UIP vs. non-uip) Gender Male 33 (55.9) 19 (54.3) 14 (58.3) Female 26 (44.1) 16 (45.7) 10 (41.7) Smoking a Never 23 (39.7) 14 (41.2) 9 (37.5) Ex-smoker 26 (44.8) 14 (41.2) 12 (50.0) Current smoker 9 (15.5) 6 (17.6) 3 (12.5) Age (y) 66 ± ± ± Serology Positive RF b 48 (84.2) 29 (85.3) 19 (82.6) Pos. ANA c 8 (17.8) 5 (20.0) 3 (15.0) Pos. anti-ccp antibody d 17 (60.8) 10 (71.4) 7 (63.6) Dyspnea b 35 (61.4) 25 (73.5) 10 (43.5) Cough e 31 (60.8) 17 (60.7) 14 (60.9) Inspiratory crackles 41 (69.5) 29 (82.9) 12 (50.0) FVC % pred 85 ± ± ± DLCO % pred 71 ± ± ± Medications Steroids, ever 54 (91.5) 32 (91.4) 22 (91.7) MTX, ever 35 (59.3) 18 (51.4) 17 (70.8) MTX, when ILD diagnosed 15 (25.4) 6 (17.1) 9 (37.5) Biological drugs, ever 14 (23.7) 7 (20.0) 7 (29.2) Data presented as n (percentage) or mean ± SD. P-values calculated using Fisher test, χ 2 - test or independent T-test RA-UIP usual interstitial pneumonia (UIP) pattern in patients with rheumatoid arthritis (RA). RA-non-UIP Rheumatoid arthritis patients with other than UIP-pattern interstitial lung disease (ILD). FVC forced vital capacity. DLCO diffusing capacity of the lung for carbon monoxide. % pred: percentage of the predicted value. RF: rheumatoid factor. ANA: anti-nuclear antibodies. MTX: methotrexate. CCP: cyclic citrullinated peptide a data missing from 1 RA-UIP patient b data missing from 2 patients (1 RA-UIP, 1 RA-non-UIP) c data missing from 14 patients (10 RA-UIP, 4 RA-non-UIP) e data missing from 34 patients (21 RA-UIP, 13 RA-non-UIP) f data missing from 8 patients (7 RA-UIP, 1 RA-non-UIP) methotrexate after the diagnosis of ILD. In two patients, either leflunomide or sulfasalazine was discontinued. In all 11 cases (9 UIP, 2 NSIP), ILD continued to progress despite the changes to their RA medication. There were no differences between RA-UIP and RA-non-UIP groups in their use of methotrexate or biological drugs (Table 1). Almost all patients (n = 54/91.5 %) had received glucocorticoids at some point. Most i.e. 6/7 (85.7 %) RA-OP patients received glucocorticoid treatment for their lung disease and the seventh patient recovered without extra treatment. Of the six steroid-treated RA-OP patients, 5 recovered completely but one did not exhibit a clear beneficial response to treatment. Five of the eight (62.5 %) RA-NSIP patients were treated with high doses of prednisolone two of them enjoying at least a partial response. Two NSIP patients received cyclophosphamide treatment, but both deteriorated despite the treatment. In five RA-UIP patients, high-dose cyclophosphamide plus high-dose steroid treatment was provided but without any positive responses. Survival Thirty-three (55.9 %) patients died with median survival of 92.0 months in the UIP and months in the non- UIP groups (p = 0.417, Table 3). Of the deceased patients, the one with RA-DAD was excluded from the survival analysis. The number of deceased patients was significantly higher in the UIP group, i.e. 23/35 patients with UIP (65.7 %) had died compared with 9/24 (37.5 %) patients with non-uip (p = 0.046, Table 4). Although the median survival in the whole group was longer in

117 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 5 of 10 Table 2 The medications of the patients with RA-ILD Medicine Ever used for RA or ILD N (%) Used at the time of ILD diagnosis N (%) Discontinued due to ILD diagnosis N (%) Prednisolone 54 (91.5) 10 (16.9) 0 (0.0) Azathioprine 42 (71.2) 8 (13.6) 0 (0.0) Methotrexate 35 (59.3) 15 (25.4) 9 out of 15 (60.0) Hydroxychloroquine 47 (79.7) 16 (27.1) 0 (0.0) Sulfasalazine 45 (76.3) 16 (27.1) 2 out of 16 (12.5) Leflunomide 12 (20.3) 2 (3.4) 2 out of 2 (100.0) Penicillamine 5 (8.5) Mycophenolate Mofetil 7 (11.9) Sodium aurothiomalate 32 (54.2) 7 (11.9) 0 (0.0) Cyclosporin 15 (25.4) Cyclophosphamide 13 (22.0) Chlorambucil 5 (8.5) Podophyllotoxin 24 (40.7) 7 (11.9) 0 (0.0) Etanercept 6 (10.2) Infliximab 2 (3.4) Golimumab 0 (0.0) Adalimumab 5 (8.5) Abatacept 1 (1.7) Rituximab 11 (18.6) 1 (1.7) 0 (0.0) Tocilizumab 1 (1.7) The first column shows the number of patients receiving each medication at any point and of any duration during their lives. The second column shows the number of patients receiving any particular medication at the time of the RA-ILD diagnosis women (152 months) than in men (87 months) this difference was not statistically significant (p = 0.305). Survival between non-smokers and current/former smokers was also similar in the whole ILD group (p = 0.525). Female and non-smoker individuals had a tendency towards longer survival than men and smokers in the non- UIP group, but not in the UIP-group (Table 3, Fig. 2a-d). Causes of death The average age at death was 75.0 ± 9.1 years, ranging from 54.8 to 91.7 years. The UIP patients died slightly younger than their non-uip counterparts (73.6 ± 9.8 vs ± 6.5, p = 0.187). According to the death certificates of the 33 deceased patients, RA-ILD was the most common primary cause of death; 13/39.4 % cases (10 UIP, 3 non-uip; p = 0.701), (Fig. 3). RA-ILD was primary cause of death equally in men and women (8 vs. 5, p = 1.000) and in non-smokers and ever-smokers (6 vs. 7, p = 0.522). Coronary artery disease (CAD) was the second most common primary cause of death in 7 individuals (21.2 %; 3 UIP, 4 non- UIP, p = 0.161). RA was the primary cause of death in 5 Table 3 Survival of the patients (months) according to gender and smoking in subgroups RA-ILD (n = 59) RA-UIP (n = 35, 59.3 %) RA-non-UIP (n = 24, 40.7 %) P-value (UIP vs. non-uip) Overall ( ) 92.0 ( ) ( ) Gender Male 87.0 ( ) 88.0 ( ) 87.0 ( ) Female ( ) 92.0 ( ) a (p = 0.305) (p = 0.777) (p = 0.093) Smoking Non-smokers ( ) 92.0 ( ) a Ever-smokers 88.0 ( ) 88.0 ( ) ( ) (p = 0.525) (p = 0.921) (p = 0.218) Data are presented as median (95 % CI). The RA-DAD patient is excluded from the survival analyses a Median survival cannot be calculated since only one death has occurred in this group

118 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 6 of 10 Table 4 Factors associating with the differential course of disease in the patients with rheumatoid arthritis associated usual interstitial pattern (RA-UIP) and non-uip patterns (RA-non-UIP) Factor RA-ILD RA-UIP RA-non-UIP P-value (n = 59) (n = 35, 59.3 %) (n = 24, 40.7 %) Oxygen therapy 8 (13.6) 8 (22.9) 0 (0) Hospitalization due to respiratory illness 1.29 ± 2.2 (0 11) 1.9 ± 2.6 (0 11) 0.5 ± 0.9 (0 4) Hospitalization due to cardiac illness 0.6 ± 1.2 (0 5) 0.7 ± 1.3 (0 5) 0.4 ± 1.2 (0 4) Latest FVC % pred 82 ± ± ± (Baseline FVC %) 85 ± ± ± 16.5 Latest DLCO % pred 61 ± ± ± (Baseline DLCO %) 71 ± ± ± 13.3 Number of deaths 33 (55.9) 23 (65.7) 9 (37.5) a Data are presented as percentage or mean ± SD and also (range) in hospitalization Hospitalization comparison performed using Mann Whitney U-test and the lung function comparison using an independent sample T-test a One RA-DAD-patient is excluded from this group, P-value calculation and survival analyses Fig. 2 a-d. Shorter survival (Kaplan-Meier, log-rank) of men was observed in the non-uip group, but the difference was not quite statistically significant (p = 0.093). Survival differences between genders in UIP group were not found (p = 0.777). In the non-uip group, the non-smoking patients seemed to survive for longer than ever-smokers i.e. current smokers and ex-smokers, but the difference did not reach statistical significance (p = 0.218). In the UIP group, no differences were found in survival between non-smokers and ever-smokers (p =0.921)

119 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 7 of 10 Primary causes of death RA-ILD CAD RA COPD lung cancer other cancer other % UIP non-uip Fig. 3 ILD was the major cause of death in the UIP group (10/43.5 %), whereas that of the non-uip group was cardiovascular disease (4/40.0 %). None of the differences reached statistical significance. CAD : coronary artery disease, COPD: chronic obstructive pulmonary disease, RA: rheumatoid arthritis cases (15.2 %). In the other cases, the primary causes of death were Alzheimer s disease, universal atherosclerotic disease with acute ischemia in legs, acute pancreatitis, intestinal tuberculosis, chronic obstructive pulmonary disease (COPD), massive bleeding due to pelvic fracture, lung cancer and suspected viral infection in the central nervous system each one case. Pneumonia and CAD were equally common as the immediate cause of death (both 10/30.3 %; 6 UIP, 4 non- UIP) and RA-ILD (5/15.2 %; 4 UIP, 1 non-uip) was also prevalent. Lung cancer, RA, diabetes, RA associated secondary amyloidosis with renal failure, acute pancreatitis, diabetes, intestinal tuberculosis and gastroenteritis represented immediate causes of death of single cases. Comorbidities The most common comorbidities were hypertension (30/50.8 %), CAD (21/35.6 %), COPD (17/28.8 %), heart insufficiency (16/27.1 %), diabetes (13/22.0 %) and asthma (12/20.3 %) (Fig. 4). Gastroesophageal reflux (GER) occurred in 6 (10.2 %) and hypothyroidism in 3 (5.1 %) patients. There were two lung cancers (3.4 %) and 9 (15.3 %) other cancers including basal cell carcinoma (n = 2), diffuse large B cell lymphoma (n =2), urinary bladder carcinoma (n =1), colon adenocarcinoma (n = 1), squamous cell carcinoma in upper lip (n =1)andin tongue (n = 1) and carcinoma in ventricle (n =1). Most of the cancers were not primary causes of death. Six (10.2 %) patients suffered from tuberculosis (five lung and one intestinal). No statistically significant differences in the comorbidities were found between UIP and non- UIP groups, although COPD was more common in the UIP group (p = 0.088). Asthma was more common in women (p =0.016) and COPD in men (p < 0.001). Comorbidities divided equally between non-smokers and ever-smokers, except for COPD (p <0.001). Comorbidites Hypertension CAD COPD Diabetes Heart failure Asthma GER Other cancer Lung cancer % UIP non-uip Fig. 4 The most common comorbidities were hypertension, coronary artery disease (CAD), COPD, diabetes and heart failure, although asthma was also relatively common. COPD occurred more often in patients with UIP (13/37.1 % UIP vs. 4/16.7 % non-uip, p = 0.088). CAD: coronary artery disease, COPD: chronic obstructive pulmonary disease, GER: gastroesophageal reflux

120 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 8 of 10 Timing of diagnosis In eight patients (13.6 %), ILD preceded RA diagnosis. In three of these cases (2 UIP, 1 OP) the RA diagnosis was made within one year after the ILD diagnosis, but in five cases (3 UIP, 1 OP, 1 unclassified) joint symptoms and RA diagnosis appeared over one year after the ILD diagnosis (range years). In two cases (3.4 %) RA and ILD were diagnosed simultaneously. The RA diagnosis date was missing in one case. ILD followed the diagnosis of RA in 48 patients after a variable period of time i.e. 6/12.5 % within a year, 13/27.1 % within 3 years and 19/39.6 % within 5 years. The longest time interval between RA and ILD was 52.1 years. Course of disease Several factors were indicative of ILD progression i.e. oxygen treatment, hospitalizations and decline of diffusion capacity to carbon monoxide (DLCO) (Table 4). All patients (n = 8) using oxygen therapy belonged to the UIP group (p = 0.016). The number of hospitalizations due to respiratory causes was significantly higher in UIP compared to non-uip (p = 0.004). The latest available DLCO results were significantly lower in UIP (p = 0.021). Forced vital capacity (% predicted) (FVC %) showed a trend towards a greater decline in the UIP group, (p = 0.091). Discussion This study revealed that the course of disease in the patients with RA-ILD was variable in subtypes categorized according to either the presence or absence of the UIPpattern in HRCT. The patients with RA-UIP used oxygen, suffered from hospitalizations due to respiratory reasons and suffered an accelerated decline of lung function more often than those with non-uip subtype. Moreover, several comorbidities were very common, and in addition to RA-ILD, CAD was a common primary cause of death. The distribution of genders was almost equal supporting previous findings that male sex is a risk factor for ILD [5, 8, 21], minding that RA is twice as common in females [22]. The proportion of patients with UIP (59.3 %) and the amount of cases (13.7 %) in which ILD preceded articular disease were similar as described recently [21]. Dyspnea and inspiratory crackles were more common in the patients with UIP, in agreement with previous results [23]. The lung disease was more progressive in the UIP group based on the number of deaths, use of oxygen, hospitalization due respiratory reasons and decline of pulmonary function, especially DLCO. Some of the hospitalizations may have been attributable to acute exacerbations, known to occur mostly in UIP patterned RA-ILD [24]. In summary, our findings support previous studies suggesting that RA-UIP follows a distinctive pathological course [13, 25]. ILD was the primary cause of death in the majority of subjects, especially in the UIP group, although this did not reach statistical significance in our small study population. A previous study also indicated that RA-ILD patients were most likely to die of ILD or RA itself [7]. A recent Finnish study revealed that CAD was responsible for 43 % of deaths of RA patients [26] whereas in Korea, malignancies were the major cause of death in these patients [27]. The high percentage (39.4 %) of ILD as a primary cause of death indicates that even though several comorbidities often coexist, ILD remains the leading cause of death. The immediate causes of deaths did not exhibit any significant differences between the UIP and non-uip groups. CAD was a major comorbidity in RA-ILD. Previously, the risk of CAD and hypertension has been shown to increase in RA already at disease onset [28]. One novel finding was that asthma was more common in females, although an association between asthma and RA has been previously detected [29]. COPD was observed in almost 30 % of patients, in line with a recent study revealing a 48 % prevalence of emphysema [30]. COPD was more common in men, although this may be attributable to different smoking habits between the genders. COPD was also more prevalent in UIP patients even though smoking was similar in both groups. GER, previously claimed to be associated with IPF [31], or hypothyroidism thought to be more commoninra[32],werenotprevalentinourstudy. Previously published studies of survival of the patients with RA-ILD have revealed variable results. Some have reported survival as being as poor as in IPF i.e. approximately 3 years [8, 33, 34] but others have revealed longer survival times i.e. 7 8 years [10, 11], durations in line with the present study. Furthermore, the lifespan of RA- UIP has been shown to be shorter than that of the other subtypes [14]. The median survival in our study was shorter in patients with UIP than in their non-uip counterparts (92 vs 137 months) but this result did not reach statistical significance. Male gender has been recognized as a risk factor for RA-ILD mortality in previous studies [35]. In our study, survival analyses revealed a tendency that non-uip, but not UIP, females and non-smokers, lived longer. Identifying the RA-ILD patients from hospital registers was challenging since two different diagnosis codes were needed and, moreover, medical records of hundreds of patients had to be reviewed before we could gather this study population, which is similar in size as the majority of published reports, except for a few multicenter studies [21]. In fact, this sample size can be considered as representative since approximately 248,400 people live in the

121 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 9 of 10 KUH region. In addition, we intentionally excluded the patients with only minor changes in HRCT since our purpose was to study the verifiably clinically relevant RA-ILD. The retrospective protocol of the data collection may have caused some inaccuracies and missing data. Categorization into either UIP or non-uip groups was based on radiological evaluation, since histological data was limited. The radiological categorization can nonetheless be considered as reasonably reliable, since a definite UIP pattern in a HRCT scan has been demonstrated to be a sensitive and specific way of detecting the histopathologic UIP pattern in both IPF and RA-ILD [36 38]. Therefore we are confident that the UIP group reliably consists of true RA-UIP patients, although it is possible that some of the patients in the NSIP or unclassified group may be suffering from histological UIP. One obvious limitation of this study is the fact that the recategorization of the patients was performed by one radiologist. However, a large proportion of the HRCT scans were evaluated in a multidisciplinary discussion. In this study, due to its retrospective nature, it was not possible to evaluate thoroughly the effects of therapeutic interventions since the patients had received highly variable treatments for RA and ILD without being followed with a standardized protocol as was also the case in a previously published investigation [39]. Conclusions In summary, we detected several differences in disease course between RA-UIP and RA-non-UIP confirming the existing impression, that the UIP patterned ILD is more severe than the other subtypes of RA-ILD. In addition, even though several comorbidities often coexist with RA-ILD, the ILD itself seems to cause the majority of the deaths in these patients. Abbreviations ANA, antinuclear antibodies; CAD, coronary artery disease; CCP, cyclic citrullinated peptide; CI, confidence interval; COPD, chronic obstructive pulmonary disease; CTD, connective tissue diseases; DAD, diffuse alveolar damage; DLCO, diffusion capacity to carbon monoxide; FVC, forced vital capacity; GER, gastro-esophageal reflux; HRCT, high-resolution computed tomography; IIP, idiopathic interstitial pneumonias; ILD, interstitial lung disease; IPF, idiopathic pulmonary fibrosis; KUH, Kuopio University Hospital; MDD, multidisciplinary discussion; NSIP, nonspecific interstitial pneumonia; OP, organizing pneumonia; RA, rheumatoid arthritis; RA-DAD, rheumatoid arthritis-associated diffuse alveolar damage; RA-ILD, rheumatoid arthritisassociated interstitial pneumonia; RA-NSIP, rheumatoid arthritis-associated nonspecific interstitial pneumonia; RA-OP, rheumatoid arthritis-associated organizing pneumonia; RA-UIP, rheumatoid arthritis-associated usual interstitial pneumonia; RF, rheumatoid factor; SD, standard deviation; UIP, usual interstitial pneumonia Acknowledgements The authors wish to thank Ewen MacDonald for providing assistance with the language. Funding The study was supported by the Finnish Anti-Tuberculosis Association, the Jalmari and Rauha Ahokas Foundation, the Väinö and Laina Kivi Foundation, The Research Foundation of the Pulmonary Diseases, The Kuopio region Respiratory Foundation and a state subsidy to the Kuopio University Hospital. Availability of data and materials We cannot share our original data. It has been gathered in a detailed manner and minding that our population is relatively small in this Eastern- Finland hospital, we could not ascertain individuals anonymity. Authors contributions HN collected the study material, analyzed the data and prepared the draft of the manuscript and takes responsibility for the integrity of the data and accuracy of the data analysis. MP contributed to the study design, analyses of data and planning of the data collection form. MK participated in planning of the data collection form. H-PK performed the radiological analyses and planned radiological data collection form. TS was responsible for the statistical analyses. RK designed and managed the study, planned the data collection form and interpreted data. All authors participated in manuscript preparation. All authors read and approved the final manuscript. Competing interests Hanna Nurmi: Consulting fees from Boehringer-Ingelheim and Roche Oy. Congress travel grants from Boehringer-Ingelheim, Lilly Oncology, Novartis, Orion Pharma and GlaxoSmithKline. Minna Purokivi: Personal fees from Boehringer-Ingelheim, Chiesi, Intermune, Orion Pharma, Roche and Takeda Leiras. Congress travel grants from Boehringer-Ingelheim and Takeda Leiras Miia Kärkkäinen: Consulting fee from Boehringer-Ingelheim. Congress travel grants from Intermune, Boehringer- Ingelheim and Roche. Hannu-Pekka Kettunen: Consulting fees from Siemens and Roche. Riitta Kaarteenaho: Congress travel grants from Intermune, Boehringer- Ingelheim, Orion Pharma and Roche. Tuomas Selander: No conflicts of interests. Ethics approval and consent to participate The study protocol was approved by the Ethical Committee of Kuopio University Hospital (statement 17/2013). In this retrospective study, the majority of the patients are deceased and no consents for publications were gathered. Author details 1 Center of Medicine and Clinical Research, Division of Respiratory Medicine, Kuopio University Hospital, POB 100, KYS Kuopio, Finland. 2 Division of Respiratory Medicine, Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland, POB 1627, Kuopio, Finland. 3 Respiratory Medicine, Internal Medicine Research Unit, Medical Research Center Oulu, Oulu University Hospital and University of Oulu, POB 20, Oulu, Finland. 4 Harjula Hospital, the Municipal Hospital of Kuopio, Niuvantie 4, Kuopio, Finland. 5 Diagnostic Imaging Center, Division of Radiology, Kuopio University Hospital, POB 100, Kuopio, Finland. 6 Science Services Center, Kuopio University Hospital, POB 100, Kuopio, Finland. Received: 19 May 2016 Accepted: 19 July 2016 References 1. O Dwyer DN, Armstrong ME, Cooke G, Dodd JD, Veale DJ, Donnelly S. Rheumatoid Arthritis (RA) associated interstitial lung disease (ILD). Eur J Intern Med. 2013;24(7): Brown KK. Rheumatoid lung disease. Proc Am Thorac Soc. 2007;4(5): Turesson C, O Fallon WM, Crowson CS, Gabriel SE, Matteson EL. Extraarticular disease manifestations in rheumatoid arthritis: Incidence trends and risk factors over 46 years. Ann Rheum Dis. 2003;62(8): Gochuico BR, Avila NA, Chow CK, Novero LJ, Wu H, Ren P, et al. Progressive preclinical interstitial lung disease in rheumatoid arthritis. Arch Intern Med. 2008;168(2): Gabbay E, Tarala R, Will R, Carroll G, Adler B, Cameron D, et al. Interstitial lung disease in recent onset rheumatoid arthritis. Am J Respir Crit Care Med. 1997;156(2 Pt 1):

122 Nurmi et al. BMC Pulmonary Medicine (2016) 16:107 Page 10 of Dawson JK, Fewins HE, Desmond J, Lynch MP, Graham DR. Predictors of progression of HRCT diagnosed fibrosing alveolitis in patients with rheumatoid arthritis. Ann Rheum Dis. 2002;61(6): Olson AL, Swigris JJ, Sprunger DB, Fischer A, Fernandez-Perez ER, Solomon J, et al. Rheumatoid arthritis-interstitial lung disease-associated mortality. Am J Respir Crit Care Med. 2011;183(3): Bongartz T, Nannini C, Medina-Velasquez YF, Achenbach SJ, Crowson CS, Ryu JH, et al. Incidence and mortality of interstitial lung disease in rheumatoid arthritis: a population-based study. Arthritis Rheum. 2010;62(6): Young A, Koduri G, Batley M, Kulinskaya E, Gough A, Norton S, et al. Mortality in rheumatoid arthritis. Increased in the early course of disease, in ischaemic heart disease and in pulmonary fibrosis. Rheumatology (UK). 2007;46(2): Navaratnam V, Ali N, Smith CJ, McKeever T, Fogarty A, Hubbard RB. Does the presence of connective tissue disease modify survival in patients with pulmonary fibrosis? Respir Med. 2011;105(12): Tsuchiya Y, Takayanagi N, Sugiura H, Miyahara Y, Tokunaga D, Kawabata Y, et al. Lung diseases directly associated with rheumatoid arthritis and their relationship to outcome. Eur Respir J. 2011;37(6): Travis WD, Costabel U, Hansell DM, King Jr TE, Lynch DA, Nicholson AG, et al. An official American Thoracic Society/European Respiratory Society statement: Update of the international multidisciplinary classification of the idiopathic interstitial pneumonias. Am J Respir Crit Care Med. 2013;188(6): Lee HK, Kim DS, Yoo B, Seo JB, Rho JY, Colby TV, et al. Histopathologic pattern and clinical features of rheumatoid arthritis-associated interstitial lung disease. Chest. 2005;127(6): Kim EJ, Elicker BM, Maldonado F, Webb WR, Ryu JH, Van Uden JH, et al. Usual interstitial pneumonia in rheumatoid arthritis-associated interstitial lung disease. Eur Respir J. 2010;35(6): Nakamura Y, Suda T, Kaida Y, Kono M, Hozumi H, Hashimoto D, et al. Rheumatoid lung disease: prognostic analysis of 54 biopsy-proven cases. Respir Med. 2012;106(8): Akira M, Sakatani M, Hara H. Thin-section CT findings in rheumatoid arthritisassociated lung disease: CT patterns and their courses. J Comput Assist Tomogr. 1999;23(6): Doyle TJ, Lee JS, Dellaripa PF, Lederer JA, Matteson EL, Fischer A, et al. A roadmap to promote clinical and translational research in rheumatoid arthritis-associated interstitial lung disease. Chest. 2014;145(3): Hakala M. Poor prognosis in patients with rheumatoid arthritis hospitalized for interstitial lung fibrosis. Chest. 1988;93(1): Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum. 1988;31(3): Raghu G, Collard HR, Egan JJ, Martinez FJ, Behr J, Brown KK, et al. An official ATS/ERS/JRS/ALAT statement: idiopathic pulmonary fibrosis: evidence-based guidelines for diagnosis and management. Am J Respir Crit Care Med. 2011;183(6): Kelly CA, Saravanan V, Nisar M, Arthanari S, Woodhead FA, Price-Forbes AN, et al. Rheumatoid arthritis-related interstitial lung disease: associations, prognostic factors and physiological and radiological characteristics a large multicentre UK study. Rheumatology (Oxford). 2014;53(9): Crowson CS, Matteson EL, Myasoedova E, Michet CJ, Ernste FC, Warrington KJ, et al. The lifetime risk of adult-onset rheumatoid arthritis and other inflammatory autoimmune rheumatic diseases. Arthritis Rheum. 2011;63(3): Dawson JK, Fewins HE, Desmond J, Lynch MP, Graham DR. Fibrosing alveolitis in patients with rheumatoid arthritis as assessed by high resolution computed tomography, chest radiography, and pulmonary function tests. Thorax. 2001;56(8): Park I, Kim DS, Shim TS, Lim C, Lee SD, Koh Y, et al. Acute exacerbation of interstitial pneumonia other than idiopathic pulmonary fibrosis. Chest. 2007;132(1): Kim EJ, Collard HR, King Jr TE. Rheumatoid arthritis-associated interstitial lung disease: the relevance of histopathologic and radiographic pattern. Chest. 2009;136(5): Kerola AM, Nieminen TV, Virta LJ, Kautiainen H, Kerola T, Pohjolainen T, et al. No increased cardiovascular mortality among early rheumatoid arthritis patients: a nationwide register study in Clin Exp Rheumatol. 2015;33(3): Kim YJ, Shim JS, Choi CB, Bae SC. Mortality and incidence of malignancy in Korean patients with rheumatoid arthritis. J Rheumatol. 2012;39(2): Kerola AM, Kerola T, Kauppi MJ, Kautiainen H, Virta LJ, Puolakka K, et al. Cardiovascular comorbidities antedating the diagnosis of rheumatoid arthritis. Ann Rheum Dis. 2013;72(11): Lai NS, Tsai TY, Koo M, Lu MC. Association of rheumatoid arthritis with allergic diseases: A nationwide population-based cohort study. Allergy Asthma Proc. 2015;36(5): Antoniou KM, Walsh SL, Hansell DM, Rubens MR, Marten K, Tennant R, et al. Smoking-related emphysema is associated with idiopathic pulmonary fibrosis and rheumatoid lung. Respirology. 2013;18(8): Ley B, Collard HR. Epidemiology of idiopathic pulmonary fibrosis. Clin Epidemiol. 2013;5: Kerola AM, Nieminen TV, Kauppi MJ, Kautiainen H, Puolakka K, Virta LJ, et al. Increased risk of levothyroxine-treated hypothyroidism preceding the diagnosis of rheumatoid arthritis: a nationwide registry study. Clin Exp Rheumatol. 2014;32(4): Koduri G, Norton S, Young A, Cox N, Davies P, Devlin J, et al. Interstitial lung disease has a poor prognosis in rheumatoid arthritis: results from an inception cohort. Rheumatology (Oxford). 2010;49(8): Solomon JJ, Ryu JH, Tazelaar HD, Myers JL, Tuder R, Cool CD, et al. Fibrosing interstitial pneumonia predicts survival in patients with rheumatoid arthritisassociated interstitial lung disease (RA-ILD). Respir Med. 2013;107(8): Assayag D, Lubin M, Lee JS, King TE, Collard HR, Ryerson CJ. Predictors of mortality in rheumatoid arthritis-related interstitial lung disease. Respirology. 2014;19(4): Raghu G, Mageto YN, Lockhart D, Schmidt RA, Wood DE, Godwin JD. The accuracy of the clinical diagnosis of new-onset idiopathic pulmonary fibrosis and other interstitial lung disease: A prospective study. Chest. 1999;116(5): Assayag D, Elicker BM, Urbania TH, Colby TV, Kang BH, Ryu JH, et al. Rheumatoid arthritis-associated interstitial lung disease: radiologic identification of usual interstitial pneumonia pattern. Radiology. 2014;270(2): Hunninghake GW, Lynch DA, Galvin JR, Gross BH, Müller N, Schwartz DA, et al. Radiologic Findings Are Strongly Associated with a Pathologic Diagnosis of Usual Interstitial Pneumonia. Chest. 2003;124(4): Solomon JJ, Chung JH, Cosgrove GP, Demoruelle MK, Fernandez-Perez ER, Fischer A, et al. Predictors of mortality in rheumatoid arthritis-associated interstitial lung disease. Eur Respir J. 2016;47(2): Submit your next manuscript to BioMed Central and we will help you at every step: We accept pre-submission inquiries Our selector tool helps you to find the most relevant journal We provide round the clock customer support Convenient online submission Thorough peer review Inclusion in PubMed and all major indexing services Maximum visibility for your research Submit your manuscript at

123 II Are risk predicting models useful for estimating survival of patients with rheumatoid arthritis-associated interstitial lung disease? Nurmi H, Purokivi M, Kärkkäinen M, Kettunen H-P, Selander T, Kaarteenaho R. BMC Pulm Med 17: , Reprinted with the kind permission of BMC Pulmonary Medicine. The original article is an open access article distributed under the terms of Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

124

125 Nurmi et al. BMC Pulmonary Medicine (2017) 17:16 DOI /s RESEARCH ARTICLE Open Access Are risk predicting models useful for estimating survival of patients with rheumatoid arthritis-associated interstitial lung disease? Hanna M. Nurmi 1,2*, Minna K. Purokivi 1, Miia S. Kärkkäinen 2, Hannu-Pekka Kettunen 4, Tuomas A. Selander 5 and Riitta L. Kaarteenaho 1,2,3 Abstract Background: Risk predicting models have been applied in idiopathic pulmonary fibrosis (IPF), but still not validated in patients with rheumatoid arthritis-associated interstitial lung disease (RA-ILD). The purpose of this study was to test the suitability of three prediction models as well as individual lung function and demographic factors for evaluating the prognosis of RA-ILD patients. Methods: Clinical and radiological data of 59 RA-ILD patients was re-assessed. GAP (gender, age, physiologic variables) and the modified interstitial lung disease (ILD)-GAP as well as the composite physiologic indexes (CPI) were tested for predicting mortality using the goodness-of-fit test and Cox model. Potential predictors of mortality were also sought from single lung function parameters and clinical characteristics. Results: The median survival was 152 and 61 months in GAP / ILD-GAP stages I and II (p = 0.017). Both GAP and ILD-GAP models accurately estimated 1-year, 2-year and 3-year mortality. CPI (p = 0.025), GAP (p = 0.008) and ILD-GAP (p = 0.028) scores, age (p = 0.002), baseline diffusion capacity to carbon monoxide (DLCO) (p = 0.014) and hospitalization due to respiratory reasons (p = 0.039), were significant predictors of mortality in the univariate analysis, whereas forced vital capacity (FVC) was not predictive. CPI score (HR 1.03, p = 0.018) and baseline DLCO (HR 0.97, p = 0.011) remained significant predictors of mortality after adjusting for age. Conclusions: GAP and ILD-GAP are applicable for evaluating the risk ofdeathofpatientswithra-ildinasimilarmanner as in those with IPF. Baseline DLCO and CPI score also predicted survival. Keywords: Mortality, Rheumatoid arthritis, Interstitial lung disease, RA-ILD, GAP, ILD-GAP, Composite physiologic index Background The course of disease in interstitial lung diseases (ILD), including rheumatoid arthritis-associated interstitial lung disease (RA-ILD), is known to be highly variable. Predicting the survival of an individual patient with ILD is challenging [1]. Several factors have, however, been proposed to predict disease progression and survival i.e. * Correspondence: hanna.nurmi@kuh.fi 1 Center of Medicine and Clinical Research, Division of Respiratory Medicine, Kuopio University Hospital, POB Kuopio, Finland 2 Division of Respiratory Medicine, Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland, POB Kuopio, Finland Full list of author information is available at the end of the article physiological, radiological and histopathological characteristics, as well as demographic variables such as age and gender [2]. Some factors reflecting the severity of the rheumatoid arthritis (RA) have also been associated with worse survival, e.g. baseline pain [3], disease activity score [4] and health-assessment questionnaire score [3, 5]. There are now several indexes which combine single factors into a multifaceted scoring system and these have proved beneficial in estimating prognosis. These models have, however, focused mainly on idiopathic pulmonary fibrosis (IPF) and some of the earliest models were rather cumbersome and therefore never achieved any widespread clinical acceptance [6]. A composite The Author(s) Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

126 Nurmi et al. BMC Pulmonary Medicine (2017) 17:16 Page 2 of 9 physiologic index (CPI) displayed some important advantages over the older models, since it contained only pulmonary function test (PFT) and gas transfer values but omitted radiological scoring or exercise testing [7]. The subsequently developed GAP model combines gender (G), age (A) and two lung physiology variables (P), i.e. forced vital capacity (FVC) and diffusion capacity to carbon monoxide (DLCO), into a multidimensional index and staging system with three stages (I-III) proposing 1-year mortality of 6, 16 and 39% [8]. This GAP model has also been utilized in the prognosis of other chronic ILDs in addition to IPF. The modified model was named as ILD-GAP, with the assumption that patients with connective tissue disease-related ILDs (CTD- ILD) enjoyed a better survival than those suffering from IPF [9]. The survival of patients with RA-ILD has been shown to be as poor as in IPF patients [10], at least in those cases with usual interstitial pneumonia (UIP) which is the most common subtype in RA-ILD and unlike the situation in the other CTD-ILDs [11]. Thus, since it is mainly UIP-typed, RA-ILD follows a distinctive disease course from the other CTD-ILDs and it remains unclear which of the prognostic indexes, GAP or ILD-GAP, would be better suited for RA-ILD. There are some reports of the benefits of using the CPI score, GAP and ILD-GAP staging systems in patients with IPF and systemic sclerosis-associated ILD [12 14]. However, as far as we are aware, neither CPI nor GAP/ILD-GAP have been previously investigated in patients with RA- ILD, if one excludes the subjects in the original ILD- GAP publication, which did include some RA-ILD patients in their CTD-ILD/idiopathic nonspecific interstitial pneumonia (insip) group of 326 patients. The aims of this study were to investigate the applicability of CPI, GAP and ILD-GAP scores for predicting the prognosis of the patients with RA-ILD treated in Kuopio University Hospital (KUH), in Eastern Finland, during the years In addition, we examined the association between individual PFT and demographic factors with the survival of the patients. Methods Data sources and search The study cohort consists of patients treated in the KUH pulmonology in-patient or out-patient clinic between and The patients were identified from the database of KUH using two International Classification of Diseases (ICD-10) codes, namely J84.X and M05.X/M06.X (Fig. 1). These searches resulted in identification of 1047 patients, and their patient records were evaluated in order to identify those patients suffering from clinically relevant RA-ILD. The search process and clinical characteristics of the patients are thoroughly described in our previous study [15]. Shortly, all patients without a certain diagnosis of RA or without HRCT confirmed ILD were excluded, as were those with mixed CTDlike symptoms. Atypical cases were debated by a multidisciplinary discussion. Finally, 59 radiologically diagnosed RA-ILD patients were identified to be studied in detail and classified adopting the year 2013 IIP classification [16]. The radiological RA-UIP criteria Fig. 1 Study protocol. Flowchart of patient enrollment into the study showing the subdivision into the different GAP / ILD-GAP groups. ILD = interstitial lung disease; RA = rheumatoid arthritis; HRCT = high-resolution computed tomography; RA-ILD = rheumatoid arthritis-associated interstitial lung disease; UIP = usual interstitial pneumonia; NSIP = nonspecific interstitial pneumonia; OP = organizing pneumonia; DAD = diffuse alveolar damage; MDD = multidisciplinary discussion; GAP = gender, age, physiologic variables

127 Nurmi et al. BMC Pulmonary Medicine (2017) 17:16 Page 3 of 9 that were applied were those for the diagnosis of IPF [17] when 32 (54.2%) of the patients had a radiological definite UIP pattern [15]. After a multidisciplinary discussion, two additional patients with a slightly upper- or mid-lung predominated distribution where included in the RA-UIP group (35/59.3%), whereas patients with a possible UIP pattern are not included in the UIP group but instead categorized in the unclassified group. In addition to RA-UIP patients, there were eight RA-NSIP (13.6%), seven RA- OP (11.9%), one RA-DAD (1.7%) and eight unclassified patients (13.6%) as previously described [15]. Gathering of demographic information Clinical information was gathered from the patient records of KUH, primary health care centers and other hospitals using a specially designed form. Demographic data and the lifelong medication history for RA were gathered comprehensively. The number of hospitalizations was also obtained and further categorized into either mainly respiratory (i.e. infections, suspected drug reactions and suspected acute exacerbations of ILD) or cardiac problems as presented previously [15]. The results of PFT, such as spirometry including FVC and forced expiratory volume (FEV1), as well as DLCO were gathered at baseline and, when available, during the follow-up annually, including also the most recent available results. The reference values of Viljanen were used when assessing PFT results [18]. Staging systems Composite physiologic index (CPI) was calculated using the formula [7]: CPI = 91 (0.65 DLCO % predicted) (0.53 FVC % predicted) + (0.34 FEV1 % predicted). GAP / ILD-GAP score was calculated by gender, age, FVC % predicted and DLCO % predicted and patients divided to GAP / ILD-GAP stages I and II as previously described [8, 9]. There were no stage III (or IV in ILD- GAP) patients in our study material. Statistical analysis The distribution of the continuous variables was verified with the Shapiro-Wilk test. If there was a normal distribution, the independent T-test was used to compare continuous variables, otherwise the Mann-Whitney U- test was used. The chi-squared test or Fisher test, when appropriate, was used for comparison of categorical variables. Gender, smoking habits, laboratory results, use of medications, comorbidities, use of oxygen and the numbers of observed deaths were calculated as percentages. Age at the time of RA-ILD diagnosis or death, lung function results and hospitalizations were expressed as mean ± SD. Survival curves were estimated using the Kaplan-Meier method and differences in survival time between GAP / ILD-GAP stages I and II were calculated by the log-rank test. Survival results are expressed as median (95% confidence interval). The observed 1-, 2-, and 3-year mortality rates were calculated and these were supplemented with an estimate of the confidence interval by using the Wilson score. Next, the observed mortality and the risk of death predicted by the GAP / ILD-GAP model were compared using Hosmer- Lemeshow goodness-of fit-test. Finally, Cox regression analysis was used to identify factors that predicted mortality. P-values <0.05 were considered significant. All data was analyzed using IBM Statistics SPSS software, version Results Patient characteristics, lung functions and CPI score The mean RA duration at the point when ILD was diagnosed was 15.6 ± 12.2 years, ranging from 0 to 52 years. The female male ratio was 1:1.27. A substantial number (39.7%) of the patients had never smoked. The mean CPI score of all RA-ILD patients was 27.2 ± 14.4 (range ) (Table 1). The detailed data of the lung function test results is shown in Table 1. Over half (30/55.6%) of the patients had a normal FVC at the time of RA-ILD diagnosis. Twenty-five (49%) of the patients had a normal baseline DLCO and furthermore, in 17 patients (33.3%) both FVC and DLCO were normal (Table 1). The clearest decline of all PFT was observed in DLCO, the final mean was 61.1 ± 21.4 (range ). Table 1 Pulmonary function test results of the patients with RA-ILD Variable Baseline results FVC Normal (>80%) 30 (55.6) Declined (50 80%) 23 (42.6) Severely declined (<50%) 1 (1.9) Normal FEV1 (>80%) 29 (53.7) Normal FEV1/FVC (>88%) 44 (81.5) Normal DLCO (>74%) 25 (49.0) Normal FVC + Normal DLCO 17 (33.3) Mean FVC ± 16.9 Mean FEV ± 16.3 Mean FEV- % ± 12.4 Mean DLCO ± 18.1 CPI score 27.2 ± 14.4 Data shown as number (%), or mean ± SD. FVC, FEV1 and FEV1/FVC results are missing from five patients. DLCO results are missing from eight patients. Both FVC and DLCO results were available for 51 patients

128 Nurmi et al. BMC Pulmonary Medicine (2017) 17:16 Page 4 of 9 GAP and ILD-GAP There was all the necessary data available from 51 patients to allow the calculation of GAP and ILD- GAP scores. The majority of the subjects i.e. 76.5% (n = 39) belonged to stage I with the remaining 23.5% categorized into the stage II group. There were no patients in stage III. The same patients who were categorized as GAP I constituted the ILD-GAP I group and the patients in GAP II group, were also the patients with ILD-GAP II (Fig. 1). GAP / ILD-GAP I and II differed significantly with respect to several clinical findings and lung function e.g. age (p = 0.024), gender (p < 0.001), smoking status (p = 0.033), baseline FVC (p <0.001), FEV1 (p = 0.013) and DLCO (p <0.001) (Table 2). The use of methotrexate was also more Table 2 Baseline characteristics of the patients with RA-ILD GAP / ILD-GAP stage I GAP/ILD-GAP stage II P-value (n = 39/76.5%) (n = 12/23.5%) Age (y) 63.4 ± ± Age at death (y) 72.6 ± ± Male sex 16 (41.0) 12 (100.0) <0.001 UIP pattern 25 (64.1) 6 (50.0) Smoking* Never 19 (48.7) 1 (9.1) a Ex-smoker 15 (38.5) 7 (63.6) Current smoker 5 (12.8) 3 (27.3) Serology Positive RF** 29 (78.4) 11 (91.7) a Positive 4 (14.3) 2 (22.2) a ANA*** Medications Steroids 36 (92.3) 10 (83.3) a MTX 25 (64.1) 4 (33.3) Biological 11 (28.2) 1 (8.3) a drugs Lung functions FVC % pred 89.8 ± ± 8.5 <0.001 FEV1 % pred 85.1 ± ± DLCO % pred 76.8 ± ± 12.7 <0.001 RA duration (y) 15.7 ± ± CPI points 22.4 ± ± 9.3 <0.001 For eight patients there was no lung function data and therefore the GAP / ILD-GAP score could not be calculated RF rheumatoid factor, ANA antinuclear antibodies, MTX methotrexate, FVC forced vital capacity, DLCO diffusing capacity of the lung for carbon monoxide, % pred percentage of the predicted value, CPI composite physiologic index *Data missing from one stage II patient **Data missing from two stage I patients with positive anti-cyclic citrullinated peptide antibodies ***Data missing from 14 patients (11 stage I, 3 stage II) a Fisher test common in stage I patients than in their stage II counterparts (64.1% vs. 33.3%), although this finding did not reach statistical significance (p = 0.060). No statistically significant differences were observed in RA serology or comorbidities. The mean CPI score was 22.4 ± 12.2 in GAP / ILD-GAP I and 42.8 ± 9.3 in stage II (p < 0.001). Patients with the UIP pattern in HRCT (RA-UIP) divided almost equally in both stages (64% in stage I, 50% in stage II, p = 0.502). The follow-up outcomes No statistically significant differences were observed between GAP / ILD-GAP I and II with regard to hospital admissions either due to respiratory or cardiologic reasons (Table 3). The use of oxygen was also similar in both groups (p = 1.000). Eighteen patients (46.2%) died due to any cause in the GAP / ILD-GAP stage I whereas there were 9 deceased patients (75.0%) in the stage II group (p = 0.080). The observed cumulative mortality rates at 1, 2 and 3 years were 7.0, 16.7 and 22.6%, respectively. The observed 1-year, 2-year or 3-year mortality did not differ significantly according to GAP / ILD-GAP stage. Survival and validation of the GAP and ILD-GAP models The median survival was 152 months in stage I but only 61 months in stage II (p = 0.017) (Fig. 2). There were no apparent differences in the observed and predicted risk of death (Table 4). Both prediction models fitted the Wilson score confidence interval of the observed mortality. The observed mortality and the risk of death predicted by these models were compared using the Hosmer- Lemeshow goodness-of-fit test (Figs. 3 and 4). Both GAP and ILD-GAP indexes predicted 1-year, 2-year and 3- year mortality accurately (all p-values were > 0.05). The ILD-GAP index was more accurate at predicting 1-year mortality (p = 0.552) than the GAP index (p = 0.254). However, the GAP index was slightly more accurate at predicting 2-year (p = 0.261) and 3-year (p = 0.595) mortality than the ILD-GAP index (2-year p = 0.139, 3-year p = 0.357). Predictors of mortality GAP and ILD-GAP indexes, as well as the CPI score were all significant predictors of mortality when assessed with the univariate Cox model. The hazard ratio (HR) of GAP was 1.56 (95% CI: ; p = 0.004), that of ILD-GAP 1.51 (95% CI: ; p = 0.026) and of CPI 1.03 (95% CI ; p = 0.015) (Table 5). Age at diagnosis (HR 1.06, 95% CI , p =0.002), baseline DLCO (HR 0.98, 95% CI , p =0.014) and hospitalization due to respiratory reasons (HR 1.12, , p = 0.039) were also significant predictors of

129 Nurmi et al. BMC Pulmonary Medicine (2017) 17:16 Page 5 of 9 Table 3 Course of disease and survival of the patients with RA-ILD GAP / ILD-GAP stage I (n = 39/76.5%) GAP / ILD-GAP stage II (n = 12/23.5%) P-value Number of deaths 18 (46.2) 9 (75.0) Hospitalization due to respiratory illness 1.0 ± 1.4 (0 5) 1.6 ± 3.1 (0 11) Hospitalization due to cardiac illness 0.5 ± 1.0 (0 5) 1.0 ± 1.8 (0 4) Use of Oxygen 6 (15.4) 1 (8.3) a Median survival ( ) 61.0 ( ) Observed 1-y deaths* 0 (0.0) 1 (8.3) Observed 2-y deaths** 5 (14.3) 1 (9.1) Observed 3-y deaths*** 6 (17.6) 3 (27.3) Categorical variables are compared using the Fisher test when marked a, otherwise χ 2 -test. Hospitalizations are compared using the Mann-Whitney U-test *Data missing from two stage I patients **Data missing from four stage I patients and one Stage II patient *** Data missing from five Stage I patients and one Stage II patient mortality in the univariate model, but neither FVC nor hospitalization due to cardiologic reasons was predictive. The UIP pattern was not an independent risk factor in this cohort, neither was smoking nor male gender. The use of either methotrexate or oxygen did not reach statistical significance as risk factors for death (Table 5). Age adjusted predictors of mortality After adjusting for age, CPI score and baseline DLCO remained as significant predictors of mortality. For every increased CPI point, the mortality risk increased by 3% (HR 1.03, 95% CI , p = 0.014) and for every increased DLCO level, the risk of death diminished by 3% (HR 0.97, 95% CI , p = 0.011). Fig. 2 Comparison of the survival curves of the RA-ILD patients categorized into either GAP / ILD-GAP stage I or II. The survival was significantly worse in GAP / ILD-GAP stage II (p = 0.017, Log Rank) The rest of the factors that were detected in the univariate Cox model lost their statistical significance after adjustment for age (Table 6). Discussion In this present study, we applied the GAP and the ILD-GAP scores in a cohort consisting of 59 patients with RA-ILD subdivided into GAP / ILD-GAP stages I and II. Both GAP systems showed significant differences in age, gender, FVC, FEV1, DLCO and CPIscore, which is understandable since GAP / ILD-GAP are mainly composed of the above-mentioned components. The median survival of the patients categorized into GAP / ILD-GAP II groups was significantly shorter than those in the GAP / ILD-GAP I group. The CPI score was an independent predictor of mortality similarly as GAP / ILD-GAP scores, age, baseline DLCO and hospitalization due respiratory reasons. However, after adjustment for age, only the CPI score and DLCO remained as statistically significant predictors. In addition to the Cox model, the applicability of GAP and ILD-GAP was tested using Table 4 Predicted and observed cumulative mortality of the patients with RA-ILD GAP/ILD-GAP stage 1-Y mortality Observed Predicted by GAP index and staging system Stage I 0.0 ( ) Stage II 8.3 ( ) Y mortality Stage I 14.3 ( ) Stage II 9.1 ( ) Y mortality Stage I 17.6 ( ) Stage II 27.3 ( ) % (95% CI calculated by Wilson score) Predicted by ILD-GAP