Rheumatology

Khan, Zeeshan;Gupta, Srishti;Jaiswal, Amit

2018 Rheumatology

doi: 10.1093/rheumatology/key137pmid: 29718373

Rheumatology, 2017;56(suppl 2):ii64. https://doi.org/10.1093/rheumatology/kex062 This abstract has been corrected online as an author was missing from the author list. The third author, Amit Jaiswal, has now been added. © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Ziegelbauer, Kathrin;Kostev, Karel;Hübinger, Maximilian;Dombrowski, Silvia;Friedrichs, Michael;Friedel, Heiko;Kachroo, Sumesh

2018 Rheumatology

doi: 10.1093/rheumatology/key099pmid: 29660105

Abstract Objective The goal of the present study was to estimate the treatment costs in immune-mediated rheumatic disease patients initiating treatment with an s.c. biologic agent based on treatment persistence. Methods This is a retrospective cohort study based on the German statutory health insurance funds database. Patients ⩾18 years of age with a diagnosis of AS, PsA or RA treated with s.c. TNF-α inhibitors (TNFis) were included. Persistence was estimated as the duration of time from s.c. TNFi therapy initiation to discontinuation, which was defined as at least 60 days without therapy. We performed 1:1 matching based on a propensity score that was constructed as the conditional probability of being persistent as a function of age, gender, index year, physician specialty and Charlson comorbidity index. Finally, the cost differences between the matched pairs were estimated using the Wilcoxon test. Results After 1:1 matching, 678 persistent and 678 non-persistent patients were available for cost analyses. Using a 2-year time period, the costs for office-based visits per patient were €2319 in the persistent cohort compared with €3094 in the non-persistent cohort (P < 0.001). Co-medication costs were €2828 in the persistent cohort compared with €5498 in the non-persistent cohort, hospitalization costs were €3551 in the persistent cohort compared with €5890 in the non-persistent cohort and sick leave costs were €717 in the persistent cohort compared with €1241 in the non-persistent cohort (all P < 0.001). Conclusion The results of this study indicate that persistence with s.c. TNFi treatment can be associated with several cost offsets for immune-mediated rheumatic disease patients. IMRD, persistence, TNF-α, blocker, treatment costs Rheumatology key messages Biologic treatment persistence is associated with several cost offsets for immune-mediated rheumatic disease patients. In immune-mediated rheumatic disease patients, direct medical costs were lower in the persistent cohort. Introduction AS, PsA and RA are chronic progressive immune-mediated rheumatic diseases (IMRDs) that can cause pain, deformity and disability and progressively impair joint structure and function [1]. IMRDs result in substantial personal as well as economic burden [2]. The primary treatment goal for all three conditions is to maximize health-related quality of life by controlling symptoms and inflammation, preventing progressive structural damage, preserving or normalizing function and social participation and targeting remission. Treatment of rheumatic diseases usually involves a multimodal approach including both pharmacological and non-pharmacological strategies [3, 4]. The advent of biologic medications such as s.c. TNF-α inhibitors (TNFis) has transformed the management of IMRDs [5]. Previous research has provided high-quality evidence that anti-TNF agents improve clinical symptoms in the treatment of IMRDs [6–8]. Taking the prescribed medication for a sufficient period of time is crucial to the success of any therapy. Non-compliance and a lack of persistence with prescription medications are major problems in the clinical management of chronic diseases and s.c. TNFi treatment is potentially a very convenient option for patients. However, patients may likely have drug interruptions or long periods of non-persistence. In IMRDs, treatment persistence can be used to represent drug effectiveness, safety and treatment satisfaction [9–11]. While associations between adherence and costs [12] and the cost impact of switching from an s.c. TNFi to different biologic classes have been examined [5, 13], studies estimating the cost impact of non-persistence in patients treated with s.c. TNFis for IMRDs are lacking. In 2016, Dalén et al. [14] analysed treatment costs in first-line (bio-naïve) IMRD patients in Sweden based on persistent and non-persistent patients. The results of their study indicated that persistence with s.c. TNFi treatment may be associated with some cost offsets for IMRD patients. In 2017, Dalén et al. [15] further showed that, in second-line therapy with s.c. TNFis, persistent patients had lower mean total costs. In the same year, Svedbom et al. [16] described and compared treatment persistence with first- and second-line s.c. TNFis as well as their corresponding costs in Sweden. Although these works are of great interest, no recent data are available about the association between non-persistence with s.c. TNFis and costs in Germany. The goal of the present study was to estimate and compare the direct and indirect treatment costs in IMRD patients initiating treatment with an s.c. biologic agent based on treatment persistence. Methods Data source This study is a retrospective cohort study based on the German statutory health insurance funds database. This database continuously provides data pertaining to about five million statutorily insured people, which corresponds to 7% coverage. Exclusively anonymous information and no personal data (in accordance with § 3 Abs. 6 German Federal Data Protection Act) are used. Each patient with defined characteristics (e.g. specific disease, number and/or durations of hospitalizations, prescriptions of specific medications or any combination of these characteristics) can be tracked within the database via a unique patient ID. Finally, the database has already been used in several epidemiological studies, including cost analyses [17]. General informed consent Informed patient consent German law allows the use of anonymous electronic medical records for research purposes under certain conditions. Because patients were only queried as aggregates and no protected health information was available for queries, no institutional review board approval was required for the use of this database or completion of this study. For this study, we only used anonymized data collected from the claims database. Therefore, this study does not require that we obtain informed consent from patients. Study population Patients ⩾18 years of age with a diagnosis of RA [International Classification of Diseases, 10th Revision (ICD-10) M05, M06], PsA (ICD-10 L40.5 in combination with M07.0, M07.1, M07.2, M07.3 or M09.0) or AS (ICD-10 M45) treated with any of the four s.c. TNF-α blockers of interest (etanercept, adalimumab, certolizumab pegol or golimumab) were selected. Patients were categorized by their main diagnosis at the index date (the diagnosis that is documented within the quarter of the index prescription date). Cases with more than one documented diagnosis were classified as a mixed group. The date of the first qualifying prescription for an s.c. biologic agent served as the patient’s index date. The 12-month period immediately preceding the index date was used as the pre-index period and the 12-month period following the final prescription within the persistence period served as the post-index period, which lasted no later than December 2015. All patients were traceable in the database for at least 12 months after their last prescription within the analysis period. The pre-index period and index date were used to measure patients’ baseline characteristics. Treatment persistence with index s.c. TNF-α blockers and health care resource utilization (such as services and direct and indirect costs) were evaluated in both the persistence and post-index periods. Persistence definition and patient stratification Persistence was defined as the duration of time from s.c. TNFi therapy initiation to discontinuation, which was identified as at least 60 days without s.c. TNFi therapy. A longitudinal dataset of medication supply was created for each individual patient and the number of days of drug supply was calculated based on the defined daily dose information associated with each prescription record. The persistence was evaluated at 24 months after the index date. Following this estimation of persistence, patients were categorized into persistent and non-persistent cohorts. The main outcome of the study was rheumatoid disease–related health care resource utilization costs, including direct costs and services. These costs were estimated for the time within 2 years (730 days) following the first s.c. TNFi prescription and separated by RA, PsA and AS indication. Direct costs included hospitalization costs, doctor and hospital visit costs, prescribed medication costs and sickness benefits (in Germany, sickness funds are obligated to provide sick pay for sick leaves of at least 43 days). Services consisted of the number of hospital stays, duration of hospitalization, medical department of the hospital involved and specialties visited. Covariates Demographic characteristics of patients at the index date included age (as a continuous variable and age groups) and gender. Further baseline variables were year of index date and the specialty of the physician initiating treatment. Furthermore, a revised version of the Charlson comorbidity index (CCI) was used as a generic marker of comorbidity. Statistical methods To control for confounding, 1:1 matching was carried out based on a propensity score that was constructed as the conditional probability of being persistent as a function of age, gender, index year, physician specialty and CCI (logistic regression). Greedy matching was used by choosing a non-persistent patient whose propensity score was closest to that of this randomly selected persistent subject for matching. Descriptive statistics were given and group differences (persistent vs non-persistent) were assessed using the Wilcoxon test or the McNemar’s tests after propensity score matching. Finally, the differences in direct costs, indirect costs and services between matched pairs were estimated using the Wilcoxon test. Results Patient characteristics A total of 2060 patients initiating s.c. TNFi therapy were included in the study. After 2 years of follow-up, 682 patients still received s.c. TNFi therapy and 1378 patients had discontinued therapy. In the cohort of non-persistent patients, the mean persistence duration was 266 days (s.d. 190). We observed no significant difference between persistent and non-persistent patients in terms of age, gender or year of therapy initiation. Non-persistent patients had a slightly worse average health status compared with the persistent cohort (CCI 2.6 vs 2.4) (Table 1). After 1:1 matching, 678 pairs (678 persistent and 678 non-persistent patients) were available for cost analyses. Both cohorts were similar in terms of age, gender, year of therapy initiation, CCI and indication (Table 1). Table 1 Baseline characteristics of study patients before and after matching Variable Before matching After matching Persistent Non-persistent P-value Persistent Non-persistent P-value Total (N) 682  1378 678  678 Indication     RA (%) 55.7 58.8 0.040 55.9 57.2 0.520     AS (%) 21.7 19.9 0.579 21.8 21.7 1.000     PsA (%) 10.7 8.1 0.277 10.3 8.7 0.454     Mixed (%) 11.9 13.3 0.369 11.9 12.4 0.803 Age, mean (s.d.), years 51.1 (13.2) 51.4 (14.1) 0.669 51.1 (13.2) 51.2 (14.0) 0.905 Gender, male, % 42.4 41.1 0.572 42.5 44.0 0.583 Index year, %     2010 29.6 25.7 0.305 29.4 29.4 0.854     2011 23.6 24.5 23.6 25.5     2012 24.5 25.9 24.6 23.6     2013 22.3 23.9 22.4 21.5 CCI, mean (s.d.) 2.4 (1.3) 2.6 (1.6) <0.001 2.0 (1.3) 2.0 (1.3) 0.9227 Variable Before matching After matching Persistent Non-persistent P-value Persistent Non-persistent P-value Total (N) 682  1378 678  678 Indication     RA (%) 55.7 58.8 0.040 55.9 57.2 0.520     AS (%) 21.7 19.9 0.579 21.8 21.7 1.000     PsA (%) 10.7 8.1 0.277 10.3 8.7 0.454     Mixed (%) 11.9 13.3 0.369 11.9 12.4 0.803 Age, mean (s.d.), years 51.1 (13.2) 51.4 (14.1) 0.669 51.1 (13.2) 51.2 (14.0) 0.905 Gender, male, % 42.4 41.1 0.572 42.5 44.0 0.583 Index year, %     2010 29.6 25.7 0.305 29.4 29.4 0.854     2011 23.6 24.5 23.6 25.5     2012 24.5 25.9 24.6 23.6     2013 22.3 23.9 22.4 21.5 CCI, mean (s.d.) 2.4 (1.3) 2.6 (1.6) <0.001 2.0 (1.3) 2.0 (1.3) 0.9227 CCI: Charlson comorbidity index. Table 1 Baseline characteristics of study patients before and after matching Variable Before matching After matching Persistent Non-persistent P-value Persistent Non-persistent P-value Total (N) 682  1378 678  678 Indication     RA (%) 55.7 58.8 0.040 55.9 57.2 0.520     AS (%) 21.7 19.9 0.579 21.8 21.7 1.000     PsA (%) 10.7 8.1 0.277 10.3 8.7 0.454     Mixed (%) 11.9 13.3 0.369 11.9 12.4 0.803 Age, mean (s.d.), years 51.1 (13.2) 51.4 (14.1) 0.669 51.1 (13.2) 51.2 (14.0) 0.905 Gender, male, % 42.4 41.1 0.572 42.5 44.0 0.583 Index year, %     2010 29.6 25.7 0.305 29.4 29.4 0.854     2011 23.6 24.5 23.6 25.5     2012 24.5 25.9 24.6 23.6     2013 22.3 23.9 22.4 21.5 CCI, mean (s.d.) 2.4 (1.3) 2.6 (1.6) <0.001 2.0 (1.3) 2.0 (1.3) 0.9227 Variable Before matching After matching Persistent Non-persistent P-value Persistent Non-persistent P-value Total (N) 682  1378 678  678 Indication     RA (%) 55.7 58.8 0.040 55.9 57.2 0.520     AS (%) 21.7 19.9 0.579 21.8 21.7 1.000     PsA (%) 10.7 8.1 0.277 10.3 8.7 0.454     Mixed (%) 11.9 13.3 0.369 11.9 12.4 0.803 Age, mean (s.d.), years 51.1 (13.2) 51.4 (14.1) 0.669 51.1 (13.2) 51.2 (14.0) 0.905 Gender, male, % 42.4 41.1 0.572 42.5 44.0 0.583 Index year, %     2010 29.6 25.7 0.305 29.4 29.4 0.854     2011 23.6 24.5 23.6 25.5     2012 24.5 25.9 24.6 23.6     2013 22.3 23.9 22.4 21.5 CCI, mean (s.d.) 2.4 (1.3) 2.6 (1.6) <0.001 2.0 (1.3) 2.0 (1.3) 0.9227 CCI: Charlson comorbidity index. Cost analyses Table 2 shows the mean per-patient costs in the persistent and non-persistent cohorts. The costs for office-based visits per patient were €2319 in the persistent cohort and €3094 in the non-persistent cohort (P < 0.001). The cost difference was the highest in the mixed indication group (Table 2). An average persistent patient received prescriptions of different co-medications worth €2828 and a non-persistent patient received prescriptions worth €5498 within the 2-year time period (P < 0.001). In this category, the highest cost difference was found for patients with RA. Hospitalization costs were €3551 per persistent patient and €5890 per non-persistent patients (Table 2), which was probably due to the longer average duration of hospitalization in the non-persistent cohort. Furthermore, sick leave costs were much lower in persistent (€717) vs non-persistent (€1241) patients. Total costs per patient (excluding the cost of s.c. TNFi therapy) were ∼€6634 lower in the persistent cohort (Table 2). Table 2 Average costs per patient within 2 years after initiation of s.c. TNFi therapy in persistent and non-persistent patients Indication Non-persistent, mean (s.d.), €; Persistent, mean (s.d.), € Cost difference P-value Office-based visits     Total 3094 (4601) 2319 (2021) 775 <0.001     RA 3388 (5642) 2426 (1884) 962 0.005     AS 2447 (2261) 1809 (1573) 638 0.001     PsA 2152 (1563) 2897 (3321) −745 0.387     Mixed 3530 (3407) 2249 (1701) 1281 0.012 Prescriptions (excluding TNF-α)     Total 5492 (9876) 2828 (13 332) 2664 <0.001     RA 6914 (10 109) 3183 (15 788) 3731 <0.001     AS 2369 (6719) 2807 (12 982) −438 0.615     PsA 4621 (12 388) 1998 (3237) 2623 0.985     Mixed 5003 (10 245) 1921 (2376) 3082 0.662 Prescriptions (including TNF-α)     Total 27 222 (13 902) 43 707 (14 254) −16 485 <0.001     RA 27 979 (13 418) 43 759 (16 403) −15 780 <0.001     AS 24 561 (13 857) 43 926 (14 772) −19 365 <0.001     PsA 30 223 (15 197) 43 629 (5319) −13 407 <0.001     Mixed 26 271 (14 676) 43 127 (4727) −16 856 <0.001 Service utilization     Total 2475 (8010) 2142 (7216) 333 0.079     RA 3125 (9920) 2142 (6760) 983 0.007     AS 1011 (2478) 2326 (9589) −1315 0.422     PsA 1722 (4572) 1165 (2904) 557 0.925     Mixed 2564 (5732) 2649 (6884) −85 0.117 Hospitalizations     Total 5890 (14 951) 3552 (11 135) 2338 <0.001     RA 6869 (13 868) 4189 (12 930) 2680 <0.001     AS 3051 (7323) 2027 (6708) 1024 0.157     PsA 8862 (31 399) 3466 (11 076) 5396 0.249     Mixed 4252 (10 753) 3429 (8027) 823 0.490 Costs associated with sick leave     Total 1241 (3774) 717 (2784) 524 0.011     RA 1342 (3918) 620 (2852) 722 0.007     AS 997 (3514) 1072 (2982) −75 0.386     PsA 1330 (4525) 594 (1689) 736 0.937     Mixed 1135 (2898) 629 (2844) 506 0.007 Total costs (excluding TNF-α)     Total 18 192 (24 628) 11 558 (20 823) 6634 <0.001     RA 21 637 (25 851) 12 561 (23 801) 9076 <0.001     AS 9875 (12 416) 10 042 (19 240) −167 0.143     PsA 18 687 (36 880) 10 119 (13 312) 8568 0.444     Mixed 16 485 (20 702) 10 877 (12 175) 5608 0.169 Total costs (including TNF-α)     Total 39 922 (25 670) 52 437 (21 521) −12 515 <0.001     RA 42 703 (26 436) 53 138 (24 119) −10 435 <0.001     AS 32 068 (17 211) 51 160 (21 153) −19 092 <0.001     PsA 44 289 (37 175) 51 751 (14 273) −7462 <0.001     Mixed 37 753 (21 575) 52 083 (12 929) −14 330 <0.001 Indication Non-persistent, mean (s.d.), €; Persistent, mean (s.d.), € Cost difference P-value Office-based visits     Total 3094 (4601) 2319 (2021) 775 <0.001     RA 3388 (5642) 2426 (1884) 962 0.005     AS 2447 (2261) 1809 (1573) 638 0.001     PsA 2152 (1563) 2897 (3321) −745 0.387     Mixed 3530 (3407) 2249 (1701) 1281 0.012 Prescriptions (excluding TNF-α)     Total 5492 (9876) 2828 (13 332) 2664 <0.001     RA 6914 (10 109) 3183 (15 788) 3731 <0.001     AS 2369 (6719) 2807 (12 982) −438 0.615     PsA 4621 (12 388) 1998 (3237) 2623 0.985     Mixed 5003 (10 245) 1921 (2376) 3082 0.662 Prescriptions (including TNF-α)     Total 27 222 (13 902) 43 707 (14 254) −16 485 <0.001     RA 27 979 (13 418) 43 759 (16 403) −15 780 <0.001     AS 24 561 (13 857) 43 926 (14 772) −19 365 <0.001     PsA 30 223 (15 197) 43 629 (5319) −13 407 <0.001     Mixed 26 271 (14 676) 43 127 (4727) −16 856 <0.001 Service utilization     Total 2475 (8010) 2142 (7216) 333 0.079     RA 3125 (9920) 2142 (6760) 983 0.007     AS 1011 (2478) 2326 (9589) −1315 0.422     PsA 1722 (4572) 1165 (2904) 557 0.925     Mixed 2564 (5732) 2649 (6884) −85 0.117 Hospitalizations     Total 5890 (14 951) 3552 (11 135) 2338 <0.001     RA 6869 (13 868) 4189 (12 930) 2680 <0.001     AS 3051 (7323) 2027 (6708) 1024 0.157     PsA 8862 (31 399) 3466 (11 076) 5396 0.249     Mixed 4252 (10 753) 3429 (8027) 823 0.490 Costs associated with sick leave     Total 1241 (3774) 717 (2784) 524 0.011     RA 1342 (3918) 620 (2852) 722 0.007     AS 997 (3514) 1072 (2982) −75 0.386     PsA 1330 (4525) 594 (1689) 736 0.937     Mixed 1135 (2898) 629 (2844) 506 0.007 Total costs (excluding TNF-α)     Total 18 192 (24 628) 11 558 (20 823) 6634 <0.001     RA 21 637 (25 851) 12 561 (23 801) 9076 <0.001     AS 9875 (12 416) 10 042 (19 240) −167 0.143     PsA 18 687 (36 880) 10 119 (13 312) 8568 0.444     Mixed 16 485 (20 702) 10 877 (12 175) 5608 0.169 Total costs (including TNF-α)     Total 39 922 (25 670) 52 437 (21 521) −12 515 <0.001     RA 42 703 (26 436) 53 138 (24 119) −10 435 <0.001     AS 32 068 (17 211) 51 160 (21 153) −19 092 <0.001     PsA 44 289 (37 175) 51 751 (14 273) −7462 <0.001     Mixed 37 753 (21 575) 52 083 (12 929) −14 330 <0.001 Table 2 Average costs per patient within 2 years after initiation of s.c. TNFi therapy in persistent and non-persistent patients Indication Non-persistent, mean (s.d.), €; Persistent, mean (s.d.), € Cost difference P-value Office-based visits     Total 3094 (4601) 2319 (2021) 775 <0.001     RA 3388 (5642) 2426 (1884) 962 0.005     AS 2447 (2261) 1809 (1573) 638 0.001     PsA 2152 (1563) 2897 (3321) −745 0.387     Mixed 3530 (3407) 2249 (1701) 1281 0.012 Prescriptions (excluding TNF-α)     Total 5492 (9876) 2828 (13 332) 2664 <0.001     RA 6914 (10 109) 3183 (15 788) 3731 <0.001     AS 2369 (6719) 2807 (12 982) −438 0.615     PsA 4621 (12 388) 1998 (3237) 2623 0.985     Mixed 5003 (10 245) 1921 (2376) 3082 0.662 Prescriptions (including TNF-α)     Total 27 222 (13 902) 43 707 (14 254) −16 485 <0.001     RA 27 979 (13 418) 43 759 (16 403) −15 780 <0.001     AS 24 561 (13 857) 43 926 (14 772) −19 365 <0.001     PsA 30 223 (15 197) 43 629 (5319) −13 407 <0.001     Mixed 26 271 (14 676) 43 127 (4727) −16 856 <0.001 Service utilization     Total 2475 (8010) 2142 (7216) 333 0.079     RA 3125 (9920) 2142 (6760) 983 0.007     AS 1011 (2478) 2326 (9589) −1315 0.422     PsA 1722 (4572) 1165 (2904) 557 0.925     Mixed 2564 (5732) 2649 (6884) −85 0.117 Hospitalizations     Total 5890 (14 951) 3552 (11 135) 2338 <0.001     RA 6869 (13 868) 4189 (12 930) 2680 <0.001     AS 3051 (7323) 2027 (6708) 1024 0.157     PsA 8862 (31 399) 3466 (11 076) 5396 0.249     Mixed 4252 (10 753) 3429 (8027) 823 0.490 Costs associated with sick leave     Total 1241 (3774) 717 (2784) 524 0.011     RA 1342 (3918) 620 (2852) 722 0.007     AS 997 (3514) 1072 (2982) −75 0.386     PsA 1330 (4525) 594 (1689) 736 0.937     Mixed 1135 (2898) 629 (2844) 506 0.007 Total costs (excluding TNF-α)     Total 18 192 (24 628) 11 558 (20 823) 6634 <0.001     RA 21 637 (25 851) 12 561 (23 801) 9076 <0.001     AS 9875 (12 416) 10 042 (19 240) −167 0.143     PsA 18 687 (36 880) 10 119 (13 312) 8568 0.444     Mixed 16 485 (20 702) 10 877 (12 175) 5608 0.169 Total costs (including TNF-α)     Total 39 922 (25 670) 52 437 (21 521) −12 515 <0.001     RA 42 703 (26 436) 53 138 (24 119) −10 435 <0.001     AS 32 068 (17 211) 51 160 (21 153) −19 092 <0.001     PsA 44 289 (37 175) 51 751 (14 273) −7462 <0.001     Mixed 37 753 (21 575) 52 083 (12 929) −14 330 <0.001 Indication Non-persistent, mean (s.d.), €; Persistent, mean (s.d.), € Cost difference P-value Office-based visits     Total 3094 (4601) 2319 (2021) 775 <0.001     RA 3388 (5642) 2426 (1884) 962 0.005     AS 2447 (2261) 1809 (1573) 638 0.001     PsA 2152 (1563) 2897 (3321) −745 0.387     Mixed 3530 (3407) 2249 (1701) 1281 0.012 Prescriptions (excluding TNF-α)     Total 5492 (9876) 2828 (13 332) 2664 <0.001     RA 6914 (10 109) 3183 (15 788) 3731 <0.001     AS 2369 (6719) 2807 (12 982) −438 0.615     PsA 4621 (12 388) 1998 (3237) 2623 0.985     Mixed 5003 (10 245) 1921 (2376) 3082 0.662 Prescriptions (including TNF-α)     Total 27 222 (13 902) 43 707 (14 254) −16 485 <0.001     RA 27 979 (13 418) 43 759 (16 403) −15 780 <0.001     AS 24 561 (13 857) 43 926 (14 772) −19 365 <0.001     PsA 30 223 (15 197) 43 629 (5319) −13 407 <0.001     Mixed 26 271 (14 676) 43 127 (4727) −16 856 <0.001 Service utilization     Total 2475 (8010) 2142 (7216) 333 0.079     RA 3125 (9920) 2142 (6760) 983 0.007     AS 1011 (2478) 2326 (9589) −1315 0.422     PsA 1722 (4572) 1165 (2904) 557 0.925     Mixed 2564 (5732) 2649 (6884) −85 0.117 Hospitalizations     Total 5890 (14 951) 3552 (11 135) 2338 <0.001     RA 6869 (13 868) 4189 (12 930) 2680 <0.001     AS 3051 (7323) 2027 (6708) 1024 0.157     PsA 8862 (31 399) 3466 (11 076) 5396 0.249     Mixed 4252 (10 753) 3429 (8027) 823 0.490 Costs associated with sick leave     Total 1241 (3774) 717 (2784) 524 0.011     RA 1342 (3918) 620 (2852) 722 0.007     AS 997 (3514) 1072 (2982) −75 0.386     PsA 1330 (4525) 594 (1689) 736 0.937     Mixed 1135 (2898) 629 (2844) 506 0.007 Total costs (excluding TNF-α)     Total 18 192 (24 628) 11 558 (20 823) 6634 <0.001     RA 21 637 (25 851) 12 561 (23 801) 9076 <0.001     AS 9875 (12 416) 10 042 (19 240) −167 0.143     PsA 18 687 (36 880) 10 119 (13 312) 8568 0.444     Mixed 16 485 (20 702) 10 877 (12 175) 5608 0.169 Total costs (including TNF-α)     Total 39 922 (25 670) 52 437 (21 521) −12 515 <0.001     RA 42 703 (26 436) 53 138 (24 119) −10 435 <0.001     AS 32 068 (17 211) 51 160 (21 153) −19 092 <0.001     PsA 44 289 (37 175) 51 751 (14 273) −7462 <0.001     Mixed 37 753 (21 575) 52 083 (12 929) −14 330 <0.001 However, as expected, when including the cost of s.c. TNFi prescriptions, the total cost per patient was higher in the persistent cohort than in the non-persistent cohort (Table 2). Discussion This retrospective study of 1356 patients found that the 2-year costs per patient for office-based visits, hospitalizations, co-medications and sick leave costs were higher in patients who discontinued their s.c. TNFi therapy after an average of 9 months compared with patients with at least 24 months of therapy. Moreover, the average duration of hospitalization was longer in non-persistent patients. Since the database used does not include data on symptoms and diagnosis severity, costs can be used as a surrogate parameter for patients’ health status. Evidence regarding the share of patients who continued to do well after discontinuation of anti-TNF therapy is lacking. However, even if only some patients who discontinue s.c. TNFi therapy experience an increase in disease activity, this can result in more physician visits, longer hospitalizations and more days of sick leave. On the one hand, this can be caused by the IMRD symptoms themselves. On the other hand, increased time-averaged disease activity in RA can be associated with a higher risk of cardiovascular events [18] as well as a higher probability of developing infections [19]. The difference in non-TNF medication in our study was ∼54%, which is in line with results of Dalen et al. [14] who reported a difference of 56%. Regardless of the diagnosis, but especially in the case of diseases accompanied by pain, better persistence and greater adherence is associated with a considerable reduction in cost. This does not refer to total costs but rather to co-therapy and hospitalizations. Therefore such cost savings provide indirect but clear evidence regarding quality of life. TNFi products are relatively costly and patients with longer treatment durations, that is, better persistence, receive a greater number of prescriptions, which leads to higher treatment costs, while at the same time reducing other costs (co-therapies, hospital, service utilization, sick leave costs). However, longer therapy can achieve an improvement in the quality of life of IMRD patients. Moreover, the value of good persistence can be considered in terms of benefits to all participants in the health care system, including health care providers and payers. Study limitations Each retrospective study is limited by the availability of data in this database. The specific markers or tests of disease severity and the reasons for discontinuation are not recorded in the database. Furthermore, analyses are based on data of the Betriebskrankenkasse insurance fund, which is only one of many insurance funds in Germany, and the patient characteristics of different funds may vary. Moreover, no privately insured patients are included in the analyses. Finally, this study used mean values for costs per patient. However, mean values are highly sensitive to outliers. In this study, service utilization costs in patients with AS were €1011 (s.d. 2478) in non-persistent and €2326 (s.d. 9589) in persistent patients, due to some outliers with very high costs in persistent patients. On the other side, the median costs were higher in non-persistent vs persistent patients (€262 vs €200), showing the role of outliers. The same statistical effect of outliers can be observed in office-based visit costs of patients with PsA. Conclusion The results of this study indicate that persistence with s.c. TNFi treatment can be associated with several cost offsets for IMRD patients, with an improvement in their quality of life. For health care providers and payers, these findings reinforce the value of persistence, given the additional economic burden associated with non-persistent patients. Funding: This study was funded by Merck & Co., Kenilworth, NJ, USA. The sponsor was involved with the writing, study design, analysis and interpretation of data. Disclosure statement: K.Z., K.K., M.H. and S.D. are employees of QuintilesIMS. QuintilesIMS were paid consultants to Merck & Co., Kenilworth, NJ USA, in conjunction with the development of this manuscript. M.F. and H.F. are employees of Team Gesundheit Gesellschaft für Gesundheitsmanagement. S.K. is an employee of Merck & Co., Kenilworth, NJ, USA, and holds stock and options. References 1 Silman AJ , Hochberg MC. Epidemiology of the rheumatic diseases . New York, USA : Oxford University Press , 2001 . 2 Smolen JS , Landewe R , Breedveld FC et al. EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological disease-modifying antirheumatic drugs . Ann Rheum Dis 2010 ; 69 : 964 – 75 . Google Scholar CrossRef Search ADS PubMed 3 Braun J , Berg R , Boehm H et al. 2010 update of the ASAS/EULAR recommendations for the management of ankylosing spondylitis . Ann Rheum Dis 2011 ; 70 : 896 – 904 . Google Scholar CrossRef Search ADS PubMed 4 Gossec L , Smolen JS , Gaujoux-Viala C et al. European League Against Rheumatism recommendations for the management of psoriatic arthritis with pharmacological therapies . Ann Rheum Dis 2012 ; 71 : 4 – 12 . Google Scholar CrossRef Search ADS PubMed 5 Stevens SR , Chang TH. History of development of TNF inhibitors . Basel, Switzerland : Springer , 2006 : 9 – 22 . Google Scholar CrossRef Search ADS 6 Maxwell LJ , Zochling J , Boonen A et al. TNF-alpha inhibitors for ankylosing spondylitis . Cochrane Database Syst Rev 2015 ; 4 : CD005468 . 7 Aaltonen KJ , Virkki LM , Malmivaara A et al. Systematic review and meta-analysis of the efficacy and safety of existing TNF blocking agents in treatment of rheumatoid arthritis . PLoS One 2012 ; 7 : e30275 . Google Scholar CrossRef Search ADS PubMed 8 Lemos LL , de Oliveira Costa J , Almeida AM et al. Treatment of psoriatic arthritis with anti-TNF agents: a systematic review and meta-analysis of efficacy, effectiveness and safety . Rheumatol Int 2014 ; 34 : 1345 – 60 . Google Scholar CrossRef Search ADS PubMed 9 Aletaha D , Smolen JS. Effectiveness profiles and dose dependent retention of traditional disease modifying antirheumatic drugs for rheumatoid arthritis. An observational study . J Rheumatol 2002 ; 29 : 1631 – 8 . Google Scholar PubMed 10 Fries JF. Effectiveness and toxicity considerations in outcome directed therapy in rheumatoid arthritis . J Rheumatol Suppl 1996 ; 44 : 102 – 6 . Google Scholar PubMed 11 Pincus T , Marcum SB , Callahan LF. Long-term drug therapy for rheumatoid arthritis in seven rheumatology private practices: II. Second line drugs and prednisone . J Rheumatol 1992 ; 19 : 1885 – 94 . Google Scholar PubMed 12 Emery P , Keystone E , Tony HP et al. IL-6 receptor inhibition with tocilizumab improves treatment outcomes in patients with rheumatoid arthritis refractory to anti-tumour necrosis factor biologicals: results from a 24-week multicentre randomised placebo-controlled trial . Ann Rheum Dis 2008 ; 67 : 1516 – 23 . Google Scholar CrossRef Search ADS PubMed 13 Salaffi F , Carotti M , Gasparini S , Intorcia M , Grassi W. The health-related quality of life in rheumatoid arthritis, ankylosing spondylitis, and psoriatic arthritis: a comparison with a selected sample of healthy people . Health Qual Life Outcomes 2009 ; 7 : 25 . Google Scholar CrossRef Search ADS PubMed 14 Dalén J , Svedbom A , Black CM et al. Treatment persistence among patients with immune-mediated rheumatic disease newly treated with subcutaneous TNF-alpha inhibitors and costs associated with non-persistence . Rheumatol Int 2016 ; 36 : 987 – 95 . Google Scholar CrossRef Search ADS PubMed 15 Dalén J , Svedbom A , Black CM , Kachroo S. Second-line treatment persistence and costs among patients with immune-mediated rheumatic diseases treated with subcutaneous TNF-alpha inhibitors . Rheumatol Int 2017 ; 37 : 2049 – 58 . Google Scholar CrossRef Search ADS PubMed 16 Svedbom A , Dalén J,M , Black CM , Kachroo S. Persistence and costs with subcutaneous TNF-alpha inhibitors in immune-mediated rheumatic disease stratified by treatment line . Patient Pref Adherence 2017 ; 11 : 95 – 106 . Google Scholar CrossRef Search ADS 17 Friedel H , Delges A , Clouth J , Trautvetter DT. Expenditures of the German statutory health insurance system for patients suffering from acute coronary syndrome and treated with percutaneous coronary intervention . Eur J Health Econ 2010 ; 11 : 449 – 55 . Google Scholar CrossRef Search ADS PubMed 18 Solomon DH , Reed GW , Kremer JM et al. Disease activity in rheumatoid arthritis and the risk of cardiovascular events . Arthritis Rheumatol 2015 ; 67 : 1449 – 55 . Google Scholar CrossRef Search ADS PubMed 19 Au K , Reed G , Curtis JR et al. High disease activity is associated with an increased risk of infection in patients with rheumatoid arthritis . Ann Rheum Dis 2011 ; 70 : 785 – 91 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

van der Linden, Maarten;van den Hoogen, Lucas L;Westerlaken, Geertje H A;Fritsch-Stork, Ruth D E;van Roon, Joël A G;Radstake, Timothy R D J;Meyaard, Linde

2018 Rheumatology

doi: 10.1093/rheumatology/key067pmid: 29608758

Abstract Objectives Increased release of neutrophil extracellular traps (NETs) is implicated in the activation of plasmacytoid dendritic cells, vascular disease and thrombosis in SLE and APS. However, studies comparing NET release between patients with SLE and APS are lacking. Here we evaluated plasma-induced NET release in a large cohort of patients with SLE, SLE + APS and primary APS in relation to clinical and serological parameters. Methods Neutrophils from healthy controls were exposed to plasma of heterologous healthy controls (n = 27) or SLE (n = 55), SLE + APS (n = 38) or primary APS (PAPS) (n = 28) patients and NET release was quantified by immunofluorescence. In a subset of SLE patients, NET release was assessed in longitudinal samples before and after a change in treatment. Results Plasma-induced NET release was increased in SLE and APS patients, with the highest NET release found in patients with SLE (±APS). Plasma of 60% of SLE, 61% of SLE + APS and 45% of PAPS patients induced NET release. NET release did not correlate with disease activity in SLE or APS. However, increased levels of anti-nuclear and anti-dsDNA autoantibodies were associated with increased NET release in SLE and APS. Only in SLE patients, elevated NET release and an increased number of low-density granulocytes were associated with a high IFN signature. Conclusion Increased NET release is associated with autoimmunity and inflammation in SLE and APS. Inhibition of NET release thus could be of potential benefit in a subset of patients with SLE and APS. anti-phospholipid syndrome, autoantibodies, interferon signature, neutrophil extracellular traps, systemic lupus erythematosus Rheumatology key messages Plasma of SLE, SLE + APS and primary APS patients induces neutrophil extracellular trap release. Plasma-induced neutrophil extracellular trap release is associated with increased anti-nuclear and anti-dsDNA antibodies. SLE, but not APS, plasma-induced neutrophil extracellular trap release is associated with the IFN signature. Introduction SLE and APS are overlapping autoimmune diseases that can occur separately or in the same patient. In SLE, immune complexes (ICs) of autoantibodies are deposited into tissues, leading to inflammation in several organs, including the kidney, skin and joints. In APS, aPLs activate endothelial cells and trophoblasts resulting in thrombosis and pregnancy morbidity. APS is termed primary APS (PAPS) when no underlying disease such as SLE is present. In APS, most research has focused on the pro-thrombotic role of aPLs. However, research in recent years indicates an important role for immune cells in the pathogenesis of (P)APS, often in a similar fashion as in SLE, although studies that compare immunopathology between SLE, SLE + APS and PAPS patients are scarce [1]. There is growing interest in the role of neutrophils in rheumatic diseases [2]. Neutrophils act as a first line of defence against infectious invaders by, among other strategies, the release of neutrophil extracellular traps (NETs). NETs consist of decondensed chromatin decorated with neutrophil-derived proteases and antimicrobial peptides that trap and kill pathogens [3]. Neutrophils from SLE and PAPS patients are prone to release NETs spontaneously [4, 5]. In addition, healthy neutrophils release NETs when stimulated in vitro with autoantibodies present in sera of SLE or APS patients [5, 6]. Furthermore, DNase activity is decreased in SLE and APS, resulting in increased NET exposure [2, 7], and SLE and APS patients have increased numbers of circulating low-density granulocytes (LDGs) [8, 9], a subset of neutrophils prone to undergo NET formation. As a result, SLE and APS patients have elevated levels of NET remnants in the circulation [2, 7] and NETs are present in affected tissues such as the skin or kidney in SLE or in aPL-induced thrombi [9, 10]. Uncontrolled NET release triggers a pathological cascade of events relevant for the pathophysiology of SLE and APS. NETs induce tissue damage [2], activate the clotting system to promote thrombus formation [5], induce endothelial dysfunction [11] and represent a source of autoantigens [4]. Moreover, in vitro, NETs activate plasmacytoid dendritic cells to produce IFN-α [4, 9, 12], which might explain the IFN signature in SLE and APS patients [13]. Until now, NET release has only been studied in SLE and APS separately in small-scale studies. The different methodologies to induce and quantify NET release hamper the comparison across studies. Recently we developed a novel NET assay that specifically measures NET release, as it distinguishes NET release from other forms of neutrophil death while the automatic quantification avoids subjectivity [14]. Here we employed our assay to investigate plasma-induced NET release in a large cohort of SLE, SLE + APS and PAPS patients in relation to clinical and serological parameters, including the IFN signature. Methods A detailed description of the methods is available as supplementary data at Rheumatology online. Study population SLE, SLE + APS and PAPS patients and age- and sex-matched healthy controls (HCs) were recruited from our outpatient clinic or in-house healthy donor service. PAPS patients did not meet classification criteria for SLE nor had clinical evidence of SLE. None of the patients had evidence of an ongoing infection. Patients were stratified by a high or low IFN signature as previously described [13]. The University Medical Center Utrecht medical ethical committee approved this study and all study participants signed informed consent. Quantification of NETs HC neutrophils were cultured with 10% (heterologous) plasma of patients or controls for 4 h in a 96-well plate. NET release was quantified as previously described, with or without fixation after 4 h. Sytox Green images were used to quantify NET area [14] (Fig. 1A). Fig. 1 View largeDownload slide NET release in response to plasma of SLE, SLE + APS and PAPS patients (A) Experimental approach to measure NET release at a fixed time point. Twenty fields of view were captured per condition and Sytox Green images were used to analyse NET area. (B) Live imaging showed enhanced NET release of HC neutrophils when exposed to SLE plasma compared with autologous plasma. Data points represent median ± interquartile range of four independent experiments with four plasma samples each. (C) Citrullinated histone H3 staining confirmed the presence of actual NETs and (D) fixation after 4 h did not affect NET quantification. In a pilot experiment, serum and plasma from HCs (n = 8) and SLE (n = 15) patients were used to induce NET release in HC neutrophils. The NET area of independent experiments with neutrophils from three different HC donors is presented. (E) Elevated NET release was shown in neutrophils exposed to plasma, not serum, from SLE patients compared with those exposed to plasma or serum from heterologous HCs. (F) Plasma-induced NET release correlated between independent experiments. (G) HC neutrophils exposed to heterologous plasma from SLE (n = 55), SLE + APS (n = 38) and PAPS (n = 28) patients displayed increased NET release compared with those exposed to plasma from HCs (n = 27). The NET area of independent experiments with neutrophils from four different HC donors is presented. (H) Prevalence of high NET release in patients with SLE, SLE + APS or PAPS. The data in (D), (E) and (G) are presented as means, ***P < 0.001. The images in (C) are representative of at least three experiments with neutrophils from different donors. Fig. 1 View largeDownload slide NET release in response to plasma of SLE, SLE + APS and PAPS patients (A) Experimental approach to measure NET release at a fixed time point. Twenty fields of view were captured per condition and Sytox Green images were used to analyse NET area. (B) Live imaging showed enhanced NET release of HC neutrophils when exposed to SLE plasma compared with autologous plasma. Data points represent median ± interquartile range of four independent experiments with four plasma samples each. (C) Citrullinated histone H3 staining confirmed the presence of actual NETs and (D) fixation after 4 h did not affect NET quantification. In a pilot experiment, serum and plasma from HCs (n = 8) and SLE (n = 15) patients were used to induce NET release in HC neutrophils. The NET area of independent experiments with neutrophils from three different HC donors is presented. (E) Elevated NET release was shown in neutrophils exposed to plasma, not serum, from SLE patients compared with those exposed to plasma or serum from heterologous HCs. (F) Plasma-induced NET release correlated between independent experiments. (G) HC neutrophils exposed to heterologous plasma from SLE (n = 55), SLE + APS (n = 38) and PAPS (n = 28) patients displayed increased NET release compared with those exposed to plasma from HCs (n = 27). The NET area of independent experiments with neutrophils from four different HC donors is presented. (H) Prevalence of high NET release in patients with SLE, SLE + APS or PAPS. The data in (D), (E) and (G) are presented as means, ***P < 0.001. The images in (C) are representative of at least three experiments with neutrophils from different donors. Statistical analysis The NET area of 20 different microscopic fields per well was averaged. The mean of log-transformed NET areas per plasma donor of four independent experiments was reported as the mean NET area. The J-statistic of the Youden index of the receiver operating characteristics (ROC) curve of NET release in patients (SLE, SLE + APS and PAPS) as compared with HCs was used to define a cut-off to stratify patients into high or low NET inducers. Differences between groups were tested two-sided by analysis of variance (ANOVA) and Tukey’s post-test or t test as appropriate (α = 0.05) using SPSS (v22). Results Validation of a high-throughput assay to measure plasma-induced NET release Our live-imaging assay to monitor NET release over time revealed NET release within 30 min after exposure to SLE plasma (Fig. 1B; supplementary Video S1, available at Rheumatology online). The presence of citrullinated histone H3 in the extracellular DNA confirmed the formation of NETs (Fig. 1C). To allow the measurement of >100 samples without a time difference between the first and last sample, we introduced a fixation step after 240 min, which did not affect the quantification of NETs (Fig. 1D). Pilot experiments with plasma and serum samples from SLE patients (n = 15) and HCs (n = 8) showed increased NET induction by plasma from SLE patients as compared with HCs. Although the sera of both patients and controls had higher NET induction capacity than plasma, no difference was observed between SLE patients and HCs (Fig. 1E). In these pilot experiments we observed a moderate (r = 0.4375, P = 0.002) correlation of NET release between independent experiments using different neutrophil donors (Fig. 1F). Increased plasma-induced NET release in SLE, SLE + APS and PAPS We next used plasma samples of HCs (n = 27), SLE (n = 55), SLE + APS (n = 38) and PAPS (n = 28; supplementary Table S1, available at Rheumatology online) patients to induce NET release in neutrophils of four HC donors in four independent experiments. Confirming our pilot experiments, the mean NET release of four independent experiments was higher using plasma from SLE and SLE + APS patients as compared with HC plasma (P < 0.001), with a similar trend when using PAPS plasma (P = 0.14, Fisher’s least significant difference P = 0.03; Fig. 1G; supplementary Fig. S1, available at Rheumatology online). NET release did not differ among SLE, SLE + APS or PAPS patients (ANOVA P = 0.19). Setting a threshold by ROC curve analysis (supplementary Fig. S2, available at Rheumatology online), plasma samples from 33/55 (60%) of SLE, 23/38 (61%) of SLE + APS and 13/28 (46%) of PAPS patients had high NET release (Fig. 1H), as compared with 2/27 (7%) of HC plasma samples. Thus our data show that plasma from the majority of SLE and APS patients induces NET release. NET release did not correlate with clinical measures of disease activity, including Safety of Estrogen in Lupus Erythematosus National Assessment–SLEDAI for SLE(±APS) patients (P = 0.57; supplementary Fig. S2A, available at Rheumatology online) and the adjusted global APS score (aGAPSS) for (P)APS patients (P = 0.88; supplementary Fig. S2B, available at Rheumatology online). Furthermore, there were no significant differences among clinical phenotypes including (active) LN in SLE patients or APS patients with or without arterial or venous thrombosis or pregnancy morbidity. Also, NET release did not differ between patients treated with or without prednisolone, AZA, aspirin or other immunosuppressants (P > 0.05, data not shown). Plasma-induced NET release is associated with ANA and anti-dsDNA antibodies In vitro studies implicate NET release as a source of autoantigens eliciting the production of autoantibodies against nuclear components in SLE [5]. In line with these observations, SLE patients with high NET release had increased levels of anti-dsDNA antibodies compared with patients whose plasma induced low NET release (P = 0.008; Fig. 2A). Likewise, PAPS patients with high NET release had elevated ANA staining intensities (P < 0.05; Fig. 2B). In longitudinal samples, collected from SLE patients before and after a change in immunosuppressive therapy, a decline in anti-dsDNA antibodies (P = 0.02; Fig. 2C) was paralleled by a decline in NET release (P = 0.03; Fig. 2D), whereas patients with stable or increasing anti-dsDNA antibodies (Fig. 2E) between two time points did not have a decrease in plasma-induced NET release (P = 0.48; Fig. 2F). Among SLE or APS patients, no specific association between the presence or absence of anti-β2 glycoprotein I and anti-RNP antibodies was observed. Fig. 2 View largeDownload slide Plasma-induced NET release is associated with ANA levels in plasma of SLE and PAPS (A) Plasma of SLE patients classified as high NET inducers contained elevated levels of anti-dsDNA antibodies compared with those that are classified as low NET inducers. (B) Plasma of PAPS patients classified as high NET inducers contained increased ANA staining intensities compared with those that are classified as low NET inducers (−: negative; ±: weak; +: positive; ++: strongly positive). Longitudinal samples were collected from SLE patients at the time of active disease as well as subsequent quiescent disease. The (C) anti-dsDNA antibody level and (D) NET area were increased in SLE patients (n = 5) with active disease compared with those with quiescent disease. (E, F) Patients (n = 4) with stable or increasing anti-dsDNA antibodies between two time points did not have a decrease in plasma-induced NET release. (G) NET release in response to IC was high compared with SLE and APS plasma-induced NET release. Data points represent mean (s.d.) of independent experiments with neutrophils of three HC donors and four plasma samples per group. (H) DPI and NADPH oxidase inhibitor suppressed IC-induced NET release while SLE and PAPS plasma-induced NET release was independent of NADPH oxidase. The percentage of inhibition of NET release was calculated based on the area under the curve relative to neutrophils exposed to IC, PAPS or SLE plasma in the presence of DMSO. (I) Plasma from SLE patients classified as IFN-high displayed elevated levels of NET release compared with those that were classified as IFN-low. (J) An increased amount of LDGs were present in SLE IFN-high patients compared with SLE IFN-low patients. Differences in LDG amounts between IFN-high and IFN-low patients were not seen in SLE + APS and PAPS patients. The data in (A), (B), (H), (I) and (J) are presented as means, *P < 0.05 and **P < 0.005. Fig. 2 View largeDownload slide Plasma-induced NET release is associated with ANA levels in plasma of SLE and PAPS (A) Plasma of SLE patients classified as high NET inducers contained elevated levels of anti-dsDNA antibodies compared with those that are classified as low NET inducers. (B) Plasma of PAPS patients classified as high NET inducers contained increased ANA staining intensities compared with those that are classified as low NET inducers (−: negative; ±: weak; +: positive; ++: strongly positive). Longitudinal samples were collected from SLE patients at the time of active disease as well as subsequent quiescent disease. The (C) anti-dsDNA antibody level and (D) NET area were increased in SLE patients (n = 5) with active disease compared with those with quiescent disease. (E, F) Patients (n = 4) with stable or increasing anti-dsDNA antibodies between two time points did not have a decrease in plasma-induced NET release. (G) NET release in response to IC was high compared with SLE and APS plasma-induced NET release. Data points represent mean (s.d.) of independent experiments with neutrophils of three HC donors and four plasma samples per group. (H) DPI and NADPH oxidase inhibitor suppressed IC-induced NET release while SLE and PAPS plasma-induced NET release was independent of NADPH oxidase. The percentage of inhibition of NET release was calculated based on the area under the curve relative to neutrophils exposed to IC, PAPS or SLE plasma in the presence of DMSO. (I) Plasma from SLE patients classified as IFN-high displayed elevated levels of NET release compared with those that were classified as IFN-low. (J) An increased amount of LDGs were present in SLE IFN-high patients compared with SLE IFN-low patients. Differences in LDG amounts between IFN-high and IFN-low patients were not seen in SLE + APS and PAPS patients. The data in (A), (B), (H), (I) and (J) are presented as means, *P < 0.05 and **P < 0.005. NET kinetics is similar in SLE and PAPS We previously showed that the kinetics of NET release differs between stimuli [14]. We observed no difference in the kinetics of plasma-induced NET release among high-inducing plasma samples of SLE or PAPS patients (Fig. 2G). In comparison to patient plasma, exposure of neutrophils to ICs induced abundant NETs, ∼30% compared with 3% in patient plasma samples. Diphenyleneiodonium (DPI) inhibits, among others, NADPH oxidase, and NET release in response to ICs was inhibited by 60–70% in the presence of DPI while PAPS and SLE plasma-induced NET release was not inhibited in the presence of DPI (Fig. 2H). NET release and low-density granulocytes are associated with the IFN signature in SLE In vitro experiments implicate NETs and LDGs as a trigger for IFN-α production by plasmacytoid dendritic cells [4, 9], although no studies have explored NET release in relation to the presence or absence of the IFN signature. SLE patients with a high IFN signature (IFN-high) had higher NET release than patients with a low IFN signature (IFN-low) (P < 0.01; Fig. 2I). Corroborating this finding we observed that IFN-high SLE patients had increased numbers of circulating LDGs (P < 0.01; Fig. 2J). Interestingly, these associations were not seen in APS patients, neither in SLE + APS nor in PAPS (P > 0.05). Discussion Using a novel high-throughput assay we show that plasma of SLE, SLE + APS and PAPS patients induces NET release, which is associated with ANAs in PAPS and anti-dsDNA autoantibodies and the IFN signature in SLE patients. This study highlights the potential role of NET release in relation to autoimmunity and inflammation in SLE and APS and compares NET release in a large cohort of SLE and APS patients. Previous studies have shown induction of NET release using serum of SLE or APS patients [6, 15]. In small pilot studies, when comparing serum with plasma, serum induced a higher release of NETs than plasma, both in patients and HCs, and no difference between HCs and patients was observed. The generation of serum leads to platelet activation, which is a strong NET inducer [16] and is a likely cause for the higher NET release induced by serum samples. As a result, to avoid potential effects of platelet activation, we used patient plasma to trigger NET release in further experiments. Importantly, although similar trends were observed when using different neutrophil donors, the amount of NETs formed differed between neutrophil donors and therefore our results stress the need to use different neutrophil donors when studying NET release [16]. Besides the amount of NETs, the kinetics of NET release differ between stimuli [14]. We observed a rapid release of NETs (within 30 min) upon exposure to patient plasma in both SLE and PAPS patients. As the composition of NETs differs between stimuli [17], we speculate that the content of NETs could differ between SLE and APS, since NET release was differentially associated with the IFN signature in SLE and APS. NET release in the context of SLE and APS has been mainly studied using purified antibodies or cytokines to trigger NET release, including anti-β2 glycoprotein I, anti-RNP, anti-human neutrophil protein, anti-LL37, anti-MMP9 antibodies and IL-18 and hyperacetylated microparticles [4–6, 11, 12, 18–20]. As multiple factors present in patient plasma may induce NET release, it is unknown which stimulus is responsible for NET release in our assay; however, the association of autoantibodies and the IFN signature with NET release suggest their involvement, and this is clearly different in plasma from healthy individuals. Nevertheless, NET release by purified factors should be interpreted with caution since the concentration and composition of these factors in patients’ plasma might be different. Indeed, in our study, NET release in response to ICs was much higher than in response to patient plasma. Moreover, immune complex–induced NET release is dependent on NAPDH oxidase, in contrast to plasma-induced NET release, although it is unknown whether the magnitude of NET release in vitro can be directly translated to in vivo situations. Enhanced NET release is considered a major pathogenic factor linked to tissue damage, the IFN signature and other disease manifestations in both SLE and APS [4, 5, 10]. Consistent with this, we report increased NET release in patients with elevated autoantibodies or the IFN signature in SLE and APS. Treatment options that mitigate NET release could therefore be of added clinical value. Inhibition of NET release ameliorates mouse models of SLE and APS [10, 21]. Several small inhibitory molecules reduce NET release in vivo [22], while HCQ, a treatment for SLE, inhibits NET release in vitro [2]. We previously reported that triggering signal inhibitory receptor on leucocytes-1 attenuates SLE plasma and autoantibody-induced NET release [6]. Our current results indicate that only a subset of patients [∼60% of SLE(±APS) and ∼45% of PAPS patients] would benefit from inhibiting NET release. Acknowledgements L.M. was supported by a Vici grant (91815608) from the Netherlands Organization for Scientific Research (NWO). The authors thank Dr Michiel van der Vlist and Dr Inês Ramos for helpful discussions. Funding: This work was supported by the Dutch Arthritis Foundation (grant 12-2-406). Disclosure statement: The authors have declared no conflicts of interests. Supplementary data Supplementary data are available at Rheumatology online. References 1 van den Hoogen LL , van Roon JAG , Radstake TRDJ , Fritsch-Stork RDE , Derksen RHWM. Delineating the deranged immune system in the antiphospholipid syndrome . Autoimmun Rev 2016 ; 15 : 50 – 60 . Google Scholar CrossRef Search ADS PubMed 2 Grayson PC , Schauer C , Herrmann M , Kaplan MJ. Neutrophils as invigorated targets in rheumatic diseases . Arthritis Rheumatol 2016 ; 68 : 2071 – 82 . Google Scholar CrossRef Search ADS PubMed 3 Brinkmann V , Reichard U , Goosmann C et al. Neutrophil extracellular traps kill bacteria . Science 2004 ; 303 : 1532 – 5 . Google Scholar CrossRef Search ADS PubMed 4 Lande R , Ganguly D , Facchinetti V et al. Neutrophils activate plasmacytoid dendritic cells by releasing self-DNA-peptide complexes in systemic lupus erythematosus . Sci Transl Med 2011 ; 3 : 73ra19 . Google Scholar CrossRef Search ADS PubMed 5 Yalavarthi S , Gould TJ , Rao AN et al. Antiphospholipid antibodies promote the release of neutrophil extracellular traps: a new mechanism of thrombosis in the antiphospholipid syndrome . Arthritis Rheumatol 2015 ; 67 : 2290 – 3003 . Google Scholar CrossRef Search ADS 6 Van Avondt K , Fritsch-Stork R , Derksen RHWM , Meyaard L. Ligation of signal inhibitory receptor on leukocytes-1 suppresses the release of neutrophil extracellular traps in systemic lupus erythematosus . PLoS One 2013 ; 8 : e78459 . Google Scholar CrossRef Search ADS PubMed 7 Hakkim A , Fürnrohr BG , Amann K et al. Impairment of neutrophil extracellular trap degradation is associated with lupus nephritis . Proc Natl Acad Sci USA 2010 ; 107 : 9813 – 8 . Google Scholar CrossRef Search ADS PubMed 8 van den Hoogen L , Fritsch-Stork RDE , van Roon JAG , Radstake TRDJ. Low density granulocytes are increased in the antiphospholipid syndrome and are associated with anti-β2GPI antibodies . Arthritis Rheumatol 2016 ; 68 : 1320 – 1 . Google Scholar CrossRef Search ADS PubMed 9 Villanueva E , Yalavarthi S , Berthier CC et al. Netting neutrophils induce endothelial damage, infiltrate tissues, and expose immunostimulatory molecules in systemic lupus erythematosus . J Immunol 2011 ; 187 : 538 – 52 . Google Scholar CrossRef Search ADS PubMed 10 Meng H , Yalavarthi S , Kanthi Y et al. In vivo role of neutrophil extracellular traps in antiphospholipid antibody–mediated venous thrombosis . Arthritis Rheumatol 2017 ; 69 : 655 – 67 . Google Scholar CrossRef Search ADS PubMed 11 Carmona-Rivera C , Zhao W , Yalavarthi S , Kaplan MJ. Neutrophil extracellular traps induce endothelial dysfunction in systemic lupus erythematosus through the activation of matrix metalloproteinase-2 . Ann Rheum Dis 2014 ; 74 : 1417 – 24 . Google Scholar CrossRef Search ADS PubMed 12 Garcia-Romo GS , Caielli S , Vega B et al. Netting neutrophils are major inducers of type I IFN production in pediatric systemic lupus erythematosus . Sci Transl Med 2011 ; 3 : 73ra20 . Google Scholar CrossRef Search ADS PubMed 13 van den Hoogen LL , Fritsch-Stork RD , Versnel MA et al. Monocyte type I interferon signature in antiphospholipid syndrome is related to proinflammatory monocyte subsets, hydroxychloroquine and statin use . Ann Rheum Dis 2016 ; 75 : e81 . Google Scholar CrossRef Search ADS PubMed 14 van der Linden M , Westerlaken GHA , van der Vlist M , van Montfrans J , Meyaard L. Differential signalling and kinetics of neutrophil extracellular trap release revealed by quantitative live imaging . Sci Rep 2017 ; 7 : 6529 . Google Scholar CrossRef Search ADS PubMed 15 Kraaij T , Tengström FC , Kamerling SWA et al. Autoimmunity reviews: a novel method for high-throughput detection and quantification of neutrophil extracellular traps reveals ROS-independent NET release with immune complexes . Autoimmun Rev 2016 ; 15 : 577 – 84 . Google Scholar CrossRef Search ADS PubMed 16 Hoppenbrouwers T , Autar ASA , Sultan AR et al. In vitro induction of NETosis: comprehensive live imaging comparison and systematic review . PLoS One 2017 ; 12 : 1 – 29 . Google Scholar CrossRef Search ADS 17 Khandpur R , Carmona-Rivera C , Vivekanandan-Giri A et al. NETs are a source of citrullinated autoantigens and stimulate inflammatory responses in rheumatoid arthritis . Sci Transl Med 2013 ; 5 : 178ra40 . Google Scholar CrossRef Search ADS PubMed 18 Kahlenberg JM , Carmona-Rivera C , Smith CK , Kaplan MJ. Neutrophil extracellular trap-associated protein activation of the NLRP3 inflammasome is enhanced in lupus macrophages . J Immunol 2013 ; 190 : 1217 – 26 . Google Scholar CrossRef Search ADS PubMed 19 Rother N , Pieterse E , Lubbers J , Hilbrands L , van der Vlag J. Acetylated histones in apoptotic microparticles drive the formation of neutrophil extracellular traps in active lupus nephritis . Front Immunol 2017 ; 8 : 1136 . Google Scholar CrossRef Search ADS PubMed 20 Dieker J , Tel J , Pieterse E et al. Circulating apoptotic microparticles in systemic lupus erythematosus patients drive the activation of dendritic cell subsets and prime neutrophils for NETosis . Arthritis Rheumatol 2016 ; 68 : 462 – 72 . Google Scholar CrossRef Search ADS PubMed 21 Lood C , Blanco LP , Purmalek MM et al. Neutrophil extracellular traps enriched in oxidized mitochondrial DNA are interferogenic and contribute to lupus-like disease . Nat Med 2016 ; 22 : 146 – 53 . Google Scholar CrossRef Search ADS PubMed 22 van der Linden M , Meyaard L. Fine-tuning neutrophil activation: strategies and consequences . Immunol Lett 2016 ; 178 : 3 – 9 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Masoud, Sherry;McAdoo, Stephen P;Bedi, Rachna;Cairns, Thomas D;Lightstone, Liz

2018 Rheumatology

doi: 10.1093/rheumatology/key042pmid: 29562252

Abstract Objective B cell depletion, most commonly with rituximab, is an evolving therapeutic approach in SLE. Infusion reactions after rituximab are common, and may prevent re-treatment in patients who previously demonstrated beneficial response. We have used ofatumumab, a fully humanized anti-CD20 mAb, as an alternative B cell–depleting agent in patients with SLE who are rituximab-intolerant due to severe infusion reactions. Methods A single-centre retrospective case series of 16 patients were treated with ofatumumab for SLE between 2012 and 2015. Results Ofatumumab infusion was well tolerated in 14/16 patients, in whom the median age was 34 (range 19–55) and the median duration of SLE 9.2 years (0.6–28.5). The cohort was heavily pre-treated, with 50% having prior CYC exposure, and a median cumulative dose of prior rituximab 4 g (1–6). Twelve patients were treated for LN, one for extra-renal flare and one for remission maintenance. B cell–depletion was achieved in 12/14 patients, with comparable reconstitution kinetics to a previous cohort treated with rituximab at our centre, and was associated with improvements in serological markers of disease activity, including ANA, anti-dsDNA antibody and complement levels. Half of the patients with LN achieved renal remission by 6 months. Progressive disease that was unresponsive to augmented immunosuppression with CYC was seen in five patients. During long-term follow-up (median 28 months), five grade III infections were reported, and there were no malignancies or deaths. Conclusion In this pre-treated cohort with long-standing SLE, ofatumumab was a well-tolerated, safe and effective alternative to rituximab for B cell–depletion therapy. systemic lupus erythematosus, lupus nephritis, biologic therapy, monoclonal antibody, B cell depletion, rituximab, ofatumumab, anti-CD20, human-anti-chimeric antibodies Rheumatology key messages Ofatumumab is an alternative to rituximab for B cell depletion in SLE. Reaction infusions to ofatumumab were uncommon in patients with SLE who had demonstrated intolerance to rituximab. Ofatumumab treatment was associated with serological and clinical response in the majority of patients with SLE. Introduction SLE is a heterogeneous autoimmune condition with a variable and unpredictable course. Until recently, the mainstay of therapy had remained unchanged for decades, consisting of CSs, anti-malarials and non-specific immunosuppression. However, the challenges posed by refractory disease, and the adverse effects of these conventional therapies, particularly those of long-term steroids, have led to a search for new treatments. B cell depletion, most commonly with rituximab, a chimeric monoclonal IgG1 directed against the CD20 antigen expressed on B lymphocytes, has emerged as a potential strategy [1]. While two controlled trials, the Exploratory Phase II/III SLE Evaluation of Rituximab (EXPLORER) and the Lupus Nephritis Assessment with Rituximab (LUNAR) studies [2, 3], did not show any additional benefit of rituximab as an add-on to standard therapy in non-renal and renal SLE, respectively, we have previously reported successful use of rituximab as a steroid-sparing agent when used with MMF in patients diagnosed with LN [4, 5]. In addition, there is a significant body of evidence, albeit uncontrolled, that rituximab may be beneficial in refractory or resistant disease, or in patients intolerant of standard therapy [6]. Infusion reactions after rituximab are common, occurring in 10–15% of patients in the randomized studies, and these may be sufficiently severe to preclude repeated treatment in patients who had previously demonstrated good clinical response. Fully humanized mAb may avoid these adverse immunogenic reactions. To date, three humanized anti-CD20 mAb have been developed; obinutuzumab, ocrelizumab and ofatumumab. Compared with rituximab, ofatumumab is directed against a distinct extracellular epitope of CD20, has slower dissociation kinetics, and is a more potent activator of complement-dependent cytotoxicity in vitro [7]. It is licensed for use in haematological malignancies, and has demonstrated biological activity in RA [8, 9]. We previously reported the use of ofatumumab in a cohort of patients with ANCA-associated vasculitis [10], and there is an individual case report, and one small series, describing successful use of ofatumumab in a total of four patients with SLE [11, 12]. Since 2012, we have offered ofatumumab to patients with SLE who are intolerant of rituximab and for whom B cell depletion is deemed an appropriate therapeutic strategy. Here, we report our experience in 16 patients, the majority of whom were treated for LN and had previously demonstrated beneficial responses to rituximab. Methods This is a retrospective case series of patients treated with ofatumumab for SLE/LN at the Imperial College Lupus Centre. Since 2012, we have offered ofatumumab off label on compassionate grounds to patients who are intolerant of rituximab and for whom B cell depletion is deemed a desirable treatment aim. Patients were classified as intolerant of rituximab if they developed angioedema, anaphylactoid or anaphylactic reactions despite the use of intravenous steroids and antihistamines as pre-medication. Our treatment protocol was based on our previously published regimen using rituximab and MMF, without oral steroids (the Rituxilup regimen [4]). Herein, rituximab 2 × 1 g was replaced with ofatumumab 2 × 700 mg i.v. doses given 2 weeks apart, with 250–500 mg of i.v. methylprednisolone at each infusion. First-line maintenance was with MMF, and we aimed to minimize the use of oral steroids. Clinical and laboratory data were retrospectively extracted from case records, for at least 6 months after treatment until the last encounter prior to April 2016. Renal biopsy findings were classified according to the International Society of Nephrology/Renal Pathology Society system for LN. B cell counts were determined by CD19 flow cytometry, and a threshold of 20 cells/μl was used to define B cell depletion and repopulation. Complete renal remission was defined by the combination of a urine protein: creatinine ratio (uPCR) of <50 mg/mmol and serum creatinine not >15% above baseline, or by histological evidence of inactive disease on repeat biopsy. Partial remission was defined by uPCR <300 mg/mmol and 50% reduction in uPCR from baseline, with serum creatinine not >15% above baseline. Unless otherwise stated, data are reported as median (range), and comparisons made by non-parametric testing. Graphs were constructed and statistical analysis performed using Prism 7.0 (GraphPad Software, La Jolla, CA, USA). Informed consent was provided prior to initiation of therapy in all cases. In accordance with the UK National Health Service Research Ethics Committee guidelines, ethics approval was not required for this report, since all treatment decisions were made prior to this evaluation. Results Patient characteristics and treatment received To date, 16 patients with SLE have been treated with ofatumumab at our centre. All had previously been treated with rituximab, and had demonstrated infusion reactions that were deemed sufficiently severe to preclude further rituximab challenge. Two of these 16 patients had significant infusion reactions to ofatumumab, such that they were unable to complete treatment, and are therefore excluded from this analysis. The clinical features of the 14 patients who completed treatment are summarized in Table 1. All were female, with a mean age of 34 years (range 19–55) and median disease duration of 9.2 years (0.6–28.5) before they were treated with ofatumumab. The majority of patients were of non-Caucasian ethnicity (5/14 African Caribbean; 4/14 south-Asian; 2/14 Caucasian; 3/14 other). Fifty percent of patients had previously received CYC, and the median time since last rituximab exposure was 192 days (9–1099), with a median cumulative prior dose of 4 g (1–6). The indication for ofatumumab treatment in 12 patients was active LN. Of the remaining two patients, one was already established on haemodialysis and treated for extra-renal flare; one was treated to maintain remission due to non-adherence with oral medications. Table 1 Clinical characteristics, treatment received and outcome for individual patients Case Age, gender Duration SLE, years Previous therapiesa Previous RTX dose, g Time since RTX, days Indicationb Creatinine, μmol/l uPCR, mg/mmol Albumin, g/l Ofatumumab dose, mg Maintenance after ofatumumab Outcome at 6 months 1 44 F 18.4 CS, MMF, HCQ, CYP 6 22 LN: IV-S (A/C) 79 325 32 1400 CS, MMF, HCQ CRR 2 28 F 3.3 CS, MMF, HCQ 4 53 LN: III-S (A/C) 61 91 39 1400 MMF, HCQ CRR 3 35 F 5.8 CS, MMF, HCQ 5 24 LN: V 58 41 24 1400 CS, MMF, HCQ CRR 4 23 F 1.9 CS, MMF, CYP 3 134 LN: III-S (A) + V 79 604 25 1400 CS, MMF, HCQ PRR 5 41 F 8.2 CS, MMF, HCQ 5 42 LN: IV-S (A/C) 60 845 24 1400 CS, MMF, HCQ PRR 6 26 F 1.3 MMF, HCQ 4 9 LN: III-S (A) 70 84 33 1400 MMF, HCQ NR: lost function 7 33 F 8.8 CS, MMF, AZA, MTX 4 355 LN: V 109 414 17 700 MMF, HCQ NR: persistent proteinuria 8 31 F 7.3 CS, MMF, AZA, CYP 1 1099 LN: IV-G (A/C) 90 300 22 1400 CS, MMF NR: lost function 9 55 F 1.7 CS, MMF, MTX, HCQ, CYP 5 332 LN: IV-S (A/C) 88 590 27 1400 CS, MMF, HCQ NR: lost function 10 21 F 7.4 CS, MMF, CYP 5 174 LN: IV-S (A/C) + V 85 461 19 1400 MMF, HCQ NR: persistent proteinuria 11 43 F 0.6 CS, HCQ 2 15 LN: III-S (A/C) 112 429 28 1400 MMF PRR 12 48 F 28.5 CS, MMF, AZA, HCQ, CYP, FK 4 68 LN: V 110 911 14 1400 MMF, HCQ NR: persistent proteinuria 13 34 F 11.5 CS, MMF, AZA, CYP, IVIG 5 151 Extra-renal Flare – – – 700 CS, MMF, HCQ Clinical improvement 14 19 F 15.4 MMF, HCQ 2 219 Maintenance – – – 1400 Nil Stable remission Case Age, gender Duration SLE, years Previous therapiesa Previous RTX dose, g Time since RTX, days Indicationb Creatinine, μmol/l uPCR, mg/mmol Albumin, g/l Ofatumumab dose, mg Maintenance after ofatumumab Outcome at 6 months 1 44 F 18.4 CS, MMF, HCQ, CYP 6 22 LN: IV-S (A/C) 79 325 32 1400 CS, MMF, HCQ CRR 2 28 F 3.3 CS, MMF, HCQ 4 53 LN: III-S (A/C) 61 91 39 1400 MMF, HCQ CRR 3 35 F 5.8 CS, MMF, HCQ 5 24 LN: V 58 41 24 1400 CS, MMF, HCQ CRR 4 23 F 1.9 CS, MMF, CYP 3 134 LN: III-S (A) + V 79 604 25 1400 CS, MMF, HCQ PRR 5 41 F 8.2 CS, MMF, HCQ 5 42 LN: IV-S (A/C) 60 845 24 1400 CS, MMF, HCQ PRR 6 26 F 1.3 MMF, HCQ 4 9 LN: III-S (A) 70 84 33 1400 MMF, HCQ NR: lost function 7 33 F 8.8 CS, MMF, AZA, MTX 4 355 LN: V 109 414 17 700 MMF, HCQ NR: persistent proteinuria 8 31 F 7.3 CS, MMF, AZA, CYP 1 1099 LN: IV-G (A/C) 90 300 22 1400 CS, MMF NR: lost function 9 55 F 1.7 CS, MMF, MTX, HCQ, CYP 5 332 LN: IV-S (A/C) 88 590 27 1400 CS, MMF, HCQ NR: lost function 10 21 F 7.4 CS, MMF, CYP 5 174 LN: IV-S (A/C) + V 85 461 19 1400 MMF, HCQ NR: persistent proteinuria 11 43 F 0.6 CS, HCQ 2 15 LN: III-S (A/C) 112 429 28 1400 MMF PRR 12 48 F 28.5 CS, MMF, AZA, HCQ, CYP, FK 4 68 LN: V 110 911 14 1400 MMF, HCQ NR: persistent proteinuria 13 34 F 11.5 CS, MMF, AZA, CYP, IVIG 5 151 Extra-renal Flare – – – 700 CS, MMF, HCQ Clinical improvement 14 19 F 15.4 MMF, HCQ 2 219 Maintenance – – – 1400 Nil Stable remission a Previous treatment at any time point. b Classification according to the ISN/RPS system for LN. RTX: rituximab; uPCR: urinary protein: creatinine ratio; FK: tacrolimus; CRR: complete renal remission; PRR: partial renal remission; NR: non-responder. Table 1 Clinical characteristics, treatment received and outcome for individual patients Case Age, gender Duration SLE, years Previous therapiesa Previous RTX dose, g Time since RTX, days Indicationb Creatinine, μmol/l uPCR, mg/mmol Albumin, g/l Ofatumumab dose, mg Maintenance after ofatumumab Outcome at 6 months 1 44 F 18.4 CS, MMF, HCQ, CYP 6 22 LN: IV-S (A/C) 79 325 32 1400 CS, MMF, HCQ CRR 2 28 F 3.3 CS, MMF, HCQ 4 53 LN: III-S (A/C) 61 91 39 1400 MMF, HCQ CRR 3 35 F 5.8 CS, MMF, HCQ 5 24 LN: V 58 41 24 1400 CS, MMF, HCQ CRR 4 23 F 1.9 CS, MMF, CYP 3 134 LN: III-S (A) + V 79 604 25 1400 CS, MMF, HCQ PRR 5 41 F 8.2 CS, MMF, HCQ 5 42 LN: IV-S (A/C) 60 845 24 1400 CS, MMF, HCQ PRR 6 26 F 1.3 MMF, HCQ 4 9 LN: III-S (A) 70 84 33 1400 MMF, HCQ NR: lost function 7 33 F 8.8 CS, MMF, AZA, MTX 4 355 LN: V 109 414 17 700 MMF, HCQ NR: persistent proteinuria 8 31 F 7.3 CS, MMF, AZA, CYP 1 1099 LN: IV-G (A/C) 90 300 22 1400 CS, MMF NR: lost function 9 55 F 1.7 CS, MMF, MTX, HCQ, CYP 5 332 LN: IV-S (A/C) 88 590 27 1400 CS, MMF, HCQ NR: lost function 10 21 F 7.4 CS, MMF, CYP 5 174 LN: IV-S (A/C) + V 85 461 19 1400 MMF, HCQ NR: persistent proteinuria 11 43 F 0.6 CS, HCQ 2 15 LN: III-S (A/C) 112 429 28 1400 MMF PRR 12 48 F 28.5 CS, MMF, AZA, HCQ, CYP, FK 4 68 LN: V 110 911 14 1400 MMF, HCQ NR: persistent proteinuria 13 34 F 11.5 CS, MMF, AZA, CYP, IVIG 5 151 Extra-renal Flare – – – 700 CS, MMF, HCQ Clinical improvement 14 19 F 15.4 MMF, HCQ 2 219 Maintenance – – – 1400 Nil Stable remission Case Age, gender Duration SLE, years Previous therapiesa Previous RTX dose, g Time since RTX, days Indicationb Creatinine, μmol/l uPCR, mg/mmol Albumin, g/l Ofatumumab dose, mg Maintenance after ofatumumab Outcome at 6 months 1 44 F 18.4 CS, MMF, HCQ, CYP 6 22 LN: IV-S (A/C) 79 325 32 1400 CS, MMF, HCQ CRR 2 28 F 3.3 CS, MMF, HCQ 4 53 LN: III-S (A/C) 61 91 39 1400 MMF, HCQ CRR 3 35 F 5.8 CS, MMF, HCQ 5 24 LN: V 58 41 24 1400 CS, MMF, HCQ CRR 4 23 F 1.9 CS, MMF, CYP 3 134 LN: III-S (A) + V 79 604 25 1400 CS, MMF, HCQ PRR 5 41 F 8.2 CS, MMF, HCQ 5 42 LN: IV-S (A/C) 60 845 24 1400 CS, MMF, HCQ PRR 6 26 F 1.3 MMF, HCQ 4 9 LN: III-S (A) 70 84 33 1400 MMF, HCQ NR: lost function 7 33 F 8.8 CS, MMF, AZA, MTX 4 355 LN: V 109 414 17 700 MMF, HCQ NR: persistent proteinuria 8 31 F 7.3 CS, MMF, AZA, CYP 1 1099 LN: IV-G (A/C) 90 300 22 1400 CS, MMF NR: lost function 9 55 F 1.7 CS, MMF, MTX, HCQ, CYP 5 332 LN: IV-S (A/C) 88 590 27 1400 CS, MMF, HCQ NR: lost function 10 21 F 7.4 CS, MMF, CYP 5 174 LN: IV-S (A/C) + V 85 461 19 1400 MMF, HCQ NR: persistent proteinuria 11 43 F 0.6 CS, HCQ 2 15 LN: III-S (A/C) 112 429 28 1400 MMF PRR 12 48 F 28.5 CS, MMF, AZA, HCQ, CYP, FK 4 68 LN: V 110 911 14 1400 MMF, HCQ NR: persistent proteinuria 13 34 F 11.5 CS, MMF, AZA, CYP, IVIG 5 151 Extra-renal Flare – – – 700 CS, MMF, HCQ Clinical improvement 14 19 F 15.4 MMF, HCQ 2 219 Maintenance – – – 1400 Nil Stable remission a Previous treatment at any time point. b Classification according to the ISN/RPS system for LN. RTX: rituximab; uPCR: urinary protein: creatinine ratio; FK: tacrolimus; CRR: complete renal remission; PRR: partial renal remission; NR: non-responder. Following ofatumumab treatment, all patients were treated with maintenance MMF as per unit protocol, excepting the one patient who was non-adherent to oral medications. Five patients continued on pre-established oral CS therapy, and two patients had oral CSs added to their treatment regimen. Seven patients did not receive oral CSs during this cycle of treatment. Serological and clinical response Following treatment, 12 patients achieved B cell depletion, with a median time to documented depletion of 14 days (7–75) and to subsequent reconstitution of 185 days (64–553). The kinetics of B cell depletion and reconstitution were comparable with those of our previously reported cohort treated with rituximab (‘Rituxilup’ cohort; Fig. 1A). Treatment was associated with significant improvements in the serological markers of disease activity, including ANA titres, anti-dsDNA antibody levels and circulating complement levels, in the first 6 months (Fig. 1B and C). Fig. 1 View largeDownload slide Laboratory and clinical responses in the first 6 months after ofatumumab treatment (A) Individual B cell counts in this cohort (left panel), and compared with the Rituxilup cohort (right panel). There was no significant difference in B cell count between the two cohorts at 1, 3 or 6 months. (B) ANA titre (n = 13) and anti-dsDNA levels (n = 12) for individual patients during 6 months’ therapy (ANA 1/2560 to 1/320, P = 0.03; anti-dsDNA 310 to 120 AU, P = 0.01). (C) C3 and C4 levels for the entire cohort during 6 months’ therapy (C3 0.78 vs 0.91 g/l, P = 0.03; C4 0.14 vs 0.17 g/l, P = 0.14). (D–F) Changes in renal function (D), proteinuria (E) and serum albumin (F). Left panel compares responders and non-responders: at 6 months, responders had stable serum creatinine (61 vs 67 μmol/l, P = 0.59), improved proteinuria (uPCR 325 vs 99 mg/mmol, P = 0.01) and a trend towards increased serum albumin (26 vs 31 g/l; P = 0.31) that was not seen in non-responding patients (creatinine 89 vs 113 μmol/l, P = 0.18; uPCR 438 vs 449 mg/mmol, P = 0.99; serum albumin 21 vs 22 g/l, P = 0.88). Right panel compares parameters in this cohort with those in the Rituxilup cohort: there were no significant differences in any parameter at 0, 3 or 6 months (box-and-whisker plots describe median, IQR and range. Linear plots describe median and IQR. Statistical comparison between 0 and 6 month time-points in the ofatumumab cohort is by Wilcoxin signed-rank test for repeated measures of non-parametric data. Comparison between the ofatumumab and Rituxilup cohort is by multiple t-test.). Fig. 1 View largeDownload slide Laboratory and clinical responses in the first 6 months after ofatumumab treatment (A) Individual B cell counts in this cohort (left panel), and compared with the Rituxilup cohort (right panel). There was no significant difference in B cell count between the two cohorts at 1, 3 or 6 months. (B) ANA titre (n = 13) and anti-dsDNA levels (n = 12) for individual patients during 6 months’ therapy (ANA 1/2560 to 1/320, P = 0.03; anti-dsDNA 310 to 120 AU, P = 0.01). (C) C3 and C4 levels for the entire cohort during 6 months’ therapy (C3 0.78 vs 0.91 g/l, P = 0.03; C4 0.14 vs 0.17 g/l, P = 0.14). (D–F) Changes in renal function (D), proteinuria (E) and serum albumin (F). Left panel compares responders and non-responders: at 6 months, responders had stable serum creatinine (61 vs 67 μmol/l, P = 0.59), improved proteinuria (uPCR 325 vs 99 mg/mmol, P = 0.01) and a trend towards increased serum albumin (26 vs 31 g/l; P = 0.31) that was not seen in non-responding patients (creatinine 89 vs 113 μmol/l, P = 0.18; uPCR 438 vs 449 mg/mmol, P = 0.99; serum albumin 21 vs 22 g/l, P = 0.88). Right panel compares parameters in this cohort with those in the Rituxilup cohort: there were no significant differences in any parameter at 0, 3 or 6 months (box-and-whisker plots describe median, IQR and range. Linear plots describe median and IQR. Statistical comparison between 0 and 6 month time-points in the ofatumumab cohort is by Wilcoxin signed-rank test for repeated measures of non-parametric data. Comparison between the ofatumumab and Rituxilup cohort is by multiple t-test.). At this time-point, 6 of the 12 patients who were treated for LN had achieved complete (cases 1, 2 and 3) or partial (cases 4, 5 and 11) renal remission (Fig. 1D–F, left panels). One partial-responder subsequently achieved complete remission by 12 months (case 4). Six patients were deemed non-responders at month 6: three due to persistent nephrotic-range proteinuria (cases 7, 10 and 12) and three due to deteriorating renal function (cases 6, 8 and 9). During the 6-month follow-up, we observed similar improvements in serum creatinine, albumin and proteinuria in this cohort compared with in the ‘Rituxilup’ cohort (Fig. 1D–F, right panels) One non-responder was retreated with ofatumumab, and subsequently achieved complete remission (case 6). The remaining five non-responders were subsequently treated with pulsed i.v. CYC (in the context of a systemic flare 7–9 months after initial ofatumumab treatment), although none of these five patients achieved renal remission during follow-up. Systemic flares were subsequently observed in four cases. Three of these occurred in the aforementioned non-responding patients (cases 6, 7 and 8, occurring at 10, 7 and 9 months after ofatumumab, respectively). One occurred in a patient who had achieved complete remission (case 3) 20 months after ofatumumab treatment. One patient (case 7) progressed to ESRD 21 months after ofatumumab treatment. In the remaining two patients treated for non-renal disease, resolution of clinical symptoms was observed in case 13: ECLAM [13] Index improved from 3 (fatigue, rash, arthritis, hypocomplementaemia) to 0 over 6 months, and stable remission was maintained without any oral maintenance immunosuppression in case 14. Adverse events Two of 16 patients who have received ofatumumab developed angioedema during drug administration, such that they were unable to complete treatment. They were subsequently treated with MMF and CS therapy. The median duration of the total follow-up of the 14 patients who completed ofatumumab treatment was 28 months (range 9–43). During this period, five severe infections (requiring hospital admission or treatment with i.v. antibiotics) were observed in three patients: 2× lower respiratory tract infections; 2× dialysis access infections, 1× gastroenteritis. Non-severe infections were also reported: 1× lower respiratory tract infection; 2× upper respiratory tract infections; 1× urinary tract infection; 1× folliculitis. No atypical or opportunistic infections were observed. There were no new cases of hypogammaglobulinaemia or persistent neutropenia. No malignancies, cases of progressive multifocal leucoencephalopathy or deaths were observed during follow-up. Conclusions This is the largest published series of patients treated with ofatumumab for SLE/LN. Our experience suggests that ofatumumab is a well-tolerated and effective alternative for patients who are intolerant of rituximab, but for whom B cell depletion is deemed a desirable therapeutic strategy, with 14 of 16 patients receiving treatment without infusion reactions, and 12 of these 14 patients achieving peripheral B cell depletion. This was associated with a serological and clinical response within 6 months, with half of the patients with active LN progressing to remission by this time-point. Of note, B cell kinetics and clinical response rates were comparable with those seen in our previously reported cohort of newly diagnosed patients treated with rituximab for LN. In contrast to the Rituxilup cohort, the majority of patients in this series had long-standing SLE and were heavily pre-treated. As such, it included a number of patients with aggressive or resistant disease, reflected in our observations that in those patients who did not respond to ofatumumab, the response to augmented immunosuppression with CYC was similarly limited, and of subsequent disease flares in four cases. Reassuringly, however, we did not detect any unexpected adverse events in this cohort with a significant burden of prior and concomitant immunosuppression. Infusion reactions to rituximab are common, occurring in 10–15% of patients in randomized controlled studies in SLE, and it is likely that the frequency of infusion reactions increases with repeated exposure. Indeed, all but one of our patients had more than one prior exposure to rituximab. Human anti-chimeric antibodies (HACA) develop in 5–10% of patients treated with rituximab for RA [14], though at an apparently higher frequency of 15–26% in patients with SLE [2, 3], perhaps reflecting an enhanced state of ‘immunoreactivity’ in these patients. The role of HACA in the development of infusion reactions is unclear, though the presence of HACA has been associated with incomplete B cell depletion and poorer clinical response in some SLE cohorts [15, 16]. As a fully-humanized mAb, ofatumumab has low immunogenicity, with no patients developing demonstrable anti-ofatumumab antibodies in haematological or RA studies [17, 18]. Ofatumumab may therefore be a preferred alternative to rituximab in patients with SLE who are likely to require repeated treatments due to the chronic and relapsing nature of their disease. Ofatumumab has also shown efficacy in rituximab-resistant cases of paediatric nephrotic syndrome and haematological malignancy [19–21]. It has been suggested that differences in epitope specificity, pharmacokinetics and ability to activate both complement- and antibody-dependent cell-mediated cytotoxicity may account for this differential response. Whether ofatumumab might likewise provide benefit in patients with SLE who fail to have a primary response to rituximab is unclear. Of note, a phase III study investigating the utility of ocrelizumab, another humanized anti-CD20 antibody, suggested that renal responses were numerically (though not statistically) higher in patients treated with ocrelizumab vs controls, though a higher than expected rate of infectious episodes were reported, such that the study was terminated early [22]. Clearly, not all B cell–depleting strategies are equal, and controlled studies are ideally needed to better define the safety and potential value of ofatumumab in resistant SLE. A randomized control trial investigating the use of obinutuzumab in LN is underway (NCT02550652). Our study has obvious limitations—it is small, uncontrolled, and includes a heterogeneous case mix in whom there was non-uniform use of maintenance immunosuppression and CSs. While we observed that non-responding patients tended to have poorer renal function, lower serum albumin and higher urinary protein loss at baseline, we were unable to identify clear predictors of response in this small cohort. In addition, as a retrospective series, we were unable to perform detailed analysis of B cell subsets or reconstitution, B cell survival factors or the presence of HACA, and it would be of value to investigate these in any future study. Our experience, however, suggests that, in patients who are intolerant of rituximab, ofatumumab is a potential alternative agent, which in this heavily pre-treated cohort was well-tolerated, safe and effective in inducing B cell depletion, with subsequent clinical responses in a significant proportion of patients with long-standing and aggressive SLE. Acknowledgements This work was reported in abstract form at the International Society of Nephrology Nexus Meeting on Translational Immunology in Kidney Disease in April 2016, in Berlin, Germany; and at the UK Renal Association Annual Meeting in Birmingham UK, in June 2016. S.P.M. is in receipt of a National Institute for Health Research (NIHR) clinical lectureship. This work was supported by the NIHR Imperial Biomedical Research Centre. The views expressed are those of the authors and not necessarily those of the National Health Service, the NIHR or the Department of Health. Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript. Disclosure statement: L.L. has received honoraria for advisory boards and lectures from Aurinia, Genentech, GSK, Hoffman La Roche and UCB and has received research support from Hoffman La Roche. All other authors have declared no conflicts of interest. References 1 Gregersen JW , Jayne DR. B-cell depletion in the treatment of lupus nephritis . Nat Rev Nephrol 2012 ; 8 : 505 – 14 . Google Scholar CrossRef Search ADS PubMed 2 Merrill JT , Neuwelt CM , Wallace DJ et al. Efficacy and safety of rituximab in moderately-to-severely active systemic lupus erythematosus: the randomized, double-blind, phase II/III systemic lupus erythematosus evaluation of rituximab trial . Arthritis Rheum 2010 ; 62 : 222 – 33 . Google Scholar CrossRef Search ADS PubMed 3 Rovin BH , Furie R , Latinis K et al. Efficacy and safety of rituximab in patients with active proliferative lupus nephritis: the Lupus Nephritis Assessment with Rituximab study . Arthritis Rheum 2012 ; 64 : 1215 – 26 . Google Scholar CrossRef Search ADS PubMed 4 Condon MB , Ashby D , Pepper RJ et al. Prospective observational single-centre cohort study to evaluate the effectiveness of treating lupus nephritis with rituximab and mycophenolate mofetil but no oral steroids . Ann Rheum Dis 2013 ; 72 : 1280 – 6 . Google Scholar CrossRef Search ADS PubMed 5 Pepper R , Griffith M , Kirwan C et al. Rituximab is an effective treatment for lupus nephritis and allows a reduction in maintenance steroids . Nephrol Dialysis Transplant 2009 ; 24 : 3717 – 23 . Google Scholar CrossRef Search ADS 6 Beckwith H , Lightstone L. Rituximab in systemic lupus erythematosus and lupus nephritis . Nephron Clin Pract 2014 ; 128 : 250 – 4 . Google Scholar CrossRef Search ADS PubMed 7 Zhang B. Ofatumumab . mAbs 2009 ; 1 : 326 – 31 . Google Scholar CrossRef Search ADS PubMed 8 Taylor PC , Quattrocchi E , Mallett S et al. Ofatumumab, a fully human anti-CD20 monoclonal antibody, in biological-naive, rheumatoid arthritis patients with an inadequate response to methotrexate: a randomised, double-blind, placebo-controlled clinical trial . Ann Rheum Dis 2011 ; 70 : 2119 – 25 . Google Scholar CrossRef Search ADS PubMed 9 Kurrasch R , Brown JC , Chu M et al. Subcutaneously administered ofatumumab in rheumatoid arthritis: a phase I/II study of safety, tolerability, pharmacokinetics, and pharmacodynamics . J Rheumatol 2013 ; 40 : 1089 – 96 . Google Scholar CrossRef Search ADS PubMed 10 McAdoo SP , Bedi R , Tarzi R et al. Ofatumumab for B cell depletion therapy in ANCA-associated vasculitis: a single-centre case series . Rheumatology 2016 ; 55 : 1437 – 42 . Google Scholar CrossRef Search ADS PubMed 11 Thornton CC , Ambrose N , Ioannou Y. Ofatumumab: a novel treatment for severe systemic lupus erythematosus . Rheumatology 2015 ; 54 : 559 – 60 . Google Scholar CrossRef Search ADS PubMed 12 Haarhaus ML , Svenungsson E , Gunnarsson I. Ofatumumab treatment in lupus nephritis patients . Clin Kidney J 2016 ; 9 : 552 – 5 . Google Scholar CrossRef Search ADS PubMed 13 Mosca M , Bencivelli W , Vitali C et al. The validity of the ECLAM index for the retrospective evaluation of disease activity in systemic lupus erythematosus . Lupus 2000 ; 9 : 445 – 50 . Google Scholar CrossRef Search ADS PubMed 14 van Vollenhoven RF , Emery P , Bingham CO III et al. Longterm safety of patients receiving rituximab in rheumatoid arthritis clinical trials . J Rheumatol 2010 ; 37 : 558 – 67 . Google Scholar CrossRef Search ADS PubMed 15 Albert D , Dunham J , Khan S et al. Variability in the biological response to anti-CD20 B cell depletion in systemic lupus erythaematosus . Ann Rheum Dis 2008 ; 67 : 1724 – 31 . Google Scholar CrossRef Search ADS PubMed 16 Melander C , Sallée M , Trolliet P et al. Rituximab in severe lupus nephritis: early B-cell depletion affects long-term renal outcome . Clin J Am Soc Nephrol 2009 ; 4 : 579 – 87 . Google Scholar CrossRef Search ADS PubMed 17 Wierda WG , Kipps TJ , Mayer J et al. Ofatumumab as single-agent CD20 immunotherapy in fludarabine-refractory chronic lymphocytic leukemia . J Clin Oncol 2010 ; 28 : 1749 – 55 . Google Scholar CrossRef Search ADS PubMed 18 Østergaard M , Baslund B , Rigby W et al. Ofatumumab, a human anti-CD20 monoclonal antibody, for treatment of rheumatoid arthritis with an inadequate response to one or more disease-modifying antirheumatic drugs: results of a randomized, double-blind, placebo-controlled, phase I/II study . Arthritis Rheum 2010 ; 62 : 2227 – 38 . Google Scholar CrossRef Search ADS PubMed 19 Basu B. Ofatumumab for rituximab-resistant nephrotic syndrome . New Engl J Med 2014 ; 370 : 1268 – 70 . Google Scholar CrossRef Search ADS PubMed 20 Czuczman MS , Fayad L , Delwail V et al. Ofatumumab monotherapy in rituximab-refractory follicular lymphoma: results from a multicenter study . Blood 2012 ; 119 : 3698 – 704 . Google Scholar CrossRef Search ADS PubMed 21 Matasar MJ , Czuczman MS , Rodriguez MA et al. Ofatumumab in combination with ICE or DHAP chemotherapy in relapsed or refractory intermediate grade B-cell lymphoma . Blood 2013 ; 122 : 499 – 506 . Google Scholar CrossRef Search ADS PubMed 22 Mysler EF , Spindler AJ , Guzman R et al. Efficacy and safety of ocrelizumab in active proliferative lupus nephritis: results from a randomized, double-blind, phase III study . Arthritis Rheum 2013 ; 65 : 2368 – 79 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Sato, Shinji;Masui, Kenichi;Nishina, Naoshi;Kawaguchi, Yasushi;Kawakami, Atsushi;Tamura, Maasa;Ikeda, Kei;Nunokawa, Takahiro;Tanino, Yoshinori;Asakawa, Katsuaki;Kaneko, Yuko;Gono, Takahisa;Ukichi, Taro;Kaieda, Shinjiro;Naniwa, Taio;Kuwana, Masataka;investigators, JAMI

2018 Rheumatology

doi: 10.1093/rheumatology/key060pmid: 29596687

Abstract Objective To identify initial predictors of poor survival in patients with PM/DM-associated interstitial lung disease (ILD). Methods We established a multicentre retrospective cohort of incident cases of PM/DM-associated ILD from 44 institutions across Japan (Multicentre Retrospective Cohort of Japanese Patients with Myositis-associated ILD, JAMI). Inclusion criteria were an onset age ⩾16 years; PM/DM or clinically amyopathic DM according to the published criteria; imaging evidence of ILD; and availability of serum samples for assays of autoantibodies such as anti-melanoma differentiation-associated gene 5 and anti-aminoacyl tRNA synthetase. We collected demographic data and clinical characteristics recorded at the time of diagnosis, as well as follow-up survival data. Predictors of ILD-related mortality were identified by univariate and multivariate analyses. Results JAMI enrolled a cohort of 497 patients with PM (15%), classic DM (32%) and clinically amyopathic DM (53%). During the observation period (median 20 months), 76 died of respiratory insufficiency directly related to ILD. Univariate analysis revealed several initial parameters associated with ILD mortality, including demographic, clinical, laboratory, imaging and autoantibody variables. We used multivariate analysis with a stepwise selection of parameters to generate an appropriate predictive model, and identified the following independent risk factors for ILD mortality: age at onset ⩾60 years [hazard ratio (HR) = 4.3, 95% CI: 2.4, 7.5], CRP ⩾1 mg/dl (HR = 2.6, 95% CI: 1.5, 4.8), peripheral capillary oxygen saturation <95% (HR = 2.0, 95% CI: 1.2, 3.4) and anti-melanoma differentiation-associated gene 5 antibody (HR = 7.5, 95% CI: 2.8, 20.2). Conclusion We established a large cohort of incident cases of PM/DM-associated ILD, and successfully identified independent predictors of short-term ILD mortality. myositis and muscle disease, respiratory, autoantigens and autoantibodies, outcome measures, biomarkers Rheumatology key messages Early death due to respiratory insufficiency is common in myositis-associated interstitial lung disease. Anti-melanoma differentiation-associated gene 5 antibody is the most powerful initial predictor of myositis-associated interstitial lung disease mortality. Independent predictors of myositis-associated interstitial lung disease mortality include older age and high CRP. Introduction PM and DM are autoimmune conditions classified into a group of idiopathic inflammatory myopathies that target muscle, skin, joints and lungs to various degrees [1]. Interstitial lung disease (ILD) is a life-threatening complication in PM/DM [2]. The clinical course and treatment response of PM/DM-associated ILD are quite heterogeneous, with the most devastating form being rapidly progressive ILD (RP-ILD), which is often resistant to immunosuppressive treatment [3, 4]. Approximately half of the patients who develop RP-ILD die despite intensive immunosuppressive treatment [5, 6]. Although an evidence-based treatment strategy has not been established for RP-ILD, studies suggest that it may be beneficial to introduce intensive immunosuppressive therapy, such as high-dose CS in combination with immunosuppressive agents, at an early stage of the disease, before the lungs are irreversibly damaged [7, 8]. Thus, it is critical to identify predictive factors of poor survival in patients with PM/DM-associated ILD. A number of studies have examined potential risk factors associated with poor survival in patients with PM/DM-associated ILD, and have reported associations between poor outcomes and older age, DM classification, skin ulceration, pneumomediastinum and a lack of myositis [9–11]. Among the myositis-specific autoantibodies (MSAs), anti-melanoma differentiation-associated gene 5 (MDA5) and anti-aminoacyl tRNA synthetase (ARS) antibodies are associated with ILD, but RP-ILD is more likely to be associated with anti-MDA5 antibody [5, 6, 12, 13]. A series of studies have reported that anti-MDA5 antibody, acute or subacute ILD onset, fever, higher serum ferritin levels, decreased circulating lymphocyte counts, elevated serum CRP and a lower consolidation/ground-glass attenuation (GGA) pattern on chest high-resolution CT (HRCT) are independent risk factors for predicting the course of RP-ILD or mortality [14–19]. While anti-MDA5 antibody is consistently reported as a poor survival factor in most studies [13], other reported prognostic factors are less consistent, probably due to the small sample sizes of these single-centre studies and to differences in patient populations between studies. To overcome these limitations, it was necessary to design a large-scale, well-defined, multicentre cohort. To that end, we established the Multicentre Retrospective Cohort of Japanese Patients with Myositis-associated ILD (JAMI), which involved 44 institutions across Japan with a wide range of specialties, including rheumatology, respirology and dermatology. Information for the cohort was collected from demographic, clinical, imaging and laboratory data at the time of diagnosis, as well as autoantibody profiles, and the clinical course was retrospectively and/or prospectively recorded. Using the data from 497 incident cases of adult PM/DM-associated ILD registered in the JAMI database, we identified initial factors that predict mortality from respiratory insufficiency due directly to ILD. Methods Patients JAMI is a multicentre retrospective cohort that allowed collaborative research on PM/DM-associated ILD. Investigators in individual participating centres were asked to enrol incident PM/DM-associated ILD cases who visited their centres between October 2011 and 2015. The inclusion criteria were as follows: an age at disease onset ⩾16 years; definite or probable PM/DM according to the criteria proposed by Bohan and Peter [20] or clinically amyopathic DM (CADM) according to the criteria proposed by Sontheimer [21]; the presence of ILD, based on the American Thoracic Society criteria and a multidisciplinary assessment of clinical, radiologic and pathologic findings [22]; and availability of serum samples at diagnosis for comprehensive MSA analysis. At enrolment, data were collected retrospectively in a dedicated electronic database by a reference clinician in each centre. To evaluate outcomes, follow-up survival data, any diagnosis of malignancy and the cause of death if deceased were collected prospectively. This study was approved by the Ethics Committee of the coordinating centre (Nippon Medical School, Tokyo, Japan; 26-03-434) and by individual participating centres. All clinical information was obtained after the patients had given their informed consent. The JAMI cohort is registered in the University Hospital’s Medical Information Network Clinical Trial Registry (UMIN000018663). JAMI database At enrolment, demographic, clinical, laboratory and imaging information at the time of diagnosis, before beginning immunosuppressive treatment, was inputted. The following demographic and clinical data were collected: gender, age at disease onset, body weight, height, postal code of residence at disease onset, disease duration at diagnosis, initial symptoms related to the disease, disease classification (PM, classic DM or CADM), fever, arthritis/arthralgia, Raynaud’s phenomenon, sclerodactyly, sicca, muscle weakness, myalgia, Gottron’s papules, Gottron’s sign, palmar papules, heliotrope rash, shawl sign, neck V-sign, flagellate erythema, skin ulceration, periungual erythema, mechanic’s hands and subcutaneous calcinosis. Peripheral blood laboratory results were recorded, including creatine kinase, aldolase, CRP, ESR, Krebs von den Lungen-6 (KL-6), surfactant protein D (SP-D), ferritin, Raynaud’s phenomenon and ANAs on IIF. Findings on electromyography, MRI of extremities, chest radiograph, chest HRCT, peripheral capillary oxygen saturation (SpO2), arterial blood gas analysis, forced vital capacity, forced expiratory volume in 1 s, diffusing capacity for carbon monoxide, 6-min walking distance, and pathologic findings of skeletal muscle, skin and lung were also collected when available. Chest HRCT findings were classified into four patterns according to Tanizawa et al. [23]: lower lobe consolidation/GGA, lower lobe reticulation, random GGA and others. The induction treatment regimen was also recorded, including CS (with starting dosage), CS pulse therapy, CYC, CSA, tacrolimus, MTX, AZA, MMF, IVIG and other treatments. A prednisolone equivalent of ⩾50 mg daily was regarded as high-dose CS. We examined the patient’s records from diagnosis until the last visit, and recorded the cause of death if the patient was deceased. MSA detection Serum samples were stored at −20˚C until use. MSAs were identified centrally at laboratories with the proper expertise at Tokai University, Keio University or Nippon Medical School. Anti-ARS antibodies were detected by RNA immunoprecipitation assay using K562-cell extracts as described previously [24]. Anti-ARS antibody results were judged positive when the given serum samples precipitated RNA components identical to those precipitated by the prototype sera positive for anti-Jo-1, anti-EJ, anti-OJ, anti-PL-7, anti-PL-12 or anti-KS. The anti-MDA5 antibody was measured by an in-house ELISA using recombinant MDA5 as an antigen source [25]. Statistical analysis We conducted univariate and multivariate analyses to identify predictive factors of mortality due to respiratory insufficiency directly related to PM/DM-associated ILD. Continuous values were shown as the median and 2.5–97.5 percentile. All statistical analyses were performed by an independent medical statistician (K.M.) using SPSS Statistics version 23 (IBM, Tokyo, Japan). We conducted univariate analysis by Kaplan-Meier analysis with the Breslow test to compare the equality of survival curves for individual variables for demographic, clinical, laboratory, imaging and autoantibody parameters at diagnosis. Variables with available data for fewer than 350 cases were excluded from analysis. All continuous variables were converted to dichotomous variables for analysis, and cut-off values were determined by receiver-operating characteristics [26]. Multiple dichotomous variables were used for diagnosis (PM, classic DM or CADM) or MSAs (anti-MDA5−/anti-ARS−, anti-MDA5+/anti-ARS−, anti-MDA5−/anti-ARS+ or anti-MDA5+/anti-ARS+). The cause-specific Cox proportional hazards model was used for multivariate analysis to identify an optimal model for predicting poor survival caused by respiratory insufficiency. In some cases, we applied subdistribution hazards model, and/or imputed all missing values of dichotomous variables using a multiple imputation method and resultant generation of 1000 multiple imputation data sets, to verify the original model. During the stepwise selection process, sex, age at diagnosis and anti-MDA5 and anti-ARS antibodies were first adopted into the model, and all other variables at the time of diagnosis with P < 0.2 in the univariate analysis were subsequently entered into the model as potential predictors. Backward deletion (P ⩾ 0.15) and forward inclusion (P < 0.1) were performed using the likelihood test to select the predictor variables in the model. Finally, selected induction treatment drugs were included in the final multivariate model as potential confounders. The results were shown as the hazard ratio (HR) with a 95% CI. Results Baseline characteristics Of 499 patients registered, 2 were excluded because of incomplete data during the follow-up period, and 497 incident cases were enrolled in the study. The baseline characteristics of the cohort (available data ⩾350) are shown in Table 1. The disease duration at diagnosis was 3 months (2.5–97.5 percentile 1–62), indicating that most patients were diagnosed and treated at an early stage. Our cohort consisted mainly of classic DM and CADM; only 15% were classified as PM. Muscle weakness was detected in 48%. Chest HRCT findings were available for 496 patients: a lower consolidation/GGA pattern was most prevalent, followed by lower reticulation and random GGA patterns. There were 166 patients (34%) with anti-ARS antibodies; 44 with anti-Jo-1, 41 with anti-EJ, 13 with anti-OJ, 32 with anti-PL-7, 26 with anti-PL-12 and 12 with anti-KS. Two patients had two anti-ARS specificities simultaneously: one was positive for anti-Jo-1 and anti-EJ, and the other for anti-Jo-1 and anti-PL-7. Anti-MDA5 was detected in 209 patients (42%). Although anti-ARS and anti-MDA5 antibodies were mutually exclusive in general, two patients had both (anti-PL-7 or anti-KS in one each). The detailed clinical course of one such patient is published elsewhere [27]. Table 1 Baseline characteristics and initial treatment for 497 patients with PM/DM-associated ILD in the JAMI database Variables Value Available data per outcome Demographics     Age at onset, years 57 (29–80) 497 (100)     Male 167 (34) 497 (100)     Body weight, kg 55 (39–85) 492 (99)     Disease duration at diagnosis, month 3 (1–62) 495 (100) Diagnosis     PM 76 (15) 497 (100)     Classic DM 158 (32)     CADM 263 (53) Clinical features     Fever 237 (49) 482 (97)     Raynaud’s phenomenon 69 (16) 443 (89)     Muscle weakness 233 (48) 486 (98)     Neck V-sign 92 (21) 443 (89)     Skin ulceration 48 (10) 459 (92)     Periungual erythema 260 (58) 452 (91) Laboratory parameters     CRP, mg/dl 0.7 (0.0–13.4) 486 (98)     CK, IU/l 202 (32–4267) 486 (98)     Aldolase, IU/l 9.0 (3.6–91.2) 424 (85)     KL-6, U/ml 801 (208–4431) 476 (96)     SP-D, ng/ml 91 (16–615) 380 (76)     Ferritin, ng/ml 357 (22–3846) 361 (73) Chest HRCT patterns     Lower consolidation/GGA 268 (54) 496 (100)     Lower reticulation 167 (34)     Random GGA 60 (12) SpO2, % 96 (89–99) 477 (96) MSAs     Anti-ARS−/anti-MDA5− 122 (25) 493 (99)     Anti-ARS+/anti-MDA5− 162 (33)     Anti-ARS−/anti-MDA5+ 200 (41)     Anti-ARS+/anti-MDA5+ 2 (0) Drugs used for induction treatment     Any CS 485 (98) 497 (100)     High-dose CS 289 (58) 497 (100)     Initial dose of CS (PSL equivalent), mg/day 50 (1.8–100) 497 (100)     CS pulse therapy 285 (57) 497 (100)     CYC 223 (45) 497 (100)     Intravenous CYC 218 (44) 497 (100)     CSA 238 (48) 497 (100)     Tac 226 (45) 497 (100)     MTX 17 (3) 497 (100)     AZA 21 (4) 497 (100)     MMF 6 (1) 497 (100)     IVIG 86 (17) 497 (100)     Rituximab 5 (1) 497 (100)     PMX-DHP 47 (9) 497 (100) Initial treatment regimens     High-dose CS alone 30 (6) 497 (100)     High-dose CS + CYC 5 (1) 497 (100)     High-dose CS + CSA/Tac 110 (22) 497 (100)     High-dose CS + CYC + CSA/Tac 145 (29) 497 (100) Variables Value Available data per outcome Demographics     Age at onset, years 57 (29–80) 497 (100)     Male 167 (34) 497 (100)     Body weight, kg 55 (39–85) 492 (99)     Disease duration at diagnosis, month 3 (1–62) 495 (100) Diagnosis     PM 76 (15) 497 (100)     Classic DM 158 (32)     CADM 263 (53) Clinical features     Fever 237 (49) 482 (97)     Raynaud’s phenomenon 69 (16) 443 (89)     Muscle weakness 233 (48) 486 (98)     Neck V-sign 92 (21) 443 (89)     Skin ulceration 48 (10) 459 (92)     Periungual erythema 260 (58) 452 (91) Laboratory parameters     CRP, mg/dl 0.7 (0.0–13.4) 486 (98)     CK, IU/l 202 (32–4267) 486 (98)     Aldolase, IU/l 9.0 (3.6–91.2) 424 (85)     KL-6, U/ml 801 (208–4431) 476 (96)     SP-D, ng/ml 91 (16–615) 380 (76)     Ferritin, ng/ml 357 (22–3846) 361 (73) Chest HRCT patterns     Lower consolidation/GGA 268 (54) 496 (100)     Lower reticulation 167 (34)     Random GGA 60 (12) SpO2, % 96 (89–99) 477 (96) MSAs     Anti-ARS−/anti-MDA5− 122 (25) 493 (99)     Anti-ARS+/anti-MDA5− 162 (33)     Anti-ARS−/anti-MDA5+ 200 (41)     Anti-ARS+/anti-MDA5+ 2 (0) Drugs used for induction treatment     Any CS 485 (98) 497 (100)     High-dose CS 289 (58) 497 (100)     Initial dose of CS (PSL equivalent), mg/day 50 (1.8–100) 497 (100)     CS pulse therapy 285 (57) 497 (100)     CYC 223 (45) 497 (100)     Intravenous CYC 218 (44) 497 (100)     CSA 238 (48) 497 (100)     Tac 226 (45) 497 (100)     MTX 17 (3) 497 (100)     AZA 21 (4) 497 (100)     MMF 6 (1) 497 (100)     IVIG 86 (17) 497 (100)     Rituximab 5 (1) 497 (100)     PMX-DHP 47 (9) 497 (100) Initial treatment regimens     High-dose CS alone 30 (6) 497 (100)     High-dose CS + CYC 5 (1) 497 (100)     High-dose CS + CSA/Tac 110 (22) 497 (100)     High-dose CS + CYC + CSA/Tac 145 (29) 497 (100) Values listed as n (%); continuous variables are shown as median (2.5–97.5 percentile). CADM: clinically amyopathic DM; CK: creatine kinase; KL-6: Krebs von den Lungen-6; SP-D: surfactant protein D; HRCT: high-resolution CT; GGA: ground glass attenuation; MSA: myositis-specific autoantibody; ARS: aminoacyl tRNA synthetase; MDA5: melanoma differentiation-associated gene 5; PSL: prednisolone; Tac: tacrolimus; PMX-DHP: polymyxin B-immobilized fibre column-direct haemoperfusion; JAMI: Multicentre Retrospective Cohort of Japanese Patients with Myositis-associated ILD; ILD: interstitial lung disease. Table 1 Baseline characteristics and initial treatment for 497 patients with PM/DM-associated ILD in the JAMI database Variables Value Available data per outcome Demographics     Age at onset, years 57 (29–80) 497 (100)     Male 167 (34) 497 (100)     Body weight, kg 55 (39–85) 492 (99)     Disease duration at diagnosis, month 3 (1–62) 495 (100) Diagnosis     PM 76 (15) 497 (100)     Classic DM 158 (32)     CADM 263 (53) Clinical features     Fever 237 (49) 482 (97)     Raynaud’s phenomenon 69 (16) 443 (89)     Muscle weakness 233 (48) 486 (98)     Neck V-sign 92 (21) 443 (89)     Skin ulceration 48 (10) 459 (92)     Periungual erythema 260 (58) 452 (91) Laboratory parameters     CRP, mg/dl 0.7 (0.0–13.4) 486 (98)     CK, IU/l 202 (32–4267) 486 (98)     Aldolase, IU/l 9.0 (3.6–91.2) 424 (85)     KL-6, U/ml 801 (208–4431) 476 (96)     SP-D, ng/ml 91 (16–615) 380 (76)     Ferritin, ng/ml 357 (22–3846) 361 (73) Chest HRCT patterns     Lower consolidation/GGA 268 (54) 496 (100)     Lower reticulation 167 (34)     Random GGA 60 (12) SpO2, % 96 (89–99) 477 (96) MSAs     Anti-ARS−/anti-MDA5− 122 (25) 493 (99)     Anti-ARS+/anti-MDA5− 162 (33)     Anti-ARS−/anti-MDA5+ 200 (41)     Anti-ARS+/anti-MDA5+ 2 (0) Drugs used for induction treatment     Any CS 485 (98) 497 (100)     High-dose CS 289 (58) 497 (100)     Initial dose of CS (PSL equivalent), mg/day 50 (1.8–100) 497 (100)     CS pulse therapy 285 (57) 497 (100)     CYC 223 (45) 497 (100)     Intravenous CYC 218 (44) 497 (100)     CSA 238 (48) 497 (100)     Tac 226 (45) 497 (100)     MTX 17 (3) 497 (100)     AZA 21 (4) 497 (100)     MMF 6 (1) 497 (100)     IVIG 86 (17) 497 (100)     Rituximab 5 (1) 497 (100)     PMX-DHP 47 (9) 497 (100) Initial treatment regimens     High-dose CS alone 30 (6) 497 (100)     High-dose CS + CYC 5 (1) 497 (100)     High-dose CS + CSA/Tac 110 (22) 497 (100)     High-dose CS + CYC + CSA/Tac 145 (29) 497 (100) Variables Value Available data per outcome Demographics     Age at onset, years 57 (29–80) 497 (100)     Male 167 (34) 497 (100)     Body weight, kg 55 (39–85) 492 (99)     Disease duration at diagnosis, month 3 (1–62) 495 (100) Diagnosis     PM 76 (15) 497 (100)     Classic DM 158 (32)     CADM 263 (53) Clinical features     Fever 237 (49) 482 (97)     Raynaud’s phenomenon 69 (16) 443 (89)     Muscle weakness 233 (48) 486 (98)     Neck V-sign 92 (21) 443 (89)     Skin ulceration 48 (10) 459 (92)     Periungual erythema 260 (58) 452 (91) Laboratory parameters     CRP, mg/dl 0.7 (0.0–13.4) 486 (98)     CK, IU/l 202 (32–4267) 486 (98)     Aldolase, IU/l 9.0 (3.6–91.2) 424 (85)     KL-6, U/ml 801 (208–4431) 476 (96)     SP-D, ng/ml 91 (16–615) 380 (76)     Ferritin, ng/ml 357 (22–3846) 361 (73) Chest HRCT patterns     Lower consolidation/GGA 268 (54) 496 (100)     Lower reticulation 167 (34)     Random GGA 60 (12) SpO2, % 96 (89–99) 477 (96) MSAs     Anti-ARS−/anti-MDA5− 122 (25) 493 (99)     Anti-ARS+/anti-MDA5− 162 (33)     Anti-ARS−/anti-MDA5+ 200 (41)     Anti-ARS+/anti-MDA5+ 2 (0) Drugs used for induction treatment     Any CS 485 (98) 497 (100)     High-dose CS 289 (58) 497 (100)     Initial dose of CS (PSL equivalent), mg/day 50 (1.8–100) 497 (100)     CS pulse therapy 285 (57) 497 (100)     CYC 223 (45) 497 (100)     Intravenous CYC 218 (44) 497 (100)     CSA 238 (48) 497 (100)     Tac 226 (45) 497 (100)     MTX 17 (3) 497 (100)     AZA 21 (4) 497 (100)     MMF 6 (1) 497 (100)     IVIG 86 (17) 497 (100)     Rituximab 5 (1) 497 (100)     PMX-DHP 47 (9) 497 (100) Initial treatment regimens     High-dose CS alone 30 (6) 497 (100)     High-dose CS + CYC 5 (1) 497 (100)     High-dose CS + CSA/Tac 110 (22) 497 (100)     High-dose CS + CYC + CSA/Tac 145 (29) 497 (100) Values listed as n (%); continuous variables are shown as median (2.5–97.5 percentile). CADM: clinically amyopathic DM; CK: creatine kinase; KL-6: Krebs von den Lungen-6; SP-D: surfactant protein D; HRCT: high-resolution CT; GGA: ground glass attenuation; MSA: myositis-specific autoantibody; ARS: aminoacyl tRNA synthetase; MDA5: melanoma differentiation-associated gene 5; PSL: prednisolone; Tac: tacrolimus; PMX-DHP: polymyxin B-immobilized fibre column-direct haemoperfusion; JAMI: Multicentre Retrospective Cohort of Japanese Patients with Myositis-associated ILD; ILD: interstitial lung disease. Induction treatment regimens As shown Table 1, almost all of the patients were treated with CS: 58% with high-dose CS and 57% with CS pulse therapy. Calcineurin inhibitor (CSA or tacrolimus) was preferably used (80%) as an immunosuppressant, rather than CYC, MTX, AZA or MMF. In regimens using CYC, it was almost always administered intravenously every 2–4 weeks. Patients were frequently treated with a combination of high-dose CS with immunosuppressants: ‘double combo’ with CYC in 5 (1%), ‘double combo’ with calcineurin inhibitor in 110 (22%), and ‘triple combo’ with CYC and calcineurin inhibitor in 145 (29%). IVIG was administered to 86 patients (17%), while MMF and rituximab was rarely used. Notably, 47 patients (9%) underwent polymyxin B-immobilized fibre column-direct haemoperfusion [28]. Patient outcomes The median observation period was 20 months (2.5–97.5 percentile 1–50), during which 93 patients died. The cause of death was respiratory insufficiency directly related to PM/DM-associated ILD in 76 (82%), infection in 5 (5%), malignancy in 5 (5%), and other causes such as renal insufficiency, cardiomyopathy and suicide in 7 (8%). Thus, the majority of deaths in the JAMI cohort were directly due to ILD. Initial predictors of mortality due to ILD We examined initial predictors at the time of diagnosis of mortality directly due to ILD by univariate analysis of all initial variables (excluding induction treatment regimens) with available data for 350 or more cases. Table 2 lists initial parameters significantly associated with ILD mortality. The following initial parameters were identified as predictors for poor ILD outcomes: male, older age at disease onset (⩾60 years), CADM, classic DM, fever, neck V-sign, skin ulceration, periungual erythema, higher CRP (⩾1 mg/dl), higher KL-6 (⩾1000 U/ml), higher ferritin (⩾500 ng/ml), lower consolidation/GGA and random GGA patterns on chest HRCT, lower SpO2 (<95%) and anti-MDA5 antibody. In contrast, favourable ILD outcomes were associated with Raynaud’s phenomenon, muscle weakness, higher creatine kinase, higher aldolase, higher SP-D and anti-ARS antibody. It was of note that the majority of predictors for ILD outcomes identified in univariate analysis were associated with anti-MDA5 antibody (supplementary Table S1, available at Rheumatology online). Table 2 Initial parameters significantly associated with mortality due to ILD, identified by univariate analysis Variable N (%) P-value Dead (n = 76a) Alive (n = 421a) Demographics     Male 32 (42) 135 (32) 0.044     Age at onset ≥60 years 53 (70) 167 (40) <0.001 Diagnosis     PM 2 (3) 74 (18) Reference     CADM 61 (80) 202 (48) <0.001b     Classic DM 13 (17) 145 (34) 0.009b Clinical features     Fever 59 (77) 178 (44) <0.001     Raynaud’s phenomenon 3 (5) 66 (17) 0.012     Muscle weakness 23 (33) 210 (51) 0.014     Neck V-sign 19 (29) 73 (19) 0.045     Skin ulceration 12 (17) 36 (9) 0.024     Periungual erythema 48 (69) 212 (56) 0.046 Laboratory parameters     CRP ≥1 mg/dl 55 (72) 154 (38) <0.001     CK <750 IU/l 68 (90) 293 (72) 0.001     Aldolase <17.5 IU/l 62 (94) 243 (68) <0.001     KL-6 ≥1000 U/ml 40 (53) 141 (35) 0.007     SP-D <100 ng/ml 50 (74) 154 (49) <0.001     Ferritin ≥500 ng/ml 43 (74) 95 (31) <0.001 Chest HRCT patterns     Lower consolidation/GGA 50 (66) 218 (52) 0.027     Lower reticulation 13 (17) 154 (37) 0.001     Random GGA 17 (22) 43 (10) 0.005 SpO2 <95% 38 (50) 78 (20) <0.001 MSAs     Anti-ARS−/anti-MDA5− 5 (7) 117 (28) Reference     Anti-ARS+/anti-MDA5− 4 (5) 160 (39) <0.001b     Anti-ARS−/anti-MDA5+ 64 (88) 138 (33) <0.001b Variable N (%) P-value Dead (n = 76a) Alive (n = 421a) Demographics     Male 32 (42) 135 (32) 0.044     Age at onset ≥60 years 53 (70) 167 (40) <0.001 Diagnosis     PM 2 (3) 74 (18) Reference     CADM 61 (80) 202 (48) <0.001b     Classic DM 13 (17) 145 (34) 0.009b Clinical features     Fever 59 (77) 178 (44) <0.001     Raynaud’s phenomenon 3 (5) 66 (17) 0.012     Muscle weakness 23 (33) 210 (51) 0.014     Neck V-sign 19 (29) 73 (19) 0.045     Skin ulceration 12 (17) 36 (9) 0.024     Periungual erythema 48 (69) 212 (56) 0.046 Laboratory parameters     CRP ≥1 mg/dl 55 (72) 154 (38) <0.001     CK <750 IU/l 68 (90) 293 (72) 0.001     Aldolase <17.5 IU/l 62 (94) 243 (68) <0.001     KL-6 ≥1000 U/ml 40 (53) 141 (35) 0.007     SP-D <100 ng/ml 50 (74) 154 (49) <0.001     Ferritin ≥500 ng/ml 43 (74) 95 (31) <0.001 Chest HRCT patterns     Lower consolidation/GGA 50 (66) 218 (52) 0.027     Lower reticulation 13 (17) 154 (37) 0.001     Random GGA 17 (22) 43 (10) 0.005 SpO2 <95% 38 (50) 78 (20) <0.001 MSAs     Anti-ARS−/anti-MDA5− 5 (7) 117 (28) Reference     Anti-ARS+/anti-MDA5− 4 (5) 160 (39) <0.001b     Anti-ARS−/anti-MDA5+ 64 (88) 138 (33) <0.001b P-values were calculated using the Breslow test. aMaximum number of subjects per outcome. bCompared with the reference. CADM: clinically amyopathic DM; CK creatine kinase; KL-6: Krebs von den Lungen-6; SP-D: surfactant protein D; HRCT: high-resolution CT; GGA: ground glass attenuation; MSA: myositis-specific autoantibody; ARS: aminoacyl tRNA synthetase; MDA5: melanoma differentiation-associated gene 5; ILD: interstitial lung disease. Table 2 Initial parameters significantly associated with mortality due to ILD, identified by univariate analysis Variable N (%) P-value Dead (n = 76a) Alive (n = 421a) Demographics     Male 32 (42) 135 (32) 0.044     Age at onset ≥60 years 53 (70) 167 (40) <0.001 Diagnosis     PM 2 (3) 74 (18) Reference     CADM 61 (80) 202 (48) <0.001b     Classic DM 13 (17) 145 (34) 0.009b Clinical features     Fever 59 (77) 178 (44) <0.001     Raynaud’s phenomenon 3 (5) 66 (17) 0.012     Muscle weakness 23 (33) 210 (51) 0.014     Neck V-sign 19 (29) 73 (19) 0.045     Skin ulceration 12 (17) 36 (9) 0.024     Periungual erythema 48 (69) 212 (56) 0.046 Laboratory parameters     CRP ≥1 mg/dl 55 (72) 154 (38) <0.001     CK <750 IU/l 68 (90) 293 (72) 0.001     Aldolase <17.5 IU/l 62 (94) 243 (68) <0.001     KL-6 ≥1000 U/ml 40 (53) 141 (35) 0.007     SP-D <100 ng/ml 50 (74) 154 (49) <0.001     Ferritin ≥500 ng/ml 43 (74) 95 (31) <0.001 Chest HRCT patterns     Lower consolidation/GGA 50 (66) 218 (52) 0.027     Lower reticulation 13 (17) 154 (37) 0.001     Random GGA 17 (22) 43 (10) 0.005 SpO2 <95% 38 (50) 78 (20) <0.001 MSAs     Anti-ARS−/anti-MDA5− 5 (7) 117 (28) Reference     Anti-ARS+/anti-MDA5− 4 (5) 160 (39) <0.001b     Anti-ARS−/anti-MDA5+ 64 (88) 138 (33) <0.001b Variable N (%) P-value Dead (n = 76a) Alive (n = 421a) Demographics     Male 32 (42) 135 (32) 0.044     Age at onset ≥60 years 53 (70) 167 (40) <0.001 Diagnosis     PM 2 (3) 74 (18) Reference     CADM 61 (80) 202 (48) <0.001b     Classic DM 13 (17) 145 (34) 0.009b Clinical features     Fever 59 (77) 178 (44) <0.001     Raynaud’s phenomenon 3 (5) 66 (17) 0.012     Muscle weakness 23 (33) 210 (51) 0.014     Neck V-sign 19 (29) 73 (19) 0.045     Skin ulceration 12 (17) 36 (9) 0.024     Periungual erythema 48 (69) 212 (56) 0.046 Laboratory parameters     CRP ≥1 mg/dl 55 (72) 154 (38) <0.001     CK <750 IU/l 68 (90) 293 (72) 0.001     Aldolase <17.5 IU/l 62 (94) 243 (68) <0.001     KL-6 ≥1000 U/ml 40 (53) 141 (35) 0.007     SP-D <100 ng/ml 50 (74) 154 (49) <0.001     Ferritin ≥500 ng/ml 43 (74) 95 (31) <0.001 Chest HRCT patterns     Lower consolidation/GGA 50 (66) 218 (52) 0.027     Lower reticulation 13 (17) 154 (37) 0.001     Random GGA 17 (22) 43 (10) 0.005 SpO2 <95% 38 (50) 78 (20) <0.001 MSAs     Anti-ARS−/anti-MDA5− 5 (7) 117 (28) Reference     Anti-ARS+/anti-MDA5− 4 (5) 160 (39) <0.001b     Anti-ARS−/anti-MDA5+ 64 (88) 138 (33) <0.001b P-values were calculated using the Breslow test. aMaximum number of subjects per outcome. bCompared with the reference. CADM: clinically amyopathic DM; CK creatine kinase; KL-6: Krebs von den Lungen-6; SP-D: surfactant protein D; HRCT: high-resolution CT; GGA: ground glass attenuation; MSA: myositis-specific autoantibody; ARS: aminoacyl tRNA synthetase; MDA5: melanoma differentiation-associated gene 5; ILD: interstitial lung disease. To generate an appropriate predictive model for mortality due to PM/DM-associated ILD, we conducted multivariate analysis using a stepwise selection of parameters and the cause-specific Cox proportional hazards regression model to identify independent predictors. Sex, age at diagnosis, anti-MDA5 antibody and anti-ARS antibody were first adopted into the model, since demographic features and MSAs are reported as a poor survival factor in most previous studies [9–19]. Subsequently, all other initial variables with P < 0.2 in the univariate analysis were sequentially included in the model, and the number of variables was reduced by repeating the step-up and step-down procedure. Fig. 1 shows a final predictive model of mortality derived from 349 patients in whom all variable data were available, consisting of nine independent variables. In particular, age at disease onset ⩾60 years (HR = 4.3, 95% CI: 2.4, 7.5; P < 0.001), anti-MDA5 antibody (HR = 7.5, 95% CI: 2.8, 20.2; P < 0.001), CRP ⩾1 mg/dl (HR = 2.6, 95% CI: 1.5, 4.8; P = 0.001) and SpO2 <95% (HR = 2.0, 95% CI: 1.2, 3.4; P = 0.011) were selected as independent risks for ILD mortality. The identical variables were selected when subdistribution hazards model was applied (supplementary Table S2, available at Rheumatology online). Fig. 1 View largeDownload slide Predictive model for mortality due to respiratory insufficiency in patients with PM/DM-associated ILD Initial predictors of poor prognosis due to fatal respiratory insufficiency were identified by multivariate analysis using a stepwise selection of parameters and the cause-specific Cox proportional hazards regression model. The hazard ratio with 95% CI is shown for each variable finally selected. ILD: interstitial lung disease; KL-6: Krebs von den Lungen-6; SP-D: surfactant protein; ARS: aminoacyl; MDA5: melanoma differentiation-associated gene 5. Fig. 1 View largeDownload slide Predictive model for mortality due to respiratory insufficiency in patients with PM/DM-associated ILD Initial predictors of poor prognosis due to fatal respiratory insufficiency were identified by multivariate analysis using a stepwise selection of parameters and the cause-specific Cox proportional hazards regression model. The hazard ratio with 95% CI is shown for each variable finally selected. ILD: interstitial lung disease; KL-6: Krebs von den Lungen-6; SP-D: surfactant protein; ARS: aminoacyl; MDA5: melanoma differentiation-associated gene 5. When we included the drugs used for induction treatment (including high-dose CS, the calcineurin inhibitors CSA or tacrolimus, CYC and IVIG) in the final multivariate model as potential confounders, the four independent risk factors remained statistically significant: age at onset (HR = 3.8, 95% CI: 2.2, 6.7; P < 0.001), anti-MDA5 antibody (HR = 6.5, 95% CI: 2.3, 17.9; P < 0.001), CRP (HR = 2.6, 95% CI: 1.5, 4.7; P = 0.001) and SpO2 (HR = 2.1, 95% CI: 1.2, 3.5; P = 0.005), while KL-6 ⩾1000 U/ml was additionally selected as an independent risk factor (HR = 2.0, 95% CI: 1.2, 3.3; P = 0.01). Fig. 2 shows Kaplan-Meier analyses comparing the equality of cumulative survival curves between two groups stratified by individual prognostic factors. The cumulative survival rates were significantly different (P < 0.01 for all comparisons), and this difference was apparent even early in the course of the disease because of deaths due to RP-ILD. The survival curves for the low and high SP-D groups intersected at 80 months, indicating possible violation in proportional hazards model. Exclusion of SP-D from the explanatory variables resulted in the selection of age at onset, anti-MDA5 antibody and CRP as variables for ILD mortality (supplementary Fig. S1, available at Rheumatology online). Since there were several missing variables in our database, a multiple imputation method was applied to confirm the model validity (supplementary Fig. S2, available at Rheumatology online). The significant variables obtained were principally identical, and included age at disease onset, anti-MDA5 antibody, CRP, KL-6 and SpO2. Fig. 2 View largeDownload slide Cumulative survival rates between groups stratified by independent predictors of ILD mortality Kaplan-Meier analysis with the Breslow test was used for factors selected by final model for predictors for ILD mortality. Censors were indicated in each graph. ILD: interstitial lung disease; KL-6: Krebs von den Lungen-6; SP-D: surfactant protein; ARS: aminoacyl; MDA5: melanoma differentiation-associated gene 5. Fig. 2 View largeDownload slide Cumulative survival rates between groups stratified by independent predictors of ILD mortality Kaplan-Meier analysis with the Breslow test was used for factors selected by final model for predictors for ILD mortality. Censors were indicated in each graph. ILD: interstitial lung disease; KL-6: Krebs von den Lungen-6; SP-D: surfactant protein; ARS: aminoacyl; MDA5: melanoma differentiation-associated gene 5. Because 67 (88%) deaths occurred in anti-MDA5-positive patients, our model was likely to depend heavily on the anti-MDA5 antibody. To assess this possibility, we generated predictive models for the mortality, separately in patients with and without anti-MDA5 antibody. Two patients with coexisting anti-MDA5 and anti-ARS antibodies were included in the anti-MDA5-positive group. As shown in supplementary Table S3, available at Rheumatology online, and Fig. 3A, in anti-MDA5-positive patients, age at disease onset ⩾60 years and CRP ⩾1 mg/dl were again significant explanatory variables for ILD mortality risk, while periungual erythema was identified as a favourable predictor. The same variables were selected when a multiple imputation method was applied (supplementary Fig. S3A, available at Rheumatology online). In contrast, no significant explanatory variable for ILD mortality risk was identified in anti-MDA5-negative patients, regardless of the absence (supplementary Table S4, available at Rheumatology online; Fig. 3B) or the presence of a multiple imputation method (supplementary Fig. S3B, available at Rheumatology online), but there were trends toward a correlation between ILD mortality risk and age at disease onset ⩾60 years or CRP ⩾1 mg/dl. Fig. 3 View largeDownload slide Predictive models for mortality due to respiratory insufficiency in patients with and without anti-MDA5 antibody Initial predictors of poor prognosis due to fatal respiratory insufficiency were identified in patients with anti-MDA5 antibody (A) and in those without anti-MDA5 antibody (B), by multivariate analysis using a stepwise selection of parameters and the cause-specific Cox proportional hazards regression model. The hazard ratio with 95% CI is shown for each variable finally selected. ILD: interstitial lung disease; MDA5: melanoma differentiation-associated gene 5; ARS: aminoacyl. Fig. 3 View largeDownload slide Predictive models for mortality due to respiratory insufficiency in patients with and without anti-MDA5 antibody Initial predictors of poor prognosis due to fatal respiratory insufficiency were identified in patients with anti-MDA5 antibody (A) and in those without anti-MDA5 antibody (B), by multivariate analysis using a stepwise selection of parameters and the cause-specific Cox proportional hazards regression model. The hazard ratio with 95% CI is shown for each variable finally selected. ILD: interstitial lung disease; MDA5: melanoma differentiation-associated gene 5; ARS: aminoacyl. Discussion In this study, we used the large-scale JAMI database to identify initial predictors for mortality due to respiratory insufficiency in adult patients with PM/DM-associated ILD. A number of factors at diagnosis were selected by univariate analysis, but cause-specific Cox proportional hazards models revealed the following to be predictors independently associated with future poor mortality due to ILD: age at disease onset, anti-MDA5 antibody, CRP and SpO2. This final model was verified by a variety of different models. Since age, CRP and SpO2 are readily available in routine clinical practice, and kits to measure the anti-MDA5 antibody are now commercially available [29, 30], our findings are widely applicable for identifying PM/DM patients with intractable ILD who require intensive treatment at an early stage. Our results will also assist in designing appropriate patient inclusion criteria for future clinical trials for PM/DM-associated ILD. JAMI is a nationwide, large-scale cohort established in Japan. It is the first multicentre cohort focusing on PM/DM-associated ILD, although there are several multicentre cohorts of patients with idiopathic inflammatory myopathy [31, 32], juvenile-onset myositis [33] and anti-synthetase syndrome [34, 35]. The patient population might vary between specialties, because the clinical presentation of PM/DM is quite heterogeneous; patients with progressive or advanced ILD with significant respiratory distress are likely to be referred to pulmonologists or critical care physicians, while those whose predominant symptom is skin rashes are often referred to dermatologists. Therefore, patients in the JAMI cohort are cared for by a wide variety of specialties to avoid potential patient selection bias. Finally, because serum samples obtained at the time of diagnosis were available for all patients, we were able to run assays for a full panel of MSAs. These features give the JAMI cohort several advantages for analysing various clinical aspects of PM/DM-associated ILD. This study confirmed many prognostic factors reported in previous studies, including gender, age at disease onset, CADM, classic DM, fever, skin ulceration, periungual erythema, muscle weakness, creatine kinase, ferritin, chest HRCT patterns, SpO2, anti-MDA5 antibody and anti-ARS antibody, although the populations and endpoints were somewhat different among studies [9–19]. Since many of these factors are correlated with each other, it is difficult to identify primary contributing factors from previous studies, primarily because of small sample sizes. In fact, in our cohort, the majority of these factors were not selected as independent predictors in multivariate analysis. Our predictive modelling successfully identified age at disease onset, CRP and anti-MDA5 antibody as independent initial predictors of poor ILD prognosis. In particular, anti-MDA5 antibody provided the highest HR, with a 7.5-fold increase in risk. In fact, the other prognostic factors were associated with anti-MDA5 antibody, and were mostly relevant to anti-MDA5-positive patients who dominated the analysis, since we failed to identify risk for ILD mortality in patients without anti-MDA5 antibody, probably because of the small number of deaths. This finding is consistent with a meta-analysis of 13 studies that found a pooled sensitivity and specificity of anti-MDA5 antibody for RP-ILD of 77 and 86%, respectively, in DM patients [36]. In contrast, we found that anti-ARS antibody tended to predict favourable ILD outcomes, at least in the short term, although this is likely a reflection of the strong association between anti-MDA5 antibody and poor prognosis. In this regard, a previous study found no significant increase in mortality in patients with anti-synthetase syndrome compared with the general US population [37]. In addition, Yoshida et al. [38] reported that the anti-MDA5 antibody is associated with a higher 3-month mortality compared with the anti-ARS antibody. Thus, MSA measurement at diagnosis is highly useful for predicting treatment responses and outcomes in patients with PM/DM-associated ILD. The prognostic significance of CRP in patients with PM/DM-associated ILD has not been emphasized in the literature, although a recent meta-analysis indicated that CRP is associated with an increased risk of developing ILD in PM/DM patients [39]. Notably, in a small Chinese study of 40 patients with CADM, multivariate logistic regression analysis showed that anti-MDA5 antibody combined with elevated CRP is an independent risk factor for developing RP-ILD [15]. Interestingly, CRP is a prognostic factor for other forms of ILD, such as biopsy-proven idiopathic pulmonary fibrosis [40] and SS-associated ILD [41], suggesting that this biomarker might reflect inflammation in the lung parenchyma that leads to irreversible damage. Since JAMI is a retrospective cohort, management was determined by attending physicians, who were unaware of the MSA data. In addition, the treatment regimens for the JAMI cohort were considerably different from those in previous reports. The JAMI patients were often treated with calcineurin inhibitors and a ‘triple combo’ consisting of high-dose CS, CYC and calcineurin inhibitor. IVIG and polymyxin B-immobilized fibre column-direct haemoperfusion were occasionally administered on top of immunosuppressive regimens [42, 43]. MMF or rituximab was seldom used, potentially contributing to worse survivals in our cohort. The selections of treatment regimens are different from those in other countries, primarily due to Japan’s unique health insurance system. Although these treatment regimens had the potential to influence patient outcomes, our multivariate model consistently selected the same independent prognostic factors even after adjusting for the drugs used in induction treatment as potential confounders. The present study has potential limitations due to its retrospective nature. Patients were selected mainly from tertiary referral hospitals, so there may be a bias toward more severe forms of the disease. However, the nature of a multicentre study and the inclusion of dermatology centres may minimize this bias. In addition, because the observation period after diagnosis was a median of 20 months, our current analysis only detected predictors for short-term mortality. This feature might have caused us to overestimate the importance of the anti-MDA5 antibody due to its strong association with RP-ILD. Long-term outcomes can be assessed as more follow-up data become available. Finally, information on smoking was not included in the JAMI database, because it has never been reported as a poor prognostic factor in patients with PM/DM-associated ILD. However, a recent multicentre registry of patients with idiopathic inflammatory myopathy (EuroMyositis) found an association of smoking with ILD [31], urging us to examine a potential influence of smoking habits on ILD prognosis in the JAMI cohort as a future project. In conclusion, we successfully generated a model to predict ILD mortality in patients with PM/DM-associated ILD at diagnosis, based on a newly established multicentre cohort. Our findings are easily applied to routine care, and may help to identify patients who are at high risk for ILD mortality and require immediate intensive treatment. Acknowledgements Other JAMI investigators: Yutaka Okano (Kawasaki Municipal Kawasaki Hospital), Yukie Yamaguchi (Yokohama City University Graduate School of Medicine), Yoshinori Taniguchi (Kochi Medical School Hospital, Kochi University), Jun Kikuchi (Saitama Medical Centre, Saitama Medical University), Makoto Kubo (Yamaguchi University Graduate School of Medicine), Masaki Watanabe (Graduate School of Medical and Dental Sciences, Kagoshima University), Tatsuhiko Harada (Nagasaki University School of Medicine), Taisuke Kazuyori (The Jikei University School of Medicine Katsushika Medical Centre), Hideto Kameda (Toho University Ohashi Medical Centre), Makoto Kaburaki (Toho University School of Medicine), Yasuo Matsuzawa (Toho University Medical Centre, Sakura Hospital), Shunji Yoshida (Fujita Health University School of Medicine), Yasuko Yoshioka, Takuya Hirai (Juntendo University Urayasu Hospital), Yoko Wada (Niigata University Graduate School of Medical and Dental Sciences), Koji Ishii, Sakuhei Fujiwara (Faculty of Medicine Oita University), Takeshi Saraya (Kyorin University), Kozo Morimoto (Fukujuji Hospital, Japan Anti-Tuberculosis Association), Tetsu Hara (Hiratsuka Kyosai Hospital), Hiroki Suzuki (Saiseikai Yamagata Saisei Hospital), Hideki Shibuya (Tokyo Teishin Hospital), Yoshinao Muro (Nagoya University Graduate School of Medicine), Ryoichi Aki (Kitasato University School of Medicine), Takuo Shibayama (National Hospital Organisation Okayama Medical Centre), Shiro Ohshima (National Hospital Organisation Osaka Minami Medical Centre), Yuko Yasuda (Saiseikai Kumamoto Hospital), Masaki Terada (Saiseikai Niigata Daini Hospital) and Yoshie Kawahara (Keiyu Hospital). Author contributions: conceived and designed the study: S.S., K.M., N.N., T.G., M.K.; analysed the data: K.M.; contributed data collection/analysis tools: S.S., K.M., N.N., Y.Kawaguchi, A.K., M.T., K.I., T.Nunokawa, Y.T., K.A., Y.Kaneko, T.G., T.U., S.K., T.Naniwa, M.K. and JAMI investigators; wrote the paper: S.S., K.M., N.N., Y.Kawaguchi, A.K., M.T., K.I., T.Nunokawa, Y.T., K.A., Y.Kaneko, T.G., T.U., S.K., T.Naniwa, M.K. and JAMI investigators. Funding: This work was supported by a research grant from Astellas and a research grant for intractable diseases from the Japanese Ministry of Health, Labour and Welfare. The funders had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript. Disclosure statement: S.S. and M.K. hold a patent for the anti-MDA5 antibody measurement kit. The other authors have declared no conflicts of interest. Supplementary data Supplementary data are available at Rheumatology online. References 1 Tieu J , Lundberg IE , Limaye V. Idiopathic inflammatory myositis . Best Pract Res Clin Rheumatol 2016 ; 30 : 149 – 68 . Google Scholar CrossRef Search ADS PubMed 2 Johnson C , Pinal-Fernandez I , Parikh R et al. Assessment of mortality in autoimmune myositis with and without associated interstitial lung disease . Lung 2016 ; 194 : 733 – 7 . Google Scholar CrossRef Search ADS PubMed 3 Mimori T , Nakashima R , Hosono Y. Interstitial lung disease in myositis: clinical subsets, biomarkers, and treatment . Curr Rheumatol Rep 2012 ; 14 : 264 – 74 . Google Scholar CrossRef Search ADS PubMed 4 Morisset J , Johnson C , Rich E , Collard HR , Lee JS. Management of myositis-related interstitial lung disease . Chest 2016 ; 150 : 1118 – 28 . Google Scholar CrossRef Search ADS PubMed 5 Hamaguchi Y , Kuwana M , Hoshino K et al. Clinical correlations with dermatomyositis-specific autoantibodies in adult Japanese patients with dermatomyositis: a multicenter cross-sectional study . Arch Dermatol 2011 ; 147 : 391 – 8 . Google Scholar CrossRef Search ADS PubMed 6 Moghadam-Kia S , Oddis CV , Sato S , Kuwana M , Aggarwal R. Anti-melanoma differentiation-associated gene 5 is associated with rapidly progressive lung disease and poor survival in US patients with amyopathic and myopathic dermatomyositis . Arthritis Care Res 2016 ; 68 : 689 – 94 . Google Scholar CrossRef Search ADS 7 Kotani T , Makino S , Takeuchi T et al. Early intervention with corticosteroids and cyclosporin A and 2-hour postdose blood concentration monitoring improves the prognosis of acute/subacute interstitial pneumonia in dermatomyositis . J Rheumatol 2008 ; 35 : 254 – 9 . Google Scholar PubMed 8 Horai Y , Isomoto E , Koga T et al. Early diagnosis and treatment for remission of clinically amyopathic dermatomyositis complicated by rapid progress interstitial lung disease: a report of two cases . Mod Rheumatol 2013 ; 23 : 190 – 4 . Google Scholar CrossRef Search ADS PubMed 9 Takada T , Suzuki E , Nakano M et al. Clinical features of polymyositis/dermatomyositis with steroid-resistant interstitial lung disease . Intern Med 1998 ; 37 : 669 – 73 . Google Scholar CrossRef Search ADS PubMed 10 Selva-O’Callaghan A , Labrador-Horrillo M , Muñoz-Gall X et al. Polymyositis/dermatomyositis-associated lung disease: analysis of a series of 81 patients . Lupus 2005 ; 14 : 534 – 42 . Google Scholar CrossRef Search ADS PubMed 11 Cao H , Xia Q , Pan M et al. Gottron papules and Gottron sign with ulceration: a distinctive cutaneous feature in a subset of patients with classic dermatomyositis and clinically amyopathic dermatomyositis . J Rheumatol 2016 ; 43 : 1735 – 42 . Google Scholar CrossRef Search ADS PubMed 12 Sato S , Hirakata M , Kuwana M et al. Autoantibodies to a 140-kd polypeptide, CADM-140, in Japanese patients with clinically amyopathic dermatomyositis . Arthritis Rheum 2005 ; 52 : 1571 – 6 . Google Scholar CrossRef Search ADS PubMed 13 Gono T , Kuwana M. Choosing the right biomarkers to predict ILD in myositis . Nat Rev Rheumatol 2016 ; 12 : 504 – 6 . Google Scholar CrossRef Search ADS PubMed 14 Gono T , Kawaguchi Y , Satoh T et al. Clinical manifestation and prognostic factor in anti-melanoma differentiation-associated gene 5 antibody-associated interstitial lung disease as a complication of dermatomyositis . Rheumatology (Oxford) 2010 ; 49 : 1713 – 9 . Google Scholar CrossRef Search ADS PubMed 15 Xu Y , Yang CS , Li YJ et al. Predictive factors of rapidly progressive-interstitial lung disease in patients with clinically amyopathic dermatomyositis . Clin Rheumatol 2016 ; 35 : 113 – 6 . Google Scholar CrossRef Search ADS PubMed 16 Hozumi H , Fujisawa T , Nakashima R et al. Comprehensive assessment of myositis-specific autoantibodies in polymyositis/dermatomyositis-associated interstitial lung disease . Respir Med 2016 ; 121 : 91 – 9 . Google Scholar CrossRef Search ADS PubMed 17 Koga T , Fujikawa K , Horai Y et al. The diagnostic utility of anti-melanoma differentiation-associated gene 5 antibody testing for predicting the prognosis of Japanese patients with DM . Rheumatology (Oxford) 2012 ; 51 : 1278 – 84 . Google Scholar CrossRef Search ADS PubMed 18 Tanizawa K , Handa T , Nakashima R et al. The prognostic value of HRCT in myositis-associated interstitial lung disease . Respir Med 2013 ; 107 : 745 – 52 . Google Scholar CrossRef Search ADS PubMed 19 Fujiki Y , Kotani T , Isoda K et al. Evaluation of clinical prognostic factors for interstitial pneumonia in anti-MDA5 antibody-positive dermatomyositis patients . Mod Rheumatol 2017 ; 11 : 1 – 8 . 20 Bohan A , Peter JB. Polymyositis and dermatomyositis (first of two parts) . N Engl J Med 1975 ; 292 : 344 – 7 . Google Scholar CrossRef Search ADS PubMed 21 Sontheimer RD. Would a new name hasten the acceptance of amyopathic dermatomyositis (dermatomyositis siné myositis) as a distinctive subset within the idiopathic inflammatory dermatomyopathies spectrum of clinical illness? J Am Acad Dermatol 2002 ; 46 : 626 – 36 . Google Scholar CrossRef Search ADS PubMed 22 American Thoracic Society, European Respiratory Society . American Thoracic Society/European Respiratory Society international multidisciplinary consensus classification of the idiopathic interstitial pneumonias . Am J Respir Crit Care Med 2002 ; 165 : 277 – 304 . CrossRef Search ADS PubMed 23 Tanizawa K , Handa T , Nakashima R et al. HRCT features of interstitial lung disease in dermatomyositis with anti-CADM-140 antibody . Respir Med 2011 ; 105 : 1380e7 . Google Scholar CrossRef Search ADS 24 Forman MS , Nakamura M , Mimori T , Gelpi C , Hardin JA. Detection of antibodies to small nuclear ribonucleoproteins and small cytoplasmic ribonucleoproteins using unlabeled cell extracts . Arthritis Rheum 1985 ; 28 : 1356 – 61 . Google Scholar CrossRef Search ADS PubMed 25 Sato S , Hoshino K , Satoh T et al. RNA helicase encoded by melanoma differentiation-associated gene 5 is a major autoantigen in patients with clinically amyopathic dermatomyositis: association with rapidly progressive interstitial lung disease . Arthritis Rheum 2009 ; 60 : 2193 – 200 . Google Scholar CrossRef Search ADS PubMed 26 Zweig MH , Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine . Clin Chem 1993 ; 39 : 561 – 77 . Google Scholar PubMed 27 Naniwa T , Tamechika S , Okazaki Y , Maeda S , Kuwana M. Coexistence of anti-melanoma differentiation-associated gene 5 and anti-aminoacyl-transfer RNA synthetase antibodies in a patient with dermatomyositis and rapidly progressive and relapsing interstitial lung disease . Mod Rheumatol Case Rep 2017 ; 1 : 3 – 8 . Google Scholar CrossRef Search ADS 28 Abe S , Azuma A , Mukae H et al. Polymyxin B-immobilized fiber column (PMX) treatment for idiopathic pulmonary fibrosis with acute exacerbation: a multicenter retrospective analysis . Intern Med 2012 ; 51 : 1487 – 91 . Google Scholar CrossRef Search ADS PubMed 29 Sato S , Murakami A , Kuwajima A et al. Clinical utility of an enzyme-linked immunosorbent assay for detecting anti-melanoma differentiation-associated gene 5 autoantibodies . PLoS One 2016 ; 11 : e0154285 . Google Scholar CrossRef Search ADS PubMed 30 Cavazzana I , Fredi M , Ceribelli A et al. Testing for myositis specific autoantibodies: comparison between line blot and immunoprecipitation assays in 57 myositis sera . J Immunol Methods 2016 ; 433 : 1 – 5 . Google Scholar CrossRef Search ADS PubMed 31 Lilleker JB , Vencovsky J , Wang G et al. The EuroMyositis registry: an international collaborative tool to facilitate myositis research . Ann Rheum Dis 2018 ; 77 : 30 – 9 . Google Scholar CrossRef Search ADS PubMed 32 Nuño-Nuño L , Joven BE , Carreira PE et al. Mortality and prognostic factors in idiopathic inflammatory myositis: a retrospective analysis of a large multicenter cohort of Spain . Rheumatol Int 2017 ; 37 : 1853 – 61 . Google Scholar CrossRef Search ADS PubMed 33 Martin N , Krol P , Smith S et al. A national registry for juvenile dermatomyositis and other paediatric idiopathic inflammatory myopathies: 10 years' experience; the Juvenile Dermatomyositis National (UK and Ireland) Cohort Biomarker Study and Repository for Idiopathic Inflammatory Myopathies . Rheumatology (Oxford) 2011 ; 50 : 137 – 45 . Google Scholar CrossRef Search ADS PubMed 34 Cavagna L , Nuño L , Scirè CA et al. Clinical spectrum time course in anti Jo-1 positive antisynthetase syndrome: results from an international retrospective multicenter study . Medicine 2015 ; 94 : e1144 . Google Scholar CrossRef Search ADS PubMed 35 Trallero-Araguás E , Grau-Junyent JM , Labirua-Iturburu A et al. Clinical manifestations and long-term outcome of anti-Jo1 antisynthetase patients in a large cohort of Spanish patients from the GEAS-IIM group . Semin Arthritis Rheum 2016 ; 46 : 225 – 31 . Google Scholar CrossRef Search ADS PubMed 36 Chen Z , Cao M , Plana MN et al. Utility of anti-melanoma differentiation-associated gene 5 antibody measurement in identifying patients with dermatomyositis and a high risk for developing rapidly progressive interstitial lung disease: a review of the literature and a meta-analysis . Arthritis Care Res 2013 ; 65 : 1316 – 24 . Google Scholar CrossRef Search ADS 37 Pinal-Fernandez I , Casal-Dominguez M , Huapaya JA et al. A longitudinal cohort study of the anti-synthetase syndrome: increased severity of interstitial lung disease in black patients and patients with anti-PL7 and anti-PL12 autoantibodies . Rheumatology (Oxford) 2017 ; 56 : 999 – 1007 . Google Scholar CrossRef Search ADS PubMed 38 Yoshida N , Okamoto M , Kaieda S et al. Association of anti-aminoacyl-transfer RNA synthetase antibody and anti-melanoma differentiation-associated gene 5 antibody with the therapeutic response of polymyositis/dermatomyositis-associated interstitial lung disease . Respir Investig 2017 ; 55 : 24 – 32 . Google Scholar CrossRef Search ADS PubMed 39 Zhang L , Wu G , Gao D et al. Factors associated with interstitial lung disease in patients with polymyositis and dermatomyositis: a systematic review and meta-analysis . PLoS One 2016 ; 11 : e0155381 . Google Scholar CrossRef Search ADS PubMed 40 Lee SH , Shim HS , Cho SH et al. Prognostic factors for idiopathic pulmonary fibrosis: clinical, physiologic, pathologic, and molecular aspects . Sarcoidosis Vasc Diffuse Lung Dis 2011 ; 28 : 102 – 12 . Google Scholar PubMed 41 Liu X , Mayes MD , Pedroza C et al. Does C-reactive protein predict the long-term progression of interstitial lung disease and survival in patients with early systemic sclerosis? Arthritis Care Res 2013 ; 65 : 1375 – 80 . Google Scholar CrossRef Search ADS 42 Suzuki Y , Hayakawa H , Miwa S et al. Intravenous immunoglobulin therapy for refractory interstitial lung disease associated with polymyositis/dermatomyositis . Lung 2009 ; 187 : 201 – 6 . Google Scholar CrossRef Search ADS PubMed 43 Teruya A , Kawamura K , Ichikado K et al. Successful polymyxin B hemoperfusion treatment associated with serial reduction of serum anti-CADM-140/MDA5 antibody levels in rapidly progressive interstitial lung disease with amyopathic dermatomyositis . Chest 2013 ; 144 : 1934 – 6 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Bakker, Pauline A;van den Berg, Rosaline;de Hooge, Manouk;van Lunteren, Miranda;Ez-Zaitouni, Zineb;Fagerli, Karen M;Landewé, Robert;van Oosterhout, Maikel;Ramonda, Roberta;Reijnierse, Monique;van Gaalen, Floris A;van der Heijde, Désirée

2018 Rheumatology

doi: 10.1093/rheumatology/kex532pmid: 29584927

Abstract Objectives To investigate in patients with chronic back pain of a short duration, the utility of adding structural MRI lesions of the SI joints to the imaging criterion of the Assessment of SpondyloArthritis International Society (ASAS) axial SpA (axSpA) criteria and the utility of replacement of radiographic sacroiliitis by structural MRI lesions. Methods MRI STIR (inflammation, MRI-SI), MRI T1-weighted images (structural lesions, MRI-SI-s) and radiographs of the SI joints of patients in the SPondyloArthritis Caught Early-cohort (chronic back pain: ⩾3 months, ⩽2 years; onset <45 years) were scored by two well-calibrated readers. Previously proposed cut-offs for a positive MRI-SI-s were used (based on <5% prevalence in no-SpA patients): erosions ⩾3, fatty lesions ⩾3, fatty lesions and/or erosions (erosions/fatty lesions) ⩾5. Using the definitions of MRI-SI-s, patients were classified according to the ASAS axSpA criteria. Results Twenty-nine of 294 patients were modified New York (mNY) positive and 32 were MRI-SI-s positive (erosions/fatty lesions ⩾5). Agreement between mNY and MRI-SI-s (erosions/fatty lesions ⩾5) was moderate (κ: 0.58). Using the erosions/fatty lesions ⩾5 cut-off, 3/294 additional patients were classified as axSpA (adding MRI). Using this cut-off instead of mNY (replacing mNY), classification did not change in 286 patients (97.3%), but 5 patients (1.7%) would not be classified as axSpA and 3 previously unclassified patients (1.0%) would be classified as axSpA. Similar results were seen for the other cut-offs (erosions ⩾3 and fatty lesions ⩾3). Conclusion Assessment of structural lesions (fatty lesions and erosions) on MRI-SI instead of or in addition to conventional radiographs does not lead to a different ASAS axSpA classification in most of the patients with early disease onset. This suggests that structural lesions (fatty lesions and erosions) can be reliably used in the ASAS axSpA classification of patients, as both addition and replacement of radiographs of the SI joints. axial spondyloarthritis, sacroiliitis, structural lesions, MRI, radiographs Rheumatology key messages Structural MRI lesions in SI joints can be reliably used to classify patients with axial SpA. Adding or replacing radiographic sacroiliitis by structural MRI lesions led to minor changes in axial SpA classification. Introduction Recently, we have demonstrated that in the DEvenir des Spondyloarthrites Indifférenciées Récentes (DESIR) cohort of patients with inflammatory back pain with a high suspicion of having axial SpA (axSpA), structural lesions on MRI can be used reliably either as an addition to or as a substitute for radiographs in the Assessment of SpondyloArthritis International Society (ASAS) axSpA classification of patients spondyloarthritis [1]. Since this is a new concept with potential consequences for daily clinical practice, replication in other cohorts is highly warranted. The current study investigates the same research question in an entirely independent cohort, the SPondyloArthritis Caught Early (SPACE) cohort of patients with chronic back pain of a short duration. The rationale behind using structural lesions on MRI instead of conventional radiographs is mainly based on the fact that a reliable detection of sacroiliitis on conventional radiographs is notoriously difficult. Sacroiliitis is the hallmark of AS and the presence of sacroiliitis on conventional radiography is obligatory in the modified New York (mNY) criteria for AS [2]. However, it has been shown that substantial observer variation (interreader, intrareader) exists for both radiologists and rheumatologists, and training does not lead to improvement [2]. Recent data from the DESIR cohort have also revealed a poor to moderate agreement between different readers leading to substantial misclassification of patients [3]. With the help of MRI it has become possible to detect active sacroiliitis (bone marrow oedema, BME) in early stages of axSpA. The ASAS definition for a positive MRI is solely based on identification of BME highly suggestive for SpA, but structural lesions can be seen [fatty lesions, erosions, sclerosis and (partial) ankylosis] are visible on MRI as well [4]. The 3D (albeit tomographic) character of MRI may be an advantage as compared with the 2D projection of conventional radiographs [5]. Another advantage of MRI above radiography is that patients are not exposed to ionizing radiation. Both the SPACE and the DESIR are cohorts of early disease. Since it often takes 6–8 years from the onset of symptoms before radiographic sacroiliitis can be detected on plain radiographs, conventional radiography is a suboptimal imaging technique especially in patients with complaints of a short duration (in addition to the earlier described reliability issues). The DESIR cohort only includes patients with inflammatory back pain (IBP) with up to 3 years of symptoms, whereas the SPACE cohort includes patients with chronic back pain (CBP) with up to 2 years of symptoms. Different reader pairs were involved in the scoring process of the two cohorts, to rule out a reader-dependent effect. In this study we aim to evaluate the usefulness of structural lesions with regard to the ASAS axSpA classification criteria in the SPACE cohort. First of all, we determine the agreement between sacroiliitis seen on radiographs (mNY) and structural lesions seen on MRI. Then we aim to evaluate the potential impact of adding structural lesions on MRI to the definition of a positive MRI or replacing radiographic sacroiliitis by MRI structural lesions on the ASAS axSpA classification of patients. Methods Study population For this analysis, baseline data from the SPACE cohort were used. An extensive description of the SPACE cohort has been given elsewhere [6]. In short, SPACE is an inception cohort with on-going inclusion and follow-up of patients with CBP of short duration (⩾3 months but ⩽2 years, with onset <45 years). Patients were recruited from five participating centres in the Netherlands (Leiden, Amsterdam, Gouda), Norway (Oslo) and Italy (Padua). Approval by the local medical ethics committees was obtained, as well as written informed consent from all patients in accordance with the Declaration of Helsinki. A full diagnostic work-up was performed in all patients including: HLA-B27 testing, conventional radiographs and MRI of the SI joints (X-SI and MRI-SI, respectively), and the assessment of all other SpA features, in agreement with the descriptions supplied by the ASAS group [7]. Imaging and scoring methods MRI-SI was performed on a 1.5 Tesla machine (Philips Medical Systems, Best, Netherlands). Coronal oblique T1-weighted TSE (TR 550/TE 10) and STIR (TR 2500/TE 60) with a slice thickness of 4 mm were the acquired sequences used. All available baseline MRI-SI and X-SI were read independently, in different reading sessions, by two trained, well-calibrated readers (P.A.B., M.d.H.). Readers were blinded to the score of the other reader and other modality as well as the clinical information, throughout the scoring process. The mNY criteria were used to assess radiographs of the SI joints; radiographic sacroiliitis was defined as bilateral grade ⩾2 or unilateral grade ⩾3 [8]. In consensus with the ASAS definition, an MRI-SI was defined positive if one BME lesion highly suggestive of SpA was present on two or more consecutive slices, or otherwise if several BME lesions highly suggestive of SpA were visible on a single slice [9]. A third reader served as adjudicator (R.vd.B.) in case of disagreement among the two initial readers regarding a positive MRI (ASAS definition) or the presence of sacroiliitis (mNY criteria). The presence of structural lesions, namely fatty lesions, erosions, sclerosis and (partial) ankylosis, was assessed on MRI T1-weighted images in conjunction with the STIR images. A scoring system with similarities to the methodology outlined in the Spondyloarthritis Research Consortium of Canada (SPARCC) online training module, as described by Weber et al. [10] was used. Weber et al. proposed a scoring system to quantify structural lesions called SPARCC SI structural lesion score. This method is founded on the assessment of lesions (present vs absent) counting them in each quadrant on six consecutive slices through the SI joints. The starting point is the slice on which at least 1 cm of vertical height of the cartilage compartment can be seen, from anterior to posterior, evaluating the cartilaginous compartment of the SI joints and the antero-inferior portion of the SI joint. Each SI joint is split into four quadrants. Structural lesions were taken into account only if present on at least two consecutive slices. This is reflected by a maximum score of 40 per lesion (5 lesions per quadrant × 4 quadrants × 2 SI joints) except for (partial) ankylosis. As an exception, (partial) ankylosis was considered sufficient when seen on a single slice reflected by a maximum score of 24 per patient. As shown earlier by our group, an MRI positive for structural lesions (MRI-SI-s) was determined by the use of different cut-offs [11]. These chosen cut-offs were anchored on ⩽5% false positives whereby the false positives were specified as structural lesions among patients not having axSpA according to the ASAS axSpA criteria. The described cut-offs have pointed out to be: erosions ⩾3, fatty lesions ⩾3, fatty lesions and/or erosions ⩾5. In this early cohort, the prevalence of sclerosis and (partial) ankylosis was so low that there was no cut-off that could clearly differentiate between SpA and no-SpA patients. As a consequence, these types of lesions were not further considered. Classification criteria Patients were classified according to the ASAS axSpA criteria. Subsequently, patients were grouped based on the way they met the criteria: through the imaging arm of the ASAS axSpA criteria alone (either by mNY criteria and/or by positive MRI); through the clinical arm of the ASAS axSpA criteria alone; or through both. If patients fulfilled more than one category, they were classified in that way, reflected by seven possible combinations (Figs 1 and 2). The no-axSpA group is made up of patients not fulfilling the ASAS axSpA criteria. Fig. 1 View largeDownload slide Addition of MRI-SI-s to the ASAS axSpA criteria (combination fatty lesions and/or erosions): scenario 1 axSpA: axial SpA; ASAS: Assessment of SpondyloArthritis International Society; mNY: modified New York criteria; MRI-SI-s: MRI of the SI joints assessed for structural lesions. Fig. 1 View largeDownload slide Addition of MRI-SI-s to the ASAS axSpA criteria (combination fatty lesions and/or erosions): scenario 1 axSpA: axial SpA; ASAS: Assessment of SpondyloArthritis International Society; mNY: modified New York criteria; MRI-SI-s: MRI of the SI joints assessed for structural lesions. Fig. 2 View largeDownload slide Replacement of mNY criteria by MRI-SI-s in the ASAS axSpA criteria (combination fatty lesions and/or erosions): scenario 2 axSpA: axial SpA; ASAS: Assessment of SpondyloArthritis International Society; mNY: modified New York criteria; MRI-SI-s: MRI of the SI joints assessed for structural lesions. Fig. 2 View largeDownload slide Replacement of mNY criteria by MRI-SI-s in the ASAS axSpA criteria (combination fatty lesions and/or erosions): scenario 2 axSpA: axial SpA; ASAS: Assessment of SpondyloArthritis International Society; mNY: modified New York criteria; MRI-SI-s: MRI of the SI joints assessed for structural lesions. Data analysis Descriptive statistics were used to calculate disease characteristics of the included patients. In the analysis, regarding structural lesions the mean score of the two readers in agreement of a positive MRI (ASAS definition) was used. In case of disagreement, the mean of the scores of the adjudicator and the reader in agreement with the adjudicator’s judgement regarding a positive MRI for that particular case were used. Agreement about the absence or presence of structural lesions using both imaging modalities (X-SI and MRI-SI-s) was assessed by cross-tabulation and expressed as Cohen’s κ. Percentage positive agreement was calculated in order to leave out patients labelled as negative by the two readers, which could lead to an artificially high agreement. Subsequently, using the various definitions of MRI-SI-s, patients were classified by the ASAS axSpA criteria. MRI-SI-s was added to the imaging criterion of the ASAS axSpA criteria in the first place, resulting in an additional possibility to fulfil the imaging arm (scenario 1). Second, replacement of the mNY criterion by MRI-SI-s (scenario 2) was applied as if only an MRI was performed. Disease probabilities were calculated by the positive likelihood ratio (LR) product, for those patients changing ASAS axSpA classification in scenario 2. This was done by multiplying the individual LRs of all identified SpA features [12]. In patients with CBP with an assumed disease prevalence of axSpA of 5%, an LR product of 200 results in a positive predictive value of 90%. These patients were further described by their clinical phenotype: gender, age and whether a patient was diagnosed as axSpA according to the treating rheumatologist. The analyses were performed in STATA 12.0 (StataCorp LP, College Station, Texas, USA). Results Included in the analysis are 294 patients with complete imaging data at baseline (both MRI-SI and X-SI present). Table 1 describes patient characteristics. Patients had a mean (s.d.) age of 31.6 years of age (10.7 years) and a mean (s.d.) duration of back pain of 13.1 months (7.3 months). Of these patients, 34.6% were men and 34.6% were HLA-B27 positive. Table 1 Disease characteristics of the included patients Disease characteristics of the included patients Total number (n = 294) Age at inclusion, mean (s.d.), years 31.6 (10.7) Male, n (%) 102 (34.6) Symptom duration at first visit, mean (s.d.), months 13.1 (7.3) Good response to NSAIDs, n (%) 92 (31.6) IBP, n (%) 184 (62.6) Positive family history of SpA, n (%) 112 (38.1) Peripheral arthritis, n (%) 37 (12.6) Dactylitis, n (%) 10 (3.4) Enthesitis, n (%) 43 (14.6) Uveitis, n (%) 21 (7.1) IBD, n (%) 26 (8.8) Psoriasis, n (%) 29 (9.7) Elevated CRP, n (%) 54 (18.8) HLA-B27 positive, n (%) 100 (34.6) Sacroiliitis present on radiograph, n (%) 29 (9.8) Positive MRI (ASAS definition), n (%) 37 (12.5) Disease characteristics of the included patients Total number (n = 294) Age at inclusion, mean (s.d.), years 31.6 (10.7) Male, n (%) 102 (34.6) Symptom duration at first visit, mean (s.d.), months 13.1 (7.3) Good response to NSAIDs, n (%) 92 (31.6) IBP, n (%) 184 (62.6) Positive family history of SpA, n (%) 112 (38.1) Peripheral arthritis, n (%) 37 (12.6) Dactylitis, n (%) 10 (3.4) Enthesitis, n (%) 43 (14.6) Uveitis, n (%) 21 (7.1) IBD, n (%) 26 (8.8) Psoriasis, n (%) 29 (9.7) Elevated CRP, n (%) 54 (18.8) HLA-B27 positive, n (%) 100 (34.6) Sacroiliitis present on radiograph, n (%) 29 (9.8) Positive MRI (ASAS definition), n (%) 37 (12.5) ASAS: Assessment of SpondyloArthritis International Society; IBP: inflammatory back pain; SPACE: SPondyloArthritis Caught Early. Table 1 Disease characteristics of the included patients Disease characteristics of the included patients Total number (n = 294) Age at inclusion, mean (s.d.), years 31.6 (10.7) Male, n (%) 102 (34.6) Symptom duration at first visit, mean (s.d.), months 13.1 (7.3) Good response to NSAIDs, n (%) 92 (31.6) IBP, n (%) 184 (62.6) Positive family history of SpA, n (%) 112 (38.1) Peripheral arthritis, n (%) 37 (12.6) Dactylitis, n (%) 10 (3.4) Enthesitis, n (%) 43 (14.6) Uveitis, n (%) 21 (7.1) IBD, n (%) 26 (8.8) Psoriasis, n (%) 29 (9.7) Elevated CRP, n (%) 54 (18.8) HLA-B27 positive, n (%) 100 (34.6) Sacroiliitis present on radiograph, n (%) 29 (9.8) Positive MRI (ASAS definition), n (%) 37 (12.5) Disease characteristics of the included patients Total number (n = 294) Age at inclusion, mean (s.d.), years 31.6 (10.7) Male, n (%) 102 (34.6) Symptom duration at first visit, mean (s.d.), months 13.1 (7.3) Good response to NSAIDs, n (%) 92 (31.6) IBP, n (%) 184 (62.6) Positive family history of SpA, n (%) 112 (38.1) Peripheral arthritis, n (%) 37 (12.6) Dactylitis, n (%) 10 (3.4) Enthesitis, n (%) 43 (14.6) Uveitis, n (%) 21 (7.1) IBD, n (%) 26 (8.8) Psoriasis, n (%) 29 (9.7) Elevated CRP, n (%) 54 (18.8) HLA-B27 positive, n (%) 100 (34.6) Sacroiliitis present on radiograph, n (%) 29 (9.8) Positive MRI (ASAS definition), n (%) 37 (12.5) ASAS: Assessment of SpondyloArthritis International Society; IBP: inflammatory back pain; SPACE: SPondyloArthritis Caught Early. One hundred and three out of 294 patients (35.0%) fulfilled the ASAS axSpA criteria using the standard definition. Of these 103 patients, 50 patients fulfilled the imaging arm (48.5%) and 53 patients fulfilled the clinical arm only (51.5%). The most prevalent SpA features were IBP (62.6%), a positive family history for SpA (38.1%) and a good response to NSAIDs (31.6%). Regarding the prevalence of structural lesions, 20/294 patients (6.8%) showed three or more fatty lesions; 16 of these 20 patients (80%) already formally fulfilled the ASAS axSpA criteria. Thirty-four patients (11.6%) had three or more erosions, and 27 of these 34 (79.4%) already formally fulfilled the ASAS axSpA criteria. Thirty-one patients (10.5%) had 5 or more fatty and/or erosive lesions (combination), of which 26 patients (83.9%) already formally fulfilled the ASAS axSpA criteria. The agreement regarding the presence/absence of radiographic sacroiliitis and the presence/absence of structural lesions on MRI was moderate (Table 2). Subtle differences were seen between the various definitions used: κ: 0.51 (erosions ⩾3); κ: 0.45 (fatty lesions ⩾3); and κ: 0.58 (combination of fatty lesions and erosions ⩾5). Table 2 Agreement between sacroiliitis on conventional radiographs (mNY criteria) and a positive MRI-SI based on structural lesions mNY (adjudicated) κ: 0.51 Erosions: cut-off ≥3; mean 2 out of 3 readers Positive Negative Total Positive 18 17 35 Negative 11 248 259 Total 29 265 294 mNY (adjudicated) κ: 0.45 Fatty lesions: cut-off ≥3; mean 2 out of 3 readers Positive Negative Total Positive 12 8 20 Negative 17 257 274 Total 29 265 294 mNY (adjudicated) κ: 0.58 Erosions/fatty lesions: cut-off ≥5; mean 2 out of 3 readers Positive Negative Total Positive 19 13 32 Negative 10 252 262 Total 29 265 294 mNY (adjudicated) κ: 0.51 Erosions: cut-off ≥3; mean 2 out of 3 readers Positive Negative Total Positive 18 17 35 Negative 11 248 259 Total 29 265 294 mNY (adjudicated) κ: 0.45 Fatty lesions: cut-off ≥3; mean 2 out of 3 readers Positive Negative Total Positive 12 8 20 Negative 17 257 274 Total 29 265 294 mNY (adjudicated) κ: 0.58 Erosions/fatty lesions: cut-off ≥5; mean 2 out of 3 readers Positive Negative Total Positive 19 13 32 Negative 10 252 262 Total 29 265 294 Agreement based on MRI-SI structural lesions using the three different definitions. mNY: modified New York criteria; MRI-SI: MRI of the SI joints. Table 2 Agreement between sacroiliitis on conventional radiographs (mNY criteria) and a positive MRI-SI based on structural lesions mNY (adjudicated) κ: 0.51 Erosions: cut-off ≥3; mean 2 out of 3 readers Positive Negative Total Positive 18 17 35 Negative 11 248 259 Total 29 265 294 mNY (adjudicated) κ: 0.45 Fatty lesions: cut-off ≥3; mean 2 out of 3 readers Positive Negative Total Positive 12 8 20 Negative 17 257 274 Total 29 265 294 mNY (adjudicated) κ: 0.58 Erosions/fatty lesions: cut-off ≥5; mean 2 out of 3 readers Positive Negative Total Positive 19 13 32 Negative 10 252 262 Total 29 265 294 mNY (adjudicated) κ: 0.51 Erosions: cut-off ≥3; mean 2 out of 3 readers Positive Negative Total Positive 18 17 35 Negative 11 248 259 Total 29 265 294 mNY (adjudicated) κ: 0.45 Fatty lesions: cut-off ≥3; mean 2 out of 3 readers Positive Negative Total Positive 12 8 20 Negative 17 257 274 Total 29 265 294 mNY (adjudicated) κ: 0.58 Erosions/fatty lesions: cut-off ≥5; mean 2 out of 3 readers Positive Negative Total Positive 19 13 32 Negative 10 252 262 Total 29 265 294 Agreement based on MRI-SI structural lesions using the three different definitions. mNY: modified New York criteria; MRI-SI: MRI of the SI joints. Scenario 1 The addition of MRI-SI-s to the imaging criterion of the ASAS axSpA criteria was investigated first (Fig. 1). Classification did not change in the majority of the patients (96.3%) using a cut-off value of 5 for the combination of fatty lesions and/or erosions. If a cut-off value of 3 for erosions or fatty lesions only was used, comparable results were seen: classification did not change in 94.9 and 97.6% of the patients, respectively. Considering the combination of five fatty lesions and/or erosions, three patients would be classified additionally axSpA if structural lesions on MRI were taken into consideration. The positive LR products of the three additionally classified patients were 15.8, 5.1 and 2.5, corresponding to post-test probabilities of 44, 20 and 11%, respectively. The rheumatologist diagnosed only one of these three patients (all female and HLA-B27 negative) with axSpA. Regarding the definition based on erosions, five patients would be additionally classified (all female and HLA-B27 negative). Only two of them were diagnosed as axSpA by the rheumatologist. Regarding the definition based on fatty lesions, three patients would be additionally classified as axSpA (one was described above, the other two were male patients, HLA-B27 negative, of whom one was diagnosed as having axSpA by the rheumatologist). Some patients changed subgroups within the ASAS classification, without a change in ASAS axSpA positivity or negativity. Eight patients would be classified via different arms due to the presence of structural lesions, using the combination of fatty lesions and/or erosions. Five patients also showed structural lesions on MRI (Fig. 1) but were already classified via the imaging arm based on inflammatory lesions on MRI. Three other patients fulfilled the clinical arm only, but fulfilled the imaging arm as well based on a positive MRI-SI-s. Using the cut-offs for fatty lesions or erosions only (both cut-off values were 3), the same trends were seen (data not shown). Scenario 2 Second, the replacement of radiographic sacroillitis by structural lesions on MRI (Fig. 2) and its impact on the ASAS axSpA classification was assessed. Using the same cut-off values of 5 for the combination of fatty lesions and/or erosions, classification did not change in the large majority of the patients (93.5%). A similar result was seen at a cut-off value of 3 for erosions and fatty lesions only: classification did not change in 91.8 and 92.9% of the patients, respectively. The same patients would be additionally classified as axSpA as in scenario 1. But assuming that only an MRI was performed, five patients (1.7%) would not be classified axSpA anymore if radiographic sacroliliitis was replaced by structural lesions on MRI using the combination of fatty lesions and/or erosions. The SpA-features of the patients that are newly classified by the ASAS axSpA criteria are described in Fig. 3, and the SpA-features of the patients that are no longer classified by the ASAS axSpA criteria are described in Fig. 4. One of these patients had four axSpA features, leading to a positive LR product of 192.4 and a corresponding high disease probability of 91%. One other patient had a disease probability of 75%. The disease probabilities of the other three patients turned out to be much lower: between 14 and 32%. Three of these five patients were female and two of the five patients were HLA-B27 positive (one male, one female). Of the five patients who would not be classified as axSpA anymore in this scenario, three patients were diagnosed axSpA by the rheumatologist. Regarding the definition based on erosions only, one additional patient would not be classified as ASAS axSpA anymore. This female patient was HLA-B27 negative, and was not diagnosed axSpA by the treating rheumatologist. Regarding the combination of fatty lesions and/or erosions, 11 patients changed arms within the criteria under this scenario but all stayed ASAS axSpA positive. Similar results were found when using a cut-off for fatty lesions only and erosions only (both cut-off values of 3) (data not shown). Fig. 3 View largeDownload slide ASAS axSpA positive patients (mNY positive) becoming ASAS axSpA negative (MRI-SI-s negativity) in scenario 2 Five patients marked in blue in Fig. 2. aProbability of axSpA is based on the positive likelihood ratio product [13]. IBP: inflammatory back pain; ASAS: Assessment of SpondyloArthritis International Society; mNY: modified New York criteria; MRI-SI-s: MRI of the SI joints assessed for structural lesions. Fig. 3 View largeDownload slide ASAS axSpA positive patients (mNY positive) becoming ASAS axSpA negative (MRI-SI-s negativity) in scenario 2 Five patients marked in blue in Fig. 2. aProbability of axSpA is based on the positive likelihood ratio product [13]. IBP: inflammatory back pain; ASAS: Assessment of SpondyloArthritis International Society; mNY: modified New York criteria; MRI-SI-s: MRI of the SI joints assessed for structural lesions. Fig. 4 View largeDownload slide ASAS axSpA negative patients becoming ASAS axSpA positive (due to MRI-SI-s positive) in scenario 2 The patients marked in purple in Fig. 2. aProbability of axSpA is based on the positive likelihood ratio product [13]. IBP: inflammatory back pain; ASAS: Assessment of SpondyloArthritis International Society; mNY: modified New York criteria; MRI-SI-s: MRI of the SI joints assessed for structural lesions. Fig. 4 View largeDownload slide ASAS axSpA negative patients becoming ASAS axSpA positive (due to MRI-SI-s positive) in scenario 2 The patients marked in purple in Fig. 2. aProbability of axSpA is based on the positive likelihood ratio product [13]. IBP: inflammatory back pain; ASAS: Assessment of SpondyloArthritis International Society; mNY: modified New York criteria; MRI-SI-s: MRI of the SI joints assessed for structural lesions. Discussion In this cohort of patients with CBP of short duration we have shown that adding structural lesions to the imaging criterion of the ASAS axSpA criteria has a limited effect on the classification of patients. Also, when sacroiliitis on radiographs was replaced by structural lesions on MRI, only minor changes in the ASAS axSpA classification of patients were seen. In patients with CBP suspicious for axSpA, structural lesions are associated with other SpA features and the majority of patients with structural lesions in our study fulfil the ASAS axSpA criteria anyway. If conventional radiographs were replaced by MRI-SI-s, using a combination of fatty lesions and erosions, only 3/294 (1.0%) patients would be additionally classified as ASAS axSpA. On the other hand, only 5/294 (1.7%) patients would lose their ASAS axSpA classification. Three out of these five patients with a state change had relatively low ratio products and corresponding probabilities of axSpA, suggesting that they did not have classic disease presentations. However, the two other patients have high disease probabilities and therefore could be missed erroneously. Most changes that are seen are a change in a subcategory between the various ASAS axSpA criteria rather than a change in classification per se. In other words, patients do not lose their ASAS classification solely by changing the content of the criteria. This characteristic adds to the credibility of the ASAS criteria. Only a few patients lose ASAS axSpA classification in scenario 2, which may justify replacement of conventional radiographs by MRI. However, it is important to realize that this is purely data-driven and feasibility issues should not be overlooked. MRI is an expensive imaging technique, especially in certain areas of the world. Furthermore, rheumatologists and radiologists are worldwide familiar with the mNY criteria, and evaluation of structural lesions on an MRI of the SI joints is a new concept. Education and time is needed in order to become familiarized with it. At this point in time, we therefore favour the addition of structural lesions to the imaging criterion of the ASAS axSpA criteria above replacement. These results are in line with recent data from the DESIR cohort, investigating the same research question in another cohort. In the SPACE cohort, even more patients do not change classification while replacing radiographic sacroiliitis by structural lesions seen on MRI compared with DESIR (SPACE: 95.3% vs DESIR: 80%). In general, a notable difference between the two cohorts is that in DESIR only patients with IBP are included, whereas in SPACE 62.6% of the patients have IBP. Other SpA features are also more common in the DESIR cohort, among which HLA-B27 positivity and radiographic sacroiliitis on conventional radiography and inflammation on MRI are highlighted. This is reflected in the fact that in the DESIR cohort, 71.8% fulfil the ASAS axSpA classification criteria, compared with 35.0% in the SPACE cohort. In DESIR all patients included are of French origin, whereas the SPACE cohort recruits patients from three European countries (The Netherlands, Italy, Norway). Though there might not be a big disparity between the prevalence of axSpA and CBP in general between these countries, it is important to observe these results in two populations of a different origin. Reliability of structural lesions remains a difficult issue and in a research setting often two or more well-calibrated readers are involved, which may cause difficulties translating this to clinical practice. So we should also be informed on the agreement between evaluation of MRI-SI-s in clinical practice before this can be advocated for use in a clinical setting. A limitation of the study is the absence of a gold standard to assess structural changes in the SI joint, by means of CT. Although the agreement regarding the presence/absence of radiographic sacroiliitis and the presence/absence of structural lesions on MRI in this study is only moderate, it is slightly better compared with the similar study in the DESIR cohort [1]. In general, it is reassuring to see consistent findings in two independent cohorts with different (though well-trained) reader pairs. This increases confidence regarding the generalizability. Although the research question has now been investigated in two cohorts, both cohorts include patients with short-standing back pain complaints and it would be very interesting to see replication of these findings in cohorts with advanced disease before possible far-reaching conclusions can be drawn on potentially changing ASAS axSpA classification criteria. The ideal cut-off could potentially be different in patients with longer symptom duration and more structural lesions in the SI joints. In general, more data are warranted on the prevalence of structural lesions in advanced disease, and it is beyond the scope of this study how lesions develop over time. Although the focus of this study is the impact of structural lesions on the ASAS axSpA classification criteria, we could speculate about possible implications for the diagnostic process. The modified Berlin algorithm advises that all patients suspected of axSpA should have a plain radiograph of the pelvis to check for sacroiliitis as a first step [13]. In patients without evidence of radiological sacroiliitis in whom axSpA still is suspected, an MRI of the SI joints (assessed for inflammation only) may support a diagnosis of non-radiographic axial spondyloarthritis when inflammation is present. Our data suggest that there is no solid indication to change the strategy of first asking a pelvic radiograph, since the classification remains very similar when replacing pelvic X-rays by MRI-s. However, in young patients MRI can be obtained as an alternative to plain radiography. This is in line with the recently published EULAR recommendations [14]. Similarly, if an MRI (STIR and T1 sequence) is present in a clinical setting, but there is no pelvic radiograph, this MRI may suffice and there is no reason to obtain radiographs. A strength of the study is the intensive scoring process by two well-calibrated readers with an adjudication process in place, which adds to the reliability of our findings. Another strength of this study is the SPACE cohort itself. The SPACE cohort is one of early disease, and comprises a control group of (chronic) back pain patients, just as in daily practice where a distinction between axSpA and no-axSpA should be made in every patient presenting with a suspicion of axSpA. In conclusion, our study has confirmed the earlier promising finding that the assessment of structural lesions on MRI instead of or in addition to conventional radiographs does not lead to a different ASAS axSpA classification in most of these patients with symptoms of an early disease onset. Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript. Disclosure statement: The authors have declared no conflicts of interest. References 1 Bakker PA , van den Berg R , Lenczner G et al. Can we use structural lesions seen on MRI of the sacroiliac joints reliably for the classification of patients according to the ASAS axial spondyloarthritis criteria? Data from the DESIR cohort . Ann Rheum Dis 2017 ; 76 : 392 – 8 . Google Scholar CrossRef Search ADS PubMed 2 van TA , Heuft-Dorenbosch L , Schulpen G et al. Radiographic assessment of sacroiliitis by radiologists and rheumatologists: does training improve quality? Ann Rheum Dis 2003 ; 62 : 519 – 25 . Google Scholar CrossRef Search ADS PubMed 3 van den Berg R , Lenczner G , Feydy A et al. Agreement between clinical practice and trained central reading in reading of sacroiliac joints on plain pelvic radiographs. Results from the DESIR cohort . Arthritis Rheumatol 2014 ; 66 : 2403 – 11 . Google Scholar CrossRef Search ADS PubMed 4 Rudwaleit M , van der Heijde D , Landewe R et al. The development of Assessment of SpondyloArthritis International Society classification criteria for axial spondyloarthritis (part II): validation and final selection . Ann Rheum Dis 2009 ; 68 : 777 – 83 . Google Scholar CrossRef Search ADS PubMed 5 Poddubnyy D , Gaydukova I , Hermann KG et al. Magnetic resonance imaging compared to conventional radiographs for detection of chronic structural changes in sacroiliac joints in axial spondyloarthritis . J Rheumatol 2013 ; 40 : 1557 – 65 . Google Scholar CrossRef Search ADS PubMed 6 van den Berg R , de HM , van GF et al. Percentage of patients with spondyloarthritis in patients referred because of chronic back pain and performance of classification criteria: experience from the Spondyloarthritis Caught Early (SPACE) cohort . Rheumatology (Oxford) 2013 ; 52 : 1492 – 9 . Google Scholar CrossRef Search ADS PubMed 7 Rudwaleit M , van der Heijde D , Landewe R et al. The Assessment of SpondyloArthritis International Society classification criteria for peripheral spondyloarthritis and for spondyloarthritis in general . Ann Rheum Dis 2011 ; 70 : 25 – 31 . Google Scholar CrossRef Search ADS PubMed 8 van der Linden S , Valkenburg HA , Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria . Arthritis Rheum 1984 ; 27 : 361 – 8 . Google Scholar CrossRef Search ADS PubMed 9 Rudwaleit M , Jurik AG , Hermann KG et al. Defining active sacroiliitis on magnetic resonance imaging (MRI) for classification of axial spondyloarthritis: a consensual approach by the ASAS/OMERACT MRI group . Ann Rheum Dis 2009 ; 68 : 1520 – 7 . Google Scholar CrossRef Search ADS PubMed 10 Weber U , Lambert RG , Østergaard M et al. The diagnostic utility of magnetic resonance imaging in spondylarthritis: an international multicenter evaluation of one hundred eighty-seven subjects . Arthritis Rheum 2010 ; 62 : 3048 – 58 . Google Scholar CrossRef Search ADS PubMed 11 de Hooge M , van den Berg R , Navarro-Compán V et al. Patients with chronic back pain of short duration from the SPACE cohort: which MRI structural lesions in the sacroiliac joints and inflammatory and structural lesions in the spine are most specific for axial spondyloarthritis? Ann Rheum Dis 2016 ; 75 : 1308 – 14 . Google Scholar CrossRef Search ADS PubMed 12 Rudwaleit M , Khan MA , Sieper J. The challenge of diagnosis and classification in early ankylosing spondylitis: do we need new criteria? Arthritis Rheum 2005 ; 52 : 1000 – 8 . Google Scholar CrossRef Search ADS PubMed 13 van den Berg R , de HM , Rudwaleit M et al. ASAS modification of the Berlin algorithm for diagnosing axial spondyloarthritis: results from the SPondyloArthritis Caught Early (SPACE)-cohort and from the Assessment of SpondyloArthritis international Society (ASAS)-cohort . Ann Rheum Dis 2013 ; 72 : 1646 – 53 . Google Scholar CrossRef Search ADS PubMed 14 Mandl P , Navarro-Compan V , Terslev L et al. EULAR recommendations for the use of imaging in the diagnosis and management of spondyloarthritis in clinical practice . Ann Rheum Dis 2015 ; 74 : 1327 – 39 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Rodríguez-Pintó, Ignasi;Espinosa, Gerard;Erkan, Doruk;Shoenfeld, Yehuda;Cervera, Ricard;Group, CAPS Registry Project

2018 Rheumatology

doi: 10.1093/rheumatology/key082pmid: 29660074

Abstract Objectives The objective of this study was to assess the effect that triple therapy (anticoagulation plus CS plus plasma exchange and/or IVIGs) has on the mortality risk of patients with catastrophic APS (CAPS) included in the CAPS Registry. Methods Patients from the CAPS Registry were grouped based on their treatments: triple therapy; drugs included in the triple therapy but in different combinations; and none of the treatments included in the triple therapy. The primary endpoint was all-cause mortality. Multivariate logistic regression models were used to compare mortality risk between groups. Results The CAPS Registry cohort included 525 episodes of CAPS accounting for 502 patients. After excluding 54 episodes (10.3%), a total of 471 patients with CAPS were included [mean (s.d.) age 38.5 years (17); 68.2% female; primary APS patients 62%]. Overall, 174 (36.9%) patients died. Triple therapy was prescribed in 189 episodes (40.1%), other combinations in 270 (57.3%) and none of those treatments in 12 episodes (2.5%); the mortality rate in the three groups was 28.6, 41.1 and 75%, respectively. Triple therapy was positively associated with a higher chance of survival when compared with non-treatment [adjusted odds ratio (OR) = 9.7, 95% CI: 2.3, 40.6] or treatment with other combinations of drugs included in the triple therapy (adjusted OR = 1.7, 95% CI: 1.2, 2.6). No statistical differences were found between patients that received triple therapy with plasma exchange or IVIGs (P = 0.92). Conclusion Triple therapy is independently associated with a higher survival rate among patients with CAPS. catastrophic antiphospholipid syndrome, antiphospholipid syndrome, mortality, survival rate, treatment, systemic lupus erythematosus, triple therapy, plasma exchange, immunoglobulins Rheumatology key messages Triple therapy is associated with a higher survival rate in catastrophic APS. Adminsitration of plasma exchange or immunoglobulins with triple therapy showed no difference in catastrophic APS. Introduction The combination of anticoagulation plus CS plus plasma exchange and/or IVIG, the so-called triple therapy, has been claimed for a long time as the best treatment for catastrophic APS (CAPS) [1, 2]. Despite the fact that strong biological reasons have been proposed to support the beneficial effect of triple therapy in patients with this devastating condition [3], the rarity of its presentation has precluded conducting clinical trials. Moreover, the analysis of epidemiologic data has provided contradictory results [2, 4]. To date, the best available evidence comes from a study that showed a better prognosis of those patients diagnosed after 2001, when a consensus agreement proposing the use of triple therapy to treat patients with CAPS was published [5]. To evaluate the clinical benefit of triple therapy in the treatment of CAPS, we conducted a retrospective cohort study comparing the mortality rate among patients who received triple therapy treatment (i.e. anticoagulation, CS and plasma exchange and/or IVIG) to those who were treated with other therapeutic combination. Methods Patients The study is based on the data of patients included in the CAPS Registry. This is a database that contains data for patients with CAPS from April 1992 (when CAPS was first described) to December 2015. Patients included in the CAPS Registry had either been reported to the CAPS Registry coordinators or their case had been published in the literature. Data are standardized by anonymized data forms fulfilled by their treating physicians. Received cases are checked by the CAPS Registry data manager with the database and eventually with the sending physician to rule out duplicates. Published case reports are identified through a periodic systematic review of all cases available in MEDLINE in any language using the following MeSH terms: CAPS, or catastrophic APS. Additionally, references were reviewed to identify further cases. All patients had to meet current accepted classification criteria of CAPS [1, 6]. The study was designed by the authors and approved by the clinic investigation ethics committee of the Hospital Clinic of Barcelona, which waived the requirement for individual informed consent. Demographic, clinical and immunological features of each patients with CAPS were registered. CAPS episodes were classified in three groups according to the treatment received: those that received the so-called triple therapy (combination of anticoagulation plus CS plus plasma exchange and/or IVIG); those that received any other combination of these treatments; and those that received none of these treatments. The study endpoint was survival as reported by their treating physician. Statistical analysis Descriptive analysis of demographical, clinical and immunological variables of the cases was performed. Numerical variables are described with mean (s.d.). For categorical variables, sample size and percentage are presented. Survival was the dependent variable in all analysis. A single patient could have suffered multiple episodes of CAPS. However, only first episodes of CAPS were included in the analysis and, thus, the endpoint occurred only once. Therefore, the assumption of statistical independence is not violated. The odds ratio (OR) of death between the groups, defined by the treatment used, was calculated by logistic regression models including adjustment variables estimated to be possible confounders in a stepwise method. The threshold probability for entering variables into the model was <0.05. The removal threshold was P > 0.10. Results are presented in terms of OR and 95% CI. An extensive set of clinical variables was evaluated to rule out possible confusion. This set of variables evaluated included: age, year of diagnosis, sex, SLE, presence and type of precipitating factors, clinical manifestations and number of organs affected, type of aPL detected, ANA and anti-ENA, platelet count, haemolytic features, presence of schistocytes and treatment with CYC. The existence of a modifier was ruled out by their statistical significance interaction with the treatment. Possible confounders were primary selected when they were statistically associated with the treatment and with survival in those that did not received the treatment. All variables with an eta value higher than 0.1 were included to build up the logistic regression model when treatment was included as a nominal variable [7, 8]. When the treatment was evaluated as a binominal variable, variables were selected when they showed an OR <0.67 or >1.5 for categorical variables or when they had an Spearman’s rank correlation coefficient higher than 0.1 for continuous variables [7, 8]. Final models were build up through logistic regression models by an stepwise method [9]. All analyses were performed using SPSS version 22 with R extension when needed. Results Characteristics of the study cohort The CAPS Registry cohort includes 525 episodes of CAPS accounting for 502 patients. Five hundred and two (95.6%) patients had one episode, 18 (3.4%) patients had two episodes and 5 (1.0%) patients had three episodes. Thus, 23 episodes were excluded because they were recurrences. Additionally, 31 (6.1%) episodes were excluded due to missing information regarding the outcome or the treatment prescribed. Therefore, 471 episodes of CAPS were included in the study. Patients were mainly female (68.2%) and had a mean (s.d.) age of 38.5 years (17.0). Most CAPS episodes developed in patients with primary APS (62%). Treatments Anticoagulation (usually intravenous heparin followed by oral anticoagulation) was the treatment prescribed most often, accounting for 403 out of 471 (85.6%) patients, followed by CS, used in 372 out of 471 (79.0%) patients, plasma exchange in 166 out of 471 (35.2%) and IVIGs in 126 patients of CAPS out of 471 (26.8%). Triple therapy was prescribed in 189 (40.1%) CAPS patients; other combinations of treatments included in the triple therapy were prescribed in 270 (57.3%) patients, and none of these treatments in 12 patients (2.5%) (detailed treatments prescribed in Table 1). Table 1 Treatment combinations in the patients included in the CAPS Registry Treatment CAPS episodes, n (%) N = 471 Individual treatments     AC 403 (85.6)     CS 372 (79.0)     PE 166 (35.2)     IVIG 126 (26.8) Treatment combinations     Triple therapy AC + CS + PE and/or IVIG 189 (40.1) Any other combination 270 (57.3)     AC + CS 127 (27.0)     AC alone 67 (14.2)     CS alone 26 (5.5)     CS + PE 19 (4.0)     AC + PE 11 (2.3)     CS + IVIG 7 (1.5)     AC + IVIG + PE 5 (1.1)     CS + IVIG + PE 4 (0.8)     AC + IVIG 4 (0.8) None of these treatments 12 (2.5) Treatment CAPS episodes, n (%) N = 471 Individual treatments     AC 403 (85.6)     CS 372 (79.0)     PE 166 (35.2)     IVIG 126 (26.8) Treatment combinations     Triple therapy AC + CS + PE and/or IVIG 189 (40.1) Any other combination 270 (57.3)     AC + CS 127 (27.0)     AC alone 67 (14.2)     CS alone 26 (5.5)     CS + PE 19 (4.0)     AC + PE 11 (2.3)     CS + IVIG 7 (1.5)     AC + IVIG + PE 5 (1.1)     CS + IVIG + PE 4 (0.8)     AC + IVIG 4 (0.8) None of these treatments 12 (2.5) Results presented as percentages from the total number of episodes. AC: Anticoagulants, CAPS: catastrophic antiphospholipid syndrome, PE: Plasma exchange. Table 1 Treatment combinations in the patients included in the CAPS Registry Treatment CAPS episodes, n (%) N = 471 Individual treatments     AC 403 (85.6)     CS 372 (79.0)     PE 166 (35.2)     IVIG 126 (26.8) Treatment combinations     Triple therapy AC + CS + PE and/or IVIG 189 (40.1) Any other combination 270 (57.3)     AC + CS 127 (27.0)     AC alone 67 (14.2)     CS alone 26 (5.5)     CS + PE 19 (4.0)     AC + PE 11 (2.3)     CS + IVIG 7 (1.5)     AC + IVIG + PE 5 (1.1)     CS + IVIG + PE 4 (0.8)     AC + IVIG 4 (0.8) None of these treatments 12 (2.5) Treatment CAPS episodes, n (%) N = 471 Individual treatments     AC 403 (85.6)     CS 372 (79.0)     PE 166 (35.2)     IVIG 126 (26.8) Treatment combinations     Triple therapy AC + CS + PE and/or IVIG 189 (40.1) Any other combination 270 (57.3)     AC + CS 127 (27.0)     AC alone 67 (14.2)     CS alone 26 (5.5)     CS + PE 19 (4.0)     AC + PE 11 (2.3)     CS + IVIG 7 (1.5)     AC + IVIG + PE 5 (1.1)     CS + IVIG + PE 4 (0.8)     AC + IVIG 4 (0.8) None of these treatments 12 (2.5) Results presented as percentages from the total number of episodes. AC: Anticoagulants, CAPS: catastrophic antiphospholipid syndrome, PE: Plasma exchange. A comparison of clinical characteristics and laboratory features of the three different treatment groups is presented in Table 2. Groups were generally comparable since no important differences were found between them. Patients that received triple therapy were younger than patients that received other combinations (36 vs 40 years; P = 0.046), while no difference was found in sex distribution or in their association with SLE. Additional differences were observed between groups regarding the association of a malignancy as a trigger because they were more often seen in patients that received other combinations than those that received triple therapy or none of these treatments (14.9 vs 5.1 vs 8.3%; P = 0.006). Additionally, haemolysis features (47.6 vs 30.3 vs 0%; P < 0.001), thrombocytopaenia (76.8 vs 59.5 vs 50%; P = 0.001) and schistocytes (30.8 vs 12.1 vs 0%; P < 0.001) were more often seen during the catastrophic event in those patients that received triple therapy than in the other groups. Triple positivity of aPL was observed in 26.2% among those treated with triple therapy, 18.2% among those treated with other combinations and 0% among those treated with other combinations (Table 2). Additionally, those that received triple therapy were more often treated with CYC (34.6 vs 19.3 vs 0%; P < 0.01). All these statistically significant differences between groups were considered to build up multivariate logistic regression models. Table 2 Clinical and laboratory features of CAPS episodes according to the treatment receiveda Clinical/laboratory feature N Triple therapy, n = 197 Other combinations, n = 278 None, n = 12 Age, mean (s.d.), years 460 36.2 (15.8) 40.0 (17.0) 29.5 (16.5) Sex, female 466 129 (68.6) 180 (67.7) 11 (91.7) SLE 447 53 (29.3) 85 (33.2) 3 (30.0) Identified precipitating factor 442 117 (66.5) 160 (63.0) 9 (75.0) Infection 443 60 (34.1) 70 (27.5) 5 (41.7) Malignancy 442 9 (5.1) 38 (14.9) 1 (8.3) Surgery 441 19 (10.9) 29 (11.4) 2 (16.7) Organ involvement     Renal involvement 471 147 (77.8) 188 (69.6) 10 (83.3)     Neurologic involvement 469 114 (60.6) 147 (54.6) 6 (50.0)     Lung involvement 471 114 (60.3) 165 (61.1) 9 (75.0)     Cardiac involvement 469 100 (53.5) 130 (48.1) 8 (66.7)     Hepatic involvement 470 72 (38.3) 105 (38.9) 2 (16.7)     Peripheral vessels 466 62 (33.2) 104 (39.0) 4 (33.3)     GI involvement 469 49 (26.2) 57 (21.1) 3 (25.0)     Number of organs affected (>3) 456 164 (91.1) 235 (89.0) 11 (91.7) Laboratory features     Thrombocytopaenia 426 136 (76.8) 141 (59.5) 6 (50.0)     Haemolysis 355 70 (47.6) 60 (30.3) 0 (0.0)     Schistocytes 287 40 (30.8) 18 (12.1) 0 (0.0) Immunological features     ANAs 368 94 (59.9) 112 (54.9) 4 (57.1)     aPL     LA 405 139 (84.8) 187 (81.0) 8 (80.0)     IgG anticardiolipin 432 144 (83.2) 202 (81.5) 7 (63.6)     IgM anticardiolipin 358 65 (48.1) 108 (50.5) 3 (33.3)     Triple positivity 168 17 (26.2) 18 (18.2) 0 (0.0) Clinical/laboratory feature N Triple therapy, n = 197 Other combinations, n = 278 None, n = 12 Age, mean (s.d.), years 460 36.2 (15.8) 40.0 (17.0) 29.5 (16.5) Sex, female 466 129 (68.6) 180 (67.7) 11 (91.7) SLE 447 53 (29.3) 85 (33.2) 3 (30.0) Identified precipitating factor 442 117 (66.5) 160 (63.0) 9 (75.0) Infection 443 60 (34.1) 70 (27.5) 5 (41.7) Malignancy 442 9 (5.1) 38 (14.9) 1 (8.3) Surgery 441 19 (10.9) 29 (11.4) 2 (16.7) Organ involvement     Renal involvement 471 147 (77.8) 188 (69.6) 10 (83.3)     Neurologic involvement 469 114 (60.6) 147 (54.6) 6 (50.0)     Lung involvement 471 114 (60.3) 165 (61.1) 9 (75.0)     Cardiac involvement 469 100 (53.5) 130 (48.1) 8 (66.7)     Hepatic involvement 470 72 (38.3) 105 (38.9) 2 (16.7)     Peripheral vessels 466 62 (33.2) 104 (39.0) 4 (33.3)     GI involvement 469 49 (26.2) 57 (21.1) 3 (25.0)     Number of organs affected (>3) 456 164 (91.1) 235 (89.0) 11 (91.7) Laboratory features     Thrombocytopaenia 426 136 (76.8) 141 (59.5) 6 (50.0)     Haemolysis 355 70 (47.6) 60 (30.3) 0 (0.0)     Schistocytes 287 40 (30.8) 18 (12.1) 0 (0.0) Immunological features     ANAs 368 94 (59.9) 112 (54.9) 4 (57.1)     aPL     LA 405 139 (84.8) 187 (81.0) 8 (80.0)     IgG anticardiolipin 432 144 (83.2) 202 (81.5) 7 (63.6)     IgM anticardiolipin 358 65 (48.1) 108 (50.5) 3 (33.3)     Triple positivity 168 17 (26.2) 18 (18.2) 0 (0.0) a Percentages measured over the total of episodes with valid information. Data are presented as n (%) unless otherwise indicated. CAPS: catastrophic APS; GI: gastrointestinal. Table 2 Clinical and laboratory features of CAPS episodes according to the treatment receiveda Clinical/laboratory feature N Triple therapy, n = 197 Other combinations, n = 278 None, n = 12 Age, mean (s.d.), years 460 36.2 (15.8) 40.0 (17.0) 29.5 (16.5) Sex, female 466 129 (68.6) 180 (67.7) 11 (91.7) SLE 447 53 (29.3) 85 (33.2) 3 (30.0) Identified precipitating factor 442 117 (66.5) 160 (63.0) 9 (75.0) Infection 443 60 (34.1) 70 (27.5) 5 (41.7) Malignancy 442 9 (5.1) 38 (14.9) 1 (8.3) Surgery 441 19 (10.9) 29 (11.4) 2 (16.7) Organ involvement     Renal involvement 471 147 (77.8) 188 (69.6) 10 (83.3)     Neurologic involvement 469 114 (60.6) 147 (54.6) 6 (50.0)     Lung involvement 471 114 (60.3) 165 (61.1) 9 (75.0)     Cardiac involvement 469 100 (53.5) 130 (48.1) 8 (66.7)     Hepatic involvement 470 72 (38.3) 105 (38.9) 2 (16.7)     Peripheral vessels 466 62 (33.2) 104 (39.0) 4 (33.3)     GI involvement 469 49 (26.2) 57 (21.1) 3 (25.0)     Number of organs affected (>3) 456 164 (91.1) 235 (89.0) 11 (91.7) Laboratory features     Thrombocytopaenia 426 136 (76.8) 141 (59.5) 6 (50.0)     Haemolysis 355 70 (47.6) 60 (30.3) 0 (0.0)     Schistocytes 287 40 (30.8) 18 (12.1) 0 (0.0) Immunological features     ANAs 368 94 (59.9) 112 (54.9) 4 (57.1)     aPL     LA 405 139 (84.8) 187 (81.0) 8 (80.0)     IgG anticardiolipin 432 144 (83.2) 202 (81.5) 7 (63.6)     IgM anticardiolipin 358 65 (48.1) 108 (50.5) 3 (33.3)     Triple positivity 168 17 (26.2) 18 (18.2) 0 (0.0) Clinical/laboratory feature N Triple therapy, n = 197 Other combinations, n = 278 None, n = 12 Age, mean (s.d.), years 460 36.2 (15.8) 40.0 (17.0) 29.5 (16.5) Sex, female 466 129 (68.6) 180 (67.7) 11 (91.7) SLE 447 53 (29.3) 85 (33.2) 3 (30.0) Identified precipitating factor 442 117 (66.5) 160 (63.0) 9 (75.0) Infection 443 60 (34.1) 70 (27.5) 5 (41.7) Malignancy 442 9 (5.1) 38 (14.9) 1 (8.3) Surgery 441 19 (10.9) 29 (11.4) 2 (16.7) Organ involvement     Renal involvement 471 147 (77.8) 188 (69.6) 10 (83.3)     Neurologic involvement 469 114 (60.6) 147 (54.6) 6 (50.0)     Lung involvement 471 114 (60.3) 165 (61.1) 9 (75.0)     Cardiac involvement 469 100 (53.5) 130 (48.1) 8 (66.7)     Hepatic involvement 470 72 (38.3) 105 (38.9) 2 (16.7)     Peripheral vessels 466 62 (33.2) 104 (39.0) 4 (33.3)     GI involvement 469 49 (26.2) 57 (21.1) 3 (25.0)     Number of organs affected (>3) 456 164 (91.1) 235 (89.0) 11 (91.7) Laboratory features     Thrombocytopaenia 426 136 (76.8) 141 (59.5) 6 (50.0)     Haemolysis 355 70 (47.6) 60 (30.3) 0 (0.0)     Schistocytes 287 40 (30.8) 18 (12.1) 0 (0.0) Immunological features     ANAs 368 94 (59.9) 112 (54.9) 4 (57.1)     aPL     LA 405 139 (84.8) 187 (81.0) 8 (80.0)     IgG anticardiolipin 432 144 (83.2) 202 (81.5) 7 (63.6)     IgM anticardiolipin 358 65 (48.1) 108 (50.5) 3 (33.3)     Triple positivity 168 17 (26.2) 18 (18.2) 0 (0.0) a Percentages measured over the total of episodes with valid information. Data are presented as n (%) unless otherwise indicated. CAPS: catastrophic APS; GI: gastrointestinal. Mortality Overall mortality in CAPS accounted for 36.9% (174/471) of cases. Mortality rate was 28.6% (54/189) in patients who received triple therapy, 41.1% (111/270) in those who were treated with any other combination and 75% (9/12) in those who did not received any of these treatments. Regarding the clinical characteristics of those patients who died with respect to those who survived, CAPS patients that passed away were older than those that did not (42 vs 36 years; P < 0.001). Additionally, CAPS was more often triggered by a malignancy in those that deceased (17.8 vs 6.6%; P < 0.001) and they had more often lung (76.4 vs 52.2%; P < 0.001), cardiac (56.9 vs 47.1%; P = 0.041), neurologic (66.1 vs 51.5%; P = 0.002) or renal (81.6 vs 68.4%; P = 0.002) involvement, and had higher number of organs affected (4.6 vs 3.8; P < 0.001), while those that survived more often had thrombocytopaenia (70.2% vs 60.4%; P = 0.036). Multivariate logistic regression models showed treatment to be linked to the survival of patients with CAPS. This analysis showed the group of triple therapy to be independently statistically associated with a higher survival rate (Table 3). Table 3 Regression models on the effect of treatment on the survival of CAPS patients Clinical variable P-value OR (95% CI) Treatment     Triple therapy 0.069 3.49 (0.9, 13.4)     Others 0.009 6.2 (1.6, 24.2)     None 0.004 Renal involvement <0.001 0.4 (0.3, 0.6) Thrombocytopaenia 0.1 1.5 (0.9, 2.3) Malignancy 0.006 0.4 (0.2, 0.8) Triple therapy vs none 0.002 9.7 (2.3, 40.6) Cardiac involvement 0.072 0.53 (0.27, 1.6) Age <0.001 0.96 (0.94, 0.98) Peripheral thrombosis 0.076 0.53 (0.27, 1.1) Triple therapy vs other 0.006 1.7 (1.2, 2.6) Clinical variable P-value OR (95% CI) Treatment     Triple therapy 0.069 3.49 (0.9, 13.4)     Others 0.009 6.2 (1.6, 24.2)     None 0.004 Renal involvement <0.001 0.4 (0.3, 0.6) Thrombocytopaenia 0.1 1.5 (0.9, 2.3) Malignancy 0.006 0.4 (0.2, 0.8) Triple therapy vs none 0.002 9.7 (2.3, 40.6) Cardiac involvement 0.072 0.53 (0.27, 1.6) Age <0.001 0.96 (0.94, 0.98) Peripheral thrombosis 0.076 0.53 (0.27, 1.1) Triple therapy vs other 0.006 1.7 (1.2, 2.6) CAPS: catastrophic APS; OR: odds ratio. Table 3 Regression models on the effect of treatment on the survival of CAPS patients Clinical variable P-value OR (95% CI) Treatment     Triple therapy 0.069 3.49 (0.9, 13.4)     Others 0.009 6.2 (1.6, 24.2)     None 0.004 Renal involvement <0.001 0.4 (0.3, 0.6) Thrombocytopaenia 0.1 1.5 (0.9, 2.3) Malignancy 0.006 0.4 (0.2, 0.8) Triple therapy vs none 0.002 9.7 (2.3, 40.6) Cardiac involvement 0.072 0.53 (0.27, 1.6) Age <0.001 0.96 (0.94, 0.98) Peripheral thrombosis 0.076 0.53 (0.27, 1.1) Triple therapy vs other 0.006 1.7 (1.2, 2.6) Clinical variable P-value OR (95% CI) Treatment     Triple therapy 0.069 3.49 (0.9, 13.4)     Others 0.009 6.2 (1.6, 24.2)     None 0.004 Renal involvement <0.001 0.4 (0.3, 0.6) Thrombocytopaenia 0.1 1.5 (0.9, 2.3) Malignancy 0.006 0.4 (0.2, 0.8) Triple therapy vs none 0.002 9.7 (2.3, 40.6) Cardiac involvement 0.072 0.53 (0.27, 1.6) Age <0.001 0.96 (0.94, 0.98) Peripheral thrombosis 0.076 0.53 (0.27, 1.1) Triple therapy vs other 0.006 1.7 (1.2, 2.6) CAPS: catastrophic APS; OR: odds ratio. When we compared the group that received triple therapy with the one that received none of these treatments, triple therapy was found to be associated with a higher survival rate (OR = 9.7, 95% CI: 2.3, 40.6). When a comparison was performed between the group that received the triple therapy and that which received only some of these treatments, the multivariate logistic regression model showed that those patients that received triple therapy had an increased survival chance (adjusted OR = 1.7, 95% CI: 1.2, 2.6). Thus, triple therapy led to a 46.4% (95% CI: 21.1, 71.8%) reduction of the mortality rate compared with no therapy and a 12.5% (95% CI: 3.8, 21.3%) reduction compared with any other combination. Therefore, survival could be attributed to triple therapy in 17.6% (95% CI: 5.7, 27.9%) of cases among those who received triple therapy. SLE and primary APS group subanalysis One hundred and forty-one patients developed CAPS associated with SLE. Overall mortality accounted for 47.5% in SLE patients and 32.1% in patients with CAPS associated with primary APS. Among the group with SLE, 54 patients (37.8%) were treated with triple therapy, 86 patients (60.1%) received other combination while only 3 (2.1%) patients were not treated with any of these treatments. Statistical analysis showed differences in the survival rate between treatment groups in patients with SLE (47.4 vs 52.6 vs 0%; P < 0.001). When patients with SLE that received triple therapy were compared with those that received other combinations the OR was 2.2 (95% CI: 1.1, 4.8; P = 0.014). Although statistical analysis showed differences (P = 0.045) between the triple therapy group survival rate (66.7%) and that in those that did not receive any treatment (0%), it was not possible to calculate the survival OR because none of the three SLE patients that received none of these treatments survived. Overall, 277 patients developed CAPS associated with primary APS, 118 of them (42.6%) were treated with triple therapy, 153 (55.2%) were treated with other combinations and 6 (2.2%) patients with none of these treatments. Multivariable logistic regression models showed an OR of 1.5 (95% CI: 0.8, 2.7) when triple therapy was compared with other combinations and an OR of 3.1 (95% CI: 0.6, 16.0) when triple therapy was compared with the group of patients that did not receive any of these treatments. IVIGs, plasma exchange or both Among 189 patients that received triple therapy, 83 (43.9%) received anticoagulation, CS and plasma exchange, 62 (32.8%) received anticoagulation, CS and IVIG, and 44 (23.3%) received anticoagulation, CS, plasma exchange and IVIG. The mortality of patients that received triple therapy with IVIG and plasma exchange was 27.3% while it was 29% among those that received only triple therapy with plasma exchange or IVIGs (27.4% among those that received IVIG without plasma exchange, and 30.1% among those that received plasma exchange without IVIG). The multivariate logistic regression model did not find any differences in the survival rate between the three groups of patients (P = 0.913). No differences were found between those that received both treatments and those that received only one of them (OR = 1.1, 95% CI: 0.5, 2.3; P = 0.83). Haemolysis, thrombocytopaenia and schistocytes subgropus analysis Among the subgroup of patients that had thrombocytopaenia and the subgroup that had haemolysis, triple therapy continued to show a clear benefit over other combinations [OR (95% CI) = 1.7 (1.1, 2.9) and 2.4 (1.1, 4.9), respectively]. However, it was not possible to show this benefit in the subgroup of patients with schistocytes (OR = 1.9, 95% CI: 0.6, 6.3). Plasma exchange did not seem to add a value when face to face comparison was undertaken in the subgroups with thrombocytopaenia, haemolysis and schistocytes [OR (95% CI) = 1.1 (0.7, 1.8), 1.1 (0.5, 2.2) and 0.8 (0.2, 2.5), respectively]. These findings did not differ when multivariable regression analysis was performed adjusting by possible confounders. Discussion In the present study triple therapy was associated with a higher survival rate in patients with CAPS. The mortality rate of patients who received triple therapy was 28% whereas in patients that did not receive any treatment it was 75%. In 1998, Asherson et al. [2] proposed the combination of anticoagulation, CS and plasma exchange or IVIG as the best therapeutic approach for CAPS based on the hypothetical pathogenic mechanisms of the syndrome. Current knowledge states that CAPS is caused by a thrombotic storm that leads to a cytokine storm [10]. Based on this pathological model, anticoagulation would stop the thrombotic storm while CS would hamper cytokine excess, and plasma exchange and IVIG remove from the plasma circulating proinflammatory cytokines and aPLs [11]. The mortality rate found in this study represents an important reduction compared with previous reports [2, 4, 5]. Since the first series of Asherson et al. [2, 4], the mortality linked to CAPS was up to 50%. This high mortality rate was confirmed in succeeding series [5]. In a more recent study, the mortality rate decreased over time from 53% in patients diagnosed before 2001 to 33% in those diagnosed between 2001 and 2005 [5]. The most important and clinically relevant difference between the two periods was the higher use of the combination of anticoagulation plus CS plus plasma exchange and/or IVIG in the second period. Therefore, the authors attributed the decrease in CAPS mortality ratio to this increasing use of the so-called ‘triple therapy’ after 2001 [5]. However, this hypothesis could not be proved at that point. The current study shows a statistical association between the triple therapy and the decrease of mortality rate of patients with CAPS when compared with other combinations or with none of the treatments included in the triple therapy. Interestingly, no differences in the mortality rate were found between the three triple therapy groups: patients that received plasma exchange and IVIG, patients that received IVIG without plasma exchange and those who received plasma exchange without IVIG. This finding has an important practical implication since the plasma exchange procedure is not available in all centres, whereas IVIGs are available in many hospitals and can be administered without delay. Further, plasma exchange might be difficult to perform in paediatric or pregnant patients, while administration of IVIG is easier and safer [12]. Interestingly, it seems that there is no need to administer IVIG and plasma exchange together since it seems not to provide benefit over one of these treatments without the other. This is also true for the subgroup of patients with thrombocytopaenia, haemolysis or schistocytes. The strength of our study comes from the large sample size and the possibility to adjust for several confounders. Further, the multicentre nature of the study and its hard endpoint, that is, death, increases the external validity of our findings. There are limitations in the current study since data come from published reports and cases reported to the CAPS Registry coordinators. Therefore, some information might be incomplete or inaccurate, and results might be affected by publication bias. In this sense, is was not possible to analyse the timing of treatements or the dose used since these data are not available in the cases reported and we do not know their impact on patients’ outcomes. Moreover, due to the observational nature of our study, we cannot exclude possible residual confounding. However, although an observational study is not the ideal approach to evaluate the effect of treatment following evidence-based medicine, the rarity of CAPS precludes driving any prospective trial in these patients. In conclusion, the triple therapy is associated with a reduction of the risk of death among patients with CAPS of ∼46%. Further, no differences have been found between the use of plasma exchange and/or IVIG in patients with CAPS. Acknowledgements The Catastrophic Antiphospholipid Syndrome Registry Project Group from the European Forum on Antiphospholipid Antibodies as follows. Coordinators: R. Cervera, G. Espinosa, I. Rodríguez-Pintó, Department of Autoimmune Diseases, Hospital Clinic, Barcelona, Catalonia, Spain; Y. Shoenfeld, Sheba Medical Center, Tel-Hashomer, Israel; D. Erkan, Barbara Volcker Center for Women and Rheumatic Disease, Hospital for Special Surgery, New York, NY, USA. Collaborators: J.C. Piette, Hôpital Pitié-Salpêtrière, Paris, France; M. Jacek, Jagiellonian University School of Medicine, Krakow, Poland; B. Roca, Department of Internal Medicine, Hospital General de Castelló, Castelló, Spain; M. Tektonidou and H. Moutsopoulos, Department of Pathophysiology, Medical School, National University of Athens, Athens, Greece; J. Boffa, Department of Nephrology, Hôpital Tenon, Paris, France; J. Chapman, Neuroimmunology Service, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel; L. Stojanovich, Clinical-Hospital Center “Bezhanijska Kosa”, Belgrade, Montenegro; M.P. Veloso, Servicio de Reumatología, Hospital Universitario Clementino Fraga Filho, Rio de Janeiro, Brazil; S. Praprotnik, Department of Rheumatology, University Clinical Center, Ljubljana, Slovenia; B. Traub and R. Levy, Department of Rheumatology, Rio de Janeiro State University, Rio de Janeiro, Brazil; T. Daryl, Department of Hematology, Singapore General Hospital, Singapore; Daryl Tan, Department of Hematology, Singapore General Hospital, Singapore; M.C. Boffa, Hôpital Pitié-Salpêtrière, Paris, France; A. Makatsaria, Department of Obstetrics and Gynecology, Moscow Medical Academy, Moscow, Russia; M. Ruano, Hospital Clínico Universitario de Valencia, Valencia, Spain; A. Allievi, Department of Internal Medicine and Autoimmune Diseases Unit, Hospital Fernandez, Buenos Aires, Argentina; W. You, Department Obstetrics and Gynecology, National Naval Medical Center. Bethesda, MD, USA; M. Khamastha, The Lupus Research Unit, St Thomas’ Hospital, London, UK; S. Hughes, Liberato Nilzete, Intensive care unit and Nephrology unit, Children’s Hospital Joana de Gusmão, Florianopolis, Brazil; J. Menendez Suso, Pediatric Intensive Care Unit, Hospital Infantil La Paz, Madrid, Spain; J. Pacheco, Servicio de Reumatología, Hospital de Clínicas ‘Jose de San Martin’, Buenos Aires, Argentina; M.F. Boriotti, Servicio de Hematología, Hospital Provincial de Rosario, Rosario, Argentina; C. Dias, Autoimmune Disease Unit, Department of Medicine, S. Joao Hospital, Porto, Portugal; G. Pangtey, Department of Medicine, All Institute of Medicine, New Delhi, India; S. Miller, Department of Internal Medicine, McMaster University, Hamilton, Ontario, Canada; S. Policepatil, Internal Medicine, Gundersan Lutheran Hospital, La Crosse, WI, USA; L. Larissa, Rheumatology Division, Queen’s University, Kingston, Ontario, Canada; S. Marjatta, Department of Hematology, Tampere University Hospital, Tampere, Finland; S. Carolyn, Department of Obstetrics and Gynecology, University of Utah, City, UT, USA; T. Noortje, Department of Hematology, University Medical Center De Boelelaan, Amsterdam, The Netherlands; K. Reiner, Consultant Rheumatologist, Russells Hall Hospital, Dudley, UK; S. Arteaga, Medicina Interna, Universidad de Antioquia, Medellin, Colombia; T. Leilani, Dermatology Department, St Joseph Mercy, Ypsilanti, MI, USA; D. Langsford, Department of Nephrology, Royal Hobart Hospital, Hobart, Tasmania, Australia; M. Niedzwiecki, Department of Pediatrics Hematology, Oncology and Endocrinology, Medical University of Gdansk, Gdansk, Poland; V. Queyrel, Internal Medicine Department, Centre Hospitalier Univeritaire, Nice, France; R. Moroti-Constantinescu, Infectious Disease Department, Matei Bais National Institute for Infectious Diseases, Bucharest, Romania; C. Romero, Medicina Interna, Hospital Costa del Sol, Marbella, Spain; K. Jeremic, Department of Perinatology, Institute of Gynecology and Obstetrics, Clinical Center, Belgrade, Serbia and Montenegro; A. Urbano, Servicio de Hematologia, Hospital Virgen del Rocio, Sevilla, Spain; R. Hurtado-García, Internal Medicine Service, Hospital General Universitario de Elche, Elche, Spain; A. Kumar Das, Pontefract General Infirmary, Pontefract, UK; N. Costedoat-Chalumeau, Service de Médecine Interne II, Hopital Pitié-Salpêtrière, Paris, France; F. Yngvar, Department of Hematology, Oslo University Hospital, Oslo, Norway; J.A. Gomez-Puerta, Division of Rheumatology, Brigham and Women’s Hospital, Boston, MA, USA; E. de Meigs, National Cancer Institute RJ, Rio de Janeiro, Brazil; J.P. Smith, Pediatric Rheumatology Registrar, Bristol Royal Hospital for Children, Bristol, UK; E. Zakharova, Moscow City Clinical Hospital, Moscow, Russia; A. Nayer, Division of Nephrology and Hypertension, University of Miami Clinical Research Building, Miami, FL, USA; W. Douglas, Gundersen Lutheran Medical Center Clinic, Onalaska, WI, USA; R. Lyndsey, Department of Medicine, University of Wisconsin, Madison, WI, USA; V. Blanco, CAAMEPA, Pando, Uruguay; C. Vicent, Unidad de Cuidados Intensivos, Hospital Lluís Alcanyis de Xátiva, Valencia, Spain; K. Natalya, Clinic of Internal Medicine and Nephrology, Moscow Medical Sechenov University, Moscow, Russia; L. Damian, Emergency Department, County Hospital Cluj, Cluj-Napoca, Romania; E. Valentini, Departamento de Medicina Interna, Sanatorio de La Mujer, Rosario, Argentina; B. Giula, Unit of Nephrology and Dialysis, Ospedale Civico di Lugano, Lugano, Switzerland; M. Casal Moura, Department of Internal Medicine, Centro Hospitalar Sao Joao, Porto, Portugal; O. Araújo Loperena, Internal Medicine Department, Hospital de Sant Pau i Santa Tecla, Tarragona, Spain; Y. Ritter Susan, Department of Rheumatology, Brigham and Women’s Hospital, Boston, MA, USA; G. Guettrot Imbert, Department of Internal Medicine, University Hospital of Clermont Ferrand, Clermont-Ferrand, France; H. Almasri, Hospitalist, Union Hospital, Terre Haute, IN, USA; T. Hospach, Pediatric Rheumatology, Olgahospital Stuttgart, Stuttgart, Germany; B. Mouna, Laboratory of Immunology, Military Hospital of Tunis, Tunis, Tunisia; A. Robles, Servicio de Medicina Interna, Hospital de la Paz, Madrid, Spain; H. Wilson, Consultant Rheumatologist, Glasgow Royal Infirmary, Glasgow, UK; P. Guisado, Internal Medicine Department, Hospital Quiron San Camilo, Madrid, Spain; R. Ruiz, Internal Medicine, Colegio Médico de Bolivia, Zona Central, La Paz, Bolivia; J. Rodriguez, Complejo Hospitalario Universitario de Albacete, Albacete, Spain. Funding: No specific funding was received from anybody in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript. Disclosure statement: The authors have declared no conflicts of interest. References 1 Asherson R , Cervera R , de Groot PG et al. Catastrophic antiphospholipid syndrome: international consensus statement on classification criteria and treatment guidelines . Lupus 2003 ; 12 : 530 – 4 . Google Scholar CrossRef Search ADS PubMed 2 Asherson RA , Cervera R , Piette JC et al. Catastrophic antiphospholipid syndrome. Clinical and laboratory features of 50 patients . Medicine 1998 ; 77 : 195 – 207 . Google Scholar CrossRef Search ADS PubMed 3 Rodriguez-Pintó I , Santacreu I , Cervera R , Espinosa G. What is the best strategy in treating catastrophic antiphospholipid syndrome? Int J Clin Rheumtol 2015 ; 10 : 245 – 55 . Google Scholar CrossRef Search ADS 4 Asherson RA , Cervera R , Piette JC et al. Catastrophic antiphospholipid syndrome: clues to the pathogenesis from a series of 80 patients . Medicine 2001 ; 80 : 355 – 77 . Google Scholar CrossRef Search ADS PubMed 5 Bucciarelli S , Espinosa G , Cervera R et al. Mortality in the catastrophic antiphospholipid syndrome: causes of death and prognostic factors in a series of 250 patients . Arthritis Rheum 2006 ; 54 : 2268 – 576 . Google Scholar CrossRef Search ADS 6 Cervera R , Font J , Gómez-Puerta JA et al. Validation of the preliminary criteria for the classification of catastrophic antiphospholipid syndrome . Ann Rheum Dis 2005 ; 64 : 1205 – 9 . Google Scholar CrossRef Search ADS PubMed 7 Cohen J. Statistical power analysis for the behavioral sciences . Mahwah, NJ : Lawrence Erlbaum Associates, Publishers , 1988 . 8 Delgado M , Llorca J , Domenech J. Investigación científica: fundamentos metodológicos y estadísticos . 5th edn . Barcelona : Signo , 2012 . 9 Maldonado G , Greenland S. Simulation study of confounder-selection strategies . Am J Epidemiol 1993 ; 138 : 923 – 36 . Google Scholar CrossRef Search ADS PubMed 10 Ortega-Hernandez O-D , Agmon-Levin N , Blank M , Asherson RA , Shoenfeld Y. The physiopathology of the catastrophic antiphospholipid (Asherson’s) syndrome: compelling evidence . J Autoimmun 2009 ; 32 : 1 – 6 . Google Scholar CrossRef Search ADS PubMed 11 Cervera R , Rodríguez-Pintó I , Colafrancesco S et al. 14th international congress on antiphospholipid antibodies task force report on catastrophic antiphospholipid syndrome . Autoimmun Rev 2014 ; 13 : 699 – 707 . Google Scholar CrossRef Search ADS PubMed 12 Cox JL , Koepsell SA , Shunkwiler SM. Therapeutic plasma exchange and pregnancy: a case report and guidelines for performing plasma exchange in a pregnant patient . J Clin Apher 2017 ; 32 : 191 – 5 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Ez-Zaitouni, Zineb;Landewé, Robert;van Lunteren, Miranda;Bakker, Pauline A;Fagerli, Karen M;Ramonda, Roberta;Jacobsson, Lennart T H;van der Heijde, Désirée;van Gaalen, Floris A

2018 Rheumatology

doi: 10.1093/rheumatology/key035pmid: 29579265

Abstract Objectives To evaluate the contribution of the results of sacroiliac imaging to diagnosis and to the level of confidence in diagnosis in patients presenting with chronic back pain (CBP) and suspected of having axial spondyloarthritis (axSpA). Methods Data from 513 patients from the SPondyloArthritisCaughtEarly cohort with CBP (⩾3 months, ⩽2 years, onset <45 years) were analysed after full diagnostic work-up. Rheumatologists were asked not only to provide a diagnosis before and after the imaging results had been provided to them, but also to provide the level of confidence of this diagnosis on an 11-point numerical scale. Results Before imaging, 317/513 patients were diagnosed with axSpA. Of these patients, 178/317 (56%) did not have signs of sacroiliitis on either MRI or radiography, which led to the rheumatologist refuting the initial diagnosis of axSpA in 81/178 (46%) patients. Of the 196/513 patients without axSpA before imaging, 35/196 (18%) had signs of sacroiliitis on imaging. Subsequently, 28/35 (80%) patients were diagnosed with axSpA. Overall, imaging was incongruent with the diagnosis before imaging in 213 patients. This led to a change in diagnosis in 109/213 (51%), which corresponds to 21% (109/513) of all patients in the cohort. In general, diagnostic confidence increased by having imaging results available (from 6.2 to 7.4, P < 0.001). Conclusion In patients with CBP suspected of having axSpA, sacroiliac imaging adds to the confidence in the final diagnosis. However, the number of changes in diagnosis suggests that imaging is important but not all-decisive in early axSpA diagnosis. axial spondyloarthritis, ankylosing spondylitis, magnetic resonance imaging, clinical diagnosis Rheumatology key messages Sacroiliac imaging is widely used in the diagnostic work-up of patients suspected of axial SpA. Sacroliac imaging results increase rheumatologists’ confidence in their final diagnosis. Modest changes in numbers of axial SpA following imaging suggest imaging is important but not crucial. Introduction The diagnosis of axial spondyloarthritis (axSpA) in patients presenting with chronic back pain (CBP) is a known challenge in clinical practice as there is a broad spectrum in clinical presentations [1]. Rheumatologists may use information acquired from several sources such as a patient’s medical history, physical examination, laboratory tests and imaging (radiography and/or magnetic resonance imaging of sacroiliac joints) to make a diagnosis [2–4]. Several decades ago, conventional radiography was introduced as an imaging tool to detect sacroiliitis [5]. However, pelvic radiographs can only detect structural changes such as erosions and sclerosis. Furthermore, in early axSpA radiography of the sacroiliac joints may not show structural abnormalities, which can remain absent for many years after disease onset [6]. In recent years, MRI has become an important instrument in the visualization of both inflammation and structural damage in CBP patients especially without radiographic sacroiliitis [7–11]. Imaging of the sacroiliac joints plays a pivotal role in the early recognition of axSpA [12, 13]. We have recently shown that in CBP patients suspected of axSpA, positive imaging of the sacroiliac joints is highly associated with a diagnosis of axSpA (odds ratio = 34.3; 95% CI: 17.3, 67.7) [14]. Unfortunately, recognition of sacroiliitis on radiographs can be difficult. For example, it has been shown that agreement on radiographic sacroiliitis between readers is low and that this does not improve after training [15]. Moreover, bone marrow oedema on MRI suggestive of axSpA may not always be due to axSpA [16]. Therefore, some axSpA experts have expressed their concerns about relying solely on imaging, which may lead to incorrect diagnoses of axSpA. In turn, this may lead to unnecessary exposure to anti-inflammatory drugs with potentially severe side effects [17, 18]. Given these controversies, surprisingly little is known about how rheumatologists actually integrate sacroiliac imaging results in their diagnostic work-up of CBP patients suspected of having axSpA. Therefore, the main objectives of this study were to study the contribution of sacroiliac imaging to the rheumatologist’s diagnosis, and to quantify the contribution of sacroiliac imaging to diagnostic certainty. Methods Study design and population The data used for the current study were obtained from the SPondyloArthritis Caught Early (SPACE) cohort. The SPACE cohort is a prospective multicentre study, which was initiated in January 2009. A detailed description of the study design has been published previously [19]. Briefly, patients with CBP (⩾3 months and ⩽2 years) of unknown origin and age of onset <45 years were recruited and included from multiple European rheumatology centres in the Netherlands, Norway, Italy and Sweden. The clinical database used for the current study was locked on 11 January 2017. Approval for the study was obtained from the local medical ethical committee of the Leiden University Medical Center (reference number: P08.105); patients provided written informed consent before participation. Clinical assessments and measurements All patients were examined according to a standardized full work-up including the assessment of the presence and history of clinical SpA features according to the Assessment of Spondyloarthritis international Society (ASAS) definitions [3]: CRP/ESR, HLA-B27, inflammatory back pain, good response to NSAIDs, positive family history of SpA, peripheral arthritis, dactylitis, heel enthesitis, acute anterior uveitis, IBD and psoriasis. Imaging of the sacroiliac joints Plain radiographs of the pelvis (X-SI) were performed in anteroposterior view. MRI of sacroiliac joints (MRI-SI) was performed in coronal oblique T1-weighted turbo spin echo and short tau inversion recovery with a slice thickness of 4 mm. Interpretation of the radiographs and MRI of the sacroiliac joints (sacroiliitis yes/no) was done by each participating centre as part of routine clinical practice (local reading). Patients were classified according to the ASAS classification criteria for which data from central reading in the SPACE cohort was used [20]. Diagnosis During the diagnostic work-up, rheumatologists were asked to provide a diagnosis twice (axSpA, yes/no): based only on information available after medical history, physical examination and laboratory testing, but before taking sacroiliac imaging into account (diagnosis before imaging), and based on all previously collected information but after taking sacroiliac imaging into account (diagnosis after imaging). In case of no axSpA, rheumatologists were asked to provide the most likely alternative diagnosis. In addition, rheumatologists were requested to provide a level of confidence (LoC) regarding their diagnosis (axSpA or no axSpA) on an 11-point numerical rating scale ranging from 0 (not confident at all) to 10 (very confident). Data analysis For the present study baseline patients with complete information on sacroiliac imaging (both MRI and radiography) and diagnosis were analysed. Descriptive statistics were used to define patient characteristics for the total patient group, for each diagnosis subgroup (axSpA vs no axSpA), and for each imaging subgroup (i.e. sacroiliitis positive and negative) as means (s.d.) for continuous variables or frequencies (%) for categorical variables. Chi-square and unpaired t tests were used to compare variables between groups. The paired t test was used to compare the level of confidence regarding diagnosis before and after imaging within each diagnosis subgroup according to their imaging status. Total number of SpA features was calculated including HLA-B27 status, but without taking imaging of the sacroiliac joints into account. Any positive imaging was defined as one of a negative X-SI but positive MRI-SI (X-SI−/MRI-SI+), a positive X-SI but negative MRI-SI (X-SI+/MRI-SI−) or sacroiliitis on both modalities (X-SI+/MRI-SI+). The rheumatologist’s diagnosis and the LoC regarding diagnosis were the main outcomes. Data analysis was performed using Stata SE v.14 software (StataCorp LP, College Station, TX, USA). Results Baseline data of 583 CBP patients were available. A total of 70 (12%) patients were excluded because of incomplete or missing data regarding diagnosis or sacroiliac imaging. Baseline characteristics of these patients were similar to the remaining 513 patients with complete data (data not shown). Of these 513 CBP patients suspected of having axSpA, 188 (37%) patients were male, mean age (s.d.) was 31.0 (8.2) years, mean symptom duration was 13.3 (7.0) months and 210 (41%) patients were HLA-B27 positive (Table 1). Table 1 Baseline characteristics of CBP patients suspected of axSpA in the SPACE cohort (n = 513) Characteristic Male 188 (37) Age, mean (s.d.), years 31.0 (8.2) Symptom duration, mean (s.d.), months 13.3 (7.0) Number of SpA features incl. HLA-B27, mean (s.d.) 2.9 (1.8) HLA-B27 positivea 210 (41) IBP 346 (68) Peripheral arthritis 78 (15) Dactylitis 28 (5) Enthesitis 106 (21) Acute anterior uveitis 38 (7) IBD 35 (7) Psoriasis 58 (11) Good response to NSAIDsb 216 (43) Positive family history of SpA 222 (43) Elevated CRP/ESR 147 (29) Characteristic Male 188 (37) Age, mean (s.d.), years 31.0 (8.2) Symptom duration, mean (s.d.), months 13.3 (7.0) Number of SpA features incl. HLA-B27, mean (s.d.) 2.9 (1.8) HLA-B27 positivea 210 (41) IBP 346 (68) Peripheral arthritis 78 (15) Dactylitis 28 (5) Enthesitis 106 (21) Acute anterior uveitis 38 (7) IBD 35 (7) Psoriasis 58 (11) Good response to NSAIDsb 216 (43) Positive family history of SpA 222 (43) Elevated CRP/ESR 147 (29) Values are listed as n (%), unless otherwise stated. a n = 510 patients for HLA-B27. b n = 499 patients for good response to NSAIDs. axSpA: axial spondyloarthritis; IBP: inflammatory back pain. Table 1 Baseline characteristics of CBP patients suspected of axSpA in the SPACE cohort (n = 513) Characteristic Male 188 (37) Age, mean (s.d.), years 31.0 (8.2) Symptom duration, mean (s.d.), months 13.3 (7.0) Number of SpA features incl. HLA-B27, mean (s.d.) 2.9 (1.8) HLA-B27 positivea 210 (41) IBP 346 (68) Peripheral arthritis 78 (15) Dactylitis 28 (5) Enthesitis 106 (21) Acute anterior uveitis 38 (7) IBD 35 (7) Psoriasis 58 (11) Good response to NSAIDsb 216 (43) Positive family history of SpA 222 (43) Elevated CRP/ESR 147 (29) Characteristic Male 188 (37) Age, mean (s.d.), years 31.0 (8.2) Symptom duration, mean (s.d.), months 13.3 (7.0) Number of SpA features incl. HLA-B27, mean (s.d.) 2.9 (1.8) HLA-B27 positivea 210 (41) IBP 346 (68) Peripheral arthritis 78 (15) Dactylitis 28 (5) Enthesitis 106 (21) Acute anterior uveitis 38 (7) IBD 35 (7) Psoriasis 58 (11) Good response to NSAIDsb 216 (43) Positive family history of SpA 222 (43) Elevated CRP/ESR 147 (29) Values are listed as n (%), unless otherwise stated. a n = 510 patients for HLA-B27. b n = 499 patients for good response to NSAIDs. axSpA: axial spondyloarthritis; IBP: inflammatory back pain. In all patients, rheumatologists provided a diagnosis before and after taking sacroiliac imaging into account (Fig. 1). Before imaging, 317 (62%) patients were diagnosed with axSpA (Table 2). Most common diagnoses in the 196 patients without axSpA were non-specific back pain, degenerative disc disease and mechanical back pain (data not shown). Diagnostic confidence was moderate in patients with and without an axSpA diagnosis before imaging [mean (s.d.) LoC axSpA 6.6 (1.9) and mean LoC no axSpA 5.6 (2.0), respectively]. Patients diagnosed with axSpA before imaging were more often male (41% vs 29%) and were more often HLA-B27 positive (53% vs 22%) compared with patients who were not diagnosed with axSpA before imaging. As expected, the mean (s.d.) number of SpA features was twice as high for axSpA patients compared with the patients not diagnosed with axSpA [3.1 (1.7) and 1.5 (1.0), respectively]. Table 2 Clinical features of CBP patients stratified for diagnosis before taking sacroiliac imaging results into account Characteristic Diagnosis before imaging of sacroiliac joints AxSpA, n = 317 No axSpA, n = 196 LoC regarding diagnosis, mean (s.d.) 6.6 (1.9) 5.6 (2.0) Male 131 (41) 57 (29) Number of SpA features incl. HLA-B27, mean (s.d.) 3.6 (1.7) 1.7 (1.1) HLA-B27 positive 168 (53)a 42 (22) IBP 253 (80) 93 (48) Peripheral arthritisb 68 (22) 10 (5) Dactylitisb 27 (9) 1 (1) Enthesitisb 92 (29) 14 (7) Acute anterior uveitisb 32 (10) 6 (3) IBDb 26 (8) 9 (5) Psoriasisb 48 (15) 10 (5) Good response to NSAIDsc 176 (57)d 40 (21)e Positive family history of SpAf 160 (50) 62 (32) Elevated CRP/ESR 107 (34) 40 (20) Characteristic Diagnosis before imaging of sacroiliac joints AxSpA, n = 317 No axSpA, n = 196 LoC regarding diagnosis, mean (s.d.) 6.6 (1.9) 5.6 (2.0) Male 131 (41) 57 (29) Number of SpA features incl. HLA-B27, mean (s.d.) 3.6 (1.7) 1.7 (1.1) HLA-B27 positive 168 (53)a 42 (22) IBP 253 (80) 93 (48) Peripheral arthritisb 68 (22) 10 (5) Dactylitisb 27 (9) 1 (1) Enthesitisb 92 (29) 14 (7) Acute anterior uveitisb 32 (10) 6 (3) IBDb 26 (8) 9 (5) Psoriasisb 48 (15) 10 (5) Good response to NSAIDsc 176 (57)d 40 (21)e Positive family history of SpAf 160 (50) 62 (32) Elevated CRP/ESR 107 (34) 40 (20) Values are listed as n (%), unless otherwise stated. LoC, level of confidence regarding diagnosis: 0 (not confident at all) through 10 (very confident). a n = 315 for patients with axSpA diagnosis before imaging. b Past or present condition, either confirmed or diagnosed by a physician. c Back pain not present any more or is much better 24–48 h after a full dose of NSAID. d n = 310 for patients with axSpA diagnosis before imaging, e n = 189 for patients without axSpA diagnosis before imaging. f Presence in first- or second-degree relatives of any of the following: AS, acute anterior uveitis, ReA, psoriasis or IBD. axSpA: axial spondyloarthritis; IBP, inflammatory back pain. Table 2 Clinical features of CBP patients stratified for diagnosis before taking sacroiliac imaging results into account Characteristic Diagnosis before imaging of sacroiliac joints AxSpA, n = 317 No axSpA, n = 196 LoC regarding diagnosis, mean (s.d.) 6.6 (1.9) 5.6 (2.0) Male 131 (41) 57 (29) Number of SpA features incl. HLA-B27, mean (s.d.) 3.6 (1.7) 1.7 (1.1) HLA-B27 positive 168 (53)a 42 (22) IBP 253 (80) 93 (48) Peripheral arthritisb 68 (22) 10 (5) Dactylitisb 27 (9) 1 (1) Enthesitisb 92 (29) 14 (7) Acute anterior uveitisb 32 (10) 6 (3) IBDb 26 (8) 9 (5) Psoriasisb 48 (15) 10 (5) Good response to NSAIDsc 176 (57)d 40 (21)e Positive family history of SpAf 160 (50) 62 (32) Elevated CRP/ESR 107 (34) 40 (20) Characteristic Diagnosis before imaging of sacroiliac joints AxSpA, n = 317 No axSpA, n = 196 LoC regarding diagnosis, mean (s.d.) 6.6 (1.9) 5.6 (2.0) Male 131 (41) 57 (29) Number of SpA features incl. HLA-B27, mean (s.d.) 3.6 (1.7) 1.7 (1.1) HLA-B27 positive 168 (53)a 42 (22) IBP 253 (80) 93 (48) Peripheral arthritisb 68 (22) 10 (5) Dactylitisb 27 (9) 1 (1) Enthesitisb 92 (29) 14 (7) Acute anterior uveitisb 32 (10) 6 (3) IBDb 26 (8) 9 (5) Psoriasisb 48 (15) 10 (5) Good response to NSAIDsc 176 (57)d 40 (21)e Positive family history of SpAf 160 (50) 62 (32) Elevated CRP/ESR 107 (34) 40 (20) Values are listed as n (%), unless otherwise stated. LoC, level of confidence regarding diagnosis: 0 (not confident at all) through 10 (very confident). a n = 315 for patients with axSpA diagnosis before imaging. b Past or present condition, either confirmed or diagnosed by a physician. c Back pain not present any more or is much better 24–48 h after a full dose of NSAID. d n = 310 for patients with axSpA diagnosis before imaging, e n = 189 for patients without axSpA diagnosis before imaging. f Presence in first- or second-degree relatives of any of the following: AS, acute anterior uveitis, ReA, psoriasis or IBD. axSpA: axial spondyloarthritis; IBP, inflammatory back pain. Fig. 1 View largeDownload slide Rheumatologist’s diagnosis before and after considering imaging of sacroiliac joints (n = 513) Any positive imaging defined as sacroiliitis on MRI and/or radiographic sacroiliitis; imaging negative defined as no abnormalities on MRI and radiographs of sacroiliac joints. Boxes in bold represent patients with a change in diagnosis due to imaging results discordant with the diagnosis before imaging. axSpA: axial spondyloarthritis. Fig. 1 View largeDownload slide Rheumatologist’s diagnosis before and after considering imaging of sacroiliac joints (n = 513) Any positive imaging defined as sacroiliitis on MRI and/or radiographic sacroiliitis; imaging negative defined as no abnormalities on MRI and radiographs of sacroiliac joints. Boxes in bold represent patients with a change in diagnosis due to imaging results discordant with the diagnosis before imaging. axSpA: axial spondyloarthritis. In total, 317/513 (62%) of CBP patients were initially diagnosed with axSpA, and this figure decreased to 269/513 (52%) patients with a final diagnosis of axSpA after imaging. Of these 269 axSpA patients, 55% were male, 59% were HLA-B27 positive, 62% had positive imaging and the mean number of SpA features was 3.2 (1.7). A total of 172/269 (64%) patients fulfilled the ASAS classification criteria [52% male, 87% HLA-B27 positive and mean number of SpA features 3.2 (1.5)]. Overall, the mean diagnostic confidence increased by having imaging results available (from 6.2 to 7.4, P < 0.001). This increase in diagnostic confidence was observed in both axSpA patients [from 7.1 (1.7) to 7.7 (2.1), P < 0.001] and non-axSpA patients [from 5.7 (2.0) to 7.5 (2.3), P < 0.001]. Of the 317 patients who were diagnosed with axSpA before imaging, 139 (44%) had positive imaging (Table 3). In these 139 patients, sacroiliitis was seen in 85 (61%) patients only on MRI-SI, in 10 (7%) patients only on X-SI and in 44 (32%) patients on both modalities (X-SI+/MRI-SI+). After imaging, the axSpA diagnosis was maintained in all of these 139 patients and the mean LoC in the diagnosis of axSpA increased significantly [from 7.4 (1.8) to 8.6 (1.7), P < 0.001]. Table 3 Diagnosis and imaging status after sacroiliac imaging in patients with axSpA diagnosis before sacroiliac imaging AxSpA diagnosis before imaging, n = 317 Any imaging positive, n = 139 Negative imaging, n = 178 Imaging status  Only MRI-SI+ 85 (61) N/A  Only X-SI+ 10 (7) N/A  MRI-SI+/X-SI+ 44 (32) N/A Diagnosis after imaging Diagnosis after imaging AxSpA, n = 139 No axSpA, n = 0 AxSpA, n = 97 No axSpA, n = 81 LoC regarding diagnosis before imaging, mean (s.d.) 7.4 (1.8) — 6.7 (1.6) 5.1 (1.7) LoC regarding diagnosis after imaging, mean (s.d.) 8.6 (1.7)a — 6.4 (2.1)b 6.4 (2.0)c Imaging status  Only MRI-SI+ 85 (61) — N/A N/A  Only X-SI+ 10 (7) — N/A N/A  MRI-SI+/X-SI+ 44 (32) — N/A N/A Clinical features  Male 71 (51) — 32 (33) 28 (35)  Number of SpA featuresd, mean (s.d.) 3.4 (1.7) — 3.5 (1.6) 2.2 (1.2) HLA-B27+ 90 (66)e — 51 (53) 27 (33) AxSpA diagnosis before imaging, n = 317 Any imaging positive, n = 139 Negative imaging, n = 178 Imaging status  Only MRI-SI+ 85 (61) N/A  Only X-SI+ 10 (7) N/A  MRI-SI+/X-SI+ 44 (32) N/A Diagnosis after imaging Diagnosis after imaging AxSpA, n = 139 No axSpA, n = 0 AxSpA, n = 97 No axSpA, n = 81 LoC regarding diagnosis before imaging, mean (s.d.) 7.4 (1.8) — 6.7 (1.6) 5.1 (1.7) LoC regarding diagnosis after imaging, mean (s.d.) 8.6 (1.7)a — 6.4 (2.1)b 6.4 (2.0)c Imaging status  Only MRI-SI+ 85 (61) — N/A N/A  Only X-SI+ 10 (7) — N/A N/A  MRI-SI+/X-SI+ 44 (32) — N/A N/A Clinical features  Male 71 (51) — 32 (33) 28 (35)  Number of SpA featuresd, mean (s.d.) 3.4 (1.7) — 3.5 (1.6) 2.2 (1.2) HLA-B27+ 90 (66)e — 51 (53) 27 (33) Values are listed as n (%), unless otherwise stated. a Mean difference 1.2, 95% CI: 0.9, 1.6, P < 0.001; b mean difference −0.3, 95% CI: −0.5, 0.01, P = 0.06; c Mean difference 1.3, 95% CI: 0.7, 1.8, P < 0.001. d Total number of SpA features after medical history taking, physical examination and measurement of acute phase reactants but before HLA-B27 testing and imaging. e n = 137 for patients with axSpA diagnosis before and after imaging. axSpA: axial spondyloarthritis; LoC: level of confidence regarding diagnosis: 0 (not confident at all) through 10 (very confident); MRI-SI: MRI of sacroiliac joints; N/A: not applicable; X-SI: radiographs of sacroiliac joints. Table 3 Diagnosis and imaging status after sacroiliac imaging in patients with axSpA diagnosis before sacroiliac imaging AxSpA diagnosis before imaging, n = 317 Any imaging positive, n = 139 Negative imaging, n = 178 Imaging status  Only MRI-SI+ 85 (61) N/A  Only X-SI+ 10 (7) N/A  MRI-SI+/X-SI+ 44 (32) N/A Diagnosis after imaging Diagnosis after imaging AxSpA, n = 139 No axSpA, n = 0 AxSpA, n = 97 No axSpA, n = 81 LoC regarding diagnosis before imaging, mean (s.d.) 7.4 (1.8) — 6.7 (1.6) 5.1 (1.7) LoC regarding diagnosis after imaging, mean (s.d.) 8.6 (1.7)a — 6.4 (2.1)b 6.4 (2.0)c Imaging status  Only MRI-SI+ 85 (61) — N/A N/A  Only X-SI+ 10 (7) — N/A N/A  MRI-SI+/X-SI+ 44 (32) — N/A N/A Clinical features  Male 71 (51) — 32 (33) 28 (35)  Number of SpA featuresd, mean (s.d.) 3.4 (1.7) — 3.5 (1.6) 2.2 (1.2) HLA-B27+ 90 (66)e — 51 (53) 27 (33) AxSpA diagnosis before imaging, n = 317 Any imaging positive, n = 139 Negative imaging, n = 178 Imaging status  Only MRI-SI+ 85 (61) N/A  Only X-SI+ 10 (7) N/A  MRI-SI+/X-SI+ 44 (32) N/A Diagnosis after imaging Diagnosis after imaging AxSpA, n = 139 No axSpA, n = 0 AxSpA, n = 97 No axSpA, n = 81 LoC regarding diagnosis before imaging, mean (s.d.) 7.4 (1.8) — 6.7 (1.6) 5.1 (1.7) LoC regarding diagnosis after imaging, mean (s.d.) 8.6 (1.7)a — 6.4 (2.1)b 6.4 (2.0)c Imaging status  Only MRI-SI+ 85 (61) — N/A N/A  Only X-SI+ 10 (7) — N/A N/A  MRI-SI+/X-SI+ 44 (32) — N/A N/A Clinical features  Male 71 (51) — 32 (33) 28 (35)  Number of SpA featuresd, mean (s.d.) 3.4 (1.7) — 3.5 (1.6) 2.2 (1.2) HLA-B27+ 90 (66)e — 51 (53) 27 (33) Values are listed as n (%), unless otherwise stated. a Mean difference 1.2, 95% CI: 0.9, 1.6, P < 0.001; b mean difference −0.3, 95% CI: −0.5, 0.01, P = 0.06; c Mean difference 1.3, 95% CI: 0.7, 1.8, P < 0.001. d Total number of SpA features after medical history taking, physical examination and measurement of acute phase reactants but before HLA-B27 testing and imaging. e n = 137 for patients with axSpA diagnosis before and after imaging. axSpA: axial spondyloarthritis; LoC: level of confidence regarding diagnosis: 0 (not confident at all) through 10 (very confident); MRI-SI: MRI of sacroiliac joints; N/A: not applicable; X-SI: radiographs of sacroiliac joints. Of the 317 patients who were diagnosed with axSpA before imaging, 178 (56%) had negative imaging. In 97/178 patients the diagnosis axSpA was maintained. In these patients, the LoC in the diagnosis somewhat decreased [from 6.7 (1.6) to 6.4 (2.1), P = 0.06]. In 81/178 patients with negative imaging the diagnosis was changed to no axSpA after imaging. In these patients the LoC increased significantly after imaging [from 5.1 (1.7) to 6.4 (2.0), P < 0.001]. By comparison, the 97/178 patients diagnosed with axSpA after imaging had a higher number of SpA features (excluding imaging and HLA-B27) than the 81/178 patients without axSpA [mean 3.5 (1.6) vs 2.2 (1.2), P < 0.001]. Moreover, these 97 axSpA patients were also more often HLA-B27 positive than the 81 patients without axSpA (53% vs 33%, P = 0.010). Of the 196 patients not diagnosed with axSpA before imaging, 35 (18%) had positive imaging (Table 4). In these 35 patients, sacroiliitis was seen in 24 (68%) patients only on MRI-SI, in 2 (6%) patients only on X-SI and in 9 (26%) patients on both modalities (X-SI+/MRI-SI+). In 28 of these 35 patients (80%) with sacroiliitis, the diagnosis was changed to axSpA [18/28 (64%) patients only MRI-SI+, 1/28 (4%) only X-SI+ and 9/28 (32%) MRI-SI+/X-SI+]. The mean LoC in diagnosis increased significantly following imaging [from 4.7 (2.0) to 7.6 (2.3), P < 0.001]. In 7 of the 35 patients with positive imaging (20%), the diagnosis no axSpA remained unchanged [6/7 (86%) patients were MRI-SI+ and one (14%) patient was X-SI+]. In these seven patients, the mean LoC in diagnosis remained the same after imaging [LoC from 5.0 (1.3) to 5.0 (2.3)]. By comparison, the 28/35 patients diagnosed with axSpA after imaging had a higher number of SpA features than the 7/35 patients without axSpA [mean 1.6 (0.9) vs 1.0 (0.8), P = 0.10]. Moreover, these 28 patients diagnosed with axSpA were also more often HLA-B27 positive than the seven patients without axSpA (48% vs 29%, P = 0.35). Table 4 Diagnosis and imaging status after sacroiliac imaging in patients without axSpA diagnosis before sacroiliac imaging No axSpA diagnosis before imaging, n = 196 Any imaging positive, n = 35 Negative imaging, n = 161 Imaging status  Only MRI-SI+ 24 (68) N/A  Only X-SI+ 2 (6) N/A  MRI-SI+/X-SI+ 9 (26) N/A Diagnosis after imaging Diagnosis after imaging AxSpA, n = 28 No axSpA, n = 7 AxSpA, n = 5 No axSpA, n = 156 LoC regarding diagnosis before imaging, mean (s.d.) 4.7 (2.0) 5.0 (1.3) 5.0 (1.9) 5.8 (2.0) LoC regarding diagnosis after imaging, mean (s.d.) 7.6 (2.3)a 5.0 (2.3)b 6.4 (1.9)c 7.6 (2.2)d Imaging status  Only MRI-SI+ 18 (64) 6 (86) N/A N/A  Only X-SI+ 1 (4) 1 (14) N/A N/A  MRI-SI+/X-SI+ 9 (32) — N/A N/A Clinical features  Male 15 (54) 2 (29) 2 (40) 38 (24)  Number of SpA featurese, mean (s.d.) 1.6 (0.9) 1 (0.8) 1.8 (0.8) 1.4 (1.0) HLA-B27+ 13 (48) 2 (29) 2 (40) 25 (16) No axSpA diagnosis before imaging, n = 196 Any imaging positive, n = 35 Negative imaging, n = 161 Imaging status  Only MRI-SI+ 24 (68) N/A  Only X-SI+ 2 (6) N/A  MRI-SI+/X-SI+ 9 (26) N/A Diagnosis after imaging Diagnosis after imaging AxSpA, n = 28 No axSpA, n = 7 AxSpA, n = 5 No axSpA, n = 156 LoC regarding diagnosis before imaging, mean (s.d.) 4.7 (2.0) 5.0 (1.3) 5.0 (1.9) 5.8 (2.0) LoC regarding diagnosis after imaging, mean (s.d.) 7.6 (2.3)a 5.0 (2.3)b 6.4 (1.9)c 7.6 (2.2)d Imaging status  Only MRI-SI+ 18 (64) 6 (86) N/A N/A  Only X-SI+ 1 (4) 1 (14) N/A N/A  MRI-SI+/X-SI+ 9 (32) — N/A N/A Clinical features  Male 15 (54) 2 (29) 2 (40) 38 (24)  Number of SpA featurese, mean (s.d.) 1.6 (0.9) 1 (0.8) 1.8 (0.8) 1.4 (1.0) HLA-B27+ 13 (48) 2 (29) 2 (40) 25 (16) Values are listed as n (%), unless otherwise stated. a Mean difference 2.9, 95% CI: 1.6, 4.2, P < 0.001; b mean difference 0; c mean difference 1.4, 95% CI: −3.1, 5.9, P = 0.43; d mean difference 1.8, 95% CI: 1.5, 2.1, P < 0.001. e Total number of SpA features after medical history taking, physical examination and measurement of acute phase reactants but before HLA-B27 testing and imaging. axSpA: axial spondyloarthritis; LoC: level of confidence regarding diagnosis: 0 (not confident at all) through 10 (very confident); MRI-SI: MRI of sacroiliac joints; N/A: not applicable; X-SI: radiographs of sacroiliac joints. Table 4 Diagnosis and imaging status after sacroiliac imaging in patients without axSpA diagnosis before sacroiliac imaging No axSpA diagnosis before imaging, n = 196 Any imaging positive, n = 35 Negative imaging, n = 161 Imaging status  Only MRI-SI+ 24 (68) N/A  Only X-SI+ 2 (6) N/A  MRI-SI+/X-SI+ 9 (26) N/A Diagnosis after imaging Diagnosis after imaging AxSpA, n = 28 No axSpA, n = 7 AxSpA, n = 5 No axSpA, n = 156 LoC regarding diagnosis before imaging, mean (s.d.) 4.7 (2.0) 5.0 (1.3) 5.0 (1.9) 5.8 (2.0) LoC regarding diagnosis after imaging, mean (s.d.) 7.6 (2.3)a 5.0 (2.3)b 6.4 (1.9)c 7.6 (2.2)d Imaging status  Only MRI-SI+ 18 (64) 6 (86) N/A N/A  Only X-SI+ 1 (4) 1 (14) N/A N/A  MRI-SI+/X-SI+ 9 (32) — N/A N/A Clinical features  Male 15 (54) 2 (29) 2 (40) 38 (24)  Number of SpA featurese, mean (s.d.) 1.6 (0.9) 1 (0.8) 1.8 (0.8) 1.4 (1.0) HLA-B27+ 13 (48) 2 (29) 2 (40) 25 (16) No axSpA diagnosis before imaging, n = 196 Any imaging positive, n = 35 Negative imaging, n = 161 Imaging status  Only MRI-SI+ 24 (68) N/A  Only X-SI+ 2 (6) N/A  MRI-SI+/X-SI+ 9 (26) N/A Diagnosis after imaging Diagnosis after imaging AxSpA, n = 28 No axSpA, n = 7 AxSpA, n = 5 No axSpA, n = 156 LoC regarding diagnosis before imaging, mean (s.d.) 4.7 (2.0) 5.0 (1.3) 5.0 (1.9) 5.8 (2.0) LoC regarding diagnosis after imaging, mean (s.d.) 7.6 (2.3)a 5.0 (2.3)b 6.4 (1.9)c 7.6 (2.2)d Imaging status  Only MRI-SI+ 18 (64) 6 (86) N/A N/A  Only X-SI+ 1 (4) 1 (14) N/A N/A  MRI-SI+/X-SI+ 9 (32) — N/A N/A Clinical features  Male 15 (54) 2 (29) 2 (40) 38 (24)  Number of SpA featurese, mean (s.d.) 1.6 (0.9) 1 (0.8) 1.8 (0.8) 1.4 (1.0) HLA-B27+ 13 (48) 2 (29) 2 (40) 25 (16) Values are listed as n (%), unless otherwise stated. a Mean difference 2.9, 95% CI: 1.6, 4.2, P < 0.001; b mean difference 0; c mean difference 1.4, 95% CI: −3.1, 5.9, P = 0.43; d mean difference 1.8, 95% CI: 1.5, 2.1, P < 0.001. e Total number of SpA features after medical history taking, physical examination and measurement of acute phase reactants but before HLA-B27 testing and imaging. axSpA: axial spondyloarthritis; LoC: level of confidence regarding diagnosis: 0 (not confident at all) through 10 (very confident); MRI-SI: MRI of sacroiliac joints; N/A: not applicable; X-SI: radiographs of sacroiliac joints. Of the 196 patients not diagnosed with axSpA before imaging, 161 (82%) had negative imaging. Despite having negative imaging, the diagnosis changed to axSpA after imaging in 5/196 patients (3%). In these five patients, LoC in diagnosis increased [from 5.0 (1.9) to 6.4 (1.9), P = 0.43]. In 156/161 (97%) patients with negative imaging, the diagnosis no axSpA remained unchanged. The LoC in diagnosis increased after imaging [from 5.8 (2.0) to 7.6 (2.2), P < 0.001]. By comparison, the 5/161 patients diagnosed with axSpA after imaging had a higher number of SpA features than the 156/161 patients without axSpA [mean 1.8 (0.8) vs 1.4 (1.0), P = 0.43]. Moreover, these 5 axSpA patients were also more often HLA-B27 positive than the 156 patients without axSpA (40% vs 16%, P = 0.16). In the entire cohort, the diagnosis changed in 21% (109/513) of the CBP patients following imaging. In patients with imaging results that were discordant with their primary diagnosis [n = 213 (178 + 35)], the diagnosis changed in 109 (51%) patients (Fig. 1, boxes in bold). This change of diagnosis more often pertained to patients in whom a clinical diagnosis was suspected but imaging was negative [81/109 (74%)] than in patients in whom a clinical diagnosis of axSpA was not suspected but imaging was positive [28/109 (26%)]. Discussion This study was performed to investigate the contribution of sacroiliac imaging to the rheumatologist’s diagnosis and its respective confidence. We have shown that the rheumatologist’s confidence in the diagnosis of patients with CBP suspected of having axSpA increases after the results of sacroiliac imaging are taken into account. However, the number of changes in diagnosis after imaging implies that imaging is indeed an important but not the all-decisive factor in axSpA diagnosis. Prior to imaging the physician’s confidence in diagnosis was already moderate to high in most patients, which corroborates the value of medical history taking, physical examination and laboratory tests in the diagnostic work-up of axSpA. Congruent imaging results significantly increased the diagnostic confidence of rheumatologists except in axSpA patients (before and after imaging) without sacroiliitis. In general, these findings are in line with results from an international survey conducted among rheumatologists throughout the world illustrating the value rheumatologists place on imaging in the diagnostic work-up of patients suspected of axSpA [21]. Nevertheless, there were patients diagnosed as no axSpA before imaging in which a positive imaging result did not influence the final diagnosis. A few patients were not diagnosed with axSpA after imaging, even though they had sacroiliitis (n = 6 for MRI-SI and n = 1 for X-SI). These patients had only a few other SpA features and rheumatologists remained uncertain about the diagnosis (mean LoC 5.0). Apparently, the diagnosing rheumatologist considered the observed lesions—in combination with the clinical presentation—not sufficiently specific for an axSpA diagnosis, which is an indication that rheumatologists do not necessarily find that imaging is the dominant feature in establishing a diagnosis of axSpA. An even smaller number of patients (n = 5) without a diagnosis of axSpA before imaging and without signs of sacroiliitis were still diagnosed with axSpA after imaging. Since these five patients have a low number of SpA features, it is difficult to understand these diagnoses. Moreover, our findings also stress the importance of imaging in rejecting an axSpA diagnosis as in 81/178 (46%) patients with axSpA before imaging, the diagnosis was dismissed when imaging turned out to be negative. This study has several limitations. First, rheumatologists were asked to provide CBP patients with a diagnosis before and after taking imaging into account. We cannot, however, rule out the possibility that rheumatologists may have already looked at the imaging results before filling out the diagnosis in the case report form. However, this does not explain why change in diagnosis still occurred. More importantly, the overall diagnostic confidence clearly increased after taking imaging results into account, which most likely would not have happened (or to a lesser extent) if rheumatologists had indeed looked at the imaging results before filling out the diagnosis before imaging. Although in line with clinical practice, the fact that each patient was diagnosed by one rheumatologist and the imaging was read by one radiologist might be regarded as a limitation. Furthermore, the rheumatologist may also have used information from other sources such as spinal imaging in the diagnostic process, which may have (additionally) contributed to the LoC regarding the diagnosis. In addition, due to the high overlap of positive imaging (i.e. X-SI+/MRI-SI+), we cannot elaborate on the individual value of each modality (X-SI or MRI-SI) on the final diagnosis and the diagnostic confidence. In conclusion, in CBP patients suspected of having axSpA, sacroiliac imaging increases diagnostic confidence. However, the number of changes in diagnosis suggests that imaging is important but not all-decisive in the early diagnosis of axSpA. Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript. Disclosure statement: L.T.H.J. has received advisory board fees from Abbvie, Celegen, Novartis and Pfizer. R.R. has received honoraria and speaker fees from AbbVie, MSD, Pfizer, Celgene, Janssen, Bristol-Myers Squibb and Mylan. All other authors have declared no conflicts of interest. References 1 Dougados M , Baeten D. Spondyloarthritis . Lancet 2011 ; 377 : 2127 – 37 . Google Scholar CrossRef Search ADS PubMed 2 Rudwaleit M , van der Heijde D , Khan MA , Braun J , Sieper J. How to diagnose axial spondyloarthritis early . Ann Rheum Dis 2004 ; 63 : 535 – 43 . Google Scholar CrossRef Search ADS PubMed 3 Rudwaleit M , van der Heijde D , Landewe R et al. The development of Assessment of SpondyloArthritis international Society classification criteria for axial spondyloarthritis (part II): validation and final selection . Ann Rheum Dis 2009 ; 68 : 777 – 83 . Google Scholar CrossRef Search ADS PubMed 4 van Tubergen A , Weber U. Diagnosis and classification in spondyloarthritis: identifying a chameleon . Nat Rev Rheumatol 2012 ; 8 : 253 – 61 . Google Scholar CrossRef Search ADS PubMed 5 van der Linden S , Valkenburg HA , Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria . Arthritis Rheum 1984 ; 27 : 361 – 8 . 6 Oostveen J , Prevo R , den Boer J , van de Laar M. Early detection of sacroiliitis on magnetic resonance imaging and subsequent development of sacroiliitis on plain radiography. A prospective, longitudinal study . J Rheumatol 1999 ; 26 : 1953 – 8 . Google Scholar PubMed 7 Braun J , Baraliakos X , Golder W et al. Analysing chronic spinal changes in ankylosing spondylitis: a systematic comparison of conventional x rays with magnetic resonance imaging using established and new scoring systems . Ann Rheum Dis 2004 ; 63 : 1046 – 55 . Google Scholar CrossRef Search ADS PubMed 8 Landewe RB , Hermann KG , van der Heijde DM et al. Scoring sacroiliac joints by magnetic resonance imaging. A multiple-reader reliability experiment . J Rheumatol 2005 ; 32 : 2050 – 5 . 9 Rudwaleit M , Jurik AG , Hermann KG et al. Defining active sacroiliitis on magnetic resonance imaging (MRI) for classification of axial spondyloarthritis: a consensual approach by the ASAS/OMERACT MRI group . Ann Rheum Dis 2009 ; 68 : 1520 – 7 . Google Scholar CrossRef Search ADS PubMed 10 Weber U , Lambert RG , Ostergaard M et al. The diagnostic utility of magnetic resonance imaging in spondylarthritis: an international multicenter evaluation of one hundred eighty-seven subjects . Arthritis Rheum 2010 ; 62 : 3048 – 58 . Google Scholar CrossRef Search ADS PubMed 11 Lambert RG , Bakker PA , van der Heijde D et al. Defining active sacroiliitis on MRI for classification of axial spondyloarthritis: update by the ASAS MRI working group . Ann Rheum Dis 2016 ; 75 : 1958 – 63 . Google Scholar CrossRef Search ADS PubMed 12 Mandl P , Navarro-Compan V , Terslev L et al. EULAR recommendations for the use of imaging in the diagnosis and management of spondyloarthritis in clinical practice . Ann Rheum Dis 2015 ; 74 : 1327 – 39 . Google Scholar CrossRef Search ADS PubMed 13 Baraliakos X , Maksymowych WP. Imaging in the diagnosis and management of axial spondyloarthritis . Best Pract Res Clin Rheumatol 2016 ; 30 : 608 – 23 . Google Scholar CrossRef Search ADS PubMed 14 Ez-Zaitouni Z , Bakker PA , van Lunteren M et al. Presence of multiple spondyloarthritis (SpA) features is important but not sufficient for a diagnosis of axial spondyloarthritis: data from the SPondyloArthritis Caught Early (SPACE) cohort . Ann Rheum Dis 2017 ; 76 : 1086 – 92 . Google Scholar CrossRef Search ADS PubMed 15 van Tubergen A , Heuft-Dorenbosch L , Schulpen G et al. Radiographic assessment of sacroiliitis by radiologists and rheumatologists: does training improve quality? Ann Rheum Dis 2003 ; 62 : 519 – 25 . Google Scholar CrossRef Search ADS PubMed 16 Baraliakos X , Hermann KG , Braun J. Imaging in axial spondyloarthritis: diagnostic problems and pitfalls . Rheum Dis Clin North Am 2012 ; 38 : 513 – 22 . Google Scholar CrossRef Search ADS PubMed 17 Deodhar A , Reveille JD , van den Bosch F et al. The concept of axial spondyloarthritis: joint statement of the spondyloarthritis research and treatment network and the Assessment of SpondyloArthritis international Society in response to the US Food and Drug Administration's comments and concerns. Arthritis Rheumatol 2014 ; 66 : 2649 – 56 . 18 Deodhar A. Sacroiliac joint magnetic resonance imaging in the diagnosis of axial spondyloarthritis: “A Tiny Bit of White on Two Consecutive Slices” may be objective, but not specific . Arthritis Rheumatol 2016 ; 68 : 775 – 8 . Google Scholar CrossRef Search ADS PubMed 19 van den Berg R , de Hooge M , van Gaalen F et al. Percentage of patients with spondyloarthritis in patients referred because of chronic back pain and performance of classification criteria: experience from the Spondyloarthritis Caught Early (SPACE) cohort . Rheumatology 2013 ; 52 : 1492 – 9 . Google Scholar CrossRef Search ADS PubMed 20 Ez-Zaitouni Z , Bakker PA , van Lunteren M et al. The yield of a positive MRI of the spine as imaging criterion in the ASAS classification criteria for axial spondyloarthritis: results from the SPACE and DESIR cohorts . Ann Rheum Dis 2017 ; 76 : 1731 – 6 . 21 van der Heijde D , Sieper J , Elewaut D et al. Referral patterns, diagnosis, and disease management of patients with axial spondyloarthritis: results of an international survey . J Clin Rheumatol 2014 ; 20 : 411 – 7 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Peyronel, Francesco;Fenaroli, Paride;Benigno, Giuseppe D;Palumbo, Alessandro A;Martella, Eugenia M;Biagi, Federico;Vaglio, Augusto

2018 Rheumatology

doi: 10.1093/rheumatology/key073pmid: 29579249

Rheumatology key message Whipple’s disease with systemic involvement can mimic granulomatosis with polyangiitis. Sir, Whipple’s disease is an extremely rare condition (3 cases/1 000 000 inhabitants in Northern Italy [1]), often presenting with non-specific systemic symptoms such as fatigue, weight loss and arthralgia. This is also why most patients suffering from this disease are correctly diagnosed only after several months to years [2]. Herein, we report the case of a 43-year-old man who was referred to our unit with a diagnosis of refractory ANCA-associated granulomatous vasculitis. The patient had a history of childhood-onset asthma and was a heavy smoker. In 2011 he started complaining of diffuse low-back pain and arthralgia, particularly involving the IP joints of the hands, often with an asymmetrical and migratory pattern. He received a tentative diagnosis of HLA B27-negative spondyloarthritis and was treated with MTX, steroids and NSAIDs. In March 2016 he developed palpable purpura of the lower limbs of suspected vasculitic origin (no skin biopsy was performed), generalized lymphadenopathy, abdominal pain and fever, with blood tests showing an increase in neutrophil count and inflammatory markers. Autoimmunity tests were negative. A lymph node biopsy and an oesophagogastroduodenoscopy were performed, the first demonstrating a granulomatous necrotizing lymphadenitis (Fig. 1B), the latter showing mild gastritis. He was thus diagnosed with ANCA-negative granulomatous vasculitis and treated with prednisolone (initial dose, 75 mg/day, tapered during the following months) and MTX (15 mg/week); MTX was stopped in December 2016 and AZA was prescribed, but the patient refused to take it and continued with prednisone alone. Fig. 1 View largeDownload slide Radiological and histological findings Whole-body contrast-enhanced CT (coronal view) with thoracic and abdominal axial slices: CT allows a reliable and comprehensive visualization of lymphadenopathies. Arrowheads indicate round enlarged lymph nodes in the mediastinum (i) and (ii), as well as in the peritoneal space (iii) and (iv) (A). Granulomatous necrotizing lymphadenitis—original magnification 10×, haematoxylin eosin staining (B). PAS-positive macrophages infiltrating the lymph node—original magnification 10×, PAS staining (C). Massive infiltration by foamy PAS-positive macrophages in the lamina propria of the duodenum—original magnification 10×, PAS staining (D) and 40×, PAS staining (E). PAS: periodic acid–Schiff. Fig. 1 View largeDownload slide Radiological and histological findings Whole-body contrast-enhanced CT (coronal view) with thoracic and abdominal axial slices: CT allows a reliable and comprehensive visualization of lymphadenopathies. Arrowheads indicate round enlarged lymph nodes in the mediastinum (i) and (ii), as well as in the peritoneal space (iii) and (iv) (A). Granulomatous necrotizing lymphadenitis—original magnification 10×, haematoxylin eosin staining (B). PAS-positive macrophages infiltrating the lymph node—original magnification 10×, PAS staining (C). Massive infiltration by foamy PAS-positive macrophages in the lamina propria of the duodenum—original magnification 10×, PAS staining (D) and 40×, PAS staining (E). PAS: periodic acid–Schiff. In March 2017 the patient was hospitalized for fever and severe weight loss. He underwent colonoscopy, which was reported as negative. Autoimmunity tests showed C-ANCA positivity (1:10 titre) with positive anti-PR3 antibodies (14 U/ml with EliA method; negative if <7 U/ml, positive if >10 U/ml), while all the other autoimmune tests (anti-nuclear, anti-mitochondrial, anti-smooth muscle, anti-liver/kidney-microsome, anti-dsDNA, anti-gliadin, anti-transglutaminase antibodies and cryoglobulins) were negative. Serum IgM for Rubella, CMV, HSV and EBV, and IgM for Toxoplasma were negative. Given the ANCA-positivity, the initial suspicion of granulomatous vasculitis became stronger and the patient was therefore diagnosed as having ANCA-positive granulomatosis with polyangiitis (GPA). Prednisone (initial dose, 50 mg/day) and CYC (50 mg/day) were started. Since the patient’s symptoms worsened after a month of treatment, in April 2017 he came to our centre for further evaluation. His major complaints were weight loss (20 kg over the previous 6 months), abdominal pain, diarrhoea, arthralgia and low-grade fever. Reassessing his clinical history, we considered that his clinical, serological and histological data were not consistent with GPA; we thus discontinued CYC and reduced the prednisone dose. Our lab tests demonstrated thrombocytosis (698 000/µl) and confirmed the increase in neutrophil count (white blood cell count 20 210/µl, neutrophils 78%); CRP was 55 mg/l and ESR 66 mm/h; ANCA were negative, using both immunofluorescence and ELISA for PR3 and MPO antibodies. Whole-body contrast-enhanced CT showed diffuse lymphadenopathies (Fig. 1A) and thus we repeated a lymph node biopsy (right supra-clavear lymph node), which on histological examination showed massive infiltration by foamy periodic acid–Schiff-positive macrophages (Fig. 1C), while the granulomatous and the necrotizing features that characterized the first biopsy (Fig. 1B) had disappeared. Oesophagogastroduodenoscopy was therefore repeated: marked infiltration by foamy periodic acid–Schiff-positive macrophages was found at the level of the intestinal mucosa (Fig. 1D and E). Based on these histological (intestinal and lymph node) findings, we eventually made a diagnosis of Whipple’s disease. We also considered that clinical manifestations such as arthralgia, systemic symptoms, diarrhoea and weight loss, as well as lymphadenopathies, were all compatible with Whipple’s disease. Arthralgia started a few years before the correct diagnosis, as previously reported [3]. The patient was started on intramuscular ceftriaxone (2 g/day) for 2 weeks followed by oral co-trimoxazole (800–160 mg twice a day). Since the patient was previously treated with immunosuppressive drugs, prednisone was increased to 25 mg/day and slowly tapered down, to reduce the risk of an immune reconstitution inflammatory syndrome [2]. His symptoms dramatically subsided. At last follow-up visit, 5 months after the beginning of therapy, he had no diarrhoea and had gained 15 kg. His blood count and ESR were normal, while his CRP level was still increased. Arthralgia persisted, requiring prednisone (10 mg/day). The diagnosis of GPA may be difficult when classical manifestations such as upper respiratory tract or renal involvement are absent. Our patient’s signs and symptoms could be interpreted as atypical manifestations of the disease: he had multi-organ manifestations compatible with vasculitis (palpable purpura, arthralgia, fever, gastrointestinal symptoms including chronic diarrhoea and marked weight loss), a granulomatous necrotizing process involving multiple lymph nodes and low titre PR3-ANCA. However, since ANCA can be positive in several conditions other than GPA [4], clinicians should not use them as diagnostic of vasculitis. As demonstrated by our case and by previous reports [5, 6], Whipple’s disease can mimic an atypical form of GPA, and in patients suffering from the disease C-ANCA may be present. Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript. Disclosure statement: The authors have declared no conflicts of interest. References 1 Biagi F , Balduzzi D , Delvino P et al. Prevalence of Whipple’s disease in North-Western Italy . Eur J Clin Microbiol Infect Dis 2015 ; 34 : 1347 – 8 . Google Scholar CrossRef Search ADS PubMed 2 Marth T , Moos V , Müller C , Biagi F , Schneider T. Tropheryma whipplei infection and Whipple’s disease . Lancet Infect Dis 2016 ; 16 : e12 – 21 . Google Scholar CrossRef Search ADS 3 Puéchal X. Whipple’s disease . Ann Rheum Dis 2013 ; 72 : 797 – 803 . Google Scholar CrossRef Search ADS PubMed 4 Gaffo AL. Diagnostic approach to ANCA-associated vasculitides . Rheum Dis Clin North Am 2010 ; 36 : 491 – 506 . Google Scholar CrossRef Search ADS PubMed 5 Relandison S , Fabre S , Colcombet C , Cohen JD , Jorgensen C. ANCA positive polyarthritis revealing Whipple’s disease . Clin Exp Rheumatol 2008 ; 26 (Suppl 49) : 154 . 6 Agard C , Brisseau JM , Grossi O et al. Two cases of atypical Whipple’s disease associated with cytoplasmic ANCA of undefined specificity . Scand J Rheumatol 2012 ; 41 : 246 – 8 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Nikiphorou, Elena;Norton, Sam;Young, Adam;Dixey, Josh;Walsh, David;Helliwell, Henrietta;Kiely, Patrick;Network, Early Rheumatoid Arthritis Study and the Early Rheumatoid Arthritis

2018 Rheumatology

doi: 10.1093/rheumatology/key066pmid: 29590474

Abstract Objectives To examine associations between BMI and disease activity, functional ability and quality of life in RA. Methods Data from two consecutive, similarly designed UK multicentre RA inception cohorts were used: the Early RA Study (ERAS) and the Early RA Network (ERAN). Recruitment figures/median follow-up for the ERAS and ERAN were 1465/10 years (maximum 25 years), and 1236/6 years (maximum 10 years), respectively. Standard demographic and clinical variables were recorded at baseline and annually. Multilevel piecewise longitudinal models with a change point at 2 years were used with the 28-joint DAS (DAS28), ESR, HAQ and 36-item Short Form Health Survey (SF-36) physical (PCS) and mental (MCS) components as dependent variables. BMI was examined in separate models as both continuous and categorical variables (based on World Health Organization definitions) and up to 5 years from disease onset. Results BMI data from 2386 newly diagnosed RA patients (11 348 measures) showed an increase in BMI of 0.27 U annually (95% CI 0.21, 0.33). Baseline obesity was associated with a significant reduction in the odds of achieving a low year 2 DAS28 [OR 0.52 (95% CI 0.41, 0.650)]. At year 2, HAQ and SF-36 PCS scores were significantly worse but not at year 5 in patients obese at baseline. Obesity at year 2 was associated with higher DAS28 scores at year 2, but not at year 5, and also associated with significantly higher HAQ and SF-36 PCS scores at years 2 and 5. Conclusion Obesity prevalence is rising in early RA and associates with worse disease activity, function and health-related quality of life, with a significant negative impact on achieving a low DAS28. The data argue strongly for obesity management to become central to treatment strategies in RA. DAS28, disease activity, early rheumatoid arthritis, rheumatoid arthritis Rheumatology key messages Obesity adversely affects disease activity, functional ability and quality of life in RA patients. Obesity management should form a central part of all treatment strategies for patients with RA. Introduction Obesity is increasing in prevalence [1], has been implicated as a risk factor for developing RA [2–5] and is an increasingly prevalent comorbidity seen on first presentation of RA [6]. There is growing recognition that common mechanistic pathways are shared between the inflammatory states mediated by obesity and those by inflammatory rheumatic diseases [7–10]. Indeed, the immunomodulatory properties of adipose tissue are suggestive of obesity being a low-grade, chronic inflammatory condition [11]. A recent meta-analysis indicates that the BMI category of obesity, but not overweight, reduces the chances of achieving minimal disease activity in people with RA compared with those with normal BMI [12]. However, there is contradictory evidence linking higher BMI and other adverse outcomes such as slower radiographic progression [13, 14] but higher rates of total joint replacement [15, 16]. These paradoxical data raise the question whether the negative impact of obesity on composite DAS is driven by inflammation (ESR and swollen joints) or patient-reported factors (tender joints and patient global assessment). Obesity has been associated with decreased health-related quality of life (HRQoL) and depression across various chronic conditions [17–21], but this has been less well studied in RA [15, 22, 23]. Using two RA inception cohorts, this study investigated the association between BMI and disease activity, functional ability and quality of life at diagnosis, in the short term at 2 years and in the medium term at 5 years. Methods Study design and patient recruitment The study used data from the Early RA Study (ERAS, 1986–2001) and Early RA Network (ERAN, 2002–12), two multicentre early RA inception cohorts recruiting, respectively, from 9 centres in England and 23 centres in England, Wales and Ireland. Information on the two cohorts has been previously described in detail [24]. ERAS and ERAN recruited a total of 2701 patients: ERAS, n = 1465, maximum follow-up 25 years and ERAN, n = 1236, maximum follow-up 11 years. Combined analysis of ERAS and ERAN is possible since they are consecutive inception cohorts with a similar design, including the variables captured, timing of assessments and patient recruitment, with a median time from symptom onset to first rheumatology outpatient visit being 6 months. All centres managed RA according to local practice, influenced by contemporary UK guidelines for the management of RA [25], with treatment choice and strategy at the discretion of the treating clinician [26]. The median time to first synthetic DMARD was 2 months after presentation in ERAS and 1 month after presentation in ERAN. Recruiting centres generally favoured SSZ as the first DMARD choice in ERAS, with a gradual switch to MTX being observed, such that SSZ and MTX were used in equal proportions at the start of ERAN (2002) and then MTX became the most frequent first-choice DMARD thereafter [26]. In ERAS, all patients were DMARD naïve, and in ERAN a small proportion (13.5%) of patients used synthetic DMARDs prior to baseline assessment. Ethical approval The ERAS study received ethical approval from the West Hertfordshire Local Research Ethics Committee and subsequently from the Caldicott Guardian. The ERAN study received ethical approval from the Trent Research Ethics Committee. No additional ethical approval was required for this study. Clinical, laboratory and radiographic data Standard demographic and clinical variables were recorded at baseline and repeated once between 3 and 6 months, again at 12 months and then annually until the patient left the study (deceased, moved away, declined) or the recruiting centre closed to follow-up. Variables recorded in both cohorts included patient demographics (age at disease onset, gender), baseline RF and/or anti-CCP, haemoglobin, ESR, smoking status (past, current, never) and the HAQ disability index [27]. Comorbidities were recorded at every visit and coded using the International Classification of Disease, 10th edition system. Height and weight were recorded at each visit in ERAN and converted to BMI. In ERAS, weight was recorded annually, although height was only available at baseline. Based on the World Health Organization definitions, patients were subsequently categorized into underweight (BMI <18.5), normal (BMI 18.5–24.99), overweight (BMI 25–29.99) and obese (BMI ⩾30). Based on preliminary analyses and supported by a previous meta-analysis [12], the normal and overweight BMI groups were combined and used as the reference group in the analysis since there were no substantive differences in outcomes between these groups. In ERAS, disease activity was calculated based on the original three-variable DAS [28], excluding patient global assessment and using a 44 joint count. In ERAN a four-variable 28-joint DAS (DAS28) ESR-based score was used [29]. ERAS DASs were converted to the DAS28 metric to allow combined analysis across cohorts [30]. Data on HRQoL were only available in ERAN, measured using the 36-item Short Form Health Survey (SF-36) [31]. The SF-36 consisted of eight domains with responses subsequently grouped into two higher-order constructs: a physical component summary (PCS) and a mental component summary (MCS). The PCS is based on domains assessing physical function, pain, physical role functioning and general health, whereas the MCS is based on domains assessing mental health, vitality, social functioning and social role functioning. Statistical analyses The impact of BMI on the DAS28, ESR, HAQ and SF-36 PCS and MCS scores at baseline and over time was assessed. Data for the two cohorts were combined for analysis and patients with a BMI available at baseline and at least one other time point were included. Due to the differing length of follow-up between cohorts, data analysis was restricted to 5 years to retain sufficient balance between the cohorts contributing data across time points. Standardized morbidity ratios (SMRs) were calculated, adjusted for age, gender and year of visit with respect to population rates of underweight, overweight or obese and obese using data from the Health Survey for England 1993–2013 [32]. As population data prior to 1993 were unavailable, data relating to visits before that time were excluded from the calculations. The association between treatment use and change in BMI was examined using a propensity score approach due to the likelihood of confounding by indication between BMI and treatment. Separate models were estimated for steroid use and DMARD on the change in BMI from baseline to 12 months. The analysis used augmented inverse probability of treatment weights, since this doubly robust estimator protects against potential model misspecification. Propensity scores were conditioned on age, gender, comorbidity, DAS28, HAQ, seropositive status, cohort, prior steroid or DMARD use and symptom duration. The covariates where observed to be well balanced between treatment groups after weighting by the propensity score. The analyses used longitudinal linear mixed effects regression models with a random intercept for each patient to account for repeated assessments. Separate models were estimated for each outcome (DAS28 and its components, HAQ and SF-36 PCS and MCS) to explore changes in BMI over time. ESR was log transformed for the modelling and back-transformed for presentation in the results. Changes in the outcome over time were accounted for by including covariates relating to a linear spline for time since baseline with a change point at 2 years. This allowed for different estimates of the average yearly change in the outcome between baseline and 2 years and between 2 and 5 years, which was necessary given that changes were non-linear over time capturing initial treatment response [33]. A random slope for each time covariate allowed the rate of change in the outcome to vary across patients. Missing outcome data were allowed under the assumption that data were missing at random conditional on the variables included in the model. BMI category (normal/overweight vs underweight and obese) was initially entered as a predictor in the models, reflecting the level at baseline. Interaction terms with the time covariates allowed the impact of BMI at baseline to moderate the rate of change in the outcome and allow for category-level estimates of outcomes at baseline, 2 and 5 years. Subsequently the BMI category at 2 years was entered into the model again, with interaction terms with the time covariates allowing estimates of the impact of BMI category on outcomes at 2 and 5 years. Prevalence rates for discrete DAS28 categories were estimated using the distributional approach [34]. All models controlled for potential confounding due to age at disease onset, gender, recruitment year, RF or anti-CCP positive at baseline, smoking status, haemoglobin, baseline DAS28 and the baseline level of the outcome (i.e. HAQ, SF-36 PCS, SF-36 MCS). Estimates are presented unadjusted and adjusted for putative confounders. The results in the main text are model-expected mean levels at baseline, 2 and 5 years based on complete case analysis for covariates, as sensitivity analysis using multiple imputation revealed minimal differences. The model-estimated yearly rate of change in the outcome by BMI category for the complete case and imputed data (data not shown). Results Baseline demographics and disease activity Patient demographics and clinical variables at baseline are shown in Table 1. Age, gender and serological status were similar across both cohorts. In ERAN the baseline mean BMI was higher, more patients were current smokers and the mean DAS was lower. The median HAQ was the same in both cohorts. Table 1 Demographic and clinical characteristics for ERAS and ERAN at the baseline visit Variable Cohort ERAS/ ERAN combined (N = 2701) ERAS (1986–2001) (n = 1465) ERAN (2002–12) (n = 1236) Age at disease onset, mean (s.d.), years 56 (14) 55 (15) 57 (14) Symptom duration, median (IQR), months 6 (8) 6 (7) 6 (9) Female, n (%) 1812 (67) 973 (66) 839 (68) RF and/or anti-CCP positive, n (%) N = 2513 N = 1456 N = 1057 1553 (62) 914 (63) 639 (60) BMI, mean (s.d.) N = 2386 N = 1266 N = 1120 26.5 (50) 26 (5) 27.6 (5.3) Smoker, n (%) N = 2124 N = 907 N = 1217     Current 602 (28) 199 (22) 403 (33)     Ex 511 (24) 175 (19) 336 (28) DAS, mean (s.d.) N = 2642 N = 1452 N = 1190 4.8 (1.4) 5.0 (1.2) 4.5 (1.6) MCS, mean (s.d.) — N = 950 47 (12) PCS, mean (s.d.) — N = 950 29 (12) HAQ, median (IQR) N = 2659 N = 1460 N = 1199 1 (1.1) 1 (1.3) 1 (1.1) Haemoglobin, mean (s.d.) N = 2687 N = 1460 N = 1227 13.6 (8.1) 12.6 (1.6) 14.7 (11.7) Erosions, n (%) N = 2555 N = 1433 N = 1122 698 (27) 368 (26) 330 (29) Steroid use pre-recruitment, n (%) — 0 (0) 125 (10) DMARD use pre-recruitment, n (%) — 0 (0) 168 (14) Variable Cohort ERAS/ ERAN combined (N = 2701) ERAS (1986–2001) (n = 1465) ERAN (2002–12) (n = 1236) Age at disease onset, mean (s.d.), years 56 (14) 55 (15) 57 (14) Symptom duration, median (IQR), months 6 (8) 6 (7) 6 (9) Female, n (%) 1812 (67) 973 (66) 839 (68) RF and/or anti-CCP positive, n (%) N = 2513 N = 1456 N = 1057 1553 (62) 914 (63) 639 (60) BMI, mean (s.d.) N = 2386 N = 1266 N = 1120 26.5 (50) 26 (5) 27.6 (5.3) Smoker, n (%) N = 2124 N = 907 N = 1217     Current 602 (28) 199 (22) 403 (33)     Ex 511 (24) 175 (19) 336 (28) DAS, mean (s.d.) N = 2642 N = 1452 N = 1190 4.8 (1.4) 5.0 (1.2) 4.5 (1.6) MCS, mean (s.d.) — N = 950 47 (12) PCS, mean (s.d.) — N = 950 29 (12) HAQ, median (IQR) N = 2659 N = 1460 N = 1199 1 (1.1) 1 (1.3) 1 (1.1) Haemoglobin, mean (s.d.) N = 2687 N = 1460 N = 1227 13.6 (8.1) 12.6 (1.6) 14.7 (11.7) Erosions, n (%) N = 2555 N = 1433 N = 1122 698 (27) 368 (26) 330 (29) Steroid use pre-recruitment, n (%) — 0 (0) 125 (10) DMARD use pre-recruitment, n (%) — 0 (0) 168 (14) Table 1 Demographic and clinical characteristics for ERAS and ERAN at the baseline visit Variable Cohort ERAS/ ERAN combined (N = 2701) ERAS (1986–2001) (n = 1465) ERAN (2002–12) (n = 1236) Age at disease onset, mean (s.d.), years 56 (14) 55 (15) 57 (14) Symptom duration, median (IQR), months 6 (8) 6 (7) 6 (9) Female, n (%) 1812 (67) 973 (66) 839 (68) RF and/or anti-CCP positive, n (%) N = 2513 N = 1456 N = 1057 1553 (62) 914 (63) 639 (60) BMI, mean (s.d.) N = 2386 N = 1266 N = 1120 26.5 (50) 26 (5) 27.6 (5.3) Smoker, n (%) N = 2124 N = 907 N = 1217     Current 602 (28) 199 (22) 403 (33)     Ex 511 (24) 175 (19) 336 (28) DAS, mean (s.d.) N = 2642 N = 1452 N = 1190 4.8 (1.4) 5.0 (1.2) 4.5 (1.6) MCS, mean (s.d.) — N = 950 47 (12) PCS, mean (s.d.) — N = 950 29 (12) HAQ, median (IQR) N = 2659 N = 1460 N = 1199 1 (1.1) 1 (1.3) 1 (1.1) Haemoglobin, mean (s.d.) N = 2687 N = 1460 N = 1227 13.6 (8.1) 12.6 (1.6) 14.7 (11.7) Erosions, n (%) N = 2555 N = 1433 N = 1122 698 (27) 368 (26) 330 (29) Steroid use pre-recruitment, n (%) — 0 (0) 125 (10) DMARD use pre-recruitment, n (%) — 0 (0) 168 (14) Variable Cohort ERAS/ ERAN combined (N = 2701) ERAS (1986–2001) (n = 1465) ERAN (2002–12) (n = 1236) Age at disease onset, mean (s.d.), years 56 (14) 55 (15) 57 (14) Symptom duration, median (IQR), months 6 (8) 6 (7) 6 (9) Female, n (%) 1812 (67) 973 (66) 839 (68) RF and/or anti-CCP positive, n (%) N = 2513 N = 1456 N = 1057 1553 (62) 914 (63) 639 (60) BMI, mean (s.d.) N = 2386 N = 1266 N = 1120 26.5 (50) 26 (5) 27.6 (5.3) Smoker, n (%) N = 2124 N = 907 N = 1217     Current 602 (28) 199 (22) 403 (33)     Ex 511 (24) 175 (19) 336 (28) DAS, mean (s.d.) N = 2642 N = 1452 N = 1190 4.8 (1.4) 5.0 (1.2) 4.5 (1.6) MCS, mean (s.d.) — N = 950 47 (12) PCS, mean (s.d.) — N = 950 29 (12) HAQ, median (IQR) N = 2659 N = 1460 N = 1199 1 (1.1) 1 (1.3) 1 (1.1) Haemoglobin, mean (s.d.) N = 2687 N = 1460 N = 1227 13.6 (8.1) 12.6 (1.6) 14.7 (11.7) Erosions, n (%) N = 2555 N = 1433 N = 1122 698 (27) 368 (26) 330 (29) Steroid use pre-recruitment, n (%) — 0 (0) 125 (10) DMARD use pre-recruitment, n (%) — 0 (0) 168 (14) BMI changes over time In total, 2386 individuals with data on BMI at baseline were included in the analysis (ERAS 1266, ERAN 1120). Over the 5 years of follow-up examined, BMI was recorded on a total of 11 348 occasions (ERAS 6582, ERAN 4766) relating to a mean of 4.8 occasions (range 1–7) per patient. The mean BMI at baseline was 25.5 in ERAS and 27.6 in ERAN. The mean BMI increased from disease onset to 5 years disease duration, with a quadratic trend providing the best fit to the data (Fig. 1). For both cohorts, BMI increased by 0.27 U/year (95% CI 0.21, 0.33), decelerating at a rate of −0.03 U/year (95% CI −0.04, −0.01). For a typical British woman (height 1.62 m) and man (height 1.75 cm) this relates to an average weight gain of 0.71 and 0.83 kg, respectively, in the first year of disease. Using a propensity score approach, steroid use was associated with an increase in BMI of 0.13 U by 12 months (P = 0.104; 95% CI −0.03, 0.29). While non-significant, this indicates that approximately half of the change in BMI during the first year may be attributable to steroid use. DMARD use was not associated with a change in BMI (0.01 U difference; P = 0.879; 95% CI −0.18, 0.21). Fig. 1 View largeDownload slide Observed BMI during follow up for ERAS and ERAN ERAS (triangles) and ERAN (circles) with model-estimated quadratic trends and 95% CIs (lines and shaded areas) shown. Fig. 1 View largeDownload slide Observed BMI during follow up for ERAS and ERAN ERAS (triangles) and ERAN (circles) with model-estimated quadratic trends and 95% CIs (lines and shaded areas) shown. The prevalence of obesity at baseline was 14.3% in ERAS and 25.7% in ERAN, representing an 80% increase in prevalence between the two cohorts [risk ratio 1.79 (95% CI 1.61, 2.02)]. The prevalence of obesity rose at years 2 and 5, respectively, in ERAS to 20.1 and 22.5% and in ERAN to 35.5 and 37.2% (Fig. 2). The prevalence of underweight at baseline was 2.4% in ERAS and 1.0% in ERAN, remaining relatively stable over the follow-up. Fig. 2 View largeDownload slide Change in the distribution of BMI categories for ERAS and ERAN over the first 5 years Fig. 2 View largeDownload slide Change in the distribution of BMI categories for ERAS and ERAN over the first 5 years SMRs for overweight and obesity indicated that rates across the period of follow-up were in line with the general population for England, adjusting for age, gender and calendar year of visit (SMR range 0.91–1.03; see supplementary Table S1, available at Rheumatology online). Underweight prevalence was significantly higher at 2 and 5 years, with adjusted rates increased by 76 and 123%, respectively. Impact of BMI at baseline on outcomes at baseline, 2 and 5 years Disease activity At baseline, DAS28 was significantly higher in patients in the obese BMI category (mean 4.78) compared with those in the normal/overweight category (mean 4.50) in both crude and adjusted analyses (P < 0.001) (see Table 2). Table 2 Baseline, 2 and 5 year outcomes by BMI category Model Time Normal/overweight Obese Underweight Mean LCL UCL Mean LCL UCL P-value* Mean LCL UCL P-value* DAS28 (N = 2386) Crude Baseline 4.49 4.43 4.55 4.71 4.60 4.83 0.001 5.10 4.72 5.48 0.002 2 years 3.47 3.39 3.55 3.83 3.67 3.98 0.000 3.88 3.38 4.37 0.115 5 years 3.77 3.68 3.85 3.85 3.67 4.03 0.409 3.72 3.16 4.29 0.886 Adjusted Baseline 4.50 4.44 4.56 4.78 4.66 4.89 0.000 4.67 4.27 5.08 0.414 2 years 3.53 3.44 3.62 3.85 3.68 4.03 0.001 3.89 3.29 4.50 0.243 5 years 3.81 3.71 3.90 3.85 3.64 4.05 0.727 3.35 2.68 4.01 0.182 HAQ (N = 2386) Crude Baseline 1.01 0.98 1.04 1.20 1.14 1.26 0.000 1.08 0.88 1.28 0.509 2 years 0.80 0.76 0.84 1.06 0.98 1.14 0.000 0.91 0.65 1.16 0.405 5 years 1.00 0.96 1.05 1.18 1.09 1.27 0.001 1.03 0.74 1.32 0.842 Adjusted Baseline 1.02 0.99 1.06 1.15 1.09 1.21 0.000 0.97 0.75 1.18 0.615 2 years 0.83 0.79 0.87 0.98 0.89 1.07 0.003 0.88 0.57 1.18 0.764 5 years 0.99 0.94 1.04 1.07 0.97 1.17 0.165 0.84 0.51 1.18 0.405 PCS (N = 1030) Crude Baseline 30.20 29.34 31.05 27.64 26.28 28.99 0.002 27.87 21.17 34.57 0.500 2 years 34.39 33.28 35.50 29.67 27.88 31.45 0.000 34.55 25.85 43.25 0.972 5 years 33.30 32.02 34.58 30.94 28.83 33.06 0.062 34.50 24.78 44.22 0.811 Adjusted Baseline 29.94 29.07 30.81 28.81 27.40 30.22 0.182 31.22 23.67 38.77 0.741 2 years 33.66 32.44 34.88 30.76 28.79 32.73 0.014 32.24 21.33 43.15 0.800 5 years 33.00 31.61 34.38 31.83 29.51 34.15 0.394 38.17 27.19 49.15 0.360 MCS (N = 1030) Crude Baseline 47.84 47.03 48.66 46.67 45.37 47.96 0.132 43.56 37.17 49.96 0.193 2 years 49.78 48.77 50.80 46.95 45.33 48.58 0.004 44.44 36.50 52.39 0.191 5 years 50.09 48.99 51.19 47.95 46.13 49.76 0.048 52.29 44.09 60.48 0.602 Adjusted Baseline 47.88 46.98 48.77 47.16 45.70 48.61 0.411 48.74 40.98 56.50 0.829 2 years 49.25 48.13 50.37 47.20 45.40 49.01 0.059 41.29 31.24 51.33 0.123 5 years 50.09 48.85 51.34 48.75 46.67 50.82 0.274 55.16 45.54 64.77 0.306 Model Time Normal/overweight Obese Underweight Mean LCL UCL Mean LCL UCL P-value* Mean LCL UCL P-value* DAS28 (N = 2386) Crude Baseline 4.49 4.43 4.55 4.71 4.60 4.83 0.001 5.10 4.72 5.48 0.002 2 years 3.47 3.39 3.55 3.83 3.67 3.98 0.000 3.88 3.38 4.37 0.115 5 years 3.77 3.68 3.85 3.85 3.67 4.03 0.409 3.72 3.16 4.29 0.886 Adjusted Baseline 4.50 4.44 4.56 4.78 4.66 4.89 0.000 4.67 4.27 5.08 0.414 2 years 3.53 3.44 3.62 3.85 3.68 4.03 0.001 3.89 3.29 4.50 0.243 5 years 3.81 3.71 3.90 3.85 3.64 4.05 0.727 3.35 2.68 4.01 0.182 HAQ (N = 2386) Crude Baseline 1.01 0.98 1.04 1.20 1.14 1.26 0.000 1.08 0.88 1.28 0.509 2 years 0.80 0.76 0.84 1.06 0.98 1.14 0.000 0.91 0.65 1.16 0.405 5 years 1.00 0.96 1.05 1.18 1.09 1.27 0.001 1.03 0.74 1.32 0.842 Adjusted Baseline 1.02 0.99 1.06 1.15 1.09 1.21 0.000 0.97 0.75 1.18 0.615 2 years 0.83 0.79 0.87 0.98 0.89 1.07 0.003 0.88 0.57 1.18 0.764 5 years 0.99 0.94 1.04 1.07 0.97 1.17 0.165 0.84 0.51 1.18 0.405 PCS (N = 1030) Crude Baseline 30.20 29.34 31.05 27.64 26.28 28.99 0.002 27.87 21.17 34.57 0.500 2 years 34.39 33.28 35.50 29.67 27.88 31.45 0.000 34.55 25.85 43.25 0.972 5 years 33.30 32.02 34.58 30.94 28.83 33.06 0.062 34.50 24.78 44.22 0.811 Adjusted Baseline 29.94 29.07 30.81 28.81 27.40 30.22 0.182 31.22 23.67 38.77 0.741 2 years 33.66 32.44 34.88 30.76 28.79 32.73 0.014 32.24 21.33 43.15 0.800 5 years 33.00 31.61 34.38 31.83 29.51 34.15 0.394 38.17 27.19 49.15 0.360 MCS (N = 1030) Crude Baseline 47.84 47.03 48.66 46.67 45.37 47.96 0.132 43.56 37.17 49.96 0.193 2 years 49.78 48.77 50.80 46.95 45.33 48.58 0.004 44.44 36.50 52.39 0.191 5 years 50.09 48.99 51.19 47.95 46.13 49.76 0.048 52.29 44.09 60.48 0.602 Adjusted Baseline 47.88 46.98 48.77 47.16 45.70 48.61 0.411 48.74 40.98 56.50 0.829 2 years 49.25 48.13 50.37 47.20 45.40 49.01 0.059 41.29 31.24 51.33 0.123 5 years 50.09 48.85 51.34 48.75 46.67 50.82 0.274 55.16 45.54 64.77 0.306 Crude and adjusted means are presented. *P-values compared with normal/overweight baseline BMI. UCL: upper control limit; LCL: lower control limit. Table 2 Baseline, 2 and 5 year outcomes by BMI category Model Time Normal/overweight Obese Underweight Mean LCL UCL Mean LCL UCL P-value* Mean LCL UCL P-value* DAS28 (N = 2386) Crude Baseline 4.49 4.43 4.55 4.71 4.60 4.83 0.001 5.10 4.72 5.48 0.002 2 years 3.47 3.39 3.55 3.83 3.67 3.98 0.000 3.88 3.38 4.37 0.115 5 years 3.77 3.68 3.85 3.85 3.67 4.03 0.409 3.72 3.16 4.29 0.886 Adjusted Baseline 4.50 4.44 4.56 4.78 4.66 4.89 0.000 4.67 4.27 5.08 0.414 2 years 3.53 3.44 3.62 3.85 3.68 4.03 0.001 3.89 3.29 4.50 0.243 5 years 3.81 3.71 3.90 3.85 3.64 4.05 0.727 3.35 2.68 4.01 0.182 HAQ (N = 2386) Crude Baseline 1.01 0.98 1.04 1.20 1.14 1.26 0.000 1.08 0.88 1.28 0.509 2 years 0.80 0.76 0.84 1.06 0.98 1.14 0.000 0.91 0.65 1.16 0.405 5 years 1.00 0.96 1.05 1.18 1.09 1.27 0.001 1.03 0.74 1.32 0.842 Adjusted Baseline 1.02 0.99 1.06 1.15 1.09 1.21 0.000 0.97 0.75 1.18 0.615 2 years 0.83 0.79 0.87 0.98 0.89 1.07 0.003 0.88 0.57 1.18 0.764 5 years 0.99 0.94 1.04 1.07 0.97 1.17 0.165 0.84 0.51 1.18 0.405 PCS (N = 1030) Crude Baseline 30.20 29.34 31.05 27.64 26.28 28.99 0.002 27.87 21.17 34.57 0.500 2 years 34.39 33.28 35.50 29.67 27.88 31.45 0.000 34.55 25.85 43.25 0.972 5 years 33.30 32.02 34.58 30.94 28.83 33.06 0.062 34.50 24.78 44.22 0.811 Adjusted Baseline 29.94 29.07 30.81 28.81 27.40 30.22 0.182 31.22 23.67 38.77 0.741 2 years 33.66 32.44 34.88 30.76 28.79 32.73 0.014 32.24 21.33 43.15 0.800 5 years 33.00 31.61 34.38 31.83 29.51 34.15 0.394 38.17 27.19 49.15 0.360 MCS (N = 1030) Crude Baseline 47.84 47.03 48.66 46.67 45.37 47.96 0.132 43.56 37.17 49.96 0.193 2 years 49.78 48.77 50.80 46.95 45.33 48.58 0.004 44.44 36.50 52.39 0.191 5 years 50.09 48.99 51.19 47.95 46.13 49.76 0.048 52.29 44.09 60.48 0.602 Adjusted Baseline 47.88 46.98 48.77 47.16 45.70 48.61 0.411 48.74 40.98 56.50 0.829 2 years 49.25 48.13 50.37 47.20 45.40 49.01 0.059 41.29 31.24 51.33 0.123 5 years 50.09 48.85 51.34 48.75 46.67 50.82 0.274 55.16 45.54 64.77 0.306 Model Time Normal/overweight Obese Underweight Mean LCL UCL Mean LCL UCL P-value* Mean LCL UCL P-value* DAS28 (N = 2386) Crude Baseline 4.49 4.43 4.55 4.71 4.60 4.83 0.001 5.10 4.72 5.48 0.002 2 years 3.47 3.39 3.55 3.83 3.67 3.98 0.000 3.88 3.38 4.37 0.115 5 years 3.77 3.68 3.85 3.85 3.67 4.03 0.409 3.72 3.16 4.29 0.886 Adjusted Baseline 4.50 4.44 4.56 4.78 4.66 4.89 0.000 4.67 4.27 5.08 0.414 2 years 3.53 3.44 3.62 3.85 3.68 4.03 0.001 3.89 3.29 4.50 0.243 5 years 3.81 3.71 3.90 3.85 3.64 4.05 0.727 3.35 2.68 4.01 0.182 HAQ (N = 2386) Crude Baseline 1.01 0.98 1.04 1.20 1.14 1.26 0.000 1.08 0.88 1.28 0.509 2 years 0.80 0.76 0.84 1.06 0.98 1.14 0.000 0.91 0.65 1.16 0.405 5 years 1.00 0.96 1.05 1.18 1.09 1.27 0.001 1.03 0.74 1.32 0.842 Adjusted Baseline 1.02 0.99 1.06 1.15 1.09 1.21 0.000 0.97 0.75 1.18 0.615 2 years 0.83 0.79 0.87 0.98 0.89 1.07 0.003 0.88 0.57 1.18 0.764 5 years 0.99 0.94 1.04 1.07 0.97 1.17 0.165 0.84 0.51 1.18 0.405 PCS (N = 1030) Crude Baseline 30.20 29.34 31.05 27.64 26.28 28.99 0.002 27.87 21.17 34.57 0.500 2 years 34.39 33.28 35.50 29.67 27.88 31.45 0.000 34.55 25.85 43.25 0.972 5 years 33.30 32.02 34.58 30.94 28.83 33.06 0.062 34.50 24.78 44.22 0.811 Adjusted Baseline 29.94 29.07 30.81 28.81 27.40 30.22 0.182 31.22 23.67 38.77 0.741 2 years 33.66 32.44 34.88 30.76 28.79 32.73 0.014 32.24 21.33 43.15 0.800 5 years 33.00 31.61 34.38 31.83 29.51 34.15 0.394 38.17 27.19 49.15 0.360 MCS (N = 1030) Crude Baseline 47.84 47.03 48.66 46.67 45.37 47.96 0.132 43.56 37.17 49.96 0.193 2 years 49.78 48.77 50.80 46.95 45.33 48.58 0.004 44.44 36.50 52.39 0.191 5 years 50.09 48.99 51.19 47.95 46.13 49.76 0.048 52.29 44.09 60.48 0.602 Adjusted Baseline 47.88 46.98 48.77 47.16 45.70 48.61 0.411 48.74 40.98 56.50 0.829 2 years 49.25 48.13 50.37 47.20 45.40 49.01 0.059 41.29 31.24 51.33 0.123 5 years 50.09 48.85 51.34 48.75 46.67 50.82 0.274 55.16 45.54 64.77 0.306 Crude and adjusted means are presented. *P-values compared with normal/overweight baseline BMI. UCL: upper control limit; LCL: lower control limit. At 2 years following DMARD initiation, the mean DAS28 decreased in all baseline BMI categories but remained significantly higher in patients with baseline obesity (mean 3.85) compared with those in the normal/overweight category (mean 3.53, P = 0.001). This association was lost at 5 years (P = 0.727). In the case of ESR, this was significantly higher at baseline in patients in the obese BMI category (mean 26.9) compared with those in the normal/overweight category (mean 22.4) in adjusted analyses (P < 0.001). At 2 years, the mean ESR decreased in all baseline BMI categories but remained significantly higher in patients with baseline obesity (mean 19.5) compared with those in the normal/overweight category (mean 14.3, P = 0.001), with this association persisting at 5 years (P = 0.028). No significant differences were observed between all other DAS28 components (swollen and tender joint counts and patient global assessment) and BMI categories at any of the time points after adjustment for potential confounders (see supplementary Table S2, available at Rheumatology online). In both ERAS and ERAN, DAS28 reduced from baseline to 2 years, with 624 (49.3%) in ERAS and 596 (53.2%) in ERAN achieving the EULAR DAS28 low disease activity (LDAS) target (DAS28 <3.2) on at least one occasion by 2 years. Fig. 3 illustrates, for ERAS and ERAN combined, the prevalence of discrete DAS28 status at baseline, 2 and 5 years in patients categorized at baseline as obese or normal/overweight. At 2 years, 32.1% of people in the obese category at baseline achieved LDAS compared with 43.2% in the normal/overweight category [risk ratio 0.74 (95% CI 0.67, 0.83)], but this difference was lost at 5 years (Fig. 3). Fig. 3 View largeDownload slide DAS28 categories at baseline, 2 and 5 years by baseline obesity status Fig. 3 View largeDownload slide DAS28 categories at baseline, 2 and 5 years by baseline obesity status Logistic regression indicated that being obese at baseline was related to a statistically significant 43% reduction in the odds of achieving LDAS on at least one occasion by year 2 compared with those in the normal/overweight category [OR 0.57 (95% CI 0.61, 0.90)]. After adjusting for potential confounders the effect was more pronounced [OR 0.52 (95% CI 0.41, 0.65)]. Functional ability and HRQoL HAQ scores at baseline, 2 and 5 years were significantly worse for those who were obese at baseline vs normal/overweight in the unadjusted analysis (all P < 0.001) (see Table 2). After adjustment, the magnitude of the difference at each time point was attenuated and remained significant at baseline and 2 years (P < 0.01). SF-36 PCS scores showed similar trends to HAQ but attenuated to non-significance except at 2 years in the adjusted model. SF-36 MCS scores at 2 and 5 years were significantly worse for those who were obese at baseline vs normal/overweight in the unadjusted analysis (all P < 0.01). However, after adjusting for potential confounders, the differences were attenuated and non-significant. Impact of BMI at 2 years on years 2 and 5 outcomes Obesity at 2 years was associated with a significantly higher DAS28 at 2 years compared with those in the normal/overweight category in both crude and adjusted models, but was attenuated and non-significant at 5 years (Table 3). Using discrete DAS28 categories, 32.6% in the obese category at 2 years had LDAS at 2 years compared with 40.7% in the normal/overweight category [risk ratio 0.80 (95% CI 0.72, 0.89)]. At 5 years, 33.6% in the obese category at 2 years had LDAS compared with 32.6% in the normal/overweight category [risk ratio 1.03 (95% CI 0.93, 1.15)]. Patients in the underweight category at year 2 also had a significantly higher DAS28 at year 2, but not year 5, compared with the normal/overweight category in both crude and adjusted models (Table 3). Table 3 Crude and adjusted means by BMI category at 2 years for outcomes at 2 and 5 years follow-up Model Time Normal/Overweight Obese Underweight Mean LCL UCL Mean LCL UCL P-value* Mean LCL UCL P-value* DAS28 (N = 2386) Crude 2 years 3.47 3.39 3.55 3.78 3.63 3.93 0.000 4.42 3.95 4.90 0.000 5 years 3.78 3.69 3.87 3.82 3.65 3.98 0.701 3.91 3.39 4.43 0.615 Adjusted 2 years 3.53 3.44 3.62 3.83 3.66 3.99 0.002 4.17 3.62 4.73 0.024 5 years 3.83 3.73 3.93 3.79 3.60 3.98 0.759 3.69 3.07 4.31 0.676 HAQ (N=2386) Crude 2 years 0.91 0.88 0.94 1.15 1.09 1.21 0.000 1.19 1.00 1.39 0.006 5 years 0.89 0.85 0.93 1.17 0.93 1.42 0.000 1.12 1.05 1.20 0.699 Adjusted 2 years 0.93 0.89 0.96 1.08 1.02 1.14 0.000 1.07 0.85 1.29 0.218 5 years 0.91 0.87 0.96 1.05 0.96 1.13 0.008 1.04 0.74 1.34 0.982 PCS (N = 1030) Crude 2 years 32.89 32.05 33.74 28.89 27.64 30.14 0.000 29.23 23.59 34.88 0.209 5 years 34.61 33.48 35.74 30.82 29.13 32.51 0.000 28.13 20.68 35.58 0.491 Adjusted 2 years 32.59 31.73 33.45 29.50 28.22 30.78 0.000 31.34 24.51 38.17 0.722 5 years 34.50 33.28 35.72 31.66 29.84 33.47 0.011 29.73 20.31 39.14 0.693 MCS (N = 1030) Crude 2 years 49.09 48.34 49.83 47.53 46.43 48.63 0.022 45.38 40.43 50.33 0.147 5 years 50.27 49.32 51.22 48.04 46.62 49.46 0.011 46.48 40.27 52.68 0.630 Adjusted 2 years 48.84 48.03 49.66 48.22 47.01 49.43 0.403 49.71 43.24 56.18 0.794 5 years 50.07 49.02 51.12 48.86 47.30 50.41 0.206 48.35 40.38 56.31 0.902 Model Time Normal/Overweight Obese Underweight Mean LCL UCL Mean LCL UCL P-value* Mean LCL UCL P-value* DAS28 (N = 2386) Crude 2 years 3.47 3.39 3.55 3.78 3.63 3.93 0.000 4.42 3.95 4.90 0.000 5 years 3.78 3.69 3.87 3.82 3.65 3.98 0.701 3.91 3.39 4.43 0.615 Adjusted 2 years 3.53 3.44 3.62 3.83 3.66 3.99 0.002 4.17 3.62 4.73 0.024 5 years 3.83 3.73 3.93 3.79 3.60 3.98 0.759 3.69 3.07 4.31 0.676 HAQ (N=2386) Crude 2 years 0.91 0.88 0.94 1.15 1.09 1.21 0.000 1.19 1.00 1.39 0.006 5 years 0.89 0.85 0.93 1.17 0.93 1.42 0.000 1.12 1.05 1.20 0.699 Adjusted 2 years 0.93 0.89 0.96 1.08 1.02 1.14 0.000 1.07 0.85 1.29 0.218 5 years 0.91 0.87 0.96 1.05 0.96 1.13 0.008 1.04 0.74 1.34 0.982 PCS (N = 1030) Crude 2 years 32.89 32.05 33.74 28.89 27.64 30.14 0.000 29.23 23.59 34.88 0.209 5 years 34.61 33.48 35.74 30.82 29.13 32.51 0.000 28.13 20.68 35.58 0.491 Adjusted 2 years 32.59 31.73 33.45 29.50 28.22 30.78 0.000 31.34 24.51 38.17 0.722 5 years 34.50 33.28 35.72 31.66 29.84 33.47 0.011 29.73 20.31 39.14 0.693 MCS (N = 1030) Crude 2 years 49.09 48.34 49.83 47.53 46.43 48.63 0.022 45.38 40.43 50.33 0.147 5 years 50.27 49.32 51.22 48.04 46.62 49.46 0.011 46.48 40.27 52.68 0.630 Adjusted 2 years 48.84 48.03 49.66 48.22 47.01 49.43 0.403 49.71 43.24 56.18 0.794 5 years 50.07 49.02 51.12 48.86 47.30 50.41 0.206 48.35 40.38 56.31 0.902 *Mean difference compared with normal/overweight. UCL: upper control limit; LCL: lower control limit. Table 3 Crude and adjusted means by BMI category at 2 years for outcomes at 2 and 5 years follow-up Model Time Normal/Overweight Obese Underweight Mean LCL UCL Mean LCL UCL P-value* Mean LCL UCL P-value* DAS28 (N = 2386) Crude 2 years 3.47 3.39 3.55 3.78 3.63 3.93 0.000 4.42 3.95 4.90 0.000 5 years 3.78 3.69 3.87 3.82 3.65 3.98 0.701 3.91 3.39 4.43 0.615 Adjusted 2 years 3.53 3.44 3.62 3.83 3.66 3.99 0.002 4.17 3.62 4.73 0.024 5 years 3.83 3.73 3.93 3.79 3.60 3.98 0.759 3.69 3.07 4.31 0.676 HAQ (N=2386) Crude 2 years 0.91 0.88 0.94 1.15 1.09 1.21 0.000 1.19 1.00 1.39 0.006 5 years 0.89 0.85 0.93 1.17 0.93 1.42 0.000 1.12 1.05 1.20 0.699 Adjusted 2 years 0.93 0.89 0.96 1.08 1.02 1.14 0.000 1.07 0.85 1.29 0.218 5 years 0.91 0.87 0.96 1.05 0.96 1.13 0.008 1.04 0.74 1.34 0.982 PCS (N = 1030) Crude 2 years 32.89 32.05 33.74 28.89 27.64 30.14 0.000 29.23 23.59 34.88 0.209 5 years 34.61 33.48 35.74 30.82 29.13 32.51 0.000 28.13 20.68 35.58 0.491 Adjusted 2 years 32.59 31.73 33.45 29.50 28.22 30.78 0.000 31.34 24.51 38.17 0.722 5 years 34.50 33.28 35.72 31.66 29.84 33.47 0.011 29.73 20.31 39.14 0.693 MCS (N = 1030) Crude 2 years 49.09 48.34 49.83 47.53 46.43 48.63 0.022 45.38 40.43 50.33 0.147 5 years 50.27 49.32 51.22 48.04 46.62 49.46 0.011 46.48 40.27 52.68 0.630 Adjusted 2 years 48.84 48.03 49.66 48.22 47.01 49.43 0.403 49.71 43.24 56.18 0.794 5 years 50.07 49.02 51.12 48.86 47.30 50.41 0.206 48.35 40.38 56.31 0.902 Model Time Normal/Overweight Obese Underweight Mean LCL UCL Mean LCL UCL P-value* Mean LCL UCL P-value* DAS28 (N = 2386) Crude 2 years 3.47 3.39 3.55 3.78 3.63 3.93 0.000 4.42 3.95 4.90 0.000 5 years 3.78 3.69 3.87 3.82 3.65 3.98 0.701 3.91 3.39 4.43 0.615 Adjusted 2 years 3.53 3.44 3.62 3.83 3.66 3.99 0.002 4.17 3.62 4.73 0.024 5 years 3.83 3.73 3.93 3.79 3.60 3.98 0.759 3.69 3.07 4.31 0.676 HAQ (N=2386) Crude 2 years 0.91 0.88 0.94 1.15 1.09 1.21 0.000 1.19 1.00 1.39 0.006 5 years 0.89 0.85 0.93 1.17 0.93 1.42 0.000 1.12 1.05 1.20 0.699 Adjusted 2 years 0.93 0.89 0.96 1.08 1.02 1.14 0.000 1.07 0.85 1.29 0.218 5 years 0.91 0.87 0.96 1.05 0.96 1.13 0.008 1.04 0.74 1.34 0.982 PCS (N = 1030) Crude 2 years 32.89 32.05 33.74 28.89 27.64 30.14 0.000 29.23 23.59 34.88 0.209 5 years 34.61 33.48 35.74 30.82 29.13 32.51 0.000 28.13 20.68 35.58 0.491 Adjusted 2 years 32.59 31.73 33.45 29.50 28.22 30.78 0.000 31.34 24.51 38.17 0.722 5 years 34.50 33.28 35.72 31.66 29.84 33.47 0.011 29.73 20.31 39.14 0.693 MCS (N = 1030) Crude 2 years 49.09 48.34 49.83 47.53 46.43 48.63 0.022 45.38 40.43 50.33 0.147 5 years 50.27 49.32 51.22 48.04 46.62 49.46 0.011 46.48 40.27 52.68 0.630 Adjusted 2 years 48.84 48.03 49.66 48.22 47.01 49.43 0.403 49.71 43.24 56.18 0.794 5 years 50.07 49.02 51.12 48.86 47.30 50.41 0.206 48.35 40.38 56.31 0.902 *Mean difference compared with normal/overweight. UCL: upper control limit; LCL: lower control limit. For HAQ and the SF-36 PCS, obesity at 2 years was associated with significantly worse scores at both 2 and 5 years in both the crude and adjusted models (all P < 0.05). However, while unadjusted differences in SF-36 MCS at 2 and 5 years were significantly worse for those who were obese at 2 years compared with those who were normal/overweight, after adjustment for potential confounders the differences were attenuated and non-significant. Discussion We report from two large unique UK inception cohorts of early RA recruited at the time of diagnosis and managed according to contemporary practice and followed for 5 years. Obesity not only was an increasingly prevalent comorbid condition at RA diagnosis from 1986 to 2012, but also weight continued to increase over the first 5 years after recruitment, with the prevalence of obesity growing at each year of follow-up. Thus whereas 14.3% of people at recruitment were obese in ERAS (1986–2002), 37.2% were obese at year 5 after enrolment into ERAN (2007–12). Across both inception cohorts, obesity had a significant negative impact on baseline and early year 2 composite DAS28 outcomes. This translated into those obese at baseline having a 48% reduction in the odds of achieving LDAS by year 2 compared with the normal/overweight category [OR 0.52 (95% CI 0.41, 0.65)], a difference that was both statistically significant and likely to be clinically important. These findings are generally supportive of those reported from previous cross-sectional analyses and meta-analyses [12, 35, 36]. Our study extends previous findings to demonstrate that baseline obesity is also associated with higher baseline and years 2 and 5 ESR, suggesting that the effect on DAS is at least driven in part by this component. Obesity might confound assessment of disease activity in RA through soft tissue (adiposity) around joints or effects on pain processing, such as reduced pressure pain thresholds and direct effects on patient global assessment of health [37, 38]. Previous studies have indicated that high BMI is associated with lower rates of radiographic progression after adjusting for DAS28 [14, 39], suggesting that DAS28 might overestimate disease activity in obese participants. This is supported by data showing that RA patients with obesity have lower rates of DAS28 remission but similar rates of low MRI-detected inflammation as patients without obesity, suggesting that obesity can bias composite disease activity measures [39]. We have found that associations of obesity with DAS28 were replicated with laboratory measures of inflammation (ESR), suggesting direct effects on inflammatory mechanisms. However, on exploring other DAS28 components (swollen and tender joint counts and patient global assessment) and their association with obesity, no significant associations were seen, suggesting that central pain sensitization is unlikely to be a main driver for the DAS28 in obese patients. The non-significant association between obesity and swollen joint count also suggests that obesity did not bias the clinical examination and recording of this component. Similarly, no association with obesity and the SF-36 MCS was found in the adjusted analyses. Our results contrast with those of other studies that suggest obesity is associated with increased pain sensitivity and central pain augmentation [37, 38]. Taken together these findings suggest an immediacy of effect on outcomes from obesity. We have found that baseline obesity has a negative effect on DAS28 and functional measures at baseline that persists in the short term to year 2 but is lost by year 5. Similarly in patients who were obese at year 2, worse outcomes were found at that time for DAS28, HAQ and SF-36 PCS and in short-term follow-up at year 5 for function and SF-36 PCS. This would be in keeping with the concept of a real-time effect of obesity, potentially mediated by adipokines, influencing inflammatory mediators, pain and other patient-reported outcomes with immediate measurable consequences [36]. It follows that strategies to encourage and support patients to lose weight at any stage of the disease should lead to immediate RA-specific benefits as well as longer-term cardiovascular and general health benefits that might also be expected in people without RA. Indeed, strategies to lose weight either by diet and exercise or bariatric surgery [40] have shown promise in suppressing RA disease activity. Our study has many strengths, including enrolment of patients with early RA over 3 decades, its longitudinal nature, large patient numbers and long patient follow-up. However, it also has limitations; importantly, that many patients migrated into the obesity category over time. This may have biased the observations relating to the impact of baseline obesity on 5 year outcomes. To determine whether this was likely, we also examined BMI at 2 years as a predictor of future outcome, noting broadly equivalent results. The HRQoL analysis was also limited by the availability of SF-36 data in ERAN but not ERAS. Finally, our models did not adjust for treatment use (e.g. glucocorticoids or DMARDs) at each time point, but instead controlled for disease activity, which captures the impact of treatment. This was considered the most appropriate approach, as our models assess the impact of BMI on future outcomes in general, which may be partially mediated by treatment since BMI may influence treatment decisions (e.g. steroid use) and BMI may itself be related to past treatment. As a result, it is difficult to draw strong inferences about whether the association between BMI and future outcome is due to BMI and associated factors or to differential treatment selection across the range of BMIs. That is, both explanations are likely but we cannot determine the magnitude of the effect via each pathway. In conclusion, we have demonstrated in early RA the increasing prevalence of obesity and its negative consequences on DAS28, achieving a treat-to-target LDAS goal, function and HRQoL outcomes in the short term. This effect is synchronous, with current obesity status at baseline and year 2 having an immediate and short-term effect not persisting in the medium-term. These data argue strongly for the screening and management of obesity to become a central part of all treatment strategies for patients with RA. Acknowledgements We are indebted to all the nurses and rheumatologists from both cohorts for their participation and contributions and to our study coordinator, Marie Hunt. ERAS: Dr Paul Davies and Lynn Hill (Chelmsford), Dr Andrew Gough, Dr Joe Devlin, Prof Paul Emery and Lynn Waterhouse (Birmingham), Dr David James and Helen Tate (Grimsby), Dr Peter Prouse and Cathy Boys (Basingstoke), Dr Peter Williams and Dora White (Medway), Helen Dart (Oswestry), Dr Nigel Cox and Sue Stafford (Winchester), Dr John Winfield (Sheffield) and Annie Seymour (St Albans). ERAN: Annie Seymour (City Hospital, St Albans), Dr Richard Williams, Karina Blunn and Jackie McDowell (Hereford County Hospital), Dr Peter Prouse and Sheryl Andrews (North Hampshire Hospital), Deborah Wilson (King’s Mill Hospital), Dr Malgorzata Magliano and Ursula Perks (Stoke Mandeville Hospital), Dr Amanda Coulson (Withybush General Hospital), Dr Andrew Hassle and Michele Kirwan (Haywood Hospital), Francesca Leone (St George’s Hospital), Dr Ciaran Dunne and Lindsey Hawley (Christchurch Hospital), Dr Paul Creamer, Julie Taylor and Wendy Wilmott (Southmead Hospital), Dr Sally Knights and Rebecca Rowland-Axe (Yeovil District Hospital), Dr Sandra Green and Dawn Simmons (Weston-Super-Mare General Hospital), Dr Joel David and Maureen Cox (Nuffield Orthopaedic Centre), Dr Marwan Bukhari and Bronwen Evans (Lancaster Royal Infirmary), Dr Michael Batley and Catherine Oram (Maidstone Hospital) and Dr Tanya Potter (Coventry University Hospital). The authors acknowledge the Office of National Statistics, the Medical Research Information Service and the Health & Social Care Information Service, as well as Hospital Episode Statistics and the National Joint Registry for providing the orthopaedic episode data. ERAN received funding from the British Society of Rheumatology and a grant from the Healthcare Commission. E.N. received a grant from the Essex & Hertfordshire Clinical Research Network. Funding: This work was supported by grants from the Arthritis Research Campaign and the British United Providence Association (BUPA) Foundation for ERAS and also by the National Institute for Health Research Clinical Research Network Essex & Hertfordshire. Disclosure statement: D.A.W. received an investigator-led grant from Pfizer addressing pain in early RA. All other authors have declared no conflicts of interest. Supplementary data Supplementary data are available at Rheumatology online. References 1 Forouzanfar MH , Afshin A , Alexander LT et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015 . Lancet 2016 ; 388 : 1659 – 724 . Google Scholar CrossRef Search ADS PubMed 2 Voigt LF , Koepsell TD , Nelson JL , Dugowson CE , Daling JR. Smoking, obesity, alcohol consumption, and the risk of rheumatoid arthritis . Epidemiology 1994 ; 5 : 525 – 32 . Google Scholar PubMed 3 Lu B , Hiraki LT , Sparks JA et al. Being overweight or obese and risk of developing rheumatoid arthritis among women: a prospective cohort study . Ann Rheum Dis 2014 ; 73 : 1914 – 22 . Google Scholar CrossRef Search ADS PubMed 4 Crowson CS , Matteson EL , Davis JM , Gabriel SE. Contribution of obesity to the rise in incidence of rheumatoid arthritis . Arthritis Care Res 2013 ; 65 : 71 – 7 . Google Scholar CrossRef Search ADS 5 de Hair MJH , Landewé RBM , van de Sande MGH et al. Smoking and overweight determine the likelihood of developing rheumatoid arthritis . Ann Rheum Dis 2013 ; 72 : 1654 – 8 . Google Scholar CrossRef Search ADS PubMed 6 Nikiphorou E , Norton S , Carpenter L et al. Secular changes in clinical features at presentation of rheumatoid arthritis . Arthritis Care Res 2017 ; 69 : 21 – 7 . Google Scholar CrossRef Search ADS 7 Di Minno MND , Peluso R , Iervolino S et al. Hepatic steatosis, carotid plaques and achieving MDA in psoriatic arthritis patients starting TNF-α blockers treatment: a prospective study . Arthritis Res Ther 2012 ; 14 : R211 . Google Scholar CrossRef Search ADS PubMed 8 Russolillo A , Iervolino S , Peluso R et al. Obesity and psoriatic arthritis: from pathogenesis to clinical outcome and management . Rheumatology 2013 ; 52 : 62 – 7 . Google Scholar CrossRef Search ADS PubMed 9 di Minno MND , Peluso R , Iervolino S et al. Obesity and the prediction of minimal disease activity: a prospective study in psoriatic arthritis . Arthritis Care Res 2013 ; 65 : 141 – 7 . Google Scholar CrossRef Search ADS 10 Di Minno MND , Ambrosino P , Peluso R et al. Lipid profile changes in patients with rheumatic diseases receiving a treatment with TNF-α blockers: a meta-analysis of prospective studies . Ann Med 2014 ; 46 : 73 – 83 . Google Scholar CrossRef Search ADS PubMed 11 Cottam DR , Mattar SG , Barinas-Mitchell E et al. The chronic inflammatory hypothesis for the morbidity associated with morbid obesity: implications and effects of weight loss . Obes Surg 2004 ; 14 : 589 – 600 . Google Scholar CrossRef Search ADS PubMed 12 Lupoli R , Pizzicato P , Scalera A et al. Impact of body weight on the achievement of minimal disease activity in patients with rheumatic diseases: a systematic review and meta-analysis . Arthritis Res Ther 2016 ; 18 : 297 . Google Scholar CrossRef Search ADS PubMed 13 Westhoff G , Rau R , Zink A. Radiographic joint damage in early rheumatoid arthritis is highly dependent on body mass index . Arthritis Rheum 2007 ; 56 : 3575 – 82 . Google Scholar CrossRef Search ADS PubMed 14 Baker JF , Ostergaard M , George M et al. Greater body mass independently predicts less radiographic progression on X-ray and MRI over 1–2 years . Ann Rheum Dis 2014 ; 73 : 1923 – 8 . Google Scholar CrossRef Search ADS PubMed 15 Wolfe F , Michaud K. Effect of body mass index on mortality and clinical status in rheumatoid arthritis . Arthritis Care Res 2012 ; 64 : 1471 – 9 . Google Scholar CrossRef Search ADS 16 Nikiphorou E , Carpenter L , Norton S et al. Can rheumatologists predict eventual need for orthopaedic intervention in patients with rheumatoid arthritis? Results of a systematic review and analysis of two UK inception cohorts . Curr Rheumatol Rep 2017 ; 19 : 12 . Google Scholar CrossRef Search ADS PubMed 17 Schoffman DE , Wilcox S , Baruth M. Association of body mass index with physical function and health-related quality of life in adults with arthritis . Arthritis 2013 ; 2013 : 190868 . Google Scholar CrossRef Search ADS PubMed 18 Luppino FS , de Wit LM , Bouvy PF et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies . Arch Gen Psychiatry 2010 ; 67 : 220 – 9 . Google Scholar CrossRef Search ADS PubMed 19 Popkin BM , Kim S , Rusev ER , Du S , Zizza C. Measuring the full economic costs of diet, physical activity and obesity-related chronic diseases . Obes Rev 2006 ; 7 : 271 – 93 . Google Scholar CrossRef Search ADS PubMed 20 Volz A , Schmid J-P , Zwahlen M et al. Predictors of readmission and health related quality of life in patients with chronic heart failure: a comparison of different psychosocial aspects . J Behav Med 2011 ; 34 : 13 – 22 . Google Scholar CrossRef Search ADS PubMed 21 Afendy A , Kallman JB , Stepanova M et al. Predictors of health-related quality of life in patients with chronic liver disease . Aliment Pharmacol Ther 2009 ; 30 : 469 – 76 . Google Scholar CrossRef Search ADS PubMed 22 García-Poma A , Segami MI , Mora CS et al. Obesity is independently associated with impaired quality of life in patients with rheumatoid arthritis . Clin Rheumatol 2007 ; 26 : 1831 – 5 . Google Scholar CrossRef Search ADS PubMed 23 Colmegna I , Hitchon CA , Bardales MCB , Puri L , Bartlett SJ. High rates of obesity and greater associated disability among people with rheumatoid arthritis in Canada . Clin Rheumatol 2016 ; 35 : 457 – 60 . Google Scholar CrossRef Search ADS PubMed 24 Nikiphorou E , Carpenter L , Morris S et al. Hand and foot surgery rates in rheumatoid arthritis have declined from 1986 to 2011, but large-joint replacement rates remain unchanged . Arthritis Rheumatol 2014 ; 66 : 1081 – 9 . Google Scholar CrossRef Search ADS PubMed 25 Joint Working Group of the British Society for Rheumatology and the Research Unit of the Royal College of Physicians . Guidelines and audit measures for the specialist supervision of patients with rheumatoid arthritis . J R Coll Physicians Lond 1992 ; 26 : 76 – 82 . PubMed 26 Young A , Dixey J , Williams P et al. An evaluation of the strengths and weaknesses of a register of newly diagnosed rheumatoid arthritis, 1986-2010 . Rheumatology 2011 ; 50 : 176 – 83 . Google Scholar CrossRef Search ADS PubMed 27 Fries JF , Spitz P , Kraines RG , Holman HR. Measurement of patient outcome in arthritis . Arthritis Rheum 1980 ; 23 : 137 – 45 . Google Scholar CrossRef Search ADS PubMed 28 van der Heijde DM , van ’t Hof M , van Riel PL , van de Putte LB. Development of a disease activity score based on judgment in clinical practice by rheumatologists . J Rheumatol 1993 ; 20 : 579 – 81 . Google Scholar PubMed 29 Prevoo ML , van ’t Hof MA , Kuper HH et al. Modified disease activity scores that include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of patients with rheumatoid arthritis . Arthritis Rheum 1995 ; 38 : 44 – 8 . Google Scholar CrossRef Search ADS PubMed 30 van Gestel AM , Haagsma CJ , van Riel PL. Validation of rheumatoid arthritis improvement criteria that include simplified joint counts . Arthritis Rheum 1998 ; 41 : 1845 – 50 . Google Scholar CrossRef Search ADS PubMed 31 Jenkinson C , Stewart-Brown S , Petersen S , Paice C. Assessment of the SF-36 version 2 in the United Kingdom . J Epidemiol Community Health 1999 ; 53 : 46 – 50 . Google Scholar CrossRef Search ADS PubMed 32 Neave A. Health Survey for England, 2015: trend tables . London : NHS Digital , 2016 . 33 Norton S , Sacker A , Dixey J et al. Trajectories of functional limitation in early rheumatoid arthritis and their association with mortality . Rheumatology 2013 ; 52 : 2016 – 24 . Google Scholar CrossRef Search ADS PubMed 34 Sauzet O , Kleine M. Distributional estimates for the comparison of proportions of a dichotomized continuous outcome . Stata J 2016 ; 16 : 880 – 99 . 35 Finckh A , Turesson C. The impact of obesity on the development and progression of rheumatoid arthritis . Ann Rheum Dis 2014 ; 73 : 1911 – 3 . Google Scholar CrossRef Search ADS PubMed 36 Liu Y , Hazlewood GS , Kaplan GG , Eksteen B , Barnabe C. Impact of obesity on remission and disease activity in rheumatoid arthritis: a systematic review and meta-analysis . Arthritis Care Res 2017 ; 69 : 157 – 65 . Google Scholar CrossRef Search ADS 37 McWilliams DF , Zhang W , Mansell JS et al. Predictors of change in bodily pain in early rheumatoid arthritis: an inception cohort study . Arthritis Care Res 2012 ; 64 : 1505 – 13 . Google Scholar CrossRef Search ADS 38 Ahmed S , Magan T , Vargas M , Harrison A , Sofat N. Use of the painDETECT tool in rheumatoid arthritis suggests neuropathic and sensitization components in pain reporting . J Pain Res 2014 ; 7 : 579 – 88 . Google Scholar PubMed 39 George MD , Østergaard M , Conaghan PG , Emery P , Baker DG , Baker JF. Obesity and rates of clinical remission and low MRI inflammation in rheumatoid arthritis . Ann Rheum Dis 2017 ; 76 : 1743 – 6 . Google Scholar CrossRef Search ADS PubMed 40 Sparks JA , Halperin F , Karlson JC , Karlson EW , Bermas BL. Impact of bariatric surgery on patients with rheumatoid arthritis . Arthritis Care Res 2015 ; 67 : 1619 – 26 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)