BRIEF PAPER Clinical and Experimental Rheumatology 2019; 37: 858-861. Efficacy and adverse ABSTRACT (AZA) are common first effects of methotrexate Objective. To study the efficacy in terms line therapies in ASyS patients (4), but of muscle strength, and corticosteroid mycophenolate (5, 6), the calcineurin compared with azathioprine tapering as well as the prevalence of inhibitors, cyclosporine and tacrolimus in the antisynthetase adverse effects in patients with the an- (7-9), as well as cyclosphospamide syndrome tisynthetase syndrome (ASyS) treated (10) and rituximab (11, 12), have also with azathioprine (AZA) compared to been used for the treatment of these pa- M. Casal-Dominguez1,2, those treated with methotrexate (MTX). tients with good results. I. Pinal-Fernandez1-3, J. Huapaya2, Methods. We compared the clinical The high risk of patients with the ASyS J. Albayda2, J.J. Paik2, C. Johnson2, outcomes in ASyS patients treated with developing ILD makes MTX use con- L. Silhan2, A.L. Mammen1,2, AZA versus MTX including change in troversial due its potential to induce S.K. Danoff2, L. Christopher-Stine2 corticosteroid dose, strength, and cre- hypersensitivity pneumonitis, which atine kinase (CK) as well as the preva- may be mistaken for ILD related to the 1National Institute of Arthritis and lence of adverse effects. underlying ASyS. Unlike ASyS associ- Musculoskeletal and Skin Diseases, Results. Among 169 patients with ated ILD, MTX pneumonitis is typically National Institutes of Health, Bethesda, MD; ASyS, 102 were treated at some point reversible with MTX cessation. It has 2 Johns Hopkins University School of exclusively with either AZA or MTX been suggested that MTX may be more Medicine, Baltimore, MD, USA; (± corticosteroids). There were no sig- 3Faculty of Health Sciences, Universitat beneficial than AZA in some groups Oberta de Catalunya, Barcelona, Spain. nificant differences in the rate of mus- of patients who are refractory to pred- Maria Casal-Dominguez, MD, PhD* cle strength recovery, CK decrease or nisone (13). Although MTX and azathi- Iago Pinal-Fernandez, MD, PhD* corticosteroid tapering between those oprine are two of the most widely used Julio Huapaya, MD ASyS patients treated with MTX versus immunosuppressant drugs for the ASyS, Jemima Albayda, MD AZA. The prevalence of adverse events the efficacy to treat the manifestations Julie J. Paik, MD in patients treated with AZA and MTX of the disease, comparative efficacy as Cheilonda Johnson, MD was similar (29% vs. 25%, p>0.05); el- steroid-sparing drugs and secondary ef- Leann Silhan, MD evated liver enzymes (17% AZA vs. 12% fects are, to a large extent, unknown. Andrew L. Mammen, MD, PhD Sonye K. Danoff, MD, PhD MTX) and gastrointestinal involvement Our main objective was to study the Lisa Christopher-Stine, MD, MPH (10% AZA vs. 8% MTX) were the most differences in muscle strength and *These authors contributed equally. common adverse events. While no pa- changes in the dose of corticosteroids, Please address correspondence to: tients treated with AZA developed lung as well as the profile of adverse effects Dr Lisa Christopher-Stine, complications, two of the patients treat- between ASyS patients treated with 5200 Eastern Avenue, ed with MTX experienced reversible AZA versus those treated with MTX. MFL Center Tower 4500, pneumonitis with MTX cessation. Baltimore, MD 21224, USA. Conclusion. AZA and MTX showed Materials and methods E-mail: [email protected] similar efficacy and adverse events In this longitudinal cohort study, we and to: Dr Sonye K. Danoff, in patients with ASyS. Pneumonitis is included all Johns Hopkins Myositis th 1800 E. Monument Street, 5 Floor, a rare but important event in patients Center patients who were positive for Baltimore, MD 21287, USA. receiving MTX. one of the ASyS antibodies (anti-Jo1, E-mail: [email protected] anti-PL7, anti-PL12, anti-OJ or anti-EJ) Received on January 28, 2019; accepted in revised form on April 1, 2019. Introduction and presented with at least two of the The antisynthetase syndrome (ASyS), following clinical manifestations: my- © Copyright Clinical and Experimental Rheumatology 2019. first described as an entity in 1990 (1), ositis, ILD, polyarthritis or mechanic’s is characterised by the presence of an hands. All the treatments administered Key words: antisynthetase syndrome, antisynthetase antibody which targets at each clinical evaluation were re- methotrexate, azathioprine, myositis cytoplasmic enzymes that catalyse the corded, and those patients treated with formation of the aminoacyl-tRNA com- AZA or MTX without concomitant use Funding: this research was supported in plex. Clinically, this syndrome is char- of other steroid sparing agents were in- part by the Intramural Research Program acterised by myositis, interstitial lung cluded for analysis. The sera from all of the National Institute of Arthritis and disease (ILD) or both. Other features patients was screened for anti-Jo1, anti- Musculoskeletal and Skin Diseases of the including Raynaud’s phenomenon, ar- PL7, anti-PL12, anti-EJ, and anti-OJ by National Institutes of Health. thritis, fever and mechanic’s hands are ELISA, line blotting (Euroline Myositis S.K. Danoff, L. Christopher-Stine and the also common clinical features of the Profile 4; Euroimmun), by immunopre- Myositis Research Database are supported by The Huayi and Siuling Zhang Discovery ASyS syndrome (2, 3). cipitation at the Oklahoma Medical Re- Fund. Corticosteroids are considered first line search Foundation and/or using Quest I. Pinal-Fernandez’ research is supported by treatment in the ASyS, but most of the Diagnostics myositis panels. a Fellowship from the Myositis Association. time, other immunosuppressive agents This study was approved by the Johns Competing interests: see page 861. are needed. Methotrexate (MTX) or Hopkins Institutional Review Board,

858 Clinical and Experimental Rheumatology 2019 Efficacy and adverse effects of MTXvs. AZA in ASyS / M. Casal-Dominguez et al. and written informed consent was ob- Table I. Adverse effects of methotrexate and azathioprine in patients with the antisynthetase tained from each participant. syndrome.

The change in strength, CK and dose of Azathioprine Methotrexate p-value corticosteroids during the period that (n=89) (n=52) patients were exposed exclusively to AZA or MTX (± corticosteroids) were Elevated liver function tests 17% (15) 12% (6) 0.4 analysed using multilevel regression Gastrointestinal 10% (9) 8% (4) 0.8 Nausea 7% (6) 8% (4) 1.0 models adjusted for age at onset, sex, Diarrhoea 2% (2) 0% (0) 0.5 race, dose of corticosteroids, type of Abdominal pain 1% (1) 0% (0) 1.0 antisynthetase antibody and time from Leukopenia 4% (4) 2% (1) 0.7 the onset to the clinical evaluation. Pancytopenia 1% (1) 2% (1) 1.0 Other adverse effects 6% (5) 13% (7) 0.1 At each visit, arm abduction and hip Methotrexate pneumonitis 0% (0) 4% (2) 0.1 flexion strength, were evaluated by the Rate of adverse effects 29% (26) 25% (13) 0.5 examining physician using the Medical Research Council (MRC) scale. This Dichotomous variables were compared using Chi-squared or Fisher’s test as appropriate. scale was transformed to Kendall’s 0–10 scale for analysis purposes as pre- using Stata/MP 14.1. A two-sided (10% AZA vs. 8% MTX) and cytope- viously described (14). Several investi- p-value of 0.05 or less was considered nias (6% AZA vs. 4% MTX), but none gators examined the patients, but serial significant with no correction for mul- of these were significantly different be- strength measurements for each patient tiple comparisons. tween both drugs. Of note, most report- were made by the same physician. ed adverse effects were mild. While no Adverse effects were recorded as report- Results patient with AZA experienced pulmo- ed by the attending physician. Accord- Of 169 patients with the ASyS (73% nary adverse effects related to the use of ingly, laboratory abnormalities, like el- women), 124 of them were positive for the immunosuppressant treatment, four evation of the liver enzymes, leukope- anti-Jo1, 23 for anti-PL12, 16 for anti- patients treated with MTX (8%) report- nia or pancytopenia were based on the PL7 and 3 for anti-EJ and anti-OJ re- ed pulmonary events (p=0.02) but only normality cut-off of the corresponding spectively. Of these patients, 63 (37%) two presented clear evidence of MTX facility where the tests were performed. were treated with AZA exclusively, 26 pneumonitis (p=0.1). These two pa- Also, MTX associated pneumonitis was (15%) were treated with MTX exclu- tients did not have pre-existing lung in- defined by the occurrence of cough or sively and 26 (15%) were treated with volvement. One had cough and lung CT dyspnea in a time course consistent both AZA and MTX at some point of involvement that reverted rapidly after with exposure to MTX which resolved their evolution (total of 115 patients). MTX discontinuation and the other one with stopping this medication. All the The average length of exposure to these was challenged twice with MTX de- episodes of possible MTX pneumonitis medications was 24 months for AZA veloping cough and shortness of breath were reviewed by three of the authors and 29 months for MTX. In general, that reverted quickly after stopping the (SD, MCD and IPF). The probability of AZA was administered to patients with drug both times. (Table I). Both patients the adverse effect was quantitated using less muscle involvement (lower CK and had a Naranja’s score (15) of 5 which Naranjo’s method (15). higher strength) and more severe lung corresponds to a probable adverse ef- Dichotomous variables were expressed involvement (lower FVC) while MTX fect. Complementarily, the two other as percentages and absolute frequen- was given to patients with milder lung patients that reported pulmonary events cies, and continuous features were re- involvement. MTX and AZA in combi- were patients with pre-existing ILD re- ported as means and standard deviations nation were used in patients with more porting worsening of their respiratory (SD). Pairwise comparisons for dichot- severe muscle disease (lower strength symptoms (one cough and one dysp- omous variables between groups were and higher CK). Patients treated with nea) during MTX treatment. However, made using chi-square test or Fisher’s MTX were mostly white and presented the time course was considered incon- exact test, as appropriate. Student’s t- anti-Jo1 autoantibodies more common- sistent with MTX pneumonitis and test was used to compare continuous ly than the other treatment groups (Sup- there were no objective tests available variables among groups. CK, a highly plementary Table I). to show worsening of the ILD. positively skewed variable, was com- Twenty-nine percent of all the patients Of the 115 patients treated with AZA or pared using Wilcoxon rank-sum test for who were treated with AZA showed MTX, 102 received either of the drugs the univariate analysis and transformed adverse effects to this drug, compared combined with no other immunosup- through a base-10 logarithm for regres- with 25% of the patients that were treat- pressant drug than corticosteroids (59 sion analysis. Locally weighted regres- ed with MTX (p>0.05). The most com- AZA, 20 MTX and 23 AZA and MTX sion was applied to analyse graphically mon adverse effects with both drugs at different time points). These patients the evolution of the strength and dose were elevated liver function tests (17% accounted for 450 visits under treatment of corticostestoids over time. AZA vs. 12% MTX), gastrointestinal with AZA or MTX ± corticosteroids All statistical analyses were performed symptoms such as nausea and diarrhoea (mean of 4.4 visits per patient) that were

Clinical and Experimental Rheumatology 2019 859 Efficacy and adverse effects of MTXvs. AZA in ASyS / M. Casal-Dominguez et al. A

B

Fig. 1. Strength recovery (A) and corticosteroid tapering (B) in patients with the antisynthetase syndrome treated with azathioprine and methotrexate.

860 Clinical and Experimental Rheumatology 2019 Efficacy and adverse effects of MTXvs. AZA in ASyS / M. Casal-Dominguez et al. used to compare the rate of change of ily in a routine manner for all patients. 6. SWIGRIS JJ, OLSON AL, FISCHER A et al.: strength, and corticosteroid tapering. Moreover, the small number of patients Mycophenolate mofetil is safe, well toler- ated, and preserves lung function in patients There were no significant differences in each group precludes a cautious in- with connective tissue disease-related inter- in the rate of muscle strength recovery terpretation of our results. stitial lung disease. Chest 2006; 130: 30-6. (p=0.9), CK decrease (p=0.6) or corti- In conclusion, in our real-world clini- 7. KOTANI T, TAKEUCHI T, MAKINO S et al.: costeroid tapering (p=0.9) during treat- cal experience, we found that compared Combination with corticosteroids and cyclo- sporin-A improves pulmonary function test ment with AZA or MTX (Fig. 1). with AZA, MTX had a similar preva- results and chest HRCT findings in dermato- lence of adverse effects and efficacy. myositis patients with acute/subacute inter- Discussion MTX pneumonitis occurred in 4% of stitial pneumonia. Clin Rheumatol 2011; 30: This study demonstrated that MTX and patients started on this medication, but 1021-8. 8. ODDIS CV, SCIURBA FC, ELMAGD KA, STAR- AZA are comparable in terms of rate was entirely reversible with stopping ZL TE: Tacrolimus in refractory of muscle strength recovery, rate of therapy, thus, attention to this potential with interstitial lung disease. Lancet 1999; corticosteroid tapering, and rate of CK adverse event is important with rapid 353: 1762-3. 9. WILKES MR, SEREIKA SM, FERTIG N, LUCAS decrease with similar rates of adverse discontinuation of therapy if symptoms MR, ODDIS CV: Treatment of antisynthetase- events. We did identify two episodes of occur. associated interstitial lung disease with tac- MTX pneumonitis which reversed with rolimus. Arthritis Rheum 2005; 52: 2439-46. discontinuation of therapy. Competing interests 10. YAMASAKI Y, YAMADA H, YAMASAKI M et al.: Intravenous cyclophosphamide therapy MTX has been reported to cause pneu- S. Danoff has received grant/research for progressive interstitial pneumonia in pa- monitis in 4–8% of the patients exposed support from Bristol-Meyers Squibb, tients with polymyositis/. to this drug (16) and this may dissuade Genentech/Roche and Boehringer- Rheumatology (Oxford) 2007; 46: 124-30. clinicians from prescribing MTX in pa- Inghelheim; L. Christopher-Stine has 11. ANDERSSON H, SEM M, LUND MB et al.: Long-term experience with rituximab in anti- tients with ASyS autoantibodies or pre- received royalties related to the licen- synthetase syndrome-related interstitial lung existing ILD (17). Our study confirms sure of anti-HMGCR antibody testing disease. Rheumatology (Oxford) 2015; 54: previous data regarding the low preva- to Inova Diagnostics, and receives re- 1420-8. lence of MTX pneumonitis (4%). search support from CSL Behring and 12. KEIR GJ, MAHER TM, MING D et al.: Rituxi- mab in severe, treatment-refractory intersti- Some authors have suggested an in- Novartis; the other co-authors have tial lung disease. Respirology 2014; 19: 353- creased efficacy of MTX over AZA in declared no competing interests. 9. selected groups of patients (13). Our 13. JOFFE MM, LOVE LA, LEFF RL et al.: Drug References therapy of the idiopathic inflammatory myo- study found that MTX was comparable pathies: predictors of response to prednisone, 1. MARGUERIE C, BUNN CC, BEYNON HL et to AZA in terms of efficacy in patients azathioprine, and methotrexate and a com- al.: Polymyositis, and au- parison of their efficacy. Am J Med 1993; 94: with the ASyS. toantibodies to aminoacyl-tRNA synthetase 379-87. The data that we report is based on a enzymes. Q J Med 1990; 77: 1019-38. 14. PINAL-FERNANDEZ I, CASAL-DOMINGUEZ 2. TRALLERO-ARAGUAS E, GRAU-JUNYENT cohort study followed longitudinally in M, HUAPAYA JA et al.: A longitudinal cohort the context of routine clinical care and JM, LABIRUA-ITURBURU A et al.: Clinical manifestations and long-term outcome of study of the anti-synthetase syndrome: in- not a clinical trial. The assignment of anti-Jo1 antisynthetase patients in a large co- creased severity of interstitial lung disease therapy to the individual patient was hort of Spanish patients from the GEAS-IIM in black patients and patients with anti-PL7 and anti-PL12 autoantibodies. Rheumatology based on physician preference, thus, it group. Semin Arthritis Rheum 2016; 46: 225- 31. (Oxford) 2017; 56: 999-1007. is possible that some of the analyses 3. MARASCO E, CIOFFI E, COMETI L et al.: One 15. NARANJO CA, BUSTO U, SELLERS EM et al.: were subject to unaccounted bias. Pa- year in review 2018: idiopathic inflammatory A method for estimating the probability of tients underwent PFTs and CT imaging myopathies. Clin Exp Rheumatol 2018; 36: adverse drug reactions. Clin Pharmacol Ther 1981; 30: 239-45. as part of clinical care, therefore, we 937-47. 4. ODDIS CV: Update on the pharmacological 16. SATHI N, CHIKURA B, KAUSHIK VV, cannot comment on the appearance of treatment of adult myositis. J Intern Med WISWELL R, DAWSON JK: How common is some features such as ILD except as de- 2016; 280: 63-74. methotrexate pneumonitis? A large prospec- tected based on clinical symptoms and 5. SAKETKOO LA, ESPINOZA LR: Experience tive study investigates. Clin Rheumatol 2012; of mycophenolate mofetil in 10 patients with 31: 79-83. findings. Likewise, adverse events were autoimmune-related interstitial lung disease 17. AMATO AA, GRIGGS RC: Treatment of idio- both patient-reported and surveyed by demonstrates promising effects. Am J Med pathic inflammatory myopathies. Curr Opin the treating clinicians but not necessar- Sci 2009; 337: 329-35. Neurol 2003; 16: 569-75.

Clinical and Experimental Rheumatology 2019 861