Vinson et al. Arthritis Research & Therapy (2020) 22:97 https://doi.org/10.1186/s13075-020-02188-x RESEARCH ARTICLE Open Access Impact of tapering targeted therapies (bDMARDs or JAKis) on the risk of serious infections and adverse events of special interest in patients with rheumatoid arthritis or spondyloarthritis: a systematic analysis of the literature and meta-analysis D. Vinson1* , L. Molet-Benhamou2, Y. Degboé3, A. den Broeder4, F. Ibrahim5, C. Pontes6, R. Westhovens7, J. Závada8, T. Pham1, T. Barnetche9, A. Constantin2 and A. Ruyssen-Witrand2 Abstract Objectives: To systematically review the impact of tapering targeted therapies (bDMARDs or JAKis) on the risk of serious infections and severe adverse events (SAEs) in patients with rheumatoid arthritis (RA) or axial spondyloarthritis (axSpA) in remission or low disease activity (LDA) state. Materials and methods: A meta-analysis based on a systematic review of PubMed, Embase, Cochrane, until August 2019, as well as relevant databases of international conferences, was used to evaluate the risk difference (RD) at 95% confidence interval (95% CI) of incidence density of serious infections, SAEs, malignancies, cardiovascular adverse events (CV AEs), or deaths after tapering (dose reduction or spacing) compared to continuation of targeted therapies. Results: Of the 1957 studies initially identified, 13 controlled trials (9 RA and 4 SpA trials) were included in the meta-analysis. 1174 patient-years were studied in the tapering group (TG) versus 1086 in the usual care group (UC). There were 1.7/100 patient-year (p-y) serious infections in TG versus 2.6/100 p-y in UC (RD (95% CI) 0.01 (0.00 to 0.02), p = 0.13) and 7.4/100 p-y SAEs in TG versus 6.7/100 p-y in UC (RD 0.00 (− 0.02 to 0.02), p = 0.82). The risk of malignancies, CV AEs, or deaths did not differ between the tapering and the usual care groups. Subgroup analysis (RA and SpA) detected no significant differences between the two groups. Conclusion: We could not show significant impact of tapering bDMARD or JAKi over continuation concerning the risk of serious infections, SAEs, malignancies, CV AEs, or deaths in RA and SpA patients in remission or LDA state. Keywords: Rheumatoid arthritis, Spondyloarthritis, Biological therapies, DMARDs, Infection bacteria, Viruses infection, Systematic review * Correspondence: dorothee.vinson@gmail.com; http://orcid.org/0000-0002- 9734-3522; 1 Rheumatology Department, CHU Sainte Marguerite, 270 Boulevard de Sainte-Marguerite, 13009 Marseille, France Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 2 of 11 Key message Several studies [19–22] have highlighted that discon- What is already known about this subject? tinuation of targeted therapies, more so in RA than in Current guidelines recommend tapering of targeted SpA, leads to an increased risk of relapse and radio- therapies for RA or SpA patients in remission. graphic progression, while tapering strategies does not What does this study add? seem to increase this risk when compared with continu- No change in serious infection risk, SAEs, CV AEs, ing the initial treatment. malignancies while tapering targeted therapies. Although there is extensive evidence to support that How might this impact on clinical practice? tapering of targeted therapies does not significantly in- Physicians should still try to taper targeted therap- crease the risk of relapse nor of radiographic progres- ies for efficiency and financial advantages. sion, a beneficial effect on the rate of infectious AEs, and most importantly on serious infectious, is yet to be proven. A recent meta-analysis [23] reassessed the ef- Introduction fectiveness of down-titration compared with the con- Current best practice in the treatment of chronic inflam- tinuation of the standard dose of anti-TNF for RA matory rheumatisms such as rheumatoid arthritis (RA) treatment in patients with LDA. Their study evaluated and spondyloarthritis (SpA) focusses on a tight control safety events as a secondary outcome and concluded that strategy (treat-to-target strategy) to prevent joint de- it was uncertain whether anti-TNF tapering influenced struction and improve patients’ functional prognosis. the number of SAE observed due to the very low cer- The objective of this management strategy is to establish tainty of the evidence obtained. This study was restricted remission or a low disease activity (LDA) ((a) RA: Dis- to RA patients treated with anti-TNF and did not extend ease Activity Score28 ESR (DAS28-ESR) < 3.2 for LDA to SpA patients or any other targeted therapies. and < 2.6 for remission [1], (b) axSpA: Ankylosing Spon- The aim of our study was therefore to assess the im- dylitis Disease Activity Score (ASDAS) < 2.1 for LDA pact of tapering targeted therapies (bDMARDs or and < 1.3 for inactive disease [2]). JAKis), compared to continuation of the initial treatment The first-line treatment for RA consists of conven- regimen, on the risk of serious infectious and AEs of tional synthetic disease-modifying anti-rheumatic drugs special interest (SAEs, malignancies, CV AEs, and (csDMARDs) [3], whereas initial treatment for SpA is deaths) in patients with RA or SpA, in remission or based on non-steroidal anti-inflammatory drugs LDA, by conducting a systematic analysis of the litera- (NSAIDs) [4, 5]. Thereafter, in both diseases, biological ture and a meta-analysis. DMARDs (bDMARDS) are employed, and the use of bDMARDs and more recently targeted synthetic Materials and methods DMARDs (tsDMARDs) has been part of the clear im- This meta-analysis is reported in accordance with the provement in the therapeutic management of these dis- Preferred Reporting Items for Systematic Review and eases, leading to sustained LDA or remission in a large Meta-Analysis Protocols (PRISMA-P) 2015 statement number of patients [6, 7]. [24] (see Supplementary Text 1). Nonetheless, many studies have drawn attention to the increased AE risk occurring as a result of these treat- Selection of articles ments compared to csDMARD or no treatment, of We carried out a systematic analysis of the literature to which infectious disorders [8–10] are the most fre- identify controlled trials, preferentially prospective and quently reported. Patients on bDMARDs do not have an randomized trials, which compared tapering targeted increased risk of malignancies in general [11, 12], but therapies (bDMARDs or JAKis) versus continuation of the risk of melanoma may be slightly increased. As for the initial treatment regimen, in patients with RA or cardiovascular AE (CV AE), it is well-established [13, 14] SpA in remission or LDA. The PubMed, Embase, and that certain targeted therapies help improve cardiovascu- Cochrane databases were searched from their date of in- lar comorbidities. The economic burden of these expen- ception to August 2019 using a Boolean association of sive therapies [15] should also not be underestimated keywords (see Supplementary Text 2). Abstracts from and ought to be taken into consideration by the pre- articles submitted to international conferences (ACR, scribing physician. The European League Against Rheu- EULAR, and SFR) between 2016 and 2019 were also matisms (EULAR) and the American College of queried. Rheumatology (ACR) have addressed these issues by This search was carried out independently by two publishing recommendations for the management of investigators (DV and LMB). The title and abstract of chronic inflammatory rheumatisms once remission or articles identified from database searches were subse- LDA has been established, based on a tapering strategy quently reviewed for the following inclusion criteria: (1) [3–5, 16–18]. controlled trials, randomized or not; (2) involving Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 3 of 11 rheumatoid arthritis (RA) or spondyloarthritis (SpA) pa- the I2 statistic in which high values of I2, ranging from 0 tients; (3) treated by targeted therapies: bDMARDs to 100%, represent strong heterogeneity. In case of a (anti-TNF (adalimumab, certolizumab, etanercept, goli- significant heterogeneity, a random effect model was ap- mumab, infliximab) or abatacept or anti-IL6 (sarilumab, plied to take into account heterogeneity. Publication bias tocilizumab) or rituximab or anti IL 12/23 (ustekinu- was searched using funnel plots and Egger test, no mab) or anti IL 17 (secukinumab, ixekizumab) or anti-IL significant bias was found (Supplementary Figure 2). All 23) or JAKis (tofacitinib, baricitinib or upadacitinib); (4) computations were performed using the RevMan V.5.3 in remission or LDA under targeted therapies; and (5) software package developed by Nordic Cochrane Centre comparing tapering (dose reduction or spacing [3]) tar- (Review Manager (computer program), V.5.3. Copenhagen: geted therapies (tapering group (TG)) versus continu- The Nordic Cochrane Centre, the Cochrane Collaboration, ation of the initial treatment regimen (usual care group 2011). P values lower than 0.05 were considered significant. (UC)). The other inclusion criteria applied after full text read- Results ing were (1) description of targeted therapies tapering Study selection protocol and (2) assessing at least one of the following 1957 records were screened in our systematic analysis of AE: serious infections, SAEs, CV AEs, malignancies, or the literature. 1854 records were excluded based on title death. We did not include any restrictions concerning and abstract reading, leaving a total of 103 relevant disease duration, period of remission or LDA, duration articles to be further examined. Full-text assessments of treatment, or concomitant use of csDMARDs. The excluded an additional 89 references. Among the limits were English or French language. The exclusion remaining 14 studies, one study [26], corresponding to criteria were (1) retrospective trials, (2) case reports, (3) an extension phase of another included study which trials without tapering of targeted therapies, (4) trials already described AEs, was excluded in order to avoid without control arms, and (5) trials with no data on AE. duplicated data. Thirteen references were finally in- cluded in our systematic review and meta-analysis Data extraction (Fig. 1). Data was collated using a standardized grid. For each se- We contacted 11 corresponding authors by e-mail to lected study, predefined data were extracted (see Supple- complete safety data that was omitted from the corre- mentary Text 3). sponding publications (values and details). Five authors If data were missing in the article, the corresponding [27–31] responded and were able to provide us with the authors were contacted by e-mail. Details of data col- necessary information to complete our data set. lected are available in Supplementary Table 1 and 2. Patient and public involvement was not appropriate in our study. Population characteristics Among the 13 studies included in the analysis, there Study quality assessment were 9 RA trials [27, 28, 30, 32–37] and 4 SpA trials [29, Risk of bias was assessed using the Cochrane Risk of Bias 31, 38, 39], more precisely axial SpA. All were controlled Tool [25] and is available in Supplementary Figure 1. trials, 11 were randomized controlled trials [26–28, 30– 38], whereas 2 studies followed a longitudinal observa- Analyses tional design using a propensity score matching method For the meta-analysis, the primary endpoint was the in- [29, 39]. cidence density of serious infections in each treatment A total of 2196 patients were included. Disease dur- group (TG or UC). The secondary endpoints were the ation extended from 2.2 to 16.6 years, the sex ratio was incidence density of SAEs, CV AEs, malignancies, or 65% female, and mean patient age in both groups ranged deaths. from 30 to 59 years. Disease activity was low in both A risk difference (RD) was calculated for each study groups (TG and UC): DAS 28-CRP ranging from 1.6 to included in the meta-analysis. All meta-analyses were 2.3 in RA patients and BASDAI SpA patients from 1 to performed using the inverse variance approach, which 2. assumes a fixed effect model, to determine the weight given to each study. This provided a common weighted Duration of the trials RD estimate with a 95% CI, taking into account weight- 1174 patient-years were studied in the targeted therapies ing of the different samples. RD and 95% CIs are tapering group (TG) versus 1086 in the usual care group expressed as forest plots. Statistical heterogeneity of the (UC). The study period extended from 6 months in 2 tri- selected studies was tested using the Q test (χ2), applying als [28, 37] to 12 months in 8 trials [29–35, 39] and to a 0.05 statistical significance cut-off, and reported with approximately 18 months in 3 trials [27, 36, 38]. Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 4 of 11 Fig. 1 Flow chart of systematic review and meta-analysis Studied drugs In bDMARD-naive patients, the duration of targeted In RA patients, the studied targeted therapies were pre- therapies ranged from 37 weeks to 6 years, with a pre- dominantly anti-TNF: certolizumab [32], adalimumab ponderance of patients taking targeted therapies for [27, 28, 36, 37], and etanercept [27, 28, 34–36]. Abata- more than 3 years. For patients previously treated with cept, a selective modulator of T cell co-stimulation, was bDMARDs prior to their inclusion in the trial, data re- studied in one trial [30], and baricitinib, a targeted syn- lating to the total duration of the targeted therapy could thetic DMARD (JAK-inhibitor), was also evaluated in not be retrieved. one trial [33]. SpA patient drug treatments were based on anti-TNF: adalimumab [29, 31], etanercept [29, 31, Evaluating the primary outcome: serious infections in 38, 39], infliximab [29, 31], and golimumab [31]. studies comparing tapering of targeted therapies (TG) versus usual care (UC) Previous treatment and duration of targeted therapies Our meta-analysis comparing tapering of targeted ther- Patients were bDMARD-naive prior to their inclusion in apies (TG) versus continuation of usual care (UC) was 6 trials [27, 29, 30, 32, 35, 38], whereas they received performed on 13 trials and showed no significant differ- bDMARDs before the investigation in 4 trials [31, 33, ence in relation to the incidence density of serious infec- 36, 37]. This information was not available for 3 studies tions reported between TG (20 patients presented [28, 34, 39]. No study included patients treated with serious infections among 1174 patient-years, corre- tsDMARDs prior to inclusion in the trials. sponding to 1.7/100 patient-year (p-y)) and UC groups Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 5 of 11 (28 patients presented serious infections among 1087 Meta-analysis of other safety events in studies comparing patient-years (2.6/100 p-y)), with a total risk difference TG versus UC (RD) (95% CI) of 0.01 (0.00 to 0.02), p = 0.13 (I2 hetero- Six studies [27–31, 37] reported the incidence of CV geneity score, 0%) (Fig. 2). AEs. Among these, 4 studies [27, 28, 30, 37] focused on The subgroup analysis, performed separately on RA or RA patients and 2 studies [29, 31] on SpA patients. SpA trials, did not show any significant difference in the There was no significant difference in the incidence of risk of serious infections in RA patients, 0.01 (0.00 to CV AE (RD (95% CI), 0.00 (− 0.02 to 0.02), p = 0.84 (I2, 0.02), p = 0.10 (I2, 0%), nor in SpA patients, 0.00 (− 0.02 3%)), when comparing 5 patients with CV AE among to 0.02), p = 0.95 (I2, 0%) (Fig. 2). 383 patients-years (1.3/100 p-y) in TG, to 7 patients with CV AE among 331 patients-years (2.1/100 p-y) in UC (Fig. 4). There was no significant difference in each sub- Evaluation of secondary outcomes: SAEs group either: in RA patients, RD (95% CI) was 0.02 (− Our meta-analysis did not show a decreased risk in 0.02 to 0.05), p = 0.37 (I2, 23%), whereas in SpA patients SAEs (RD (95% CI), 0.00 (− 0.02 to 0.02), p = 0.82 (I2, RD (95% CI) was − 0.01 (− 0.03 to 0.02), p = 0.68 (I2, 0%) 0%)), when comparing 87 patients with SAEs among (Fig. 4). 1174 patient-years (7.4/100 p-y) in the TG group to Eight studies [27–32, 35, 37] reported malignancies 73 patients with SAEs among 1085 patient-years (6.7/ which developed during the study period. Among these, 100 p-y) in the UC group (Fig. 3). There was also no 6 studies [27, 28, 30, 32, 35, 37] involved RA patients statistically significant difference between the two and 2 studies [29, 31] SpA patients; 13 patients with ma- subgroups: RD in RA patients (95% CI) was 0.00 (− lignancies among 712 patients-years (1.8/100 p-y) in TG 0.02 to 0.02), p = 0.93 (I2, 24%), compared to RD in and 5 patients with malignancies among 616 patients- SpA patients (95% CI) 0.00 (− 0.03 to 0.03), p = 0.79 years (0.8/100 p-y) in UC. Our meta-analysis did not de- (I2, 0%) (Fig. 3). tect any significant differences when all 8 studies were Fig. 2 Forest plot of serious infections: difference of risk of serious infections in TG versus UC Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 6 of 11 Fig. 3 Forest plot of severe adverse events: risk difference of TG versus UC Fig. 4 Forest plot of cardiovascular adverse events: risk difference of TG versus UC Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 7 of 11 examined (RD (95% CI), − 0.01 (− 0.02 to 0.01), p = 0.33 all studies included in our meta-analysis were prospect- (I2, 0)) and also not in the subgroup analysis (RA pa- ive, controlled trials, and eleven [27, 28, 30–38] out of tients: RD (95% CI), − 0.01 (− 0.03 to 0.00), p = 0.17 (I2, thirteen studies randomly assigned patients into groups, 0%); SpA patients: RD (95% CI), 0.01 (− 0.02 to 0.03), thereby granting the absence of heterogeneity. p = 0.68, (I2, 0%)) (Fig. 5). We expected to see a decrease of the risk of serious in- Three deaths were reported in the trials, 2 in one RA fections due to a reduction of the therapeutic pressure. trial [35] in the UC group (0.2/100 p-y) and 1 in one RA However, although we observed a numerical trend with trial (0.1/100 p-y) in the TG group [30], with no signifi- less serious infections in tapering versus continuation of cant differences detected (RD (95% CI), 0.00 (0.00 to the initial treatment schedule (1.7/100 p-y versus 2.6/ 0.01), p = 0.76 (I2, 0%)) (Fig. 6). 100 p-y), our study was not able to show a statistical dif- ference. The range of serious infections observed in our Discussion meta-analysis is consistent with a previous one focused Our meta-analysis focused on controlled trials in RA on bDMARDs in RA patients [10]. A recent meta- and SpA patients that achieved remission or LDA and analysis [23] published in 2019, which focused on down- showed that tapering targeted therapies was not associ- titration compared with continuation of the standard ated with a statistically significant difference in risk of dose of TNFi for RA in patients with LDA, evaluated developing serious infections, SAEs, CV AEs, malignan- safety events as secondary outcomes and concluded that cies, or death when compared with continuation of the it was uncertain whether dose tapering affected the initial treatment regimen. This result was also confirmed number of SAEs observed due to the low certainty of in the subgroup analyses (of RA and SpA). Our results the evidence. Another meta-analysis [10] assessing the are consistent with the only other meta-analysis [23] risk of serious infections in RA patients treated with tar- published on the issue. geted therapies was conducted in 2015 and also did not Our meta-analysis has several strengths. Compared to detect any differences. Indeed, the initial course of treat- the previously published meta-analysis, it includes SpA ment with targeted therapies may represent an increased patients next to RA patients and expands on the number risk of infectious events [40]. However, the absence of and type of drugs used (bDMARDS and JAKis) and fi- statistical difference may be explained by a lack of nally assesses specific subgroups of AE outcomes. Also, power. Indeed, we calculated the statistical power of our Fig. 5 Forest Plot of Malignancy: Risk difference of TG versus UC Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 8 of 11 Fig. 6 Forest Plot of Deaths: Risk difference of TG versus UC meta-analysis, based on the primary endpoint (serious may potentially influence the risk of infection [41, 42], infections), which was about 26.6%. This is relatively low most of all in RA patients, this risk may be lowered. and can be explained by the assessed endpoint being Accordingly, it may become more difficult to point rare and by the fact that the clinical trials included in out a difference between patient populations that are our meta-analysis were not designed to achieve an ob- already closely followed for any signs of AEs, during jective of tolerance. To increase the power, it would re- a longer follow-up and with better tolerance. More- quire a greater number of patients and would over, patients in these trials had been taking targeted jeopardize the feasibility of these types of trials. Indeed, therapies for at least more than a year, before initiat- to reach a statistical power of 80% in our meta-analysis, ing the tapering phase. However, the critical phase it would require 10,084 patient-year. In addition, the for AE is in the first months [43–45]. Indeed, a safety difference in frequency is very low between the two issue (serious infection and SAE) is more likely to groups (TG vs UC) for each tolerance endpoints, and happen in the initiating phase, rather than after a this also contributes to the requirement for greater pa- long period under treatment, thus leading to the ex- tient numbers in order to show a significant difference. clusion of the patient. The population included in the Despite its good internal validity, our meta-analysis has trials is therefore highly selected and might explain several limitations. the absence of any difference in risk between the TG Only one study analyzing a tsDMARD (Baricitinib) and UC group. was included in our meta-analysis, which makes the The same argument holds for the selection of subjects evaluation of infectious risk in this therapeutic subclass regarding cardiovascular and tumoral status. precarious, even though the results were consistent with As for CV AE, it is well-established [13, 14] that cer- a non-significant difference. tain targeted therapies help improve cardiovascular co- Another concern is the healthy survivor bias. After ex- morbidities in patients, yet no increase in CV AE was cluding subjects with numerous comorbidities, which detected in the TG strategy. This may potentially be Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 9 of 11 explained by the persistence of the remission/LDA state, AdB, FI, CP, RW, JZ, TP, and TB: contributed to the results interpretation and imparting lower cardiovascular risk [46]. manuscript approval. TB: was in charge of the statistical analyses. All authors take responsibility for the integrity of the work as a whole, from inception to Regarding malignancy, the duration of the studies was published article, and they should indicate that they had full access to all the too short and the events too rare to detect any statistical data in the study and take responsibility for the integrity of the data and the difference. Trials with a longer follow-up period would accuracy of the data analysis. They give permission to reproduce published material, to report sensitive personal information, to use illustrations of iden- be needed to be able to point out a statistical difference. tifiable persons, or to name persons for their contributions. Conclusion Funding The authors have not declared a specific grant for this research from any In conclusion, our meta-analysis highlights no remark- funding agency in the public, commercial, or not-for-profit sectors. able difference in the rate of infectious events, in pa- tients in the tapering treatment group or patients Availability of data and materials continuing their initial treatment schedule. This study involved data that are available in published papers used for this Nevertheless, other benefits of a tapering strategy, in- systematic review. Additional data were also got directly from authors. cluding alleviating patient burden due to self-injection, Ethics approval and consent to participate medication costs, and the safety of these strategies in re- This study, as a meta-analysis, does not involve directly human participants lation to flare-ups leads us to support this therapeutic but used data from trials performed on human participants. Indeed, the Eth- ics Committee approvals of each trial were obtained for all the studies se- approach. lected in this meta-analysis. Supplementary information Competing interests Supplementary information accompanies this paper at https://doi.org/10. The authors declare that they have no competing interest. 1186/s13075-020-02188-x. Author details 1 Additional file 1 : Supplementary text 1: Preferred Reporting Items Rheumatology Department, CHU Sainte Marguerite, 270 Boulevard de for Systematic review and Meta-Analysis Protocols (PRISMA-P) 2015 state- Sainte-Marguerite, 13009 Marseille, France. 2Rheumatology Department, ment [24]. Supplementary text 2: Boolean association of keywords Hôpital Purpan, Toulouse, France. 3INSERM U1043, CPTP, Toulouse, France. 4 used in PubMed (Medline database). Supplementary text 3: List of data Department of Rheumatology, Sint Maartenskliniek, 6500 GM Nijmegen, The extracted for each selected study. Supplementary Table 1: Population Netherlands. 5Centre for Rheumatic Diseases, Department of Inflammation characteristics of studies included in the Meta-Analysis. Supplementary Biology, Faculty of Life Sciences and Medicine, School of Immunology and Table 2: Adverse Event (AE) characteristics of studies included in the Microbial Sciences, King’s College London, Cutcombe Road, London SE5 9RJ, Meta-Analysis. Supplementary Figure 1: Cochrane Risk of Bias Tool [25]. UK. 6Department of Pharmacology, Therapeutics and Toxicology, Unitat Supplementary Figure 2: Publication bias and Funnel plots. docent Parc Taulí—Medical School—Universitat Autònoma de Barcelona, Sabadell, Barcelona, Spain. 7Skeletal Biology and Engineering Research Center, Department of Development and Regeneration, Rheumatology Abbreviations University Hospitals Leuven, KULeuven, Leuven, Belgium. 8First Faculty of ACR: American College of Rheumatology; ASDAS: Ankylosing Spondylitis Medicine, Institute of Rheumatology and Department of Rheumatology, Disease Activity Score; axSpA: Axial spondyloarthritis; CV AE: Cardiovascular Charles University, Na Slupi 4, Prague, Czech Republic. 9Rheumatology adverse events; DAS28-ESR: Disease Activity Score28 ESR; DMARDs: Disease- Department, Pellegrin Hospital, Bordeaux, France. modifying anti-rheumatic drugs; bDMARDs: Biological DMARDs; csDMARDs: Conventional synthetic disease-modifying anti-rheumatic drugs; Received: 20 November 2019 Accepted: 17 April 2020 tsDMARDs: Targeted synthetic DMARDs; EULAR: European League Against Rheumatisms; JAKi: JAKinibs; LDA: Low disease activity; NSAIDs: Non-steroidal anti-inflammatory drugs; p-y: Patient-years; RA: Rheumatoid arthritis; RD: Risk References difference; SAE: Severe adverse events; TG: Tapering group; UC: Usual care 1. Fransen J, Creemers MCW, Van Riel PLCM. Remission in rheumatoid arthritis: group agreement of the disease activity score (DAS28) with the ARA preliminary remission criteria. Rheumatol (Oxford). 2004;43(10):1252–5. Acknowledgements 2. Machado P, Landewé R, Lie E, Kvien TK, Braun J, Baker D, et al. Ankylosing We are grateful to the professors and organizers of ASLER seminary (for Spondylitis Disease Activity Score (ASDAS): defining cut-off values for Systematic Analysis of the Literature in Rheumatology) for their useful advice disease activity states and improvement scores. Ann Rheum Dis. 2011;70(1): in the writing of this manuscript. We wish to thank AbbVie who provided 47–53. logistic support in the organization of sessions about the implementation of 3. Josef S Smolen, Robert Landewé, Johannes Bijlsma, et al. EULAR meta-analysis and remained independent of the collection, analysis, and in- recommendations for the management of rheumatoid arthritis with terpretation of data. We are grateful to Mrs. Weill C. who helped us in the synthetic and biological disease-modifying antirheumatic drugs: 2016 Embase bibliographic research. update | Annals of the Rheumatic Diseases. Disponible sur: https://ard.bmj. com/content/76/6/960. [cité 15 juin 2019]. Patient and public involvement 4. van der Heijde D, Ramiro S, Landewe R, Baraliakos X, Van den Bosch F, Was not appropriate in our study Sepriano A, et al. 2016 update of the ASAS-EULAR management recommendations for axial spondyloarthritis. Ann Rheum Dis. 2017;76(6): Patient consent 978–91. Not required. 5. Gossec L. European League Against Rheumatism (EULAR) recommendations for the management of psoriatic arthritis with pharmacological therapies: Authors’ contributions 2015 update. - PubMed - NCBI. Disponible sur: https://www.ncbi.nlm.nih. DV and LM-B: contributed to the conception of the study, the article selec- gov/pubmed/26644232. [cité 18 juin 2019]. tion process, the data collection, the data analysis, the results interpretation, 6. van Tuyl LH, Sadlonova M, Davis B, Flurey C, Goel N, Hewlett SE, et al. and the manuscript writing and approval. AC and AR-W: contributed to the Remission in rheumatoid arthritis: working toward incorporation of the conception of the study, results interpretation, and manuscript approval. YD, patient perspective at OMERACT 12. J Rheumatol janv. 2016;43(1):203–7. Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 10 of 11 7. Lubrano E, Mesina F, Caporali R. Clinical remission in rheumatoid arthritis 25. Higgins JP, Altman DG, Gøtzsche PC, et al. The Cochrane Collaboration’s and psoriatic arthritis. Clin Exp Rheumatol. 2018;36(5):900–10. tool for assessing risk of bias in randomised trials | The BMJ. Disponible sur: 8. Singh JA, Wells GA, Christensen R, Tanjong Ghogomu E, Maxwell L, https://www.bmj.com/content/343/bmj.d5928. [cité 15 juin 2019]. Macdonald JK, et al. Adverse effects of biologics: a network meta-analysis 26. Bouman CA, van Herwaarden N, van den Hoogen FH, Fransen J, van and Cochrane overview. Cochrane Database Syst Rev. 2011;(2):CD008794. Vollenhoven RF, Bijlsma JW, et al. Long-term outcomes after disease https://doi.org/10.1002/14651858.CD008794.pub2. activity-guided dose reduction of TNF inhibition in rheumatoid arthritis: 3- 9. Minozzi S, Bonovas S, Lytras T, Pecoraro V, González-Lorenzo M, Bastiampillai year data of the DRESS study - a randomised controlled pragmatic non- AJ, et al. Risk of infections using anti-TNF agents in rheumatoid arthritis, inferiority strategy trial. Ann Rheum Dis. 2017;76(10):1716–22. psoriatic arthritis, and ankylosing spondylitis: a systematic review and meta- 27. van Herwaarden N, van der Maas A, Minten MJM, van den Hoogen FHJ, analysis. Expert Opin Drug Saf. 2016;15(sup1):11–34. Kievit W, van Vollenhoven RF, et al. Disease activity guided dose reduction 10. Singh JA, Cameron C, Noorbaloochi S, Cullis T, Tucker M, Christensen R, and withdrawal of adalimumab or etanercept compared with usual care in et al. Risk of serious infection in biological treatment of patients with rheumatoid arthritis: open label, randomised controlled, non-inferiority trial. rheumatoid arthritis: a systematic review and meta-analysis. Lancet Lond BMJ. 2015;350:h1389. Engl. 2015;386(9990):258–65. 28. Ibrahim F, Lorente-Canovas B, Dore CJ, Bosworth A, Ma MH, Galloway JB, 11. Ramiro S, Sepriano A, Chatzidionysiou K, Nam JL, Smolen JS, van der Heijde et al. Optimizing treatment with tumour necrosis factor inhibitors in D, et al. Safety of synthetic and biological DMARDs: a systematic literature rheumatoid arthritis-a proof of principle and exploratory trial: is dose review informing the 2016 update of the EULAR recommendations for tapering practical in good responders? Rheumatol (Oxford). 2017;56(11): management of rheumatoid arthritis. Ann Rheum Dis. 2017;76(6):1101–36. 2004–14. 12. Dreyer L, Cordtz RL, Hansen IMJ, Kristensen LE, Hetland ML, Mellemkjaer L. 29. Zavada J, Uher M, Sisol K, Forejtova S, Jarosova K, Mann H, et al. A tailored Risk of second malignant neoplasm and mortality in patients with approach to reduce dose of anti-TNF drugs may be equally effective, but rheumatoid arthritis treated with biological DMARDs: a Danish population- substantially less costly than standard dosing in patients with ankylosing based cohort study. Ann Rheum Dis. 2018;77(4):510–4. spondylitis over 1 year: a propensity score-matched cohort study. Ann 13. Ursini F, Ruscitti P, Caio GPI, Manfredini R, Giacomelli R, De Giorgio R. The Rheum Dis. 2016;75(1):96–102. effect of non-TNF-targeted biologics on vascular dysfunction in rheumatoid 30. Westhovens R, Robles M, Ximenes AC, Wollenhaupt J, Durez P, Gomez- arthritis: a systematic literature review. Autoimmun Rev. 2019;18(5):501–9. Reino J, et al. Maintenance of remission following 2 years of standard 14. England BR, Thiele GM, Anderson DR, Mikuls TR. Increased cardiovascular treatment then dose reduction with abatacept in patients with early risk in rheumatoid arthritis: mechanisms and implications. BMJ. 2018;361: rheumatoid arthritis and poor prognosis. Ann Rheum Dis. 2015;74(3):564–8. k1036. 31. Gratacos J, Pontes C, Juanola X, Sanz J, Torres F, Avendano C, et al. Non- 15. Stevenson M, Archer R, Tosh J, Simpson E, Everson-Hock E, Stevens J, et al. inferiority of dose reduction versus standard dosing of TNF-inhibitors in Adalimumab, etanercept, infliximab, certolizumab pegol, golimumab, axial spondyloarthritis. Arthritis Res Ther. 2019;21(1):11. tocilizumab and abatacept for the treatment of rheumatoid arthritis not 32. Weinblatt ME, Bingham CO 3rd, Burmester G-R, Bykerk VP, Furst DE, Mariette previously treated with disease-modifying antirheumatic drugs and after the X, et al. A phase III study evaluating continuation, tapering, and withdrawal failure of conventional disease-modifying antirheumatic drugs only: systematic of certolizumab pegol after one year of therapy in patients with early review and economic evaluation. Health Technol Assess. 2016;20(35):1–610. rheumatoid arthritis. Arthritis Rheumatol. 2017;69(10):1937–48. 16. Singh JA, Saag KG, Bridges SLJ, Akl EA, Bannuru RR, Sullivan MC, et al. 2015 33. Takeuchi T, Genovese MC, Haraoui B, Li Z, Xie L, Klar R, et al. Dose reduction American College of Rheumatology Guideline for the treatment of of baricitinib in patients with rheumatoid arthritis achieving sustained rheumatoid arthritis. Arthritis Rheumatol. 2016;68(1):1–26. disease control: results of a prospective study. Ann Rheum Dis. 2019;78(2): 17. Singh JA, Guyatt G, Ogdie A, Gladman DD, Deal C, Deodhar A, et al. 2018 171–8. American College of Rheumatology/National Psoriasis Foundation guideline 34. van Vollenhoven RF, Ostergaard M, Leirisalo-Repo M, Uhlig T, Jansson M, for the treatment of psoriatic arthritis. Arthritis Care Res. 2019;71(1):2–29. Larsson E, et al. Full dose, reduced dose or discontinuation of etanercept in 18. Ward MM, Deodhar A, Akl EA, Lui A, Ermann J, Gensler LS, et al. American rheumatoid arthritis. Ann Rheum Dis. 2016;75(1):52–8. College of Rheumatology/Spondylitis Association of America/ 35. Smolen JS, Nash P, Durez P, Hall S, Ilivanova E, Irazoque-Palazuelos F, et al. Spondyloarthritis Research and Treatment Network 2015 recommendations Maintenance, reduction, or withdrawal of etanercept after treatment with for the treatment of ankylosing spondylitis and nonradiographic axial etanercept and methotrexate in patients with moderate rheumatoid Spondyloarthritis. Arthritis Rheumatol. 2016;68(2):282–98. arthritis (PRESERVE): a randomised controlled trial. Lancet Lond. 2013; 19. Henaux S, Ruyssen-Witrand A, Cantagrel A, Barnetche T, Fautrel B, Filippi N, 381(9870):918–29. et al. Risk of losing remission, low disease activity or radiographic 36. Fautrel B, Pham T, Alfaiate T, Gandjbakhch F, Foltz V, Morel J, et al. Step- progression in case of bDMARD discontinuation or tapering in rheumatoid down strategy of spacing TNF-blocker injections for established rheumatoid arthritis: systematic analysis of the literature and meta-analysis. Ann Rheum arthritis in remission: results of the multicentre non-inferiority randomised Dis. 2018;77(4):515–22. open-label controlled trial (STRASS: Spacing of TNF-blocker injections in 20. Couvaras L, Barnetche T, Constantin A, Pham T. Tapering TNF Inhibitors in Rheumatoid ArthritiS Study). Ann Rheum Dis. 2016;75(1):59–67. Axial Spondyloarthritis: Systematic analysis of the literature and meta- 37. l’Ami MJ, Krieckaert CL, Nurmohamed MT, van Vollenhoven RF, Rispens T, Analysis. ACR Meeting Abstracts. 2018. Disponible sur: https://acrabstracts. Boers M, et al. Successful reduction of overexposure in patients with org/abstract/tapering-tnf-inhibitors-in-axial-spondyloarthritis- rheumatoid arthritis with high serum adalimumab concentrations: an open- systematicanalysisoftheliteratureandmeta-analysis/. [cité 3 juill 2019]. label, non-inferiority, randomised clinical trial. Ann Rheum Dis. 2018;77(4): 21. Park JW, Kim H-A, Shin K, Park Y-B, Kim T-H, Song YW, et al. Effects of 484–7. tapering tumor necrosis factor inhibitor on the achievement of inactive 38. Cantini F, Niccoli L, Cassara E, Kaloudi O, Nannini C. Duration of remission disease in patients with axial spondyloarthritis: a nationwide cohort study. after halving of the etanercept dose in patients with ankylosing spondylitis: Arthritis Res Ther. 2019;21(1):163. a randomized, prospective, long-term, follow-up study. Biol Targets Ther. 22. Tanaka Y, Smolen JS, Jones H, Szumski A, Marshall L, Emery P. The effect of 2013;7:1–6. deep or sustained remission on maintenance of remission after dose 39. Li K-P, Jin J-Y, Yang J-S, Li Y, Zhao W, Luo G, et al. Full dose, half dose, or reduction or withdrawal of etanercept in patients with rheumatoid arthritis. discontinuation of etanercept biosimilar in early axial spondyloarthritis Arthritis Res Ther. 2019;21(1):164. patients: a real-world study in China. Arch Med Sci AMS. 2019;15(3):700–5. 23. Verhoef LM, van den Bemt BJ, van der Maas A, Vriezekolk JE, Hulscher 40. Bernatsky S, Habel Y, Rahme E. Observational studies of infections in ME, van den Hoogen FH, et al. Down-titration and discontinuation rheumatoid arthritis: a metaanalysis of tumor necrosis factor antagonists. J strategies of tumour necrosis factor-blocking agents for rheumatoid Rheumatol. 2010;37(5):928–31. arthritis in patients with low disease activity. Cochrane Database Syst 41. Jani M, Barton A, Hyrich K. Prediction of infection risk in rheumatoid arthritis Rev. 2019;5:CD010455. patients treated with biologics: are we any closer to risk stratification? Curr 24. Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Opin Rheumatol. 2019;31(3):285–92. Preferred reporting items for systematic review and meta-analysis protocols 42. Curtis JR, Winthrop K, O’Brien C, Ndlovu MN, de Longueville M, Haraoui B. (PRISMA-P) 2015 statement. Syst Rev. 2015;4(1):1. Use of a baseline risk score to identify the risk of serious infectious events in Vinson et al. Arthritis Research & Therapy (2020) 22:97 Page 11 of 11 patients with rheumatoid arthritis during certolizumab pegol treatment. Arthritis Res Ther. 2017;19(1):276. 43. Ruderman EM. Overview of safety of non-biologic and biologic DMARDs. Rheumatology. 2012;51(suppl_6):vi37–43. 44. Galloway JB, Hyrich KL, Mercer LK, Dixon WG, Fu B, Ustianowski AP, et al. Anti-TNF therapy is associated with an increased risk of serious infections in patients with rheumatoid arthritis especially in the first 6 months of treatment: updated results from the British Society for Rheumatology Biologics Register with special emphasis on risks in the elderly. Rheumatol (Oxford). 2011;50(1):124–31. 45. Askling J, Fored CM, Brandt L, Baecklund E, Bertilsson L, Feltelius N, et al. Time-dependent increase in risk of hospitalisation with infection among Swedish RA patients treated with TNF antagonists. Ann Rheum Dis. 2007; 66(10):1339–44. 46. Solomon DH, Reed GW, Kremer JM, Curtis JR, Farkouh ME, Harrold LR, et al. Disease activity in rheumatoid arthritis and the risk of cardiovascular events. Arthritis Rheumatol. 2015;67(6):1449–55. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.