Article Text
PDF
Abstract
Objectives In the second Computer-Assisted Management in Early Rheumatoid Arthritis trial, patients had started with methotrexate and 10 mg prednisone (MTX+pred) or placebo (MTX+plac). After the trial, prednisone was tapered and stopped, if possible. The objective was to compare, during the post-trial follow-up between the two former strategy groups, initiation of the first biological disease-modifying antirheumatic drug (bDMARD), radiographic outcome and onset of glucocorticoid (GC)-related comorbidities.
Methods Data on prednisone and bDMARD use and onset of GC-related comorbidities were collected retrospectively. Sharp/van der Heijde scoring was performed. Data were analysed using Fisher’s exact and Mann-Whitney U tests.
Results Of 218 patients post-trial follow-up data were available. The maximum follow-up time was 11.8 years. Fewer patients initiated a first bDMARD in the former MTX+pred compared with the former MTX+plac strategy group: 31% vs 50%, p=0.003. At the 2 year post-trial follow-up, the median erosion score was significantly lower in the former MTX+pred versus former MTX+plac strategy group: 0 (range 0–0) versus 0 (0–2), p=0.002. No significant differences between the former strategy groups in the onset of GC-related comorbidities during the post-trial follow-up were found.
Conclusion Addition of 10 mg prednisone daily to an MTX-based treatment strategy in early rheumatoid arthritis results in a lower initiation rate of a first bDMARD and significantly better radiographic outcomes, yet does not result in more GC-related comorbidities.
- rheumatoid arthritis
- glucocorticoids
- biologicals
- outcome
Statistics from Altmetric.com
Introduction
The introduction of biological disease-modifying antirheumatic drugs (bDMARDs) in the treatment of rheumatoid arthritis (RA) has, in general, led to better disease control and improved functional ability and quality of life.1 Disadvantages are higher cost and risk of severe infections of bDMARDs compared with conventional synthetic (cs) DMARDs.2 3 In about one-third of patients with RA, bDMARD use does not result in sufficient clinical improvement.4 Therefore, it is important to optimise treatment strategies based on csDMARDs before the next step, to adding a bDMARD, is taken. In this way, initiation of a bDMARD may be delayed or even prevented.
In the second Computer-Assisted Management in Early Rheumatoid Arthritis (CAMERA-II) trial, patients initiated a methotrexate (MTX)-based treatment strategy, with 10 mg prednisone (MTX+pred) or with placebo (MTX+plac) daily. Addition of 10 mg prednisone resulted in significantly faster reduction of disease activity, less erosive joint damage after 2 years and less frequent initiation of tumour necrosis factor (TNF)-inhibitor treatment,5 showing the potential of effective disease control by csDMARDs (especially with glucocorticoids (GCs)) for a large proportion of patients and of cost savings by reduced bDMARD initiation.
In this study, we wondered whether the beneficial effects of adding 10 mg prednisone to an MTX-based treatment strategy during the CAMERA-II trial persisted during a post-trial follow-up. Hypothetically, the need of initiating a bDMARD in the former MTX+pred strategy group could be increased (rebound), as after the end of the trial it was the strategy to taper and stop the prednisone therapy. To investigate this, we examined initiation of the first bDMARD among patients who had participated in CAMERA-II. In addition, we investigated whether the benefit regarding radiographic progression persisted during the post-trial follow-up. Finally, we aimed to gain insight into the long-term GC-related comorbidities after the CAMERA-II trial, as there is a paucity of systematically collected data concerning long-term adverse effects of medium-dose GC use,6 7 in contrast to those of high-dose GC use.8
Methods
In the 2-year double-blind randomised placebo-controlled CAMERA-II trial, DMARD naive patients with early RA were randomised to initiate treatment with either MTX+pred, with a stable dosage of 10 mg of prednisone daily during the whole trial period, or MTX+plac. It was a tight-controlled and treat-to-target study, aiming for remission. Depending on disease activity, subsequent treatment steps were taken, including ultimately addition of the TNF-inhibitor adalimumab.9 In the post-trial follow-up, patients were treated by their rheumatologist according to good clinical care. In the former MTX+pred group, the aim was to taper prednisone, if possible.
We retrospectively collected data from medical charts on prednisone and bDMARD use, mortality and onset of GC-related comorbidities during the post-trial follow-up. Radiographs of hands and feet were scored with Sharp/van der Heijde scoring (SHS).10 As yearly radiographs were not present in all patients during the post-trial follow-up, we arbitrarily decided to restrict these analyses up to 2 years of the post-trial follow-up, with 72% of radiographs available. For other analyses, post-trial follow-up data up to 11 years were used. Discontinuation of prednisone use in the former MTX+pred strategy group was visualised using a Kaplan-Meier survival curve, as was bDMARD initiation in both former strategy groups, using Cox’s proportional hazard regression analysis for testing.
To avoid reporting bias by beliefs in safety of prednisone among rheumatologists as well as patients during the open, post-trial follow-up, we chose to only investigate GC-related comorbidities, based on literature review and expert opinion, for which treatment was initiated. This is more objective than investigating all negative events of which the scoring would be rather subjective and of which it would be hard to discriminate whether caused by RA or GC use. Dichotomous data were tested with Fisher’s exact tests and continuous data with Mann-Whitney U tests. Erosion scores were visualised with a cumulative probability plot. For data analyses, IBM SPSS Statistics, V.22.0 was used.
The institutional review boards of the participating centres confirmed that the Medical Research Involving Human Subjects Act (WMO) was not applicable to this study.
Results
Post-trial follow-up data were available for 218 of the 236 patients of CAMERA-II; 18 patients were no longer followed up for various reasons, for example, change of hospital. Of these 218 patients, at the start of CAMERA-II, characteristics between the randomised groups were similar (table 1).
Baseline characteristics at the start of CAMERA-II trial of the groups presently investigated
The median post-trial follow-up time in the former MTX+pred strategy group was 6.7 years (range 0.1–10.1; IQR 5.2–8.3) vs 6.6 (range 0.3–11.8; IQR 5.2–8.0) in the former MTX+plac group, p=0.71. Half of the patients in the former MTX+pred strategy group had discontinued prednisone 1 year post-trial, and 79% at the end of the post-trial follow-up (online supplementary figure 1). During the post-trial follow-up, significantly fewer patients initiated a first bDMARD in the former MTX+pred strategy group compared with the former MTX+plac group (31% vs 50%, respectively, p=0.003). Also, during the combined study and the post-trial follow-up period, fewer patients had initiated a first bDMARD in the (former) MTX+pred strategy group compared with the (former) MTX+plac strategy group (figure 1, HR 0.46, 95% CI 0.30 to 0.72; p=0.001).
Supplementary Material
Supplementary data
Initiation of a first bDMARD in (former) strategy groups during the trial and post-trial follow-up. Survival curve with 95% confidence interval, showing the proportion of patients who did not initiate a first bDMARD. Both the trial and follow-up period are depicted. Vertical black line indicates end of the two year trial period. The numbers of patients at risk are shown beneath the graph with yearly intervals. *missing: n=2. 95% CI, 95% confidence interval; bDMARD, biological disease modifying anti-rheumatic drug; HR, hazard ratio; MTX, methotrexate; plac, placebo, pred: prednisone.
The median SHS of hands and feet at the 2-year post-trial follow-up was not significantly different between the two groups (former MTX+pred median 0 (IQR 0–0), former MTX+plac (IQR 0–1.5); p=0.271). Two years post-trial, 83% of patients was erosion free in the former MTX+pred strategy group versus 62% in the former MTX+plac strategy group, p=0.16 (online supplementary figure 2). The median erosion score at 2 years post-trial was 0 (IQR 0–0) in the former MTX+pred strategy group versus 0 (0–2) in the former MTX+plac strategy group, p=0.002.
The incidence of long-term GC-related comorbidities during the post-trial follow-up was not significantly different between the former strategy groups (table 2), although there were some numerical differences: more cardiovascular comorbidities in the former MTX+pred than in the former MTX+plac strategy group (n=13 vs n=8) and a higher mortality rate (n=10 vs n=6). The most frequent cause of death in both groups was malignancy (see table 2 for details). In either group, the cause of death could not be retrieved in three cases.
Incidence of new-onset comorbidity and mortality in the post-trial follow-up
Discussion
We found that during the post-trial follow-up of CAMERA-II, fewer patients initiated a first bDMARD in the former MTX+pred than in the former MTX+plac strategy group, despite less initiation of a bDMARD in the former MTX+pred strategy group during the trial and tapering and stopping of prednisone, after the trial. Furthermore, compared with the former MTX+plac strategy group, in the former MTX+pred strategy group still less erosive damage was present during the post-trial follow-up and there was no significantly increased incidence of long-term post-trial GC-related comorbidities.
Our results show no rebound of initiation of a first bDMARD during the post-trial follow-up in the former MTX+pred strategy group. These show that long-term bDMARD initiation can be reduced by a tight-control and treat-to-target strategy including prednisone. Importantly, the lower rate of bDMARD initiation in the former MTX+pred strategy group did not negatively affect post-trial radiographic outcome at 2 years. An explanation of these favourable findings could be the DMARD properties of prednisone during the window of opportunity period, as the MTX schemes in both strategy groups during the whole trial, so at least during the window of opportunity period, often interpreted as the first 3–6 months, were identical.
In comparison, in the 56-week COBRA trial treatment with either sulfasalazine or a combination of sulfasalazine, MTX and prednisone was given. A follow-up study was done up to 11 years, in which bDMARD initiation and incidence of new onset comorbidities were comparable between the groups.11 In the Better Anti-Rheumatic PharmacOTherapy (BARFOT) study, patients with early RA received 7.5 mg prednisone daily in addition to csDMARDs alone. The long-term risk of ischaemic cardiovascular events was higher in the prednisone group, with a trend towards reduced survival.12 Our finding of a higher number of cardiovascular comorbidities and mortality even in patients with early RA in the former MTX+pred strategy group, although not statistically significant, could be clinically relevant, as this result is in line with findings of the BARFOT study.
On the basis of our own findings and the current body of evidence on GC schemes in early RA, we propose initiation of treatment in early RA patients with an MTX-based tight control and treat-to-target strategy, in combination with a rapidly remission inducing agent. From the viewpoint of cost, GCs should be preferred over bDMARDs as rapidly remission inducing agent. Our finding of initiation of bDMARDs in 41% of patients with early RA in the MTX+plac strategy group during the 2-year CAMERA-II trial is similar to that of another study in which 47% of patients with early RA on subcutaneous MTX initiated a bDMARD during an average follow-up of 1.8 years (SD 1.6).13 Ours is, however, the first study that investigated a possible rebound increased initiation of a first bDMARD after tapering and stopping of prednisone.
As in the post-trial follow-up period the controlled situation was lost and treatment was open to the rheumatologists, long-term outcomes in our study may have been influenced by the use of different antirheumatic drugs. However, our study provides real-life data on daily clinical practice. There may have been reporting bias in the onset of comorbidities between the former strategy groups, and some data were missing, but we do not expect that this would have affected the main results.
In conclusion, our results indicate that initiation of treat-to-target therapy with MTX and 10 mg prednisone daily in early RA results in a persistently lower initiation rate of a first bDMARD and significantly better radiographic outcomes. This was not clearly associated with increased incidence of long-term GC-related comorbidities.
References
Footnotes
Contributors All authors made substantial contributions to the conception or design of the work; or the acquisition, analysis or interpretation of data for the work; drafted the work or revised it critically for important intellectual content and made a final approval of the version to be published.
Competing interests JMvL received honoraria from MSD, Roche, Pfizer, BMS, Eli Lilly. JWJB received honoraria from AbbVie, BMS, MSD, Pfizer, Roche, SUN, UCB. MS was supported by a research grant from Astra Zeneca. Astra Zeneca was not involved in this study.
Ethics approval The institutional review boards of the participating centres confirmed that the Medical Research Involving Human Subjects Act (WMO) was not applicable to this study.
Provenance and peer review Not commissioned; externally peer reviewed.
Collaborators Contributing SRU hospitals and contact persons: University Medical Center Utrecht, Dr J. Tekstra; Antonius Hospital Nieuwegein/Utrecht, Dr E.J. ter Borg; Diakonessen Hospital Utrecht, Drs Y. Schenk; Meander Medical Center Amersfoort, Dr S. Linn-Rasker; Sint Jansdal Hospital Harderwijk, Drs J. Peeters; Tergooi Hospital Hilversum, Dr Z.N. Jahangier; Flevo Hospital Almere, Dr C.M. Verhoef.
Correction notice This article has been corrected since it published Online First. The legend for fig 1 has been added and the author names for M Safy and MJH De Hair corrected.