• anti-TNF
• DMARDs (biologic)
• DMARDs (synthetic)
• methotrexate
• rheumatoid arthritis

We read with great interest the article on ‘Gradual tapering TNF inhibitors vs conventional synthetic DMARDs after achieving controlled disease in patients with rheumatoid arthritis: first-year results of the randomised controlled TARA study’1by van Mulligen et al. This was the first head-to-head comparison between two tapering strategies—biological versus conventional in rheumatoid arthritis. The final results favour tapering tumour necrosis factor inhibitors (TNFis) before conventional synthetic disease-modifying antirheumatic drugs (csDMARDs). However, certain points need clarifications.

First, many patients in the study used combination csDMARDs, but no data have been provided on their number and specific combinations used. Furthermore, no clarity has been given on how the tapering was done in these patients who were on combination csDMARDs—was methotrexate the only drug reduced and stopped or were all the drugs in the combination reduced and stopped. In their trial registration (NTR2754), the authors have only mentioned methotrexate tapering, whereas the study title mentions csDMARD tapering.

Second, in their statistical analysis, they mention an ‘intention-to-treat analysis’. However, in their results, the authors have presented the clinical response after 12 months for both tapering groups, for only 85 and 89 patients in table 2 (final number of patients at 12 months) rather than the 94 and 95 patients who were initially randomised to the two tapering arms.

Third, the number of patients in clinical remission (disease activity score (DAS) <1.6) in TNFi tapering arm has been reported in table 2 as 58. However, the number of patients at risk at 12 months has been mentioned as only 54 (figure 2B). Thus, it is unclear that how the number of patients in clinical remission at 12 months more than the number at risk—ideally, all patients who dropped out or required to restart biologics or csDMARDs because of flare should be not considered in remission.

Finally, in the study protocol, the use of intra-articular glucocorticoids (GCs) and one intramuscular (IM) injection of GCs during a flare (as bridging therapy) was permitted. However, four and five patients, respectively, in csDMARDs and TNFi tapering arms received oral GCs and three patients in each arm got more than one IM injections. What was the effect of excluding these patients on the analysis would be interesting to know?