Design, quality, and bias in randomized controlled trials of systemic lupus erythematosus.

OBJECTIVE

To appraise systematically the study design and quality of reporting of randomized controlled trials (RCT) on systemic lupus erythematosus (SLE) and to identify potential defects and biases.

METHODS

RCT with at least 5 patients with SLE were retrieved from MEDLINE, EMBASE, and the Cochrane Library. We analyzed study design, quality of reporting, and trial results.

RESULTS

Ninety-four trial reports (37 on lupus nephritis) were eligible with 2,257 SLE patients (n = 795 in lupus nephritis trials). Median sample size was 28 patients. Fifty-one trials (54.3%) were double blind, but only 31 (33.0%) mentioned the randomization mode, only 19 (20.2%) described allocation concealment, and only 7 (7.5%) were adequately powered. Sixty-three trials (67%) described adequately reasons for withdrawals. Nephritis trials had on average longer followup (p = 0.001) and were less likely to be double blind (p < 0.001), to describe reasons for withdrawals [both overall (p = 0.008) and per arm (p = 0.009)] and to involve a comparison against placebo or no treatment (p < 0.001). Larger trials scored higher on several quality characteristics. Significant efficacy or trend for efficacy was claimed in 72 reports (76.6%) and this was even more common in trials published in 1999-2002 (89.5%). Significant efficacy was found more frequently in trials that clearly specified withdrawals per arm (p = 0.001) and outcomes (p = 0.001) and used intention-to-treat analyses (p = 0.03). Besides outcome specification, no other quality variables seemed to improve significantly over time.

CONCLUSION

Several aspects of the design and reporting of RCT on SLE can be improved. Larger, adequately powered, and accurately reported trials are needed.