Effect modification by contextual factors of urate-lowering therapy on serum urate in people with gout: A systematic review with meta-regression analysis

Objective: To synthesize evidence of the effect of contextual factors (CFs) on efficacy of urate-lowering therapy (ULT) on serum urate (SU) as outcome in gout patients. Methods: Randomised controlled trials (RCTs) from (updated) Cochrane reviews were the starting point. RCTs were included if they explored the role of any CF on efficacy of ULT on SU in gout patients. For CFs with sufficient data (i.e. >= 3 trials), a mixed-effects meta-regression analysis was performed with trial and comparison as random effects, whereas specific CFs were modelled as fixed factors. Results: Eight RCTs were included. Effect modification by CFs was explored for age, sex, race, renal function, cardiovascular comorbidity, tophi, thiazide-diuretic use, and previous ULT use. Crude data stratified by renal function were available for four trials (36 randomised comparisons), and suitable for meta-analysis. Pooled estimates revealed that gout patients with a normal, mildly-, or moderately impaired renal function were consistently more likely to achieve SU target with ULT compared to control. Among RCTs comparing ULT to placebo (30 comparisons), effects of ULT on achieving SU target were not statistically different for those with normal (OR:66.87; [11.39-392.75]) compared to mildly (OR:28.54; [5.11-159.46]) and moderately (OR:21.45; [3.20-143.64]) impaired renal function, but seemed lower in those with severely impaired (OR:9.13; [0.96-86.97]) renal function. Data were insufficient to draw conclusions on effect modification by other CFs. Conclusion: Few RCTs report stratified analyses exploring the role of CFs. ULT seemed effective in reaching the SU target in all levels of renal function, though severely impaired renal function appeared to render a slight disadvantage.