Background The primary outcome measure for efficacy in clinical trials of urate lowering therapy (ULT) is frequently serum urate (SU), effectively acting as a surrogate for patient-centred outcomes (e.g. gout flares). It has not been clearly demonstrated that the strength of the relationship between SU and patient-centred outcomes is strong enough for SU to be considered a surrogate.
Objectives To determine the strength of the relationship between SU and the incidence of gout flares and thereby to determine whether SU is a valid surrogate endpoint according to the BSES3 framework.
Methods A systematic search of Medline via PubMed, the Cochrane Central Register of Controlled Trials, EMBase via OVID, the Institute for Scientific Information Web of Science, and other sources identified relevant studies. Standardised data elements were extracted by 2 independent reviewers (LS, MBM) and disagreements resolved by discussion (RC). Eligible trials were parallel-group, randomised trials of ULT of at least 3 months duration in people with gout. For the meta-regression analysis, a mixed linear model was used to combine the ratio of flare rates between groups (SAS software, v9.4 for Windows). Trials with multiple arms were treated as individual trials, referred to as “randomised comparisons” (i.e., 3-arm trials with 2 active interventions was handled as 2 randomised comparisons). Data was analysed by meta-regression using the between-arm difference in proportion of individuals who achieved target SU (<0.36mmol/L) as independent variable from at 3 months (or 6 and 12 months if 3 month values not available) against flare rate (dependent variable).
Results After screening 234 abstracts, 82 trials were scrutinized, of which 9 trials (with 16 comparisons) met inclusion criteria. A total of 5,696 people with gout entered into the meta-regression model. The longest RCT included was only 12 months duration. The pooled Odds Ratio (OR) suggested a small but statistically significant favourable association between the active and comparator urate lowering therapies and flare frequency (OR, 0.83; 95% CI 0.70 to 0.99]). Substantial heterogeneity was present (between trial variance: 0.07; 0.03 to 0.30). Meta-regression analysis did not reveal any statistically significant association between the proportion of individuals who achieved target SU and the observed flare rate (P=0.82); the model fit did not improve after inclusion of the covariate into the model (between trial variance: 0.08; 0.03 to 0.33) (Figure).
Conclusions Substituting surrogate endpoints (proportion achieving target SU) for the important clinical outcome (gout flares) allows conduct of shorter smaller trials. However, based on aggregate trial-level data (meta-regression) an anticipated association between SU and gout flare could not be confirmed. Trial duration may have been too short to observe a reduction in flares and further work using data from long-term extension studies is underway.
Disclosure of Interest None declared