Heterogeneity in multicentre trial participating centers: lessons from the TOPCAT trial on interpreting trial data for clinical practice

OBJECTIVES: Randomized controlled trials (RCTs) are considered a "gold standard" of evidence, provided they meet rigorous standards in design and execution. Recently, some investigators of the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist (TOPCAT) trial advocate reanalysis of results, deviating from the statistical analysis plan. We briefly review the rationale by the TOPCAT investigators and implications for interpreting trial data. STUDY DESIGN AND SETTING: Critical examination of existing literature. RESULTS: The TOPCAT trial showed variation in patient characteristics and outcomes among different geographic regions. The investigators suggest that the observed variation indicated unreliable data, warranting deviation from protocol. That lead to claims of therapeutic effectiveness for populations in select regions. We suggest that some variation is expected in multicentre RCTs and argue that discriminating between natural variation and unreliable data can be challenging. Thus, the warrant for deviation from protocol is not clear. CONCLUSION: The TOPCAT investigators highlight important concerns about heterogeneity in RCT samples and how that may impact our interpretation of the results. If we are to maintain rigor in the RCT methodology and preserve its status as a reliable form of evidence for clinical practice, we must carefully consider when it is appropriate to deviate from a protocol when analyzing and interpreting trial data.