The effect of arrhyhthmia monitoring in post-MI patients: post-hoc analyses of the BIO-GUARD-MI trial

Background/Introduction BIO|GUARD-MI was an unblinded trial to investigate whether arrhythmia monitoring with implantable cardiac monitors (ICMs) can improve the clinical outcome in high-risk patients after myocardial infarction. The primary endpoint was a composite of cardiovascular (CV) death or hospitalisation for CV reasons. A higher incidence of non-CV adverse events (AEs) found by the DSMB was assumed to be due to an unknown performance bias and caused the premature study termination. The pre-planned analysis of the primary endpoint did not show a significant benefit (HR = 0.84, 95%-CI 0.65–1.10).[1] Recently, we have identified the cause of the assumed performance bias in the data.[2] In Denmark, patients from the treatment group visited the investigational sites on their own initiative much more often than patients of the control group (24.4 vs. 2.3%, P = 0.0004 after correction for multiple testing). An increased incidence of non-CV AEs suggests that this behaviour resulted in the reporting of more clinical events. Purpose The present analysis investigates whether the identified bias also modifies the incidence of primary endpoints. Methods The analysis of the study effect was done according to the study protocol with the Cox regression. The influence of the patients’ residence in Denmark on the primary study outcome was investigated by the interaction of both variables in this analysis. The analysis of recurrent events was post-hoc defined and done according to Andersen-Gill. Results Residence in Denmark significantly interacted with the study effect (P/interaction = 0.031). After exclusion of patients from Denmark, the primary Cox regression analysis (hazard ratio (HR) = 0.71, P = 0.031) and the analysis including recurrent events (HR = 0.66, P = 0.020) found a clinical benefit in the population outside Denmark, which was not affected by the performance bias. Conclusion Our analysis shows that the BIO|GUARD-MI study failed to show a benefit because treatment group patients from Denmark frequently visited the investigational site on their own initiative. This led to more clinical events being reported, most probably including primary endpoints such as hospitalisation for overnight observation of unstable angina, heart failure or arrhythmia related symptoms. The analysis of the remaining population shows that arrhythmia monitoring offers a clinical benefit of considerable amplitude.