Optimising the analysis of vascular prevention trials: Re-Assessment of the TARDIS trial, the first prevention trial to adopt an ordinal primary outcome measure.

Background: Ordinalised vascular outcomes incorporating event severity are more informative than binary outcomes that just include event numbers. The TARDIS trial was the first vascular prevention study to use an ordinalised vascular outcome as its primary efficacy and safety measures and collected severity information for other vascular events. Methods: TARDIS was an international prospective randomised open-label blinded-endpoint trial assessing one month of intensive versus guideline antiplatelet therapy in patients with acute non-cardioembolic stroke or TIA. Vascular events and their severity were recorded up to final follow-up at 90 days post randomisation. For each outcome, statistical techniques compared ordinal/continuous (10 models) and dichotomous (5 models) analyses; results were then ranked with the smallest p-value being given the smallest rank. Outcomes were also assessed within the pre-defined subgroup of participants with mild stroke (NIHSS≤3), or TIA recruited within 24 h. Results: Ordinal versions of vascular event outcomes were created in 3096 participants for stroke, myocardial infarction, major cardiac events, bleeding events, serious adverse events and venous thromboembolism (VTE), with 32 outcomes being created overall (29 in the subgroup population due to the absence of VTE events). Overall, the tests run on ordinal outcomes tended to rank higher than tests performed on binary outcomes. 764 (24.7%) participants were recruited within 24 h of a mild stroke/TIA; again, tests run on ordinal outcomes ranked higher. Conclusions: In TARDIS, tests performed on ordinal vascular outcomes tended to attain a higher rank than those performed on binary outcomes. Trial registration: ISRCTN47823388.