Misrepresentation of results from the DANCAVAS II trial

This commentary refers to ‘Outcomes of cardiovascular screening in men aged 60–64 years: the DANCAVAS II trial’, by J.S. Lindholt et al., https://doi.org/10.1093/eurheartj/ehaf704 and the discussion piece ‘We all risk misrepresenting the results from the DANCAVAS trials’, by J.S. Lindholt and A.C. Diederichsen, https://doi.org/10.1093/eurheartj/ehaf1113. DANCAVAS II1 aimed to investigate multi-modal screening for cardiovascular disease over 10 years in men aged 60 to 64.1 The expected reduction in total mortality was not demonstrated at 7 years of follow-up [hazard ratio (HR) 0.96, 95% CI 0.88 to 1.03]. It is commendable to evaluate total mortality as the primary outcome, which requires a large trial with long follow-up. The importance was highlighted by the finding that aspirin likely caused a clinically important increase in major intracranial and gastrointestinal bleedings (HR 1.18, 95% CI 1.05 to 1.32). While not unexpected,2–4 it shows prevention can unintentionally cause harm. In a press release from the host institution, the lead author did not mention the negative primary outcome. Instead, he claimed a 17% reduction in total mortality based on a post hoc per-protocol analysis. The analysis attempted to correct for the 70% higher total mortality rate in non-attenders compared with controls. This problematic choice led to a press release that clashed with the conclusion of the abstract.1 After a critical news article,5 a revised press release was issued, which mentioned the primary outcome, but did not change the problematic claims and was also withdrawn.