Diverging associations of an intended early invasive strategy compared with actual revascularization, and outcome in patients with non-ST-segment elevation acute coronary syndrome: the problem of treatment selection bias

Aims In several observational studies, revascularization is associated with substantial reduction in mortality in patients with non-ST-segment elevation acute coronary syndrome (nSTE-ACS). This has strengthened the belief that routine early angiography would lead to a reduction in mortality. We investigated the association between actual in-hospital revascularization and long-term outcome in patients with nSTE-ACS included in the ICTUS trial. Methods and results The study population of the present analysis consists of ICTUS participants who were discharged alive after initial hospitalization. The ICTUS trial was a randomized, controlled trial in which 1200 patients were randomized to an early invasive or selective invasive strategy. The endpoints were death from hospital discharge until 4 year follow-up and death or spontaneous myocardial infarction (MI) until 3 years. Among 1189 patients discharged alive, 691 (58%) underwent revascularization during initial hospitalization. In multivariable Cox regression analyses, in-hospital revascularization was independently associated with a reduction in 4 year mortality and 3 year event rate of death or spontaneous MI: hazard ratio (HR) 0.59 [95% confidence interval (CI) 0.37–0.96] and 0.46 (95% CI 0.31–0.68). However, when intention-to-treat analysis was performed, no differences in cumulative event rates were observed between the early invasive and selective invasive strategies: HR 1.10 (95% CI 0.70–1.74) for death and 1.27 (95% CI 0.88–1.85) for death or spontaneous MI. Conclusion The ICTUS trial did not show that an early invasive strategy resulted in a better outcome than a selective invasive strategy in patients with nSTE-ACS. However, similar to retrospective analyses from observational studies, actual revascularization was associated with lower mortality and fewer MI. Whether an early invasive strategy leads to a better outcome than a selective invasive strategy cannot be inferred from the observation that revascularized patients have a better prognosis in non-randomized studies.