Background Non-ST elevation myocardial infarction (NSTEMI) is a major cause of mortality in the elderly. Many treatments including early percutaneous coronary intervention (PCI) improve survival post-NSTEMI, but the elderly receive these therapies less frequently. This may be due to the uncertainty of a favourable effect in the elderly because data on their short-term outcome from PCI for NSTEMI are limited.
Methods 165 207 NSTEMI first admissions from the Myocardial Ischaemia National Audit Project (MINAP) database (1 January 2004 to 31 December 2007). 30-day survival was assessed using a Cox proportional hazards model for subacute/elective PCI, adjusted for age group (<55, 55–64, 65–74, 75–84 and >85 years), gender, heart rate and systolic blood pressure on admission, diabetes, current smoking status, and previous acute myocardial infarction. Effect modification of PCI by age group was modelled using interaction terms adjusted for confounders. Interventions were included as time-dependent covariates and a hierarchical structure accounted for using random intercepts for hospitals. Sensitivity analysis investigated the inclusion of incomplete data.
Results Mean (SD) age 72 (13) years, 62% male, 21% diabetes. Drugs on discharge: aspirin (72%), beta-blocker (56%), statin (71%), ACE inhibitor (58%), clopidogrel (34%). 21 395 (13%) underwent PCI. The frequency of treatments (except for clopidogrel) decreased with age. 30-day mortality was 6.5%. PCI offered better 30-day mortality (unadjusted hazard ratio (HR) 0.15–0.22, p<0.001; adjusted HR 0.17–0.24, p<0.001). There was no evidence that the benefit from PCI differed by age group.
Conclusion These data suggest that elderly NSTEMI patients receive evidence-based therapies less frequently and have at least an equivalent benefit from PCI as their younger counterparts. This is important because the burden of disease in the elderly is set to increase and because early PCI for NSTEMI is recommended. However, these data should be interpreted in the light of potential selection biases.