We are not convinced that the results of the article by Labarere et al1 warrant the recommendation that patients with an acute myocardial infarction (AMI) should be admitted to a hospital with on-site percutaneous coronary intervention (PCI) facilities. In this observational study, there is an inequality in baseline characteristics of patients admitted to one or the other hospital. Most characteristics indicative for a worse prognosis were more prevalent in patients admitted to hospitals without on-site PCI. They were older and had more heart failure, more left bundle-branch block, and more chronic renal failure. This suggests a selection bias, favoring the admission of patients with a better prognosis to hospitals with on-site PCI. Although appropriate statistical methods have been used, it cannot be excluded that some unmeasured confounding factor (eg, severity of comorbid conditions and socioeconomic factors) may annihilate the adjusted residual difference in mortality between the 2 groups.
Van Brabandt H, Vrijens F. Are Thrombolytic Drugs Not Evidence-Based Medications?. Arch Intern Med. 2008;168(2):239–240. doi:10.1001/archinternmed.2007.47