Abstract 3160: Comparison Of Primary Percutaneous Coronary Intervention and Fibrinolytic Therapy in Randomized Controlled Studies and Observational Studies: Results of Hierarchical Bayesian Meta-Analyses
Background: The previous meta-analyses that compared primary percutaneous coronary intervention (PPCI) with fibrinolytic therapy (FL) in ST-segment elevation myocardial infarction (STEMI) included exclusively randomized controlled trials (RCT). The external validity of RCTs was limited by selective patients (pts) enrolment and a lesser use of rescue and elective PCIs with FL. We aim to obviate these limitations by performing meta-analyses of both RCTs and observational studies.
Methods: All RCTs and observational studies of PPCI and FL for STEMI published from 1990 to 2006 were thoroughly reviewed for potential biases (selection, confounding, detection, ascertainment, performance and attrition). Performance bias was defined as “systematic difference between intervention groups in care apart from the intervention being evaluated” (The Cochrane collaboration).
Statistical analysis: Short-term (in-hospital to 6-week) and long-term (≥1-year) deaths and myocardial infarctions were summarized by hierarchical Bayesian random-effects meta-analyses. Non-informative prior was used so that our results were based mainly on the results of the reviewed studies and did not depend on prior beliefs.
Results: There was confounding bias noted in many observational studies: pts who underwent PPCI were more likely to have anterior STEMI, cardiogenic shock and longer symptom duration, and pts who received FL were more likely to be females, older and with more co-morbidities. A performance bias was noted in many observational studies with pts who received FL had less optimal coronary revascularization and medical therapy at discharge compared to pts who underwent PPCI. Table 1⇓ presents the short and long-term results.
Conclusion: PPCI was associated with reduced short-term mortality and reinfarction in RCTs and observational studies. There was no conclusive long-term mortality difference between PPCI and FL.