Routine use of an invasive strategy (IS) has been shown to exceed a conservative strategy in reducing myocardial infarction (MI), angina and re-hospitalization rate in patients with non-ST elevation acute coronary syndrome (NSTEACS). The present study aimed to analyse, by use of randomized trials data, whether the risk profile of patients with NSTEACS influences the survival benefit of the IS over a conservative strategy from randomization to end of follow-up (range 6–24 months).
Eight studies were identified from 1970 to 2005. A fixed effect-meta-regression analysis for: (i) the log-odds ratio on death and (ii) the log-odds ratio on death/MI against the odds of death/MI in the control group was made.
IS was associated with a significant reduction in death/MI [12% versus 13.7%, odds ratio (OR) = 0.86, P = 0.009], but not in mortality (5.1% versus 5.5%, OR = 0.92, P = 0.34). There was evidence of heterogeneity in the outcome mortality (P = 0.06 for heterogeneity) and the composite of death/MI (P = 0.01 for heterogeneity). Sensitivity analysis demonstrated that the source of heterogeneity was significantly related to the outlier VANQWISH trial. When the latter was removed from the analysis, IS was related to a significant reduction of both death (3.9% versus 4.9%, OR = 0.81, P = 0.04, P heterogeneity = 0.35) and death/MI (10% versus 12.1%, OR = 0.81, P = 0.001, P heterogeneity = 0.07).
The main finding of this meta-analysis is that, compared to a conservative strategy, the benefits of IS for the management of NSTEACS in terms of death/MI reduction are related to the patient's risk profile.
Department of Cardiac, Thoracic and Vascular Sciences, University of Padova, Padova, Italy
Received 14 July, 2006
Revised 30 August, 2006
Accepted 17 September, 2006
Correspondence and requests for reprints to Giuseppe Tarantini, Cardiology Clinic, Department of Cardiac, Thoracic and Vascular Sciences, Policlinico Universitario, Via Giustiniani 2, 35128 Padova, Italy Tel: +39 0498211844; fax: +39 0498761764; e-mail: firstname.lastname@example.org