Jochen Börgermann1, Benita von Salviati2, Jan F. Gummert1, Johannes Haerting2, Oliver Kuß2.
1Department of Cardiac and Thoracic Surgery, University Hospital Jena, Jena, Germany, 2Institute of Medical Epidemiology, Biostatistics, and Informatics, Martin-Luther-University Halle-Wittenberg, Halle/Saale, Germany.
Background: After numerous observational studies (OS) and randomized studies (RS), neither off- nor on-pump revascularization comes out superior. Patient selection and small sample size limit the compilation of clinically relevant outcomes in RS; lack of randomization limits OS. Propensity score analyses (PSA) are expected to improve on at least some of these problems.
Methods: PSA on off- vs. on-pump surgery were identified from eight bibliographic data bases, citation tracking, and a free web search. Two independent reviewers abstracted data on eleven binary short-term outcomes. We used the odds ratio (OR) to describe the treatment effect. To combine ORs from different studies, the random effects inverse-variance method was applied.
Results: 35 of 58 initially retrieved PSA were included, accounting for a total of 123,137 patients. The estimated odds ratio was <1 for all outcomes in favour of off-pump surgery. This was significant for mortality, stroke, renal failure, RBC transfusion (p<0.0001), wound infection (p<0.001), prolonged ventilation (p<0.01), inotropic (p=0.02) and IABP support (p=0.05). The odds ratios for myocardial infarction and atrial fibrillation remained nonsignificant.
Conclusions: Like OS, but unlike RS, PSA find significant advantages for relevant outcomes with off-pump surgery. Study populations need to be compared to clarify if different results of PSA and RS are related to patient selection (limited external validity).
Back to 2008 Annual Meeting
Back to Program Outline