Article type
Year
Abstract
Background: Aprotinin has been shown to be effective in reducing peri-operative blood loss and the need for re-operation due to continued bleeding in cardiac surgery. The lysine analogs tranexamic acid (TXA) and epsilon aminocaproic acid (EACA) are cheaper alternatives, but it is not known if they are as effective as aprotinin. We performed meta-analyses of the comparative trials of aprotinin, TXA, and EACA and included tests of equivalence (non-inferiority).
Methods and Results: Randomized parallel group trials in elective cardiac surgery were identified by searching electronic databases and bibliographies of published articles. Data were pooled using Cochrane Review Manager 4.1. We used a Bayesian approach to calculate the posterior probability of TXA and EACA being equivalent to aprotinin using a non-inferiority boundary of a (relative) 20% increase in rate of transfusion or re-operation with the former drugs. Peri-operative blood loss was significantly greater with TXA and EACA than with aprotinin: WMD +106ml 95% CI 37,176ml and +185ml,134,235 ml respectively. The pooled relative risks (RR) of receiving an allogeneic RBC transfusion with TXA (10 trials, 1707 subjects) and EACA (6 trials, 398 subjects), compared with aprotinin, were 1.08 (95% CI 0.88, 1.32) and 1.14 (0.84, 1.55) respectively. The equivalent Bayesian posterior mean RR were 1.15 (95% Bayesian Credibility Interval 0.90, 1.68) and 1.21 (0.79, 1.82) respectively. Using a 20% non-inferiority for transfusion requirement as the threshold, the probabilities of TXA and EACA being equivalent to aprotinin were 0.71 and 0.57 respectively. Re-operation data were sparse and calculations were only made for TXA v aprotinin. The RR calculated by Review Manager was 0.98 (0.51, 1.88) compared with a posterior mean RR of 0.80 (95% BCI 0.14, 1.86). The median posterior RR was 0.71, indicating skewness. Using a 20% non-inferiority boundary the probability of equivalence of TXA to aprotinin was 0.88.
Conclusions: The available data are conflicting regarding the equivalence of lysine analogs and aprotinin in reducing peri-operative bleeding, transfusion and the need for re-operation. The Meta-view estimate of the RR for re-operation with TXA compared with aprotinin appeared to be null-biased by the corrections applied to zero cells, and may have underestimated the efficacy of the drug. The distribution of the Bayesian estimate of the posterior RR for TXA v aprotinin was wide and skewed due to sparse data, but generated a higher probability of equivalence to aprotinin than did the more plentiful transfusion data-set.
Methods and Results: Randomized parallel group trials in elective cardiac surgery were identified by searching electronic databases and bibliographies of published articles. Data were pooled using Cochrane Review Manager 4.1. We used a Bayesian approach to calculate the posterior probability of TXA and EACA being equivalent to aprotinin using a non-inferiority boundary of a (relative) 20% increase in rate of transfusion or re-operation with the former drugs. Peri-operative blood loss was significantly greater with TXA and EACA than with aprotinin: WMD +106ml 95% CI 37,176ml and +185ml,134,235 ml respectively. The pooled relative risks (RR) of receiving an allogeneic RBC transfusion with TXA (10 trials, 1707 subjects) and EACA (6 trials, 398 subjects), compared with aprotinin, were 1.08 (95% CI 0.88, 1.32) and 1.14 (0.84, 1.55) respectively. The equivalent Bayesian posterior mean RR were 1.15 (95% Bayesian Credibility Interval 0.90, 1.68) and 1.21 (0.79, 1.82) respectively. Using a 20% non-inferiority for transfusion requirement as the threshold, the probabilities of TXA and EACA being equivalent to aprotinin were 0.71 and 0.57 respectively. Re-operation data were sparse and calculations were only made for TXA v aprotinin. The RR calculated by Review Manager was 0.98 (0.51, 1.88) compared with a posterior mean RR of 0.80 (95% BCI 0.14, 1.86). The median posterior RR was 0.71, indicating skewness. Using a 20% non-inferiority boundary the probability of equivalence of TXA to aprotinin was 0.88.
Conclusions: The available data are conflicting regarding the equivalence of lysine analogs and aprotinin in reducing peri-operative bleeding, transfusion and the need for re-operation. The Meta-view estimate of the RR for re-operation with TXA compared with aprotinin appeared to be null-biased by the corrections applied to zero cells, and may have underestimated the efficacy of the drug. The distribution of the Bayesian estimate of the posterior RR for TXA v aprotinin was wide and skewed due to sparse data, but generated a higher probability of equivalence to aprotinin than did the more plentiful transfusion data-set.