This review concluded that tranexamic acid appeared to be an effective blood-conserving agent that significantly lowered risks of death in cardiac surgery patients, but it was less effective than aprotinin and both have a risk of complications. The reliability of these conclusions remains unclear given potential limitations in the literature search, included studies and review methods.

To determine the efficacy and safety of tranexamic acid in comparison to aprotinin or no treatment control for patients undergoing cardiac surgery.

MEDLINE was searched from January 1995 to July 2009. Search terms were reported. Reference lists of retrieved papers, reviews and editorials were screened for further studies. Only English-language studies were eligible for inclusion in the review.

Randomised controlled trials (RCTs) (n≥15 per arm) and controlled observational studies (n≥150 treated patients) that reported relevant outcomes after cardiac surgery with or without postoperative blood loss and blood product transfusion were eligible for inclusion in the review. Observational studies had to use tranexamic acid patients matched to aprotinin or no treatment control patients using propensity scoring or another robust matching strategy. Relevant outcomes were postoperative blood loss, packed red cell transfusion, transfusion of haemostatic blood products such as fresh frozen plasma, platelets and cryoprecipitate, re-sternotomy for bleeding, postoperative myocardial infarction, postoperative neurological events and operative mortality. Studies that reported only intensive care outcomes or that included paediatric patients were excluded.

Included studies mainly assessed intravenously administered tranexamic acid (either as single dose, infusion or pump or both infusion and single dose); other studies assessed topically applied tranexamic acid. Total intravenous doses of tranexamic acid varied from 1g to 20g administered over a period of 20 minutes to 12 hours. Topical doses ranged from 1,000mg to 2,500mg. Tranexamic acid was compared with placebo in slightly fewer than half of the included RCTs; the other studies used either aprotinin (20%) or both aprotinin and placebo (28%).

The authors did not state how papers were selected for the review.

The authors did not state that they assessed validity.

Data were extracted to enable reporting of odds ratios (ORs) with 95% confidence intervals (CIs) for dichotomous outcomes (such as mortality) and mean differences with standard deviations for continuous outcomes (such as blood loss). Where standard deviations were not reported, they were imputed where possible from other data (further details of methods were reported in the review).

The authors did not state how many reviewers performed data extraction.

Studies were grouped by outcome and then comparator; RCTs and observational studies were analysed separately and combined in one analysis. Pooled odds ratios and mean differences, with 95% CIs, were calculated using a fixed-effect model or random-effects model where there was evidence of significant heterogeneity. Heterogeneity was assessed using the I² statistic and reported as significant where I² was more than 50% or p<0.10.

The review included 25 RCTs (n=5,411 patients) and four matched observational studies (n=5,977). Totals of 4,634 patients received tranexamic acid and 4,350 received aprotinin; 2,404 had no antifibrinolytic treatment.

In comparison with aprotinin, tranexamic acid was associated with a significantly increased mean postoperative blood loss (114mL, 95% CI 47 to 180; 11 RCTs), increased requirement for packed red cell transfusion (OR 0.70, 95% CI 0.54 to 0.92; 12 RCTs) and increased re-operation rate (OR 0.70, 95% CI 0.49 to 0.99; eight RCTs). There was no statistically significant difference between aprotinin and tranexamic acid with respect to the requirement for blood product transfusion (four RCTs), incidence of myocardial infarction (seven RCTs), incidence of neurological complications (four RCTs) and mortality rate (four RCTs).

Pooled observational studies showed no significant difference between tranexamic acid and aprotinin with respect to mean postoperative blood loss (two studies), requirement for packed red blood cell transfusion (two studies), requirement for blood product transfusion (two studies), re-operation rate (three studies), incidence of neurological complications (four studies) and mortality rate (three studies). Observational studies showed that in comparison with aprotinin, tranexamic acid was associated with a decreased incidence of myocardial infarction (OR 1.34, 95% CI 1.06 to 1.68; three studies).

Pooled results for RCTs and observational studies combined showed that in comparison with aprotinin, tranexamic acid was associated with a significantly increased mean postoperative blood loss (105mL, 95% CI 42 to 169), increased requirement for packed red blood cell transfusions (OR 0.72, 95% CI 0.56 to 1.08), decreased incidence of myocardial infarction (OR 1.24, 95% CI 1.02 to 1.51) and decreased mortality rate (OR 1.39, 95% CI 1.10 to 1.77). There was no significant difference between aprotinin and tranexamic acid with respect to the requirement for blood product transfusion, re-operation rate and incidence of neurological complications.

There was evidence of significant statistical heterogeneity for some or all pooled effect sizes for a number of outcomes and particularly for postoperative blood loss and packed red blood cell transfusion, which were analysed using a random-effects model.

Data that compared tranexamic acid to placebo and/or no treatment control was reported in the review.

The authors concluded that tranexamic acid appeared to be an effective blood-conserving agent that significantly lowered the risk of death in patients who underwent cardiac surgery, but may be associated with increased risks of complications. The review findings appeared to suggest that aprotinin was better than tranexamic acid despite the risk of complications.

This review answered a clearly defined review question. There was a risk of publication and language biases as only English-language studies were included and there were no specific attempts to locate unpublished data. Risks of reviewer error and bias were unclear as the reviewers did not report whether more than one reviewer was selected studies and extracted data. Methodological quality of the included studies was not assessed, so the reliability of the data and findings were unclear. Most studies were randomised controlled trials, but some studies used observational designs and so were more prone to bias; the authors considered this in their analysis by pooling different types of study separately and in a combined analysis. Presence of statistical heterogeneity was assessed, but it was difficult to assess the level of clinical heterogeneity due to a lack of information about study populations. The authors' stated that tranexamic acid appeared to be effective as a blood conserving agent, but the review findings appeared to show that aprotinin was better and so this conclusion appeared somewhat misleading.

Overall, the reliability of this review remains unclear given potential limitations in the literature search, included studies and review methods.

Practice: The authors stated that the trend towards indiscriminate use of tranexamic acid for all cardiac patients required re-evaluation given its potential to increase neurological complications. In particular its usage should be considered in patients at increased risk of developing neurological complications, such as the elderly and those with a history of stroke.

Research: The authors stated that further studies were required in order to clarify the neurological risk, appropriate indications and dosing of tranexamic acid for cardiac surgery.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.