Beta-blocakde to prevent perioperative death in non-cardiac surgery: a trial-sequential analysis

Andrea Messori, Coordinator,

Other Contributors:

August 08, 2013

In non-cardiac surgery, the role of beta-blocakde to prevent perioperative death is extremely controversial. While it is now agreed that the results from the DECREASE family of trials cannot be trusted (1), the evidence from the remaining 9 "secure" trials is somewhat difficult to interpret. The recent meta-analysis by Bouri et al. (1) has concluded that, according to these 9 trials, beta-blockade significantly increases mortality (risk ratio=1.27; 95% confidence interval: 1.01 to 1.60; p=0.04), but this result was at the limits of statistical significance.

Since trial-sequential analysis (TSA) can contribute to better interpret controversial findings from meta-analyses (2-4), we employed this statistical technique to re-examine the results from the 9 secure trials. The end-point was mortality. We assumed two-sided testing, type- 1 error=5%, power=80%, event frequency for controls=2.5% (i.e. the arithmetic cumulative event rate in the 9 control groups), and relative risk reduction (RRR)=25%. Boundaries for superiority, inferiority or futility were calculated according to the O'Brien-Fleming alpha-spending function. The graph of TSA was plotted according to a specific software (User Manual for TSA, Copenhagen Trial Unit 2011).

Figure 1 summarises our analysis (9 trials; 10,529 patients). Our results essentially confirm the conclusion by Bouri et al. (1). On the one hand, our trial-sequential graph showed that 17,495 patients would be the optimal sample size to show a relative difference of +/-25%. On the other hand, the z-curve based on the 9 trials reached an area very close to the boundary of inferiority. Although this boundary was not crossed, the pattern of the curve indicated that these results demonstrate at least futility (i.e. proof of no effectiveness), while they also tend to approach the demonstration of inferiority.

On the basis of this TSA, the use of beta-blockade in non-cardiac surgery is statistically proven to be ineffective (futility) and, additionally, is very likely to be harmful. Since futility of this pharmacological intervention is statistically demonstrated, in our view no further trials should be carried out to study the potential benefit in these patients; furthermore, in consideration that a preliminary evidence supports the hypothesis of a harmful effect, this is another reason why beta-blockade in non-cardiac surgery should be abandoned also in clinical practice.


1. Bouri S, Shun-Shin MJ, Cole GD, Mayet J, Francis DP. Meta-analysis of secure randomised controlled trials of ?-blockade to prevent perioperative death in non-cardiac surgery. Heart. 2013 Jul 31. doi: 10.1136/heartjnl-2013-304262.

2. Wetterslev J, Thorlund K, Brok J, Gluud C. Trial sequential analysis may establish when firm evidence is reached in cumulative meta- analysis. J ClinEpidemiol. 2008 Jan;61(1):64-75.

3. Messori A, Fadda V, Maratea D, Trippoli S. ?-3 Fatty Acid Supplements for Secondary Prevention of Cardiovascular Disease: From "No Proof of Effectiveness" to "Proof of No Effectiveness". JAMA Intern Med. 2013 Jun 17:1-2.

4. Messori A, Fadda V, Maratea D, Trippoli S. Erythropoiesis- stimulating agents in heart failure: no proof of effectiveness or proof of no effectiveness? Eur J Heart Fail. 2013 Aug;15(8):944-5.

Figure 1. Trial sequential analysis of 9 "secure" randomized trials. In the z-curve (represented in blue), individual trials correspond to individual segments; trials are plotted in chronological order (from left to right). The x-axis indicates the cumulative number of patients; the starting point of the z-curve is at x=0, i.e. inclusion of no trials. Abbreviations and symbols: red lines are the boundaries for superiority or inferiority while green lines are the boundaries for futility; T=treatment; C=controls. Availability: this Figure can be downloaded from the following Internet address: 2013.pdf

Conflict of Interest:

None declared

Conflict of Interest

None declared