We thank Dr. Araujo1 for his interest on our systematic review about the perioperative use of gabapentinoids for the management of postoperative acute pain.2 The main purpose of a systematic review with meta-analyses is to synthesize treatment effects considering all available data. By pooling data from different trials, the power to detect a differential effect is increased. Trials included in a systematic review frequently present some degree of clinical heterogeneity. However, not considering these trials in pooled analyses, as suggested by Dr. Araujo, would diminish our collective ability to both answer important research questions and understand the modifiers of important treatment effects.
In our systematic review, we carefully considered potential sources of clinical heterogeneity among the included trials. First, we performed subgroup analyses to explore whether clinical heterogeneity or methodologic aspects of included trials could explain our findings and the observed statistical heterogeneity. As such, if a differential effect existed based on clinical factors, this effect would have been observed. We observed no such influence according to the type of drug used, dose regimen, duration of the intervention, type of surgery, or postoperative care pathway. We also observed that statistical heterogeneity was minimally attributable to the type of coanalgesia and the risk of bias in trials. The type of comparator was also not associated with the direction and magnitude of the treatment effect. Importantly, 90% of the trials included in our systematic review were conducted with placebo and not with an active comparator. Second, to mitigate the influence of heterogeneity, we used random effect models for our analyses. These models consider that results from individual trials may deviate from the true intervention effect due to sampling error and further assume that sources of variance may explain differences in effect across trials.
We also evaluated the certainty of the evidence based on Grading of Recommendations Assessment, Development, and Evaluation methodology, the recommended approach in knowledge synthesis methods.3 Accordingly, the summary estimates of our primary outcomes measure were considered to be of low certainty, meaning that the true effect may be different to the one observed. We must, however, understand that the certainty of the evidence in this instance was down-graded by the high proportion of trials at high or unclear risk of bias. Considering that trials at low risk of bias showed a smaller treatment effect (or none at all) of the intervention, it is thus very plausible that our results overestimate the true analgesic effect of gabapentinoids. Since the observed effect is already shown to not be clinically significant and our trial sequential analyses demonstrated that our sample size was substantially larger than the required sample size to evaluate the observed effect, additional trials evaluating the analgesic effect of the perioperative use of gabapentinoids are not required.
We also thank Drs. Su and Guan4 for their interest in our work on the perioperative use of gabapentinoids.2 The hypothesis raised in their letter explaining the absence of a clinically significant analgesic effect of gabapentinoids for the management of postoperative acute pain is interesting and is in accordance with our findings.
The main mechanism of action of gabapentinoids is hypothesized to be the binding of α2δ-1 subunit of calcium channel in the central nervous system and the dorsal root ganglion.5,6 The absent or minimal upregulation of α2δ-1 subunit in the acute perioperative period is an interesting hypothesis that could explain the lack of clinically significant analgesic effect from gabapentinoids for the management of acute pain condition. Although the exact role of the minimal upregulation from α2δ-1 in the postoperative period is unknown, we agree that this concept deserves further considerations.
The authors declare no competing interests.