Comment on “The effect of perioperative ketamine and esketamine administration on postoperative nausea and vomiting in patients undergoing general anesthesia: a systematic review and meta-analysis”

Dear Editor,

We read with great interest the systematic review and meta-analysis by Seo et al. [1] in which they evaluated the effects of perioperative ketamine and esketamine on postoperative nausea and vomiting (PONV) following general anesthesia. This topic is highly relevant in the context of opioid-sparing anesthesia and enhanced recovery pathways. The authors are to be commended for synthesizing a large body of evidence to address this clinically important question. However, several methodological issues related to their study may influence the interpretation and applicability of these findings.

Although separate analyses for PONV, PON, and postoperative vomiting (POV) were also presented, the composite primary endpoint combines clinically distinct outcome definitions, introducing substantial clinical heterogeneity and potentially complicating interpretation of the principal findings [2–4].

First, their construction of a primary outcome warrants further investigation. Seo et al. [1] defined a composite “pooled PONV” as one of their primary endpoints by combining studies reporting PONV with those reporting postoperative nausea (PON) alone; when studies reported nausea and vomiting separately, only nausea was included in this pooled analysis. As nausea and vomiting are distinct outcomes, with different underlying pathophysiological mechanisms and clinical implications, their approach introduces substantial clinical heterogeneity [2]. Although separate analyses for PONV, PON, and POV were also presented, the composite primary endpoint combines clinically distinct outcome definitions, introducing substantial clinical heterogeneity and potentially complicating interpretation of the principal findings. Established guidelines emphasize that PON, POV, and combined PONV should be clearly defined and reported as distinct outcomes, using standardized postoperative observation periods to ensure clinical interpretability [3]. In addition, an approach of selecting the earliest postoperative time point when multiple time points are available may further compromise the clinical relevance of the pooled estimates.

Second, their comparator-based subgroup analyses should be interpreted with caution. In Seo et al. [1]’s report, some studies categorized under the “placebo” subgroup appear to have used add-on designs; i.e., ketamine or esketamine was additionally administered against an active background regimen, which was used as comparator, rather than a true inert control. Moreover, the “non-opioid” comparator group included pharmacologically diverse agents, including dexmedetomidine, dexamethasone, lidocaine, propofol, pregabalin, magnesium sulfate, diclofenac, and doxapram, each of which has differing mechanisms of action and established antiemetic effects [1]. Such pooling of heterogeneous comparators may have obscured clinically meaningful differences and could have confounded the conclusions drawn from such broad comparisons. In addition, several non-opioid comparator studies were conducted in female-only and high-PONV-risk surgical populations. Nevertheless, Seo et al. did not perform sex-stratified or Apfel score-based subgroup analyses. The absence of such stratification further limits the clinical generalizability of their findings.

Third, their handling of multi-arm trials warrants further clarification. Seo et al. [1] stated that, in selected small-group situations, duplicate comparisons were performed without proportional division of a shared comparator group. Standard meta-analytical guidance recommends appropriate handling of shared control groups in multi-arm trials to avoid unit-of-analysis errors and disproportionate weighting of individual studies [4]. Even when such an approach is pragmatically justified, it may distort the effect estimates, particularly in small subgroup analyses.

Fourth, although the inclusion of nonrandomized studies broadens the evidence base, it introduces additional risks of confounding factors and reduces internal validity. Outcomes such as PONV are highly sensitive to perioperative factors, including anesthetic techniques, opioid exposure, and antiemetic prophylaxis. Accordingly, reports offering methodological guidance have suggested that care should be taken when combining randomized and nonrandomized evidence to create a single pooled estimate [5].

Finally, the adverse effect profile warrants greater emphasis in terms of clinical interpretation. The results of the analysis demonstrated significantly increased risks of hallucinations and drowsiness for both drugs, with dizziness tending to be increased with ketamine and significantly increased with esketamine [1]. Thus, even if ketamine or esketamine demonstrates benefits in specific comparator contexts, particularly when compared against opioid-based regimens, these neuropsychiatric and sedative effects may offset their potential advantages, particularly when alternative non-opioid strategies with established antiemetic efficacy are available.

In conclusion, while Seo et al. provided valuable insights, the interpretability of their findings was constrained by the use of a hybrid primary endpoint, substantial comparator heterogeneity, and methodological issues related to multi-arm and mixed-design evidence synthesis. Future systematic review and meta-analysis studies are required to delineate the role of ketamine and esketamine in perioperative care better.