Dear Editor,

Rose and colleagues present an important systematic review and meta-analysis of haemodynamic changes following dexmedetomidine administration in children with CHD undergoing cardiac procedures. ^{Reference Rose, Hassanieh and Jackson1} Their broad search strategy, adherence to Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA), restriction to cardiac procedures, and use of Risk-of-bias 2 (RoB 2) and Risk Of Bias In Non-randomized Studies - of Exposures (ROBINS E) to assess risk of bias are all notable strengths. These design choices help consolidate a scattered literature and focus on early haemodynamic effects that are clinically relevant to perioperative decision making.

Several methodological aspects, however, warrant clarification to guide interpretation. First, the included studies primarily report within-patient pre- and post-measures, yet the meta-analyses treat pre-dexmedetomidine values as “control” and post values as “intervention” groups. Standard methods for meta-analyses of repeated measures emphasise that effect sizes and their variances should account for the correlation between pre- and post-measurements; treating these as independent groups inflates the standard errors and distorts study weights, even when the mean difference itself is unbiased. ^{Reference Skvarc and Fuller-Tyszkiewicz2,Reference Jané, Harlow and Khu3} Explicit use of change-score or repeated-measures effect size formulas, ideally with sensitivity analyses across plausible pre–post correlations, would strengthen the precision and transparency of the pooled estimates.

Second, there is an internal inconsistency between the abstract and the main text for diastolic blood pressure. The abstract describes a decrease of 6.2 mmHg, whereas the diastolic blood pressure meta-analysis reports a mean difference of approximately 4.6 mmHg with 95% confidence intervals of about 3.5 to 5.7 mmHg. This difference is not trivial relative to the magnitude of the effect and could reflect a descriptive average, a different subset of studies, or an earlier model. Clarifying which estimate is primary and ensuring consistency across sections are important to avoid overstatement of the typical blood pressure reduction.

Third, the authors report very high heterogeneity across all pooled analyses, with I ² values ranging from roughly 76% to 95%. Contemporary guidance stresses that such heterogeneity should be described not only with I ² but also through prediction intervals and structured exploration of sources of between-study variability. ^{Reference Borenstein4} The protocol appropriately planned meta regression on age, procedure type, route of administration, and risk of bias, but this could not be implemented due to sparse or poorly stratified reporting. In this context, a stronger emphasis on the wide dispersion of effects, together with sensitivity analyses by study design, dosing regimen, or risk of bias strata, would help readers understand how far individual settings might depart from the pooled mean.

Finally, the authors correctly employ RoB 2 for randomised trials and ROBINS E for observational studies, and note that all included studies are at least at moderate risk of bias. Current recommendations suggest that such assessments should inform both quantitative synthesis and the strength of inferential claims. ^{Reference Sterne, Savović and Page5,Reference Higgins, Morgan and Rooney6} Explicit stratification or down weighting of high-risk studies, or at minimum a more cautious framing of the findings as exploratory rather than confirmatory evidence of haemodynamic safety, would align even more closely with these tools’ intended use.

Rose et al. substantially advance synthesis in a clinically important area, but attention to the paired nature of the data, internal numerical consistency, and the consequences of high heterogeneity and study-level bias would refine the interpretation of their reassuring conclusions.