Response to “Evaluating inconsistences, generalizability, and heterogeneity in analysis of carbetocin versus oxytocin after cesarean delivery in high-risk women”

Dear Editor,

We would like to thank Dr. Marcela da Silva Kazitani Cunha and colleagues for their thoughtful interest in our article [1]. Constructive scientific discussion plays a critical role in improving the rigor and transparency of evidence synthesis.

General comments

Our systematic review aimed to provide the most comprehensive and methodologically rigorous synthesis to date on the use of carbetocin versus oxytocin for prevention of postpartum hemorrhage (PPH) in high-risk women undergoing cesarean delivery. The focus on this population fills a critical gap in literature where previous meta-analyses included heterogeneous obstetric populations or failed to stratify by risk level. Our analysis followed PRISMA 2020 guidelines, applied an a priori registered protocol, and assessed evidence quality using the GRADE framework. The criticisms raised by Cunha et al. offer valuable opportunity to clarify key aspects of our methodology and interpretation.

Subgroup analyses and apparent sample size discrepancies

Cunha et al. expressed concern regarding the total number of participants reported across different figures, suggesting potential duplication of data within subgroup analyses. We wish to clarify that no patient or study was counted more than once in any pooled analysis. Each figure in the manuscript corresponds to a specific outcome, and the total number of participants varies according to the subset of studies reporting each outcome.

The number of participants (n = 3652) in Fig. 9, for instance, reflects the cumulative sample of all trials contributing to that particular endpoint, whereas the total sample of 3068 reported in the main text corresponds to the population included in the primary analysis (PPH incidence). This distinction is consistent with PRISMA recommendations. Thus, there was no double-counting, and the integrity of the pooled estimates remains intact.

GRADE assessment and evaluation of risk of bias

Our approach was aligned with the GRADE Working Group and Cochrane Handbook recommendations [2]. Each domain (risk of bias, inconsistency, indirectness, imprecision, and publication bias) was independently evaluated. The overall certainty of evidence for each outcome was downgraded when substantial heterogeneity (I2 > 75%) or serious methodological limitations were detected.

While several included studies demonstrated a high risk of bias in at least one domain, sensitivity analyses revealed that their exclusion did not materially affect the pooled estimates. Therefore, downgrading the certainty further would have been overly conservative and inconsistent with GRADE guidance.

Heterogeneity and sensitivity analyses

We agree that high heterogeneity requires thorough exploration. Random-effects models were applied to account for between-study variation. Subgroup analyses were conducted based on carbetocin dose, study design, and risk profile. A leave-one-out sensitivity analysis was performed, demonstrating consistent direction and magnitude of effect across studies, confirming robustness of our conclusions.

High I2 values (e.g., 97%) are common in meta-analyses addressing complex obstetric outcomes where clinical protocols and risk factors vary considerably [3]. Importantly, heterogeneity does not necessarily indicate bias but reflects real-world variability. We explicitly acknowledged this in our discussion and avoided overgeneralization.

Inclusion of the Elgafor El Sharkwy (2013) study

Cunha et al. questioned the inclusion of Elgafor ElSharkawy study in our systematic review due to concomitent use of misoprostol in the oxytocin arm [4]. This trial was included after careful evaluation since both groups received oxytocin, and the differential factor—carbetocin versus combination regimen—provided clinically relevant insight into relative efficacy in high-risk settings. Excluding this study did not alter the direction or significance of our findings (PPH incidence: pooled RR 0.68 [95% CI 0.54–0.84] including vs. 0.70 [95% CI 0.

Choice of effect measure (odds ratio vs. risk ratio)

We selected odds ratios (ORs) as the summary statistic for dichotomous outcomes to ensure consistency across studies that reported logistic regression results or lacked full 2 × 2 data. ORs provide stable estimates across varying baseline event rates and are suitable for random-effects models [5]. Nevertheless, to confirm robustness, we reanalyzed primary outcomes using risk ratios (RRs), which yielded consistent findings. Therefore, the choice of measure did not influence clinical interpretation.

Publication bias and reporting considerations

The authors suggested that publication bias assessment was omitted. In our review, fewer than 10 studies were eligible for the primary outcome, rendering funnel plot asymmetry tests statistically unreliable. We transparently reported this limitation and noted that future updates with more trials will allow for formal bias assessment.

Strengths, novelty, and clinical implications

Our systematic review represents the first meta-analysis dedicated exclusively to high-risk cesarean deliveries, a group most vulnerable to PPH yet often underrepresented in randomized trials. Our inclusion of diverse international data enhances generalizability across healthcare settings. The study employed robust methodological safeguards: comprehensive database searches, dual independent screening, and transparent protocol registration.

Clinically, our findings reinforce that carbetocin provides longer uterotonic duration and reduces need for additional agents compared to oxytocin, offering a practical advantage for PPH prevention in both high- and limited-resource settings. These results align with WHO and FIGO recommendations endorsing carbetocin.

Sincerely,

Ahmed M Maged.

Department of Obstetrics and Gynecology, Cairo University, Cairo, Egypt.

Author contributions

AM write prepared and revised the manuscript.

Funding

No funding was received.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

The authors declare no competing interests.