Reply to: Indication-dependent heterogeneity in the effect of double-suture cerclage: concerns over indication classification and a call for stratified analysis Authors: Edis Kahraman, Nadiye Köroğlu, Turgut Aydın, Mehmet Aytaç Yüksel Published: 2026 To the editor, We thank Park and Park for their constructive correspondence regarding our study, “Does suture number matter in transvaginal cervical cerclage? A propensity score-weighted cohort study” [1]. We appreciate their careful reading of our article and agree that cerclage indication represents a clinically relevant source of heterogeneity. History-indicated, ultrasound-indicated, and physical examination-indicated/emergency cerclage reflect distinct clinical scenarios, with potentially different implications for suture reinforcement, as also suggested by prior meta-analytic and review literature on single- versus double-suture cerclage [2, 3]. The authors raised two related points: whether pooling cerclage indications could obscure indication-specific effects, and whether the original supplementary indication variable adequately distinguished conventional indication categories. The primary objective of our original study was to compare single- versus double-suture cerclage with respect to delivery timing and related obstetric and neonatal outcomes, rather than to test indication-specific treatment effects. In that context, the binary “prophylactic” versus “emergency” classification was used as a pragmatic operational distinction reflecting the clinical and operative context, including urgency of cerclage placement and intraoperative assessment, rather than as a fully granular indication-based classification. We recognize that physical examination-indicated cerclage and intraoperative urgency are related but not identical concepts. The former refers to clinical findings at presentation, whereas the latter reflects operative context in a retrospective dataset. This approach also reflected the limited sample size and the unequal treatment allocation as subdivision into three indication categories would have produced sparse treatment cells and unstable subgroup estimates. To address this point, we reviewed the electronic medical records, the operative notes, and the original supplementary dataset. Relevant clinical variables were re-evaluated, and cases were re-adjudicated into three indication-based categories: history-indicated, ultrasound-indicated, and physical examination-indicated/emergency cerclage. The original binary variable was retained for audit-trail comparison, whereas the re-adjudicated three-category variable was used for post-publication sensitivity analysis. After re-adjudication, 24 women were classified as history-indicated, 59 as ultrasound-indicated, and 42 as physical examination-indicated/emergency cerclage (Table 1). Table 1Cross-tabulation of original binary indication labels and re-adjudicated three-category cerclage indicationsOriginal categoryHistory-indicatedUltrasound-indicatedPhysical examination-indicated/emergencyTotalOriginal prophylactic21131852Original emergency3462473Total245942125Original categories reflect the binary operational classification used in the published supplementary dataset. Re-adjudicated categories were assigned after re-review of the electronic medical records, operative notes, and original supplementary dataset according to predefined clinical criteria. Therefore, Table 1 should be interpreted as a post-publication clarification of indication coding rather than as a replacement for the original primary analytical framework Table 1Cross-tabulation of original binary indication labels and re-adjudicated three-category cerclage indicationsOriginal categoryHistory-indicatedUltrasound-indicatedPhysical examination-indicated/emergencyTotalOriginal prophylactic21131852Original emergency3462473Total245942125Original categories reflect the binary operational classification used in the published supplementary dataset. Re-adjudicated categories were assigned after re-review of the electronic medical records, operative notes, and original supplementary dataset according to predefined clinical criteria. Therefore, Table 1 should be interpreted as a post-publication clarification of indication coding rather than as a replacement for the original primary analytical framework Cross-tabulation of original binary indication labels and re-adjudicated three-category cerclage indications Original categories reflect the binary operational classification used in the published supplementary dataset. Re-adjudicated categories were assigned after re-review of the electronic medical records, operative notes, and original supplementary dataset according to predefined clinical criteria. Therefore, Table 1 should be interpreted as a post-publication clarification of indication coding rather than as a replacement for the original primary analytical framework Principal analysis was then repeated using this re-adjudicated indication variable. The overall association between double-suture cerclage and term delivery remained materially unchanged. The unadjusted risk ratio for term delivery comparing double- versus single-suture cerclage was 0.77 (95% CI 0.57–1.03). After inverse probability of treatment weighting, the adjusted risk ratio was 0.87 (95% CI 0.62–1.22), and in the doubly robust IPTW-adjusted model, the risk ratio was 0.87 (95% CI 0.70–1.07). Thus, double-suture cerclage was not associated with a statistically significant improvement in term delivery in the overall cohort after re-adjudication of indication. Indication-stratified analysis was also repeated. In the ultrasound-indicated subgroup, term delivery occurred in 11 of 14 single-suture cases and 26 of 45 double-suture cases, yielding a risk ratio of 0.74 (95% CI 0.51–1.07). In the physical examination-indicated/emergency subgroup, term delivery occurred in 5 of 8 single-suture cases and 16 of 34 double-suture cases, yielding a risk ratio of 0.75 (95% CI 0.40–1.43). The history-indicated subgroup could not be meaningfully analyzed by suture number as only one patient in this subgroup received a single-suture cerclage. Interaction analysis did not show a statistically significant difference in the association between suture number and term delivery across re-adjudicated indication categories (p = 0.230) (Table 2). Table 2Re-adjudicated indication-stratified term delivery according to suture numberRe-adjudicated indicationSingle sutureDouble sutureRR for double vs single95% CIHistory-indicated(Prior mid-trimester loss or spontaneous PTB + closed/intact cervix)1/1, 100.0%16/23, 69.6%Not estimableNot estimableUltrasound-indicated(Short cervix before 24 weeks + no dilation + no visible/prolapsed membranes)11/14, 78.6%26/45, 57.8%0.740.51–1.07Physical examination-indicated/emergency(Cervical dilation and/or visible/prolapsed membranes, no labor/ROM/infection)5/8, 62.5%16/34, 47.1%0.750.40–1.43CI confidence interval, RR risk ratio, ROM rupture of membranes, PTB preterm birthThe history indicated subgroup was not suitable for comparative analysis because only one patient received single-suture cerclageThe suture number × re-adjudicated indication interaction was not statistically significant (p = 0.230) Table 2Re-adjudicated indication-stratified term delivery according to suture numberRe-adjudicated indicationSingle sutureDouble sutureRR for double vs single95% CIHistory-indicated(Prior mid-trimester loss or spontaneous PTB + closed/intact cervix)1/1, 100.0%16/23, 69.6%Not estimableNot estimableUltrasound-indicated(Short cervix before 24 weeks + no dilation + no visible/prolapsed membranes)11/14, 78.6%26/45, 57.8%0.740.51–1.07Physical examination-indicated/emergency(Cervical dilation and/or visible/prolapsed membranes, no labor/ROM/infection)5/8, 62.5%16/34, 47.1%0.750.40–1.43CI confidence interval, RR risk ratio, ROM rupture of membranes, PTB preterm birthThe history indicated subgroup was not suitable for comparative analysis because only one patient received single-suture cerclageThe suture number × re-adjudicated indication interaction was not statistically significant (p = 0.230) Re-adjudicated indication-stratified term delivery according to suture number History-indicated (Prior mid-trimester loss or spontaneous PTB + closed/intact cervix) Ultrasound-indicated (Short cervix before 24 weeks + no dilation + no visible/prolapsed membranes) Physical examination-indicated/emergency (Cervical dilation and/or visible/prolapsed membranes, no labor/ROM/infection) CI confidence interval, RR risk ratio, ROM rupture of membranes, PTB preterm birth The history indicated subgroup was not suitable for comparative analysis because only one patient received single-suture cerclage The suture number × re-adjudicated indication interaction was not statistically significant (p = 0.230) These additional analyses address the indication-heterogeneity concern raised in the correspondence. Although the original binary indication variable was less granular than the three-category framework proposed by Park and Park, re-adjudicated analysis did not demonstrate a statistically significant benefit of double-suture cerclage overall, or within the ultrasound-indicated or physical examination-indicated/emergency subgroups. Therefore, additional three-category sensitivity analysis did not materially alter the interpretation of our original findings. These results should still be interpreted within the limitations acknowledged in the original article, including retrospective design, non-random treatment allocation, unequal group sizes, and temporal evolution in surgical practice. Subgroup estimates remain limited by sample size, statistical power, and treatment overlap. Accordingly, our findings should be interpreted as cohort-level observational evidence rather than definitive evidence of equivalence across all cerclage indications. The possibility of an indication-specific effect in carefully defined physical examination-indicated/emergency cerclage remains a relevant question for future prospective studies and randomized trials [4]. We thank Park and Park for their thoughtful comments. Their correspondence provided an opportunity to clarify indication classification, repeat relevant analyses, and confirm that additional three-category sensitivity analysis did not materially alter the interpretation of our findings.