Correspondence on: “Breast implant explantation and capsulectomy in symptomatic patients. Is there any improvement with the procedure? Systematic review and meta-analysis” Authors: Luigi Annacontini, Vincenzo Verdura, Luigi Cagiano Published: 2026 Dear Editor, We read with great interest the systematic review and meta-analysis by Cuenca-Pardo et al.1 addressing whether explantation and capsulectomy improve symptoms attributed to breast implants. The question is highly topical, the distinction the Authors draw between objectively defined organic pathology and patient-reported systemic symptom clusters is clinically valuable, and we commend their candid acknowledgement that no included study provided a symptomatic control group. It is precisely because we share the direction of their clinical conclusions that we believe three aspects warrant clarification, lest the reported effect sizes be cited beyond what the design can support. First, the estimand underlying the pooled odds ratios is not defined. In the absence of any comparator group, it remains unclear whether OR 0.10 (95% CI 0.03–0.33) represents a within-arm pre–post contrast, a between-group comparison, or a pooled proportion. Equally, the outcome was collected on a 5-point ordinal Likert scale, yet the rule by which these scores were dichotomised for binary meta-analysis is not reported, precluding replication. The heterogeneity parameters compound this concern. The Results state H² = 3.15 alongside I² = 0.53, whereas these two quantities are algebraically linked by definition,2 since I² is obtained as (H² − 1) divided by H². The value H² = 2.15 given in the legend of Fig. 5 is the one consistent with I² = 0.53 and with Q = 40.73 on 19 degrees of freedom. More substantively, Tau² = 3.86 on the log-odds scale corresponds to τ ≈ 1.97, which represents severe rather than moderate dispersion. Deriving the 95% prediction interval from the Authors’ own figures yields approximately 0.001 to 7.5 — an interval comfortably spanning the null, and therefore indicating that a future study could plausibly observe no benefit at all. We would suggest that a prediction interval be reported alongside the summary estimate, as recommended for random-effects synthesis.3 Second, the composition of the evidence base raises a unit-of-analysis concern. The Authors report three systematic reviews, six clinical series and twenty case reports, a total that exceeds the twenty publications analysed and, more importantly, implies that patients already counted within primary series may be counted again through the included reviews. No formal risk-of-bias appraisal is presented for any included design, and no assessment of certainty of evidence is reported, both of which are core reporting items of the guideline the Authors state they followed.4 Relatedly, the Conclusions state that the incidence of implant-associated comorbidities was unrelated to dwell time, capsular contracture or rupture. Incidence cannot be estimated from a series restricted to explanted, symptomatic patients, since no population at risk is available; and cases of capsular contracture were explicitly excluded from the study population, which precludes any inference regarding that variable. Third, comparative data do exist and point in a partly different direction. In a cohort of 398 women followed at a dedicated clinic, Spit et al.5 compared patients who underwent explantation with those who retained their implants and reported greater symptom improvement after explantation (OR 2.9, 95% CI 1.3–6.2), together with significantly better outcomes when implants were removed within ten years of implantation. This demonstrates both that the comparator the present review lacks is obtainable, and that the conclusion regarding dwell time may not be as settled as stated. We would welcome the Authors’ view on this discrepancy. We also note an internal inconsistency worth correcting: an effect described in the Results as a trend is accompanied by p = 0.00 and a confidence interval excluding unity, while the improvement in Group I is described in the Discussion as non-significant despite an associated OR of 0.18 (95% CI 0.08–0.38; p = 0.00). In conclusion, we regard the clinical message of this work as sound and important. We would simply welcome the Authors’ clarification of the estimand, the dichotomisation rule and the risk-of-bias framework, and would respectfully suggest that the pooled odds ratios be presented as exploratory descriptors rather than as effect estimates — a framing that would, in our view, strengthen rather than weaken their central argument. Luigi Annacontini, MD, PhD Vincenzo Verdura, MD Luigi Cagiano, MD Plastic and Reconstructive Surgery Unit, A.O.U. Policlinico “Ospedali Riuniti” di Foggia, Italy Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Ethical approval Not required. This correspondence involves no human participants, human data or animal subjects. Author contributions All authors contributed to the conception of the commentary, the critical appraisal of the target article, the drafting and critical revision of the manuscript, and approved the final version to be submitted. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work the authors used AI-assisted technologies in order to assist with the verification of the reported statistical parameters and the language editing of the manuscript. After using this tool, the authors reviewed and edited the content as needed, independently verified all recalculated statistical values and all cited references, and take full responsibility for the content of the published article. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this letter. Footnotes Meeting presentation: None. This work has not been presented, wholly or in part, at any meeting. References