Comment on: Sarcopenic Obesity, Not Sarcopenia or Obesity Alone, Is Independently Associated With Urinary Incontinence in Older Women

We read with interest the article by Önür et al. on sarcopenic obesity and urinary incontinence (UI) in older women [1]. The topic is timely, as an umbrella review has highlighted the uncertain but potentially important impact of sarcopenic obesity on health outcomes. However, several aspects of the statistical interpretation and phenotype definition make the central conclusion—that sarcopenic obesity, but not sarcopenia or obesity alone, is more strongly and independently associated with UI—less secure than it appears. First, the claim of a “more strongly” associated phenotype is not supported by the analyses presented. In the multivariable model, the adjusted odds ratios for UI were 1.82 (95% CI 1.16–2.85) for sarcopenic obesity, 1.57 (0.97–2.55) for sarcopenic non-obese, and 1.35 (0.91–2.00) for non-sarcopenic obese. Only the first estimate reaches the conventional p < 0.05 threshold, but the confidence intervals overlap, and no formal comparison of coefficients between phenotypes is reported. Recent guidance on p values and compatibility intervals has emphasized that treating “significant” versus “non-significant” results as evidence that two effects differ is misleading when their intervals largely overlap. A more cautious reading is that only the sarcopenic-obese category clearly reaches conventional statistical significance, while the estimated effects for sarcopenia alone and obesity alone are similar in magnitude but imprecisely estimated. Second, the covariate strategy risks attenuating the pathway through which sarcopenic obesity may influence UI. Limitations in instrumental activities of daily living (IADL) are plausibly downstream of sarcopenic obesity and upstream of UI. Treating IADL as a confounder rather than a mediator can introduce overadjustment bias and move the total exposure–outcome association towards the null. Recent meta-research has highlighted that adjustment for potential mediators remains common in observational work [2]. In Önür et al.'s models, such overadjustment could contribute to the non-significant p values for sarcopenia-only and obesity-only phenotypes, complicating the interpretation of “no independent association” as “no clinically important relationship.” Third, the operationalization of sarcopenic obesity combines a weight-adjusted skeletal muscle mass index (muscle mass divided by body weight) with a body-fat percentage criterion. Greater adiposity simultaneously increases body weight, lowering the muscle index, and raises body-fat percentage, making individuals with high fat mass more likely to be classified as both “low muscle” and “obese.” Recent work applying the ESPEN–EASO algorithm in older populations has shown that choices about muscle indexing and adiposity cut-offs can substantially change prevalence and risk estimates [3]. Large cohort data using ESPEN–EASO–based definitions further suggest that both sarcopenia and sarcopenic obesity confer excess mortality risk, underscoring the need for phenotype definitions not unduly shaped by the mathematical coupling of muscle and adiposity [4]. Sensitivity analyses using height-adjusted indices and alternative adiposity metrics (e.g., appendicular lean mass/height2 and waist circumference or BMI-based obesity) would help show that the association with UI reflects a robust biological interaction rather than, at least in part, a consequence of definitional coupling. In summary, Önür et al. provide valuable data, but the assertion that sarcopenic obesity is uniquely and more strongly associated with UI than sarcopenia or obesity alone appears stronger than the current analyses can justify. Clarifying whether phenotype coefficients genuinely differ, reconsidering adjustment for potential mediators, and testing alternative phenotype definitions in sensitivity analyses would strengthen the evidential basis for this conclusion. Y.W. conceived the idea for this letter and drafted the manuscript. H.Z. provided critical revisions and intellectual input. Both authors approved the final version and are accountable for all aspects of the work. The authors have nothing to report. The authors declare no conflicts of interest. This publication is linked to a related reply by Önür et al. To view this article, visit https://doi.org/10.1111/jgs.70322. The authors have nothing to report.