Association of obesity with thromboembolic complications after hip fracture surgery: a tertiary center retrospective cohort study — Correspondence

We read with great interest the article by Gurunathan et al reporting the HipCLOTS study, which examined the association between body mass index (BMI) and thromboembolic complications after hip fracture surgery in 1853 older adults and identified an apparent obesity paradox in 1-year mortality[1]. In the era of transparent reporting of observational and AI-enabled surgical research, as emphasized by the TITAN guideline on artificial intelligence and the revised STROCSS guideline for surgical cohort and registry studies[2,3], the HipCLOTS study is a timely and clinically relevant contribution. From our perspective as clinicians and methodologists working at the intersection of osteoporosis, nutrition, and perioperative risk, we believe the HipCLOTS data highlight three opportunities for future research and clinical implementation. First, BMI alone is an insufficient construct for “obesity” in frail hip-fracture patients. Gurunathan et al rightly discuss that malnutrition, sarcopenia, and frailty cluster in older hip-fracture patients and may modify the association between adiposity and outcomes[1]. However, the primary exposure remains a four-level BMI category without information on fat distribution, muscle mass, or functional status. Contemporary ESPEN guidelines and consensus definitions on sarcopenic obesity emphasize that phenotyping older adults should integrate screening tools, such as SARC-F, objective measures of muscle strength and performance, and body-composition techniques (DXA or CT), rather than BMI alone[4,5]. In HipCLOTS, underweight patients had the highest prevalence of malnutrition and cognitive impairment[1], yet these dimensions were not incorporated into the multivariable models. Future cohorts could pre-specify composite phenotypes – such as “malnourished obesity” or “sarcopenic obesity” – building on prior work showing that overweight or obese malnutrition (“DOOM”) markedly worsens hip-fracture outcomes[6]. This would allow disentangling a true survival advantage of metabolically healthy obesity from artifacts driven by misclassification and reverse causation. Second, individualized risk prediction will require more than traditional regression. The authors state that there is “potential opportunity to explore individualized risk prediction” using their data[1]. We fully agree. However, the current analyses – as constrained by events per variable – could only adjust for a limited set of covariates. If hip-fracture thromboprophylaxis is to be tailored according to thromboembolic risk, future studies might consider developing and validating multivariable risk-prediction models that incorporate nutritional, functional, and body-composition markers alongside BMI. In line with the updated STROCSS 2025 guideline and the newly published TITAN guideline[2,3], such models should report discrimination, calibration-in-the-large, calibration slope, and decision-curve analysis, and should undergo temporal or external validation. This would allow clinicians to use explicit risk thresholds rather than broad BMI categories when deciding on the intensity and duration of prophylaxis. Third, recommendations to intensify and extend thromboprophylaxis should be balanced against bleeding risk and evidence gaps. HipCLOTS shows that overweight and obese patients experience higher 90-day thromboembolic rates but lower 1-year mortality[1]. The authors accordingly suggest that more aggressive and prolonged thromboprophylaxis may be warranted in these groups. Yet detailed data on in-hospital regimens, post-discharge agents, and adherence are limited, and major bleeding outcomes are only briefly summarized. Existing reviews and guideline summaries already highlight that evidence for optimal prophylaxis after hip fracture, especially in patients with obesity, remains sparse and heterogeneous[7,8]. In this context, we would encourage the design of prospective, possibly registry-based, studies that simultaneously capture thrombosis, bleeding, and functional recovery, with pre-specified analyses stratified by sarcopenic-obesity phenotypes and nutritional risk. In conclusion, HipCLOTS meaningfully advances our understanding of the complex relationship between adiposity and thromboembolic risk after hip fracture. Building on their work, we suggest that future research: (1) moves beyond BMI to incorporate sarcopenia, malnutrition and body composition into obesity phenotyping; (2) develops and transparently reports validated risk-prediction models in line with STROCSS and TITAN; and (3) evaluates thromboprophylaxis strategies using risk–benefit frameworks that account for both thrombotic and bleeding outcomes. Such steps may help translate the observed obesity paradox into actionable, phenotype-specific care pathways for older adults with hip fracture.