Comment on “Cardiometabolic Multimorbidity Increases the Risk of Hip Fracture: A Longitudinal Cohort Study Based on CHARLS” Authors: Harun Sağlıcak Published: 2026 ABSTRACT In the CHARLS cohort of Tian et al., the abstract's 13% versus 8% are column percentages from Table 1: the prevalence of cardiometabolic multimorbidity (CMM) among participants with and without hip fracture. Read across rows, the same counts give a nine‐year hip fracture incidence of 6.6% (36/544) with CMM and 4.4% (251/5770) without, an absolute difference of 2.3 percentage points (95% CI 0.4–4.7) rather than 5.0. The adjusted hazard ratio of 1.70 is unaffected; its lower confidence limit reads 1.318 in the abstract and 1.18 in Table 2. STROBE item 16(c) asks for such absolute risks. I read with interest the article by Tian and colleagues entitled “Cardiometabolic Multimorbidity Increases the Risk of Hip Fracture: A Longitudinal Cohort Study Based on CHARLS.” [1] Nine years of follow‐up in a nationally representative cohort, four stepwise adjustment models, and three sensitivity analyses including propensity score matching give the association a firmer footing than most previous work. My concern is narrow. The only absolute comparison in the abstract, 13% versus 8%, is not an incidence. First, Table 1 is stratified by outcome, not by exposure. Its CMM row reads 36 (13) in the fracture column and 508 (8) in the control column, so both are prevalence figures for cardiometabolic multimorbidity within outcome groups. Read as incidences they are impossible: 13% of the 544 participants with CMM plus 8% of the 5770 without would require 532 hip fractures, and the study reports 287. The correct denominators appear three sentences earlier in the same abstract. Incidence was 36/544, or 6.6% (95% CI, 4.8 to 9.0), against 251/5770, or 4.4% (95% CI, 3.9 to 4.9), an absolute difference of 2.3 percentage points (95% CI, 0.4 to 4.7) rather than 5.0 [2, 3]. This is arithmetic, not modeling. For a 60‐year‐old with hypertension and diabetes, the authors' own counts give 2.3 not 5.0 extra hip fractures per 100 over 9 years. Against a rising national hip fracture incidence, that is the number quoted at the bedside [4]. Second, the abstract gives the adjusted hazard ratio as 1.70 (95% CI, 1.318–2.47) where Table 2 gives 1.18–2.47. Only the latter is symmetric about 1.70 on the log scale. I would welcome confirmation of which is intended. Neither point touches the regression models: the association between cardiometabolic multimorbidity and hip fracture stands as reported. I would suggest two things. The first costs the authors nothing: a corrected sentence giving incidence by CMM status, 6.6% against 4.4%, with the absolute difference and its confidence interval. Every figure needed is already in Table 1, and the correction would leave the conclusion intact while making the clinical claim usable. The second is for cohort studies of fracture risk more generally. STROBE, which this study cites and follows, asks under item 16(c) that estimates of relative risk be translated into absolute risk for a meaningful time period [5, 6]. That step is still skipped. In one audit of medication trials in high‐impact general journals, 44% of positive trials reported their abstract results exclusively in relative terms [7]. None of this diminishes the study. A 9‐year cohort linking cardiometabolic multimorbidity to hip fracture in a Chinese population is worth having, and the finding will be cited. What it needs is an absolute risk a clinician can carry into a consultation. Whether the authors confirm the corrected values or set out a reading I have missed, the record will be clear either way, and I would welcome their response. Author Contributions Harun Sağlıcak: writing – original draft, conceptualization, investigation, software, methodology, writing – review and editing. Funding The author has nothing to report. Ethics Statement The author has nothing to report. Conflicts of Interest The author declares no conflicts of interest.