Linear Association of Derivatives of Triglyceride-Glucose Index with Incident Lower Limb Joint Pain in Middle-Aged and Older Chinese Adults: A Prospective Cohort Study [Letter] Authors: JiaLe Li, ZiHao Yi, JiaWei Wu Published: 2026 Dear editor We read with great interest the article by Xue et al, which investigated the association between triglyceride-glucose (TyG)-derived indices and incident lower limb joint pain among middle-aged and older Chinese adults1 The authors should be commended for using longitudinal data from the China Health and Retirement Longitudinal Study (CHARLS) to explore the metabolic-adiposity dimension of musculoskeletal pain. Nevertheless, several methodological considerations may warrant further discussion, particularly when interpreting TyG-derived indices as potential tools for early risk identification. First, the incremental value of TyG-derived indices beyond conventional anthropometric measures remains insufficiently established. In the fully adjusted analyses, TyG-BMI and TyG-WHtR were associated with incident lower limb joint pain, whereas TyG alone was not. This pattern suggests that the observed associations may be driven predominantly by adiposity-related components rather than by the TyG index itself. Because TyG-BMI, TyG-WC, and TyG-WHtR are mathematically constructed by combining TyG with body mass index, waist circumference, or waist-to-height ratio, their independent contribution cannot be adequately determined without direct comparison with these anthropometric components. If these indices are proposed for risk stratification or early detection, additional analyses comparing their performance with BMI, waist circumference, and waist-to-height ratio alone would be informative. Evaluating discrimination, calibration, and clinical utility would further clarify whether TyG-derived indices provide meaningful incremental information beyond readily available adiposity measures.2,3 Second, attrition-related selection bias should be considered. According to the study flowchart, 1711 participants were excluded because of loss to follow-up across the five survey waves. In an aging cohort, attrition is unlikely to occur completely at random and may be associated with obesity, metabolic dysfunction, mobility limitation, pain status, hospitalization, frailty, or mortality. Restricting the analysis to participants with available follow-up data may therefore yield a selected cohort with relatively stable health status, potentially biasing the estimated association between TyG-derived indices and lower limb joint pain. A comparison of baseline characteristics between retained and lost participants, or the application of inverse probability-of-censoring weighting, would help assess the possible direction and magnitude of this bias.4 Third, the time-to-event framework may not fully correspond to the outcome ascertainment scheme. Lower limb joint pain was assessed at discrete survey waves by asking participants which body parts were currently painful. Accordingly, the exact onset time of pain was not observed, and transient pain episodes occurring and resolving between two survey waves may have been missed. Treating the first survey wave with reported pain as the event time in Cox regression and Kaplan-Meier analyses may therefore introduce interval-censoring-related uncertainty. This issue is particularly relevant for musculoskeletal pain, which commonly has a fluctuating course rather than a single irreversible onset. Discrete-time survival models, which evaluate event occurrence within prespecified follow-up intervals, or sensitivity analyses requiring repeated pain reports, may better align with the structure of CHARLS pain data.5 In conclusion, Xue et al provide valuable longitudinal evidence regarding the association between adiposity-integrated TyG derivatives and lower limb joint pain However, the uncertain incremental value beyond conventional anthropometric indicators, potential attrition-related selection bias, and interval-censored nature of pain onset should be considered when interpreting the clinical implications of the findings. Addressing these issues in future analyses would strengthen the evidence base for applying TyG-derived indices in musculoskeletal pain risk assessment.