Association of the triglyceride-to-HDL cholesterol ratio with metabolic markers among underweight women undergoing occupational health checkups: a five-year repeated cross-sectional study

Underweight women are often considered metabolically low risk, but a low body mass index does not guarantee a favorable metabolic profile. We examined the association between the triglyceride-to-high-density lipoprotein cholesterol ratio (TG/HDL-C ratio) and routine metabolic markers in underweight women undergoing occupational health examinations. This five-year repeated cross-sectional study included women aged 20–49 years with body mass index < 18.5 kg/m 2 and concurrent measured triglyceride and HDL-C values > 0. The conventional triglyceride field was used; separately recorded postprandial values were excluded, although some nonfasting measurements may have been present in the conventional field. Fasting glucose was taken only from the fasting glucose field. Individual identifiers were unavailable, so repeat participation across fiscal years could not be identified. Linear regression models estimated crude associations and associations adjusted for age, BMI, and fiscal year; estimates were expressed per doubling of the TG/HDL-C ratio. Additional analyses excluded relevant medication users, adjusted for lifestyle factors, examined individual fiscal years and fiscal-year interactions, and assessed nonlinearity with restricted cubic splines. Among 650 records, mean age was 34.8 ± 7.8 years, mean BMI was 17.4 ± 0.9 kg/m 2 , and median TG/HDL-C ratio was 0.67 (interquartile range, 0.49–0.95). A doubling of the TG/HDL-C ratio was associated with 5.53 mg/dL higher LDL-C (95% confidence interval [CI], 2.49–8.57; p < 0.001) and 1.39 mg/dL higher fasting glucose (95% CI, 0.49–2.29; p = 0.003). Both associations survived multiple-comparison correction and changed little after excluding relevant medication users or adding smoking, alcohol consumption frequency, and regular exercise to the models. No interaction with fiscal year was detected. However, after restriction to the 1st–99th percentile range, the overall spline association remained significant for LDL-C but not for fasting glucose ( p = 0.114). Among underweight working women, higher TG/HDL-C ratios were associated with modestly higher LDL-C and a less robust increase in fasting glucose. The absolute differences were small, the cross-sectional design precludes causal inference, and measurement limitations restrict clinical interpretation. The findings indicate metabolic variation within the underweight range but do not establish clinical utility.