TL;DR
This study investigated which anthropometric and metabolic indicators best predict type 2 diabetes risk in middle-aged and elderly Chinese adults. The findings reveal that traditional measures like Body Mass Index (BMI) and Waist Circumference (WC) are more strongly associated with diabetes risk than newer indicators like Waist-Height Ratio (WHtR) and the Triglyceride-Glucose (TyG) index. Importantly, combining BMI and WC provided the strongest predictive power, highlighting the value of assessing both general and abdominal obesity.
What Problem Does This Paper Address?
Type 2 diabetes mellitus (T2DM) is a growing global health crisis, with China having the largest number of affected individuals. Early identification of high-risk individuals is crucial for prevention and management. While obesity is a known risk factor, there's ongoing debate about which specific measures of obesity are most predictive of T2DM. Traditional measures like BMI assess overall adiposity, while waist circumference (WC) and waist-to-height ratio (WHtR) focus on abdominal fat distribution, which is increasingly recognized as a key contributor to metabolic disorders. The triglyceride-glucose (TyG) index has also emerged as a potential indicator of insulin resistance and T2DM risk. This study aims to clarify which of these indicators—BMI, WC, WHtR, or TyG—offers the strongest association with incident T2DM in a specific demographic (middle-aged and elderly Chinese adults) and to explore their combined effects.
Methods
This was a retrospective cohort study utilizing data from annual health examinations of individuals aged 45 years and older in Xinzheng City, China, between 2011 and 2019. The researchers employed Cox proportional hazard regression models to assess the association between baseline measurements of BMI, WC, WHtR, and TyG index with the incidence of diabetes over the follow-up period. They analyzed these indicators in various ways: as categorical variables (e.g., BMI categories, WC quartiles) and as continuous variables. To compare the predictive strength of the indicators, they also calculated hazard ratios (HRs) per standard deviation (SD) increase and used Akaike's Information Criterion (AIC) to assess model fit. Furthermore, they conducted joint analyses of BMI and WC to evaluate their combined predictive power. Sensitivity analyses were performed to ensure the robustness of the findings.
Datasets
The study used a dataset derived from annual health examination records of residents in Xinzheng City, Henan Province, China, collected between 2011 and 2019. The final cohort comprised 41,242 participants aged 45 years and older, with a total of 205,770 person-years of follow-up. This dataset included anthropometric measurements (height, weight, waist circumference), demographic information, lifestyle factors (smoking, drinking, physical activity), medical history, and laboratory data (fasting plasma glucose, total cholesterol, triglycerides). No new benchmarks were introduced; the study utilized existing data for analysis.
Key Insights and Ideas
- BMI and WC Reign Supreme: Despite the emergence of newer indices like TyG, traditional measures of general obesity (BMI) and central obesity (WC) remain the most powerful predictors of T2DM in this population. This suggests that simpler, more established metrics are still highly relevant for risk assessment.
- The Power of Combination: The study strongly emphasizes that assessing both general adiposity (BMI) and abdominal fat distribution (WC) together significantly amplifies the prediction of diabetes risk. This highlights the synergistic effect of these two indicators.
- Gender and Age Nuances: While BMI and WC showed stronger associations in men, the study also identified specific optimal cut-off points for WC and WHtR in women that could minimize diabetes risk, suggesting that gender-specific considerations might be important. The risk associated with these indicators generally decreased with increasing age, though BMI remained a strong predictor even in the oldest age group.
- Non-Linearity in Risk: For WC, WHtR, and TyG, the relationship with diabetes risk was found to be approximately non-linear, meaning the risk doesn't increase at a constant rate. BMI, however, showed a more log-linear association. This implies that the magnitude of risk increase can vary depending on the baseline level of the indicator.
Key Results
- After multivariable adjustment, the hazard ratios (HRs) for diabetes were: BMI: 1.92 (1.71-2.16), WC: 1.99 (1.78-2.23), WHtR: 1.65 (1.47-1.86), and TyG: 1.66 (1.47-1.87) for the highest versus reference groups.
- In joint analyses, participants with high BMI combined with high WC had the highest risk of developing diabetes (HR: 2.26; 95% CI: 1.98, 2.58).
- For women, the lowest HR was observed when WC was approximately 72 cm and WHtR was approximately 0.47.
- Per standard deviation increase, BMI and WC showed stronger associations with diabetes risk (HRs of 1.30 for both) compared to WHtR (1.23) and TyG (1.25) in the total population.
Limitations and Future Work
The authors acknowledge several limitations:
- Diabetes Diagnosis: The diagnosis of diabetes relied on self-reporting, medication use, and fasting glucose, which may underestimate the true incidence compared to more comprehensive diagnostic methods like 2-hour OGTT or HbA1c.
- Sample Generalizability: The study population was drawn from annual health examinations, primarily of middle-aged and elderly individuals, which may limit the generalizability of findings to younger populations or those with different healthcare access.
- Unmeasured Confounders: Despite adjusting for numerous factors, unmeasured confounders like education and detailed dietary habits could still influence the results.
- Recall Bias: While efforts were made to standardize measurements, the retrospective nature and reliance on self-reported data can introduce bias.
Future work could involve analyzing dose-response relationships using restricted cubic splines more extensively and exploring these associations in more diverse populations. The study also suggests that simple, easily measured indicators like BMI and WC, especially when considered together, are valuable tools for diabetes risk assessment and prevention efforts.
Summary updated