Beyond BMI: Chinese Visceral Adiposity Index (CVAI) improves machine learning prediction of 9-year cardiometabolic multimorbidity in the CHARLS cohort
Using data from the CHARLS cohort, this study demonstrates that machine learning models incorporating the Chinese Visceral Adiposity Index (CVAI) and depressive symptoms effectively predict 9-year cardiometabolic multimorbidity risk, outperforming BMI-based approaches and highlighting the critical roles of visceral adiposity and mental health in composite disease prediction.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the aging population of China, two silent conditions often travel together: diabetes, a disorder where the body struggles to manage sugar, and cardiovascular disease, which encompasses heart problems and strokes. When these two conditions appear in the same person, they form a cluster of illness known as cardiometabolic multimorbidity. This combination is particularly dangerous, driving disability and death among middle-aged and older adults. For decades, doctors have relied on simple measurements like body mass index, a calculation based on height and weight, to guess who might develop these diseases. However, this single number often misses the deeper truth about where fat is stored in the body and how it interacts with other factors like blood pressure, mood, and metabolism. As the population grows older, the need to predict who is at risk has become urgent, moving beyond simple guesses toward more precise tools that can guide early intervention.
A team of researchers from Xijing Hospital in Xi'an set out to build a better way to forecast these risks over a nine-year period. They turned to a massive, nationally representative survey of Chinese adults known as the China Health and Retirement Longitudinal Study. This database provided a detailed look at nearly 10,000 people, tracking their health from 2011 to 2020. The researchers did not just look at who got sick; they used advanced computer algorithms, a field known as machine learning, to analyze thousands of data points for each person. These data points included everything from blood sugar levels and cholesterol to waist circumference, grip strength, and even scores measuring symptoms of depression. The goal was to teach the computer to recognize patterns that human doctors might miss, specifically to predict who would develop diabetes, heart disease, or both over the next nine years.
The results revealed a clear hierarchy in how easy or difficult it is to predict these conditions. The computer models were most successful at predicting diabetes, achieving an AUC of 0.71. This high level of performance suggests that the signs of diabetes are often written clearly in routine blood tests and body measurements long before the disease is diagnosed. In contrast, predicting heart disease proved much harder, with the models only achieving an AUC of 0.63. The researchers found that heart disease is a more complex puzzle, influenced by sudden events and factors that are not always captured in standard check-ups. The prediction for the combined condition, where a person develops both diabetes and heart disease, fell somewhere in the middle, with an AUC of roughly 0.63.
What made these predictions work, and what did they reveal about the human body? For diabetes, the computer correctly identified that high blood sugar and obesity were the primary drivers. However, for the combined condition of cardiometabolic multimorbidity, a different set of factors rose to the top. The most powerful predictor was not a simple weight measurement, but a specialized index called the Chinese Visceral Adiposity Index. This index calculates how much fat is stored deep inside the abdomen, around the organs, by combining age, waist size, and blood lipid levels. The study showed that this deep belly fat is a far better warning sign for the combined disease than general body weight.
Even more striking was the role of mental health. The computer analysis placed symptoms of depression as the second most important predictor for the combined condition, right behind the visceral fat index. The data suggested a powerful link: when a person carries excess deep belly fat and also struggles with depressive symptoms, their risk of developing both diabetes and heart disease increases significantly. The two factors seemed to amplify each other, creating a risk profile that was greater than the sum of its parts. This finding points to a "psycho-metabolic" connection, where the stress of mental distress and the physical burden of visceral fat work together to damage the body's metabolism.
To make these findings useful for everyday doctors, the researchers tried to simplify the complex computer models into easy-to-use scorecards. For diabetes and the combined condition, this worked well. They created simple checklists based on the top risk factors, such as blood sugar, waist size, and depression scores, which retained almost all the predictive power of the full computer models. A doctor could use these scorecards to quickly screen patients in a clinic. However, the attempt to create a similar simple scorecard for heart disease failed. The simplified tool performed poorly, unable to distinguish between those who would get sick and those who would not. This failure highlights that heart disease is too complex to be reduced to a simple linear checklist; it likely requires more sophisticated tools or new types of medical data to predict accurately.
The study concludes that while we can now predict diabetes and the combined risk of diabetes and heart disease with reasonable accuracy using routine data, predicting heart disease alone remains a significant challenge. The research underscores that looking at the whole person is essential. It is not enough to measure weight or blood pressure in isolation. To truly understand the risk for middle-aged and older adults, we must look at where fat is stored in the body and acknowledge the profound impact of mental health on physical well-being. By integrating these insights, healthcare providers can move toward a more holistic approach to preventing the dual burden of metabolic and cardiovascular illness.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.