A large longitudinal study found that higher body roundness index trajectories were associated with an increased risk of cardiovascular disease in middle-aged and older Chinese adults.

In the study, published in the Journal of the American Heart Association, investigators analyzed data from 9,935 participants in the China Health and Retirement Longitudinal Study over a 9-year period. They used group-based trajectory modeling to classify participants into three distinct body roundness index (BRI) trajectory groups based on measurements taken from 2011 to 2016: low-stable, moderate-stable, and high-stable. Cardiovascular disease (CVD) events were assessed from 2017 to 2020.

Among the participants (mean age = 58.85 ± 9.09 years, 53% male), 49.81% of them were in the low-stable group, 42.35% were in the moderate-stable group, and 7.84% were in the high-stable group. The study excluded 3,807 participants because of missing BRI data, 2,257 with preexisting CVD, and 932 were lost to follow-up.

Baseline characteristics varied across the trajectory groups. In the low-stable, moderate-stable, and high-stable groups, respectively:

• Rural residents: 28.01%, 37.49%, 40.44%
• Current smokers: 42.47%, 21.44%, 11.83%
• Current drinkers: 40.84%, 27.05%, 15.92%
• Hypertension prevalence: 15.78%, 30.74%, 48.84%
• Dyslipidemia prevalence: 4.95%, 15.13%, 20.56%
• Diabetes prevalence: 3.47%, 7.27%, 12.27%.

Medication use also differed across groups:

• Antihypertensive drugs: 10.65%, 22.72%, 41.13%
• Lipid-lowering drugs: 2.43%, 6.57%, 12.63%
• Antidiabetic drugs: 2.16%, 4.68%, 7.62%.

Median clinical features across the low-stable, moderate-stable, and high-stable groups were:

• BRI: 3.18, 4.84, 6.90
• Systolic blood pressure (mmHg): 122.0, 128.0, 135.75
• Blood glucose (mg/dL): 100.62, 103.32, 105.57
• Low-density lipoprotein cholesterol (mg/dL): 110.57, 117.91, 121.78
• High-sensitivity C-reactive protein (mg/L): 0.80, 1.11, 1.66.

Compared with the low-stable group, the high-stable group had a 55% higher risk of incident CVD (hazard ratio [HR] = 1.55, 95% confidence interval [CI] = 1.26–1.90) after adjusting for multiple covariates. The moderate-stable group also showed increased CVD risk (HR = 1.22, 95% CI = 1.09–1.37).

During follow-up, 3,052 CVD events occurred, including 965 strokes and 2,477 cardiac events. In fully adjusted models, compared with the low-stable group:

• High-stable group: CVD risk: HR = 1.55 (95% CI = 1.26–1.90), stroke risk: HR = 1.46 (95% CI = 1.10–1.95), cardiac event risk: HR = 1.35 (95% CI = 1.09–1.67)
• Moderate-stable group: CVD risk: HR = 1.22 (95% CI = 1.09–1.37), stroke risk: HR = 1.29 (95% CI = 1.08–1.54), cardiac event risk: HR = 1.14 (95% CI = 1.01–1.29).

Adding BRI trajectories to conventional risk factors improved CVD risk prediction:

• Net reclassification improvement (NRI): 16.35% (P < .0001)
• Integrated discrimination improvement (IDI): 0.32% (P < .0001).

For stroke and cardiac events separately:

• Stroke: NRI = 14.79% (95% CI = 9.13–20.45%, P < .0001), IDI = 0.16% (95% CI = 0.07–0.25%, P = .0004)
• Cardiac events: NRI = 15.64% (95% CI = 9.98–21.30%, P < .0001), IDI = 0.17% (95% CI = 0.09–0.26%, P < .0001).

The association between BRI trajectories and CVD risk remained consistent across subgroups, including sex, age, smoking status, drinking status, diabetes history, antihypertensive medication use, and baseline systolic blood pressure (all P for interaction > .05).

A sensitivity analysis excluding participants who died during follow-up showed similar results. The high-stable group maintained a higher CVD risk (HR = 1.66, 95% CI = 1.32–2.08) compared with the low-stable group. For stroke and cardiac events in the high-stable group, the HRs were 1.64 (95% CI = 1.19–2.25) and 1.45 (95% CI = 1.13–1.85), respectively.

BRI was calculated using the formula: BRI = 364.2 - 365.5 × √(1 - ((WC / (2π))^2) / (0.5 × height)^2), where WC is waist circumference in cm and height is in cm.

CVD events were self-reported based on physician diagnoses of heart attack, angina, coronary heart disease, heart failure, other heart problems, or stroke. Cox proportional hazards regression models were used to assess CVD risk, adjusting for demographics, medical history, medication use, and clinical characteristics such as systolic blood pressure, lipid levels, blood glucose, uric acid, and high-sensitivity C-reactive protein.

Limitations included potential information bias from self-reported CVD diagnoses, lack of cause-specific mortality data, and possible unmeasured confounding factors like physical activity. The findings may not be generalizable to non-Chinese populations.

The authors declared having no competing interests.