Sex-specific and BMI-specific associations between visceral fat and diabetic kidney disease in patients with diabetes: a large-scale multicentre prospective cohort study.

Abstract

There are few large-scale prospective studies on directly measured visceral fat area (VFA) and diabetic kidney disease, with inconsistent findings across sex and BMI. We aimed to investigate the association between VFA and incident diabetic kidney disease risk and to examine how this association differs by sex and BMI. This prospective cohort study included adults with diabetes from 640 centres in the China Metabolic Management Center (MMC) project. We included adult participants with diabetes who had baseline and follow-up diabetic kidney disease data, excluding those with prevalent diabetic kidney disease at baseline or missing baseline VFA data. Diabetes diagnosis was confirmed by an on-site physician at enrollment. Diabetic kidney disease was defined as albuminuria (urinary albumin-to-creatinine ratio ≥3·39 mg/mmol) or estimated glomerular filtration rate less than 60 mL/min per 1·73m². VFA was measured by bioelectrical impedance analysis. Cox regression analysis was used to evaluate the association between VFA and diabetic kidney disease, with stratification by sex and BMI categories. The MMC project is registered at ClinicalTrials.gov, NCT03811470. Between June 1, 2017, and June 26, 2024, we recruited 308 512 participants aged 18 years or older with a diagnosis of diabetes who had undergone baseline diabetic kidney disease assessment and had accessible follow-up records up to June 21, 2025. After excluding participants with baseline diabetic kidney disease (n=113 383), incomplete VFA measurements (n=41 979), or missing diabetic kidney disease assessment during follow-up (n=59 618), the final analytical cohort consisted of 93 532 participants. Participants had a median follow-up of 18·4 months (IQR 9·4-32·3) and a median age of 55·0 years (IQR 46·0-62·0). 57 865 [61·9%] of 93 532 participants were male and 35 667 (38·1%) were female. 27 486 (29·4%) of 93 532 participants developed diabetic kidney disease. VFA was significantly associated with increased diabetic kidney disease risk (per IQR [52 cm²] adjusted hazard ratio [aHR] 1·08, 95% CI 1·05-1·10; p<0·0001). Sex-stratified categorical analyses showed increased diabetic kidney disease risk in men at both low VFA (≤50 cm²; aHR 1·12, 1·06-1·19; p=0·0002) and high VFA greater than 100 cm² (100-125 cm²; 1·06, 1·01-1·11; p=0·013), while women showed significantly increased risk only at VFA greater than 75 cm² (75-100 cm²; 1·09, 1·03-1·14; p=0·0014). In male participants, a U-shaped relationship was consistently observed. Male participants without overweight (BMI <24 kg/m²) had elevated risks at VFA extremes (≤50 cm² aHR 1·12, 95% CI 1·04-1·21; p=0·0045 and >125 cm² aHR 1·25, 1·09-1·44; p=0·0019). This U-shape persisted in male participants with overweight (BMI 24-28 kg/m²), with significant risks at VFA 50 cm² or less (aHR 1·15, 95% CI 1·01-1·31; p=0·039), 100-125 cm² (1·10, 1·02-1·18; p=0·010), and greater than 125 cm² (1·15, 1·06-1·24; p=0·0005). In male participants with obesity (BMI ≥28 kg/m²), risk was significantly elevated across all VFA categories, highest at both extremes (≤50 cm² aHR 2·15, 95% CI 1·44-3·23; p=0·0002 and >125 cm² aHR 1·31, 1·20-1·42; p<0·0001). In female participants, disease risk was elevated from VFA greater than 75 cm² in both participants without overweight (75-100 cm² aHR 1·10, 95% CI 1·02-1·19; p=0·011) and participants with overweight (75-100 cm² 1·10, 1·02-1·17; p=0·011), while obesity extended significant risk to lower VFA levels (50-75 cm² aHR 1·22, 95% CI 1·02-1·46; p=0·033). VFA is a significant risk factor for diabetic kidney disease, exhibiting distinct sex-specific patterns that persist across BMI strata. These findings support precise clinical risk stratification and identify low-VFA men as a previously overlooked high-risk population. Noncommunicable Chronic Diseases-National Science and Technology Major Project.