Highlights
In this large multicentre prospective cohort from the China Metabolic Management Center, directly measured visceral fat area was independently associated with higher risk of incident diabetic kidney disease in adults with diabetes.
The association was not uniform. Men showed a U-shaped pattern, with excess risk at both very low and high visceral fat levels, whereas women showed risk increases mainly above higher visceral fat thresholds.
These sex-specific patterns persisted across BMI strata. Notably, men with low visceral fat area emerged as an overlooked high-risk group, particularly in the setting of obesity, where risk was elevated across all visceral fat categories.
The findings support a move beyond BMI alone toward more precise body composition-based renal risk stratification in diabetes care.
Background
Diabetic kidney disease remains one of the most consequential complications of diabetes, driving chronic kidney disease, cardiovascular events, kidney failure, and premature mortality. Although albuminuria and reduced estimated glomerular filtration rate are standard markers for diagnosis and staging, they do not fully explain heterogeneity in risk. Clinicians increasingly recognize that metabolic phenotyping, including adipose tissue distribution, may provide additional prognostic value.
Visceral adiposity is biologically distinct from overall obesity. Compared with subcutaneous fat, visceral fat is more metabolically active and more strongly linked to insulin resistance, systemic inflammation, lipotoxicity, oxidative stress, activation of the renin-angiotensin-aldosterone system, and endothelial dysfunction. These pathways are all relevant to glomerular hyperfiltration, albuminuria, and progressive nephron injury. Yet, despite this biologic plausibility, prospective evidence linking directly measured visceral fat area to incident diabetic kidney disease has been limited, especially in large cohorts.
Another unresolved issue is whether the relationship between visceral fat and renal risk differs by sex and BMI. Men and women differ in fat distribution, hormonal milieu, adipokine biology, and susceptibility to ectopic fat deposition. BMI, meanwhile, is an imprecise marker that cannot distinguish between lean mass, subcutaneous adiposity, and visceral adiposity. A patient with a “normal” or modestly elevated BMI may still carry excess visceral fat, whereas a patient with obesity by BMI may show different patterns of metabolic risk depending on fat compartmentalization.
Against this background, Shi and colleagues conducted a very large prospective cohort study to examine how visceral fat area relates to incident diabetic kidney disease in adults with diabetes, and whether these associations differ according to sex and BMI category.
Study Design
Study population and setting
This prospective cohort study used data from the China Metabolic Management Center (MMC), a nationwide network spanning 640 centres. Adults aged 18 years or older with physician-confirmed diabetes were eligible if they had baseline assessment for diabetic kidney disease and accessible follow-up records.
From 308,512 participants recruited between June 1, 2017, and June 26, 2024, the investigators excluded 113,383 with prevalent diabetic kidney disease at baseline, 41,979 with incomplete visceral fat area measurements, and 59,618 with missing diabetic kidney disease assessment during follow-up. The final analytic cohort comprised 93,532 participants.
Exposure and outcome definitions
The main exposure was visceral fat area, measured by bioelectrical impedance analysis. While imaging methods such as CT or MRI remain the gold standard for direct visceral fat quantification, bioelectrical impedance offers practical scalability for large, real-world clinical cohorts.
The primary outcome was incident diabetic kidney disease, defined as either albuminuria, with urinary albumin-to-creatinine ratio of at least 3.39 mg/mmol, or estimated glomerular filtration rate below 60 mL/min per 1.73 m2 during follow-up.
Follow-up and statistical analysis
Participants were followed through June 21, 2025. Median follow-up was 18.4 months, with an interquartile range of 9.4 to 32.3 months. The median age was 55.0 years, and 61.9% of the cohort were male.
The investigators used Cox proportional hazards regression to evaluate the association between visceral fat area and incident diabetic kidney disease. Analyses were performed for visceral fat area as a continuous exposure and in categories, with stratification by sex and by BMI category: without overweight, BMI below 24 kg/m2; overweight, BMI 24 to 28 kg/m2; and obesity, BMI at least 28 kg/m2.
Key Findings
Overall association between visceral fat and diabetic kidney disease
Over follow-up, 27,486 of 93,532 participants, or 29.4%, developed diabetic kidney disease. In the overall cohort, higher visceral fat area was significantly associated with increased diabetic kidney disease risk. For each interquartile range increase of 52 cm2 in visceral fat area, the adjusted hazard ratio was 1.08, with a 95% confidence interval of 1.05 to 1.10 and p below 0.0001.
This effect size is modest at the individual-unit level, but it is clinically meaningful in a high-risk diabetes population, especially given the prevalence of visceral adiposity and the burden of renal complications. The more notable message, however, emerged in the stratified analyses, where the shape and threshold of risk varied substantially by sex and BMI.
Sex-specific patterns
Among men, categorical analyses revealed elevated diabetic kidney disease risk at both low and high levels of visceral fat area, suggesting a U-shaped association. Compared with the reference category, men with visceral fat area 50 cm2 or less had an adjusted hazard ratio of 1.12, with 95% confidence interval 1.06 to 1.19 and p equal to 0.0002. Men with visceral fat area above 100 cm2 also showed increased risk; for the 100 to 125 cm2 category, the adjusted hazard ratio was 1.06, with 95% confidence interval 1.01 to 1.11 and p equal to 0.013.
In women, the pattern differed. Increased risk became statistically evident only at higher visceral fat levels. Women with visceral fat area between 75 and 100 cm2 had an adjusted hazard ratio of 1.09, with 95% confidence interval 1.03 to 1.14 and p equal to 0.0014. This implies a more threshold-like pattern in women rather than the U-shaped association seen in men.
Clinically, this is one of the most important observations in the paper. It argues against a one-size-fits-all interpretation of abdominal adiposity in diabetes and suggests that low visceral fat in men should not automatically be interpreted as metabolically reassuring.
BMI-stratified findings in men
In men without overweight, defined as BMI below 24 kg/m2, the U-shaped relationship was clear. Risk was increased both at the low and high ends of visceral fat distribution. Men with visceral fat area 50 cm2 or less had an adjusted hazard ratio of 1.12, with 95% confidence interval 1.04 to 1.21 and p equal to 0.0045. Those with visceral fat area above 125 cm2 had an adjusted hazard ratio of 1.25, with 95% confidence interval 1.09 to 1.44 and p equal to 0.0019.
In men with overweight, BMI 24 to 28 kg/m2, the U-shape persisted. Significant excess risk was seen at visceral fat area 50 cm2 or less, with adjusted hazard ratio 1.15 and 95% confidence interval 1.01 to 1.31, p equal to 0.039; at 100 to 125 cm2, with adjusted hazard ratio 1.10 and 95% confidence interval 1.02 to 1.18, p equal to 0.010; and above 125 cm2, with adjusted hazard ratio 1.15 and 95% confidence interval 1.06 to 1.24, p equal to 0.0005.
In men with obesity, BMI at least 28 kg/m2, risk was elevated across all visceral fat categories and highest at the extremes. For visceral fat area 50 cm2 or less, the adjusted hazard ratio was 2.15, with 95% confidence interval 1.44 to 3.23 and p equal to 0.0002. For visceral fat area above 125 cm2, the adjusted hazard ratio was 1.31, with 95% confidence interval 1.20 to 1.42 and p below 0.0001.
This is a striking result. In obese men, even relatively low measured visceral fat area did not confer protection; rather, it identified a subgroup with particularly high renal risk. That raises important questions about body composition heterogeneity, sarcopenic obesity, underlying illness, fat redistribution, or measurement limitations in certain phenotypes.
BMI-stratified findings in women
Among women, the association was more monotonic and threshold-driven. In women without overweight, diabetic kidney disease risk increased beginning at visceral fat area above 75 cm2. For the 75 to 100 cm2 category, the adjusted hazard ratio was 1.10, with 95% confidence interval 1.02 to 1.19 and p equal to 0.011.
In women with overweight, the same threshold appeared. Visceral fat area 75 to 100 cm2 was associated with an adjusted hazard ratio of 1.10, with 95% confidence interval 1.02 to 1.17 and p equal to 0.011.
In women with obesity, lower visceral fat levels also began to carry significant risk. Visceral fat area 50 to 75 cm2 was associated with an adjusted hazard ratio of 1.22, with 95% confidence interval 1.02 to 1.46 and p equal to 0.033. This suggests that once overall adiposity reaches the obesity range, the renal risk threshold linked to visceral fat may shift downward.
Clinical Interpretation
The study adds several clinically relevant insights. First, visceral fat area appears to provide risk information beyond BMI alone. This is increasingly important in diabetes clinics, where BMI remains widely used but often fails to capture metabolically harmful fat distribution. A patient with “acceptable” BMI may still have high visceral adiposity and substantial renal risk.
Second, sex matters. Men and women should not necessarily be assessed using identical visceral fat thresholds when estimating diabetic kidney disease risk. The male U-shaped association is especially noteworthy because it challenges the intuitive assumption that less visceral fat is always better. For women, the data instead support a rising-risk pattern above a threshold, with that threshold modified by BMI.
Third, the identification of low-visceral-fat men as a high-risk group may have practical implications. In routine care, a leaner man with diabetes might be assumed to have lower cardiometabolic risk than a heavier counterpart. These findings suggest caution. Such patients may warrant closer renal surveillance, particularly if they also have obesity by BMI, reduced muscle mass, weight loss, longstanding diabetes, or other markers of frailty or adverse metabolic remodeling.
Finally, the work supports more individualized body composition assessment in diabetes management. Where feasible, adding visceral fat estimation to standard metabolic evaluation may help identify patients who merit more intensive kidney-protective strategies, including blood pressure optimization, glucose management, renin-angiotensin system blockade when indicated, and broader use of organ-protective therapies such as SGLT2 inhibitors and nonsteroidal mineralocorticoid receptor antagonists in appropriate patients.
Biological Plausibility
The adverse association between high visceral fat and diabetic kidney disease is mechanistically credible. Visceral adipose tissue contributes to insulin resistance, increased free fatty acid flux, inflammatory cytokine production, oxidative stress, and neurohormonal activation. These factors can promote glomerular hypertension, endothelial injury, podocyte dysfunction, and tubulointerstitial damage.
The explanation for elevated risk at very low visceral fat levels in men is less straightforward. Several possibilities deserve consideration. Low visceral fat may mark catabolic states, reduced muscle mass, frailty, chronic inflammation, advanced diabetes, or unmeasured comorbidity. It may also reflect metabolically unfavorable body composition despite low measured visceral area, such as sarcopenia with ectopic lipid deposition in liver or muscle. Another possibility is that low measured visceral fat in some obese men reflects limitations of bioelectrical impedance in atypical body habitus. The study cannot resolve these mechanisms, but the observation is robust enough to warrant further investigation.
Strengths and Limitations
Strengths
The study’s major strength is scale. With 93,532 participants from 640 centres, it is among the largest prospective evaluations of directly measured visceral fat and incident diabetic kidney disease in patients with diabetes. The multicentre real-world design improves clinical relevance, and the sex- and BMI-stratified analyses address a major gap in prior literature.
The exclusion of participants with prevalent diabetic kidney disease at baseline strengthens temporal inference, and the use of a composite kidney outcome based on albuminuria or reduced estimated glomerular filtration rate reflects clinically recognizable disease.
Limitations
Several limitations are important. Visceral fat area was measured by bioelectrical impedance analysis rather than CT or MRI. Although practical, this method is less precise and may be vulnerable to hydration status, device assumptions, and body composition extremes. Residual confounding is inevitable in an observational study, even with multivariable adjustment.
Follow-up was relatively short, with a median of 18.4 months. This is sufficient to capture incident albuminuria or eGFR decline in a high-risk population, but it limits conclusions about long-term kidney function trajectories, hard renal endpoints, or causality. The diabetic kidney disease definition combined albuminuria and reduced eGFR, which is clinically appropriate but may mask potentially different relationships between visceral fat and individual kidney phenotypes.
Selection bias is also possible because many participants were excluded for missing visceral fat data or missing follow-up kidney assessment. Finally, generalizability outside the Chinese diabetes population requires confirmation, especially in populations with different adiposity phenotypes, treatment patterns, and background kidney risk.
Implications for Practice and Research
For clinicians, the immediate message is that body composition deserves more attention in diabetes care. BMI should not be the sole anthropometric lens through which renal risk is viewed. If bioelectrical impedance-based visceral fat measurement is available, it may help identify patients needing closer kidney monitoring.
For researchers, the findings open several directions. First, replication in other ethnic and geographic populations is essential. Second, mechanistic studies should explore why low visceral fat in men predicts higher diabetic kidney disease risk. Third, future analyses should determine whether visceral fat improves discrimination or reclassification beyond existing kidney risk models. Fourth, interventional studies are needed to test whether reducing visceral fat specifically translates into lower renal event rates, and whether this effect differs by sex.
It will also be valuable to compare visceral fat with related measures such as waist circumference, waist-to-height ratio, hepatic steatosis, and markers of sarcopenia. The most informative renal risk model may ultimately integrate several body composition domains rather than any single metric.
Conclusion
This large prospective cohort study shows that visceral fat area is a significant predictor of incident diabetic kidney disease in adults with diabetes, but the association is strongly modified by sex and BMI. Men exhibited a U-shaped relationship, with excess risk at both low and high visceral fat levels, whereas women showed increased risk primarily above higher visceral fat thresholds. The identification of low-visceral-fat men as a high-risk group is particularly important and may alter how clinicians interpret body composition in diabetes.
Overall, the study reinforces a broader lesson in metabolic medicine: where fat is stored may matter as much as how much body weight a patient carries. More precise phenotyping of adiposity could sharpen diabetic kidney disease prevention strategies and move routine care closer to true individualized risk assessment.
Funding and ClinicalTrials.gov
Funding: Noncommunicable Chronic Diseases-National Science and Technology Major Project.
ClinicalTrials.gov: NCT03811470.
Citation
Shi J, Zhao D, Wang J, Wang Y, Dong Q, Hu J, Sha Y, Peng Y, Dai Y, Zheng Q, Xu F, Lan H, Ji B, Li L, Qiu H, Jiang R, Tang R, Qu S, Yang L, Ke T, Li S, Zhao Z, Zhang Y, Wang W, MMC and ChinaMAP Collaborative Group. Sex-specific and BMI-specific associations between visceral fat and diabetic kidney disease in patients with diabetes: a large-scale multicentre prospective cohort study. The Lancet Diabetes & Endocrinology. 2026 May;14(5):401-411. PMID: 41991218. URL: https://pubmed.ncbi.nlm.nih.gov/41991218/
Selected Contextual References
Kidney Disease: Improving Global Outcomes (KDIGO) Diabetes Work Group. KDIGO 2022 Clinical Practice Guideline for Diabetes Management in Chronic Kidney Disease. Ann Intern Med. 2023;178(3):381-391.
American Diabetes Association Professional Practice Committee. Chronic Kidney Disease and Risk Management: Standards of Care in Diabetes—2024. Diabetes Care. 2024;47(Suppl 1).
These guideline documents support the importance of early diabetic kidney disease risk stratification and aggressive multifactorial risk reduction, although they do not yet incorporate visceral fat area as a routine risk marker.
