A routine blood component that most people associate with digestion may hold an unexpected clue to one of the most common and dangerous complications of diabetes. New research from a team at Tongji University School of Medicine in Shanghai suggests that total bile acids, a family of cholesterol-derived molecules produced by the liver, behave in a strikingly non-linear way as diabetic kidney disease develops. Rather than drifting steadily upward or downward with worsening kidney function, serum bile acid levels rise sharply at the earliest clinical stage of kidney damage and then fall back as the disease advances. The finding, published in BMC Endocrine Disorders, points to a window of metabolic upheaval that could eventually help clinicians spot kidney injury before it becomes irreversible.
Diabetic kidney disease is the leading cause of end-stage renal disease worldwide, and it often progresses silently. By the time patients notice symptoms, significant nephron loss may already have occurred. Current diagnostic tools rely on the urinary albumin-to-creatinine ratio, which tracks protein leakage into the urine, and the estimated glomerular filtration rate, which measures how well the kidneys filter blood. Both markers are useful but imperfect, particularly in the earliest phase of injury when albumin excretion is just beginning to climb. The search for complementary biomarkers has therefore become a priority in nephrology and endocrinology alike, and bile acids have emerged as intriguing candidates because they are far more than digestive detergents.
Bile acids act as signaling molecules that engage receptors such as the farnesoid X receptor and TGR5, influencing glucose metabolism, lipid handling, inflammation, and even renal hemodynamics. Previous studies had hinted at altered bile acid profiles in diabetes and chronic kidney disease, but the dynamic trajectory of total bile acids across the successive stages of diabetic kidney disease remained unclear. To address this gap, Lei Xu, Bo Feng, and Xinfeng Yan conducted a retrospective cross-sectional study of 485 individuals whose clinical data were collected between August 2025 and February 2026 at Shanghai East Hospital, with ethics approval granted in April 2026 and a waiver of informed consent for the use of de-identified routine care records.
The study population was stratified into five groups: 87 people with normal glycemia, 126 with uncomplicated type 2 diabetes mellitus, and three groups of patients with diabetic kidney disease at stages 1, 2, and 3, comprising 118, 80, and 74 individuals respectively. Classification followed standard criteria based on the urinary albumin-to-creatinine ratio and estimated glomerular filtration rate, with diabetes diagnoses anchored to the 1999 World Health Organization standards. The researchers compared serum total bile acid concentrations across groups using non-parametric tests and then probed the shape of the relationship with polynomial regression, a statistical technique capable of detecting curved, rather than straight-line, patterns in data.
The results revealed a pronounced inverted U-shaped pattern that the authors confirmed statistically with a significant quadratic term, yielding a P value for non-linearity of 0.013. Total bile acid levels were modestly elevated in uncomplicated diabetes, with a median of 2.25 micromoles per liter, but jumped abruptly at stage 1 diabetic kidney disease, where the median reached 3.65 micromoles per liter, a difference highly significant against the uncomplicated diabetes group. By stage 3, the median had fallen to 1.95 micromoles per liter, dipping even below the level seen in diabetes without kidney complications. In other words, the bile acid surge is not a simple byproduct of declining renal function but appears concentrated in the early, potentially still reversible phase of injury.
To test whether this association stood on its own, the team built a multivariate logistic regression model adjusting for a long list of potential confounders, including age, diabetes duration, blood pressure, glycated hemoglobin, body mass index, lipid profiles, estimated glomerular filtration rate, fatty liver status, and the use of antidiabetic medications or statins. Even after these adjustments, total bile acid remained independently associated with early-stage diabetic kidney disease, with a fully adjusted odds ratio of 1.394 per unit increase and a 95 percent confidence interval of 1.193 to 1.629. The statistical significance was robust, with a P value below 0.001, and variance inflation factors indicated that multicollinearity among the covariates did not undermine the model.
The researchers also explored whether bile acid levels could actually discriminate between patients with early kidney disease and those with diabetes alone. Receiver operating characteristic analysis produced an area under the curve of 0.699, a value that reflects modest but real discriminative ability, and an exploratory threshold of 2.80 micromoles per liter derived from the Youden index offered a tentative cutoff for distinguishing stage 1 disease from uncomplicated diabetes. The authors are careful to stress that this threshold is exploratory: given the modest area under the curve and wide confidence intervals, it would need validation in an independent external cohort before any clinical use, and it should not be interpreted as a ready-made diagnostic rule.
Internal validation added a further layer of rigor. Using bootstrap resampling with 1,000 iterations, the team estimated a mean apparent area under the curve of 0.818, with a 95 percent confidence interval of 0.761 to 0.868. After correcting for optimism, the bias-corrected value settled at 0.731, indicating good internal reproducibility of the model while acknowledging that the correction narrows the apparent performance. This kind of bootstrap procedure is increasingly regarded as best practice in clinical prediction research because it tempers the overfitting that can inflate performance estimates when models are evaluated on the same data used to build them. Even so, the authors emphasize that internal validation is no substitute for external confirmation in a genuinely independent population.
What might explain the early bile acid spike? The authors frame their findings as hypothesis-generating, but the biology offers several plausible threads. Bile acids are synthesized in the liver from cholesterol and are normally reabsorbed in the ileum and recycled through enterohepatic circulation. Early kidney injury could perturb this recycling, alter hepatic synthesis through inflammatory or metabolic signaling, or change the expression of bile acid transporters in renal tissue. Signaling through the farnesoid X receptor and TGR5 is known to affect insulin sensitivity and renal microvascular function, so a transient rise in circulating bile acids could be both a marker and a mediator of early nephropathy. The subsequent decline by stage 3 may reflect progressive loss of renal regulatory capacity or exhaustion of the metabolic response, though the cross-sectional design cannot distinguish cause from consequence.
The limitations of the study are clearly acknowledged. As a cross-sectional analysis, it captures a snapshot rather than a trajectory, so it cannot establish whether rising bile acids precede kidney damage or accompany it. The single-center design and the modest discriminative performance of the biomarker both argue for caution. The authors recommend that total bile acid be regarded as a candidate complementary marker, one that might eventually sit alongside the urinary albumin-to-creatinine ratio and estimated glomerular filtration rate rather than replace them, pending prospective validation in external cohorts. Funded by the Chen Xiao-Ping Foundation for the Development of Science and Technology of Hubei Province, the study opens a provocative new line of inquiry into the metabolic choreography of diabetic kidney disease, suggesting that the earliest stage of renal injury carries a distinctive and detectable biochemical signature that current care largely overlooks.
Subject of Research: The non-linear relationship between serum total bile acid levels and early-stage diabetic kidney disease.
Article Title: Total bile acid as an independent indicator for early-stage diabetic kidney disease: a cross-sectional study
Article References: Total bile acid as an independent indicator for early-stage diabetic kidney disease: a cross-sectional study. (n.d.). https://doi.org/10.1186/s12902-026-02553-y
Image Credits: AI Generated
DOI: 10.1186/s12902-026-02553-y
Keywords: total bile acid, diabetic kidney disease, type 2 diabetes, biomarker, urinary albumin-to-creatinine ratio, estimated glomerular filtration rate, bile acid metabolism, nephropathy, logistic regression, ROC analysis, endocrinology, kidney disease
Cite Scienmag News
APA MLA Chicago
Copy citation Download RIS
Tags: bile acid metabolismbiomarkerdiabetic kidney diseaseendocrinologyestimated glomerular filtration ratekidney diseaselogistic regressionnephropathyROC analysistotal bile acidType 2 diabetesurinary albumin-to-creatinine ratio
