Adults newly diagnosed with type 2 diabetes who present with preserved kidney function and an absence of detectable albuminuria face a substantially higher long-term risk of adverse renal outcomes compared with matched non-diabetic peers, according to a nationwide cohort study based on data from the Korean National Health Insurance Service database. Although the overall difference in the annual rate of kidney filtration decline between individuals with type 2 diabetes and non-diabetic controls was modest across the entire population, long-term tracking revealed a pronounced divergence in cumulative clinical events. The investigation demonstrated that even when routine initial screening shows no biochemical signs of renal impairment, the clinical course of early diabetes harbors considerable underlying heterogeneity, with high-risk trajectories hidden beneath population-wide averages.
Type 2 diabetes represents the leading cause of end-stage kidney disease globally, yet characterizing the early natural trajectory of diabetic kidney disease has historically presented notable methodological challenges. In clinical practice, a significant proportion of patients already exhibit reduced kidney function or pathological albumin excretion at the moment of diagnosis, which obscures the initial progression pattern from normal physiological baselines. Furthermore, the modern therapeutic landscape incorporates effective kidney-protective agents, including sodium-glucose cotransporter 2 inhibitors, nonsteroidal mineralocorticoid receptor antagonists, and glucagon-like peptide-1 receptor agonists alongside renin-angiotensin system blockade. Because these interventions alter disease progression, clarifying the unperturbed early trajectory in patients who initially lack overt renal manifestations requires observing populations from their initial clinical diagnosis.
To overcome these limitations, a research team analyzed national healthcare data to evaluate incident albuminuria, progressive filtration decline, and major renal endpoints over an extended follow-up window. The study population was identified from the Korean National Health Insurance Service, a single-payer public insurance program that covers approximately 98 percent of South Korea’s population, with the remaining 2 percent supported by the Medical Aid program. The investigators screened adults aged 20 to 80 years who underwent standardized national health examinations between 2012 and 2013 and received a new diagnosis of type 2 diabetes during that interval. Eligible participants were required to complete at least three subsequent health screenings over the follow-up period, ensuring serial assessments of laboratory and clinical markers.
The study protocol established stringent baseline criteria to isolate participants with entirely preserved renal parameters. Incident type 2 diabetes was defined by either two or more diagnostic claims under International Classification of Diseases codes E11 through E14 within one year of initial diagnosis, or a single diagnostic claim accompanied by a prescription for antidiabetic medication. Investigators excluded individuals with an estimated glomerular filtration rate below 60 milliliters per minute per 1.73 square meters, any baseline dipstick albuminuria, prior chronic kidney disease, end-stage kidney disease, prior diabetes diagnoses, fasting glucose levels of 126 milligrams per deciliter or higher before the index year, or extreme baseline filtration outliers exceeding the 99.75th percentile. Each diabetic patient was matched with three non-diabetic controls based on age and sex, resulting in a final analytical cohort of 226,265 individuals with newly diagnosed type 2 diabetes and 678,795 matched controls.
The cohort comprised 105,762 men and 120,503 women with type 2 diabetes, alongside 317,286 male and 361,509 female control subjects. At baseline, the mean age was 54.98 years among men and 57.54 years among women across both groups. Compared with controls, individuals with newly diagnosed diabetes exhibited higher mean body mass index values, with 25.00 versus 23.74 kilograms per square meter in men and 24.43 versus 23.29 kilograms per square meter in women. Systolic blood pressure averaged 127.26 millimeters of mercury in men with diabetes compared with 123.86 millimeters of mercury in male controls, while diastolic blood pressure averaged 79.53 versus 77.48 millimeters of mercury. Baseline estimated glomerular filtration rates calculated using the Chronic Kidney Disease Epidemiology Collaboration equation were virtually identical between groups, averaging 91.40 versus 91.33 milliliters per minute per 1.73 square meters in men and 91.59 versus 91.45 in women.
Over a median follow-up period of 9.1 years, with an interquartile range spanning 8.0 to 9.9 years, the primary composite kidney outcome occurred in 12.9 percent of patients with type 2 diabetes compared with 9.4 percent of matched controls, representing an absolute risk difference of 3.5 percent. The primary composite outcome encompassed the first recorded occurrence of an estimated glomerular filtration rate below 60 milliliters per minute per 1.73 square meters, incident albuminuria defined by a semi-quantitative urine dipstick grade of 1+ or greater, chronic dialysis, or kidney transplantation. In multivariable Cox proportional hazards regression models adjusting for age, sex, body mass index, blood pressure, smoking, alcohol use, exercise, comorbidities, and baseline medications, type 2 diabetes was associated with a 20 percent higher risk of the composite outcome, corresponding to an adjusted hazard ratio of 1.20 with a 95 percent confidence interval from 1.18 to 1.22.
Evaluation of the secondary endpoints demonstrated that individual renal outcomes exhibited varying degrees of relative and absolute risk elevation among patients with diabetes. The adjusted hazard ratio for developing an estimated glomerular filtration rate below 60 milliliters per minute per 1.73 square meters was 1.16, with an absolute event rate of 10.5 percent in the diabetes cohort versus 8.1 percent in controls. For incident albuminuria, the adjusted hazard ratio was 1.45, reflecting incidence rates of 2.8 percent versus 1.5 percent. The most pronounced relative increase appeared in severe outcomes: progression to chronic dialysis or kidney transplantation yielded an adjusted hazard ratio of 2.98, with a 95 percent confidence interval of 2.59 to 3.43, occurring in 0.2 percent of diabetic individuals compared with 0.1 percent of controls. Additionally, all-cause mortality was higher in the diabetes group, exhibiting an adjusted hazard ratio of 1.48.
When analyzing kidney filtration rates continuously over time using linear mixed-effects models, patients with type 2 diabetes demonstrated a slightly steeper annual decline in estimated glomerular filtration rate compared with controls, reflecting a between-group difference of –0.03 milliliters per minute per 1.73 square meters per year. While this difference was statistically significant due to the extensive sample size, the researchers noted that a cumulative difference of less than 0.3 milliliters per minute per 1.73 square meters over nine years would appear clinically negligible if interpreted solely as an aggregated mean. However, this modest population-wide average masked striking divergence across individual clinical trajectories, prompting the application of longitudinal trajectory-based clustering using finite mixture mixed-effects regression models with cubic spline functions of time.
The trajectory analysis classified the type 2 diabetes cohort into four distinct progression clusters that exhibited radically different long-term patterns. Clusters 1 and 2, which accounted for 19.3 percent and 49.6 percent of the overall diabetes population respectively, demonstrated stable or minimally declining filtration trajectories resembling those of the non-diabetic population. In contrast, clusters 3 and 4 comprised 14.9 percent and 16.2 percent of the cohort, representing approximately 31.1 percent of patients who experienced progressive renal decline. Remarkably, although these two progressive clusters constituted less than one-third of the diabetic cohort, they accounted for 76.5 percent of all composite kidney events observed during follow-up. In the most rapid decline group, cluster 4, the adjusted hazard ratio reached 14.53 for the composite outcome, 134.01 for filtration falling below 60, and 4.77 for end-stage renal replacement therapy compared with cluster 1.
The study authors highlighted several methodological considerations and limitations inherent in national observational health registries. Because kidney function and dipstick proteinuria were assessed through periodic biennial or annual screening examinations, the operational endpoints relied on single measurements without immediate confirmatory testing, which could introduce classification error. Furthermore, semi-quantitative dipstick urinalysis exhibits limited sensitivity for moderately increased albuminuria compared with laboratory albumin-to-creatinine ratios. The dataset also lacked longitudinal hemoglobin A1c measurements to assess ongoing glycemic control, and the cohort was diagnosed in 2012–2013 prior to widespread reimbursement and utilization of newer therapeutic agents such as sodium-glucose cotransporter 2 inhibitors. Nevertheless, the findings indicate that baseline preservation of kidney function does not guarantee a benign long-term course in newly diagnosed type 2 diabetes, emphasizing the clinical value of early trajectory monitoring to identify high-risk individuals.
Subject of Research: Long-term kidney disease trajectories and risks in adults with newly diagnosed type 2 diabetes
Article Title: Long-term kidney outcomes after new-onset type 2 diabetes in adults with preserved kidney function and no albuminuria: a nationwide cohort study
Article References: Kim, H. W., Heo, S.-J., Park, S., Kim, H. J., Park, S.-H., & Han, S. Y. (2026). Long-term kidney outcomes after new-onset type 2 diabetes in adults with preserved kidney function and no albuminuria: a nationwide cohort study. Frontiers in Endocrinology, 17, Article 1944422. https://doi.org/10.3389/fendo.2026.1944422
Image Credits: AI Generated
DOI: 10.3389/fendo.2026.1944422
Keywords: type 2 diabetes, kidney disease, albuminuria, estimated glomerular filtration rate, end-stage kidney disease, epidemiology, cohort study, Korean National Health Insurance Service, chronic kidney disease, disease trajectories, trajectory-based clustering, renal health, dialysis, kidney transplantation, preventive medicine, clinical nephrology
News Source: Jerry Hayes. (November 20, 2026). Type 2 Diabetes Increases Long-Term Kidney Failure Risk Despite Early Preserved Function. Scienmag.



