Kidney Biopsy Findings Among Patients With Diabetes in the Cleveland Clinic Kidney Biopsy Epidemiology Project
Alvin G Kwon,Hanny Sawaf,Gilda Portalatin,Shruti Shettigar,Leal C Herlitz,Tariq Shafi,Hong Liang,Adam Kabuka,Scott Cohen,Surafel K Gebreselassie,Shane A Bobart
DOI: https://doi.org/10.1016/j.xkme.2024.100889
2024-08-13
Abstract:Rationale & objectives: Diabetic kidney disease (DKD) is a significant complication of diabetes mellitus, often leading to kidney failure. The absence of well-defined factors prevents distinguishing DKD from non-diabetic kidney disease (non-DKD; alternative primary diagnosis identified on kidney biopsy). Study design: Retrospective cohort study. Setting & participants: This study assessed 1,242 patients with a history of diabetes from the Cleveland Clinic Kidney Biopsy Epidemiology Project between January 2015 and September 2021. Exposure: Proteinuria, retinopathy, A1c levels, and estimated glomerular filtration rate. Outcomes: Non-DKD, defined as an alternative primary diagnosis identified on kidney biopsy other than DKD. Analytical approach: Multivariate logistic regression model with backward elimination method. Results: At the time of biopsy, the median (IQR) age was 63 (53-71 years) years, and 58.8% were men. The median hemoglobin A1c value was 6.7% (6.0%-8.1%), and the median serum creatinine level was 2.5 (1.6-3.9 mg/dL) mg/dL. Among 1,242 patients, 462 (37.2%) had DKD alone, and 780 (62.8%) had non-DKD. Among those with non-DKD, the most common diagnoses were focal segmental glomerulosclerosis (24%), global glomerulosclerosis otherwise not specified (13%), acute tubular necrosis (9%), IgA nephropathy (8%), antineutrophil cytoplasmic antibody vasculitis (7%), and membranous nephropathy (5%). Factors associated with having non-DKD on biopsy were having no retinopathy (vs retinopathy) (adjusted odds ratio [aOR], 3.98; 95% CI, 2.69-5.90), lower A1c levels (<7% vs ≥7%) (aOR, 3.08; 95% CI, 2.16-4.39), higher estimated glomerular filtration rate (≥60 vs <60 mL/min/1.73 m2) (aOR, 2.39; 95% CI 1.28-4.45), microalbuminuria (<300 vs macroalbuminuria ≥300 [mg/g]) (aOR; 2.94; 95% CI, 1.84-4.72), and lower protein-creatinine ratio on random urine sample (<3 vs ≥3 mg/mg) (aOR; 1.80; 95% CI, 1.24-2.61). Limitations: Selection bias of clinically indicated biopsies, not protocol biopsies, which likely represent a ceiling (maximum) for non-DKD. Conclusions: Among patients with diabetes undergoing kidney biopsy, 63% have findings in addition to DKD on biopsy. We identified clinical parameters associated with non-DKD in the setting of diabetes. This provides valuable information for clinicians when kidney biopsy should be considered among patients with diabetes to capture all etiologies of proteinuria and kidney dysfunction.