Patterns of multimorbidity and in-hospital kidney function decline in older adults: a latent class analysis of three acute-care cohorts.

Multimorbidity is associated with an increased risk of renal function decline in older individuals. In this study, we aimed to evaluate whether multimorbidity patterns differently contributed to this risk in three different hospital settings. We conducted a comparative cohort study of patients aged ≥ 65 years admitted to acute-care hospitals in Italy. Data were drawn from the retrospective SIN-SIGG study (geriatric units, n = 1600; nephrology units, n = 560) and the prospective REPOSI registry (internal medicine units, n = 4595). Quantitative multimorbidity was defined as the number of chronic conditions with prevalence ≥ 5% per cohort; qualitative multimorbidity was derived through latent class analysis of disease combinations. The primary outcome was change in estimated glomerular filtration rate (eGFR) from admission to discharge, calculated using the Berlin Initiative Study 1 (BIS1) equation. A relative decline was defined as ≥ 20% reduction from baseline. Multivariable models were adjusted for age, sex, baseline eGFR, acute diagnoses, and number of medications. Among 6,755 patients (mean age 80.6 years; 47% women), multimorbidity prevalence ranged from 87 to 93%. Quantitative multimorbidity showed modest and inconsistent associations with renal decline. In contrast, qualitative multimorbidity consistently predicted eGFR deterioration. Across all cohorts, the cardiorespiratory and renal cluster was associated with higher odds of in-hospital eGFR decline compared with neurogeriatric and mixed patterns (geriatrics OR 2.24 [95% CI 1.16-4.70]; internal medicine OR 1.72 [95% CI 1.22-2.45]; nephrology OR 2.11 [95% CI 1.01-4.56]). Disease-pattern-based multimorbidity better stratifies short-term renal risk than disease counts, identifying high-risk patients for targeted nephroprotective care during hospitalization.