Albumin, mNUTRIC and NRS-2002: Predicting Mortality in Elderly ICU Fracture Patients

The primary objective was to evaluate whether admission serum albumin predicts six-month all-cause mortality in older adult patients admitted to the intensive care unit (ICU) after simple fracture surgery, and to compare its predictive performance with the modified Nutrition Risk in the Critically Ill (mNUTRIC) score and the Nutrition Risk Screening 2002 (NRS-2002). The secondary objectives were to identify baseline predictors of six-month mortality and high-risk mNUTRIC classification.

This retrospective cohort study included patients aged ≥65 years admitted to the ICU of a tertiary care hospital after surgery for a simple fracture between July and December 2024. Demographic data, comorbidities, admission laboratory values (including albumin, prealbumin, and 25-hydroxy vitamin D, the latter included as an adjunctive nutritional biomarker), APACHE II, SOFA, mNUTRIC, and NRS-2002 scores were recorded. Postoperative complications and admission durations were evaluated. Binomial logistic regression models were constructed for six-month all-cause mortality and nutritional risk group classification. Receiver operating characteristic (ROC) analysis with the Youden Index was performed to determine cutoff values.

A total of 172 patients (mean age 80.84 ± 7.72 years; 67.4% female) were analyzed. Six-month all-cause mortality was 22.7%. Serum albumin (OR 0.823, 95% CI 0.729-0.928, p = 0.002) and ICU admission duration (OR 1.413, 95% CI 1.101-1.812, p = 0.007) were independent predictors of six-month all-cause mortality, whereas mNUTRIC, NRS-2002, and vitamin D were not. Neither mNUTRIC nor NRS-2002 scores differed significantly between survivors and non-survivors. In nutritional risk group analysis, age (OR 1.117, p = 0.001) and APACHE II (OR 1.694, p = 0.001) were independent predictors of high mNUTRIC risk. Head-to-head ROC analysis for the primary outcome of six-month all-cause mortality showed that admission serum albumin (AUC 0.698, 95% CI 0.604-0.793) provided significantly better discrimination than mNUTRIC (AUC 0.570, DeLong p = 0.046) and NRS-2002 (AUC 0.550, DeLong p = 0.039). In a sensitivity model restricted to admission-time variables (albumin, age, APACHE II, vitamin D, Charlson Comorbidity Index), admission albumin remained an independent predictor (OR 0.830, 95% CI 0.747-0.923, p < 0.001) and age emerged as a further independent predictor (OR 1.062, p = 0.034).

Serum albumin outperformed mNUTRIC and NRS-2002 in predicting six-month all-cause mortality among older adult post-fracture ICU patients. Because neither mNUTRIC nor NRS-2002 discriminated between survivors and non-survivors, these scores alone cannot be recommended as mortality-prediction tools in this orthogeriatric ICU population. Whether admission albumin adds incremental value to existing nutritional scoring in this setting requires prospective, adequately powered validation.