Relationship between the geriatric nutritional risk index and sepsis in elderly critically ill patients: a retrospective cohort study.

Objective: To investigate the relationship between the geriatric nutritional risk index (GNRI) and sepsis in elderly critically ill patients.

Methods: We extracted data from the MIMIC-IV database for patients aged 65 and older who were first-time admissions to the Intensive Care Unit. The GNRI, derived from serum albumin, height, and weight, served as the exposure variable. The main outcome was the occurrence of sepsis. We utilized multivariable logistic regression to evaluate the association between GNRI and sepsis. Additionally, we employed restricted cubic splines to evaluate the potential nonlinear relationship between GNRI and the incidence of sepsis.

Results: The study included 3,535 elderly patients, with 2,005 (56.7%) developing sepsis. Septic patients had significantly lower GNRI values (median: 99 vs. 101, P = 0.021) and a higher prevalence of malnutrition risk (GNRI ≤ 98) (47.5% vs. 43.5%, P = 0.018). The analysis employing restricted cubic splines identified a U-shaped association between GNRI and sepsis (P for non-linearity = 0.004), with an inflection point at 105.4. To the left of this inflection point, GNRI was significantly negatively associated with sepsis (OR = 0.987, 95% CI 0.978-0.996, P = 0.005), while the right side showed a positive correlation (OR = 1.011, 95% CI 1.000-1.022, P = 0.044).

Conclusion: The study reveals a U-shaped association between GNRI and sepsis in elderly critically ill patients. Both low and high GNRI values are associated with increased sepsis risk, with a turning point at 105.4. These findings highlight GNRI's utility as a screening tool for identifying sepsis risk in elderly ICU patients.