# Association between Geriatric Nutritional Risk Index (GNRI) and all-cause mortality in centenarians, with a focus on nonlinear and threshold effects: a multi-method observational study

## TL;DR

This study finds that a nutritional index called GNRI is linked to lower death risk in centenarians, with a key threshold at 96.

## Contribution

The study identifies a nonlinear relationship and a specific GNRI threshold for mortality risk in centenarians.

## Key findings

- GNRI is inversely associated with mortality in centenarians, with a hazard ratio of 0.97 per 1-unit increase.
- A threshold GNRI value of 96.378 was identified, below which mortality risk decreases significantly with higher GNRI.
- The inverse association between GNRI and mortality remained consistent after propensity score matching.

## Abstract

The Geriatric Nutritional Risk Index (GNRI) is a simple index for assessing nutritional status in older adults. Its association with mortality has been shown in general elderly populations, but data in centenarians are scarce. We aimed to examine the relationship between GNRI and all-cause mortality in centenarians.

We included 1,002 centenarians with complete clinical and follow-up data and categorized them into quartiles according to GNRI. All-cause mortality was the primary endpoint. Cox proportional hazards models, restricted cubic splines, subgroup analyses and sensitivity analyses were used to evaluate the association between GNRI and mortality. To enhance robustness, propensity score matching (PSM) was performed.

Among the 1,002 centenarians (18% male), the overall mortality rate was 92.7%. In multivariable Cox models, GNRI was inversely associated with mortality (per 1-unit increase: HR 0.97, 95% CI 0.96–0.98, p < 0.001). Restricted cubic spline analysis showed a nonlinear relationship between GNRI and mortality, with a gradual increase in death risk at lower GNRI values. A clear threshold was identified at GNRI = 96.378. When GNRI <96.378, each 1-unit increase in GNRI was associated with a 4% reduction in mortality risk (HR 0.96, p < 0.001), whereas when GNRI ≥96.378, mortality risk no longer changed significantly (HR 1.01, p = 0.443). Subgroup analyses showed no significant interactions between GNRI and most covariates, except sex, and this pattern persisted after PSM. PSM-based sensitivity analyses yielded consistent results: the inverse association between GNRI and mortality remained (per 1-unit increase: HR 0.98, 95% CI 0.97–0.99, p < 0.001), and the threshold effect at GNRI = 96.378 was confirmed (GNRI <96.378: HR 0.95, p < 0.001; GNRI ≥96.378: HR 1.02, p = 0.136).

GNRI is strongly and nonlinearly associated with all-cause mortality in centenarians, with a key threshold around 96. This threshold may provide a quantitative target for individualized nutritional assessment and intervention in this extremely old population.