# Sarcopenia Index trajectories predict long-term mortality in super-elderly patients with sepsis: a retrospective cohort study

## TL;DR

Tracking changes in muscle mass over time in very elderly sepsis patients can predict long-term survival better than single measurements.

## Contribution

Shows that dynamic muscle mass trajectories, not just static measures, predict mortality in super-elderly sepsis patients.

## Key findings

- Low-level sarcopenia index trajectory group had higher mortality (62.3% vs 48.9%) after adjustment.
- SI trajectory added small but significant value to risk models (NRI=0.020, p<0.05).
- Dynamic SI monitoring improves risk stratification in super-elderly sepsis patients.

## Abstract

The dynamic trajectory of muscle mass during sepsis may hold superior prognostic value over static assessments, particularly in vulnerable super-elderly patients. This study aimed to identify distinct dynamic trajectories of the Sarcopenia Index (SI) using Group-Based Trajectory Modeling (GBTM) and investigate their association with 180-day mortality.

This retrospective cohort study enrolled 210 super-elderly patients (aged >85 years) with sepsis. GBTM was employed to delineate SI trajectories over 60 days. The primary outcome was 180-day mortality. Kaplan–Meier analysis and multivariable Cox proportional hazards regression were used to assess the association between SI trajectories and mortality. The incremental predictive value of trajectory data was evaluated using C-index, Net Reclassification Improvement (NRI), and Integrated Discrimination Improvement (IDI).

Two distinct SI trajectories were identified: a “High-Level Group” (n = 88) and a “Low-Level Group” (n = 122). The Low-Level Group was characterized by lower baseline SI, poorer functional status, and higher prevalence of long-term bedridden status. The 180-day mortality rate was higher in the Low-Level Group (62.3% vs. 48.9%); however, this difference did not reach statistical significance in the unadjusted analysis (p = 0.072). After multivariable adjustment, assignment to the Low-Level trajectory remained an independent predictor of mortality (Adjusted HR = 1.64, 95% CI: 1.08–2.48, p = 0.020). Adding the SI trajectory to a clinical risk model led to a small but statistically significant improvement in risk reclassification (NRI = 0.020, p < 0.05), while discrimination gains were modest (IDI = 0.150, p = 0.078).

A low and declining SI trajectory is an independent predictor of long-term mortality in super-elderly sepsis patients. Dynamic monitoring of SI provides incremental prognostic value over static assessments, offering a novel tool for early risk stratification and targeted interventions.