# Respiratory failure, underlying acute illnesses, and hospital outcomes: the S. Giovanni-Addolorata–SIGOT GRACE Study

## TL;DR

This study finds that respiratory failure in older hospitalized patients is linked to higher death rates and longer hospital stays, regardless of other health issues.

## Contribution

The study demonstrates that respiratory failure independently predicts adverse hospital outcomes in older patients, beyond its underlying illnesses.

## Key findings

- Respiratory failure was associated with a 3.98-fold higher odds of hospital mortality.
- Respiratory failure predicted longer hospital stays in adjusted models.
- The effect of respiratory failure on mortality remained significant even after excluding chronic cases.

## Abstract

Although respiratory failure (RF) is frequent among older medical patients admitted to non-intensive hospital units, inherent data are scarce. We determined whether RF predicted adverse hospital outcomes independently of its causative illnesses. In a retrospective observational study from the Geriatric Risk Assessment and Care Evaluation database, we included 1093 patients consecutively admitted to a geriatric hospital unit (2022–2024). Study outcomes included hospital death, length of hospital stay (LOS) and post-discharge institutionalization (discharge to nursing homes or other long-term facilities). RF was diagnosed according to admission peripheral oxygen saturation ≤ 91%, or oxygen therapy/non-invasive ventilation during hospitalization, or specific diagnostic discharge codes. The predictive role of RF was investigated controlling for RF causative illnesses, frailty measured by the Multidimensional Prognostic Index (MPI), and C-reactive protein. The RF prevalence was 43%. Compared to controls, RF patients had higher hospital mortality (25.4% vs. 6.0%) and longer LOS, but comparable institutionalization rate. The fully adjusted odds ratio (OR) of RF for hospital mortality was 3.98 (95% Confidence Interval [CI] 2.53–6.28) and further increased after exclusion of 106 acute-on-chronic RF patients (4.71, CI 2.96–7.49). MPI, C-reactive protein, and sepsis emerged as additional significant predictors of mortality. RF also predicted longer LOS (F 6.78, p 0.009) in a linear regression model, along with age, MPI, pneumonia, pulmonary embolism, stroke, and sepsis. RF was highly prevalent and predicted hospital mortality and longer LOS per se, independently of its causative acute illnesses, frailty, and systemic inflammation. Older patients should be actively screened for RF during hospitalization.

The online version contains supplementary material available at 10.1007/s11739-025-04074-3.