# Lung function predicts mortality in people with serious mental illness: A 6-year follow-up study

## TL;DR

Lung function, especially FEV1, predicts higher mortality risk in people with serious mental illness over six years.

## Contribution

Shows spirometry can identify at-risk individuals with SMI for targeted prevention.

## Key findings

- Lower FEV1 and FVC z-scores were significantly linked to increased mortality risk.
- Kaplan-Meier analysis confirmed FEV1 and FVC as mortality predictors.
- FEV1 z-score independently predicted mortality after adjusting for age and abdominal circumference.

## Abstract

The population with a serious mental illness (SMI) shows a high risk of premature mortality. Overexposed to multiple health risks throughout life, their main threat is physical illness, which starts earlier and is not diagnosed in time. Developing preventive actions is a public health priority.

This longitudinal prospective study evaluated the predictive value of lung function on all-cause mortality in patients with schizophrenia (SCHIZ) or bipolar disorder. Patients aged 40–70 years, active smokers, and without preexisting respiratory disease underwent spirometry following American Thoracic Society/European Respiratory Society 2021 standards. Mortality data were collected through December 2022. Cox proportional hazards models and Kaplan–Meier survival curves analyzed the association between lung function, specifically forced expiratory volume in the first second (FEV1), forced vital capacity (FVC), and mortality, adjusting for relevant confounders (age, gender, abdominal circumference, and comorbidities).

Of 107 participants (mean age 49.3 years, 63.3% male) with SMI (72% SCHIZ) and active smokers, 8 (7.5%) died during the 6-year follow-up (5 cardiovascular and 3 cancer). Mean z-scores were −1.41 (SD = 1.22) for FEV1 and −0.99 (SD = 1.16) for FVC. Lower FEV1 and FVC z-scores were significantly associated with increased mortality risk (p = 0.002 and p = 0.009, respectively). Kaplan–Meier analysis confirmed this association for FEV1 (p = 0.039) and FVC (p = 0.007) but not for gender, comorbidities (hypertension, diabetes, and dyslipidemia), or FEV1/FVC. A multivariate Cox regression model, adjusting for age and abdominal circumference, confirmed the independent predictive value of lower FEV1 z-score for mortality (hazard regression = 0.473, 95% confidence interval: 0.220–0.979, p = 0.044).

Poorer lung function, especially lower FEV1, was independently associated with all-cause mortality in SMI. Spirometry, an easily implementable technique, could help to detect at-risk individuals and favor prevention initiatives.