Method: retrospective study with external validation.
The medical records of all patients referred for pulmonary rehabilitation at ADIR Association, France, between 2015 and 2023 will be screened for eligibility. Demographic data and performance during the 6-Minute Stepper Test will be collected directly from the patients' medical records.
Collected data include: age, weight, height, body-mass index, sex, long-term oxygen therapy, long-term non-invasive ventilation, oxygen uptake (from cardiopulmonary exercise testing), peak power output (from cardiopulmonary exercise testing), pulmonary function testing, disease stage according to the Global Initiative for Chronic Obstructive Lung Disease classification, modified Medical Research Council dyspnea score, smoking status, quality of life, number of disease exacerbations prior to the pulmonary rehabilitation program, number of pulmonary rehabilitation programs performed during the follow-up period, and performance during the 6-Minute Stepper Test.
Vital status (alive or deceased) will be collected from the database of the French National Institute of Statistics and Economic Studies (INSEE): https://arbre.app/insee. This variable will be collected during the first quarter of 2025.
A receiver operator curve analysis will be performed to determine the statistical threshold value (number of steps) with the best sensitivity and specificity for predicting all-cause mortality at 2 and 5 years.
Patients will then be divided into two groups based on the established threshold, and a Cox regression model will be used to determine the hazard ratio for mortality at 2 and 5 years.
Invidiual participant data from a prospective study will be used to assess the external validity of established thresholds.(https://pubmed.ncbi.nlm.nih.gov/27216848/).
Statistical analysis :
Continuous data will be expressed as mean (standard deviation), median (interquartile range), and categorical data as number (percentage). The distribution of data will be assessed using the Shapiro-Wilk test.
To determine the statistical threshold value of the 6-Minute Stepper Test with the best sensitivity and specificity for 2-year and 5-year mortality, receivor operator curves (95% confience interval) will be generated.
Patients will then be divided into two groups (based on the threshold value). The characteristics of the two groups will be compared using Fisher's exact test (categorical data), the independent Student's t-test (parametric data), or the Mann-Whitney U test (non-parametric data).
A Cox proportional hazards regression model will be used to calculate the "crude" and "adjusted" hazard ratio (95% confidence interval) for 2-year and 5-year mortality. The adjusted model will include the following variables: sex, age, FEV1 (L), modified Medical Research Council dyspnea score, BMI. Sensitivity analysis will be porformed adding smoking status, long term oxygen use and number of pulmonary rehabilitation during the follow-up.
The external validity of the model will be assessed using individual pariticipant data from a previously published study (https://pubmed.ncbi.nlm.nih.gov/27216848/).
