Causes of hypercapnic respiratory failure: a population-based case-control study.

OBJECTIVE: There are no population-based data on the relative importance of specific causes of hypercapnic respiratory failure (HRF). We sought to quantify the associations between hospitalisation with HRF and potential antecedent causes including chronic obstructive pulmonary disease (COPD), obstructive sleep apnea, and congestive cardiac failure. We used data on the prevalence of these conditions to estimate the population attributable fraction for each cause. METHODS: A case-control study was conducted among residents aged ≥ 40 years from the Liverpool local government area in Sydney, Australia. Cases were identified from hospital records based on PaCO2 > 45 mmHg. Controls were randomly selected from the study population using a cluster sampling design. We collected self-reported data on medication use and performed spirometry, limited-channel sleep studies, venous sampling for N-terminal pro-brain natriuretic peptide (NT-proBNP) levels, and sniff nasal inspiratory pressure (SNIP) measurements. Logistic regression analyses were performed using directed acyclic graphs to identify covariates. RESULTS: We recruited 42 cases and 105 controls. HRF was strongly associated with post-bronchodilator airflow obstruction, elevated NT-proBNP levels, reduced SNIP measurements and self-reported opioid medication use. There were no differences in the apnoea-hypopnea index or oxygen desaturation index between groups. COPD had the highest population attributable fraction (42%, 95% confidence interval 18% to 59%). CONCLUSIONS: COPD, congestive cardiac failure, and self-reported use of opioid medications, but not obstructive sleep apnea, are important causes of HRF among adults over 40 years old. No single cause accounts for the majority of cases based on the population attributable fraction.