Abstract
Background. Identification of patients who may become hypercapnic, or develop acidotic hypercapnic respiratory failure (AHRF), is important in chronic obstructive pulmonary disease (COPD) to avoid hospital admission and select patients for use of home NIV. This study aimed to identify factors associated with presence and development of hypercapnia. Methods. 1224 patients, 637 with COPD and 587 with alpha 1 antitrypsin deficiency (AATD), from 4 previously established patient cohorts, were included in cross-sectional analyses of hypercapnia (PaCO2 >= 6.5 kPa or 48.8 mmHg), focusing on phenotypic features of COPD and mortality. Longitudinal associations of rising PaCO2 were also assessed. A second cohort of 160 COPD patients underwent sleep studies and 1-year follow-up, analysing in a similar way, incorporating additional information from their sleep studies if appropriate. Results. Hypercapnia was 15 times more common in usual COPD than AATD (p <0.01) after adjustment for baseline differences by regression. Independent predictors of hypercapnia in COPD included FEV1 and current use of oxygen; these variables, together with lack of emphysema, explained 11% of variance in CO2. Increasing PaCO2 also associated with higher risk of death (p = 0.03). 44/160 patients exhibited sleep disordered breathing. The sleep study cohort also showed an association of low FEV1 with hypercapnia. Prior hospital admission for AHRF was also clinically significant, being a feature of almost double the number of hypercapnic patients in both test and sleep study COPD cohorts. Conclusion. Lower FEV1 and prior AHRF are the main associations of hypercapnia in COPD, which carries a poor prognosis, particularly worsening over time.