Abstract
BACKGROUND: The higher mortality rate in untreated patients with obesity-associated hypoventilation is a strong rationale for long-term noninvasive ventilation (NIV). The impacts of comorbidities, medications and NIV compliance on survival of these patients remain largely unexplored. METHODS: Observational cohort of hypercapnic obese patients initiated on NIV between March 2003 and July 2008. Survival curves were estimated by the Kaplan-Meier method. Anthropometric measurements, pulmonary function, blood gases, nocturnal SpO(2) indices, comorbidities, medications, conditions of NIV initiation and NIV compliance were used as covariates. Univariate and multivariate Cox models allowed to assess predictive factors of mortality. RESULTS: One hundred and seven patients (56% women), in whom NIV was initiated in acute (36%) or chronic conditions, were followed during 43±14 months. The 1, 2, 3 years survival rates were 99%, 94%, and 89%, respectively. In univariate analysis, death was associated with older age (>61 years), low FEV1 (<66% predicted value), male gender, BMI×time, concomitant COPD, NIV initiation in acute condition, use of inhaled corticosteroids, ß-blockers, nonthiazide diuretics, angiotensin-converting enzyme inhibitors and combination of cardiovascular drugs (one diuretic and at least one other cardiovascular agent). In multivariate analysis, combination of cardiovascular agents was the only factor independently associated with higher risk of death (HR = 5.3; 95% CI 1.18; 23.9). Female gender was associated with lower risk of death. CONCLUSION: Cardiovascular comorbidities represent the main factor predicting mortality in patient with obesity-associated hypoventilation treated by NIV. In this population, NIV should be associated with a combination of treatment modalities to reduce cardiovascular risk.