Background: Long-term outcome studies of home mechanical ventilation (HMV) are sparse. We analysed predictors of survival and survival per se over 26 years in patients treated with HMV.

Methods: Population-based cohort study of adult patients starting HMV in the Swedish Registry for Respiratory Failure (Swedevox) 1996?2021, cross-linked with the National Cause of Death registry. Risk factors for mortality were analysed using multivariable Cox regression models.

Results: We included 10,190 patients (50.1 % women, age 62.9±14.5 years).  Median survival in those with lung disease (n=2,264) was 2,6 years (95% confidence interval (CI) 2.4-2.8), restrictive thoracic disease (n=786) 6.9 years (6.4-7.6), obesity hypoventilation syndrome (n=3,660) 7.0 years (6.6-7.4), neuromuscular disease (n=1,030) 7.6 years (6.9-8.6), amyotrophic lateral sclerosis (n=1,444) 0.8 years (0.8-0.9), and other neurological disorders (n=1,006) 5.2 years (4.4-6.0). Lower mortality was associated with BMI, HR 0.89 (95%CI 0.87-0.92) per 5 units increase, vital capacity, HR 0.92 (0.90-0.94) per 10 percent units, and normalization of PaCO2 at follow-up after 1.3±0.9 years, HR 0.78 (0.69-0.88). Higher mortality was associated with age, HR 1.52 (1.46-1.58) per 10 years, and acute initiation of HMV, HR 1.18 (1.08-1.29). Sex, baseline PaCO2, tracheostomy, and start year did not independently predict mortality.

Conclusion: Survival differed markedly between diagnosis groups but was unchanged over the study period of 26 years. Our data suggest that control of PaCO2 at follow-up is a key treatment goal in HMV.