Admission Respiratory Support Strategies Amongst Patients with Cardiogenic Shock

Abstract Background Respiratory failure frequently complicates cardiogenic shock (CS), yet the optimal initial ventilatory strategy remains uncertain. Whether a noninvasive ventilation (NIV)–first approach is associated with worse outcomes compared to direct invasive mechanical ventilation (IMV) is unclear. Methods We performed a cohort study using a multicenter, nationally representative database including adults with CS who received NIV or IMV on the first hospital day. Patients were categorized as NIV only, direct IMV, or NIV before IMV. Inverse probability of treatment weighting balanced baseline characteristics. The primary outcome was in-hospital mortality. Sensitivity analyses excluded mechanical circulatory support and out-of-hospital cardiac arrest and additionally incorporated lactate adjustment and entropy weighting. Results Among 81,892 patients with CS, 13.7% received NIV alone, 81.9% direct IMV, and 4.4% NIV before IMV on the first day of admission. In-hospital mortality was 37.2%, 57.7%, and 52.1%, respectively. After adjustment, IMV and NIV before IMV had higher mortality than NIV (absolute risk difference 11.8%, 95% CI 10.4–13.3, and 12.9%, 95% CI 10.6–15.1; both p<0.001). Mortality did not differ between direct IMV and NIV before IMV (risk difference 1.0%, 95% CI −1.0-3.0; p=0.279). Notably, sensitivity analyses including lactate and when stratified by SCAI D shock showed higher mortality with NIV before IMV compared to direct IMV. Conclusions Direct IMV and NIV before IMV were associated with worse outcomes than NIV alone. Results for direct IMV compared to NIV before IMV were not consistent between analyses and support the urgent need for a prospective study of respiratory support strategies.