Investigation of Prolonged Mechanical Ventilation Mediating the Association Between Mechanical Power and Mortality...

Studies in both intensive care unit (ICU) and noncardiac intraoperative patients have shown that higher pulmonary mechanical power is associated with increased mortality. We sought to estimate whether intraoperative mechanical power is associated with operative mortality, and secondarily with late mortality, in cardiac surgery patients and if these associations are mediated via prolonged mechanical ventilation.

In this restrospective, single-center, cohort study of adult, cardiac surgery patients, we calculated mechanical power based on tidal volume, peak inspiratory pressure, positive end-expiratory pressure, and respiratory rate. After transforming mechanical power using fractional polynomials, we used logistic regression models to estimate the associations between mechanical power, prolonged mechanical ventilation, and the primary outcome of operative mortality, with prolonged mechanical ventilation as the mediator between mechanical power and mortality. Similar separate regressions were done for the secondary outcome of late mortality, which excluded patients with operative mortality.

Mean ± standard deviation (SD) of mechanical power for the 5907 patients was 7.9 ± 2.6 J/min. One hundred thirty-six (2.3%) patients were operative mortalities, and a total of 598 (10%) patients were dead at a median (interquartile range) follow-up of 792 (161-1496) days. Late mortality in patients who survived the perioperative period was 462 of 5771 (8.0%) patients. Operative mortality doubled from 1.9% to 3.8% between the first (mechanical power <6.13 J/min) and last (>9.23 J/min) mechanical power quartiles. Late mortality in patients who survived the perioperative period increased from 7.7% to 9.7% between the first (mechanical power <6.13 J/min) and last (mechanical power >9.23 J/min) quartiles. After adjustment for confounders, mechanical power was associated with operative mortality odds ratio (OR) = 1.102 (per J/min) (95% confidence interval [CI], 1.030-1.179], P =.005. Mechanical power via the square, MP2, and cubic, MP3, terms of the transformations was associated with prolonged mechanical ventilation (OR = 1.015 [95% CI, 1.008-1.021], P <.001 for MP2 and OR = 0.99961 [95% CI, 0.99930-0.99991], P =.012 for MP3). These correspond to a combined OR = 0.746 (95% CI, 0.622-0.891), 1.431 (1.108-1.857), and 3.033 (1.109-8.222) for mechanical power = 6, 10, and 14 J/min, respectively, compared to a patient with the mean mechanical power = 7.9 J/min. By mediation analysis, all of mechanical power's association with operative mortality was mediated via its association with prolonged mechanical ventilation.

We found that higher levels of mechanical power were associated with intraoperative and late mortality, and this association was mediated by prolonged mechanical ventilation. Further study is needed to understand how mechanical power might lead to prolonged mechanical ventilation and mortality.