Associations Between Pulmonary Artery Catheter Use and Outcomes After Cardiac Surgery: An Entropy-Balanced Cohort Study.

Abstract

BACKGROUND Pulmonary artery catheters are used widely in cardiac surgery despite conflicting associations with patient outcomes. We evaluated the associations between pulmonary artery catheter use and clinical outcomes following cardiac surgery using a large cohort of patients treated at a US academic center. METHODS We performed a retrospective entropy-balanced cohort study of consecutive adults undergoing cardiac surgery from a single tertiary center in Boston, Massachusetts, from 2008 to 2022. We used entropy balancing to achieve exact covariate balance on prespecified baseline characteristics and then estimated the average treatment effect of pulmonary artery catheter use on clinical and mechanistic outcomes. The primary outcome was mortality measured 90 days after surgery. Secondary outcomes were acute kidney injury, hospital and intensive care unit (ICU) length of stay, time in postoperative organ dysfunction measured at 7 days, prolonged inotrope use (>4 hours), significant peak vasopressor requirement (>0.1 μg/kg/min in norepinephrine equivalents), net fluid balance at 24 hours, number of fluid boluses administered, volume of allogeneic red blood cells transfused, and total duration of mechanical ventilation. RESULTS We included 10,044 patients, of whom 5850 (58.2%) were managed with a pulmonary artery catheter. Pulmonary artery catheter use was not associated with 90-day mortality (risk ratio [RR], 0.966; 95% confidence interval [CI], 0.719-1.30; P =.816) nor in-hospital mortality (RR, 0.921; 95% CI, 0.632-1.34; P =.670). There was no between-group difference in hospital length of stay (median difference [MD], 0.050 days; 95% CI, 0.0477-0.148; P =.269), but patients managed with a pulmonary artery catheter had greater ICU length of stay (MD = 13.1 hours; 95% CI, 9.74-16.4; P <.001). Pulmonary artery catheters were also associated with increased incidence of acute kidney injury (RR, 1.12; 95% CI, 1.07-1.17; P <.001). Patients who received a pulmonary artery catheter were more likely to have prolonged inotrope requirements (RR, 4.13; 95% CI, 3.42-4.98; P <.001), significant vasopressor requirements (RR, 1.37; 95% CI, 1.28-1.46; P <.001), higher positive fluid balances (MD, 566 mL; 95% CI, 453-678; P <.001), higher volumes of allogeneic RBCs transfused (MD, 226 mL; 95% CI, 183-269; P <.001), higher time in postoperative organ dysfunction (MD, 3.96 hours; 95% CI, 3.01-4.90; P <.001), and longer durations of mechanical ventilation (MD, 11.3 hours; 95% CI, 7.33-15.2; P <.001). CONCLUSIONS In a large entropy-balanced cohort study of adults undergoing cardiac surgery, pulmonary artery catheter use was not associated with mortality, but was linked with a higher treatment intensity and longer ICU stay.