Association of intraoperative end-tidal CO2 levels with postoperative outcomes: a patient-level analysis of two randomised clinical trials

Abstract

Background The relationship between intraoperative end-tidal CO2 (etCO2) levels and postoperative outcomes remains unclear. We conducted a post hoc analysis of two randomised trials in adults undergoing major surgery under general anaesthesia. Methods We re-analysed individual participant data comparing high or low positive end-expiratory pressure with low tidal volume intraoperative ventilation using a merged database derived from two randomised trials in non-obese (PROVHILO: ISRCTN70332574) and obese (PROBESE: NCT02148692) patients. The exposure of interest was low etCO2 (<4.7 kPa) vs normal-high etCO2 (≥4.7 kPa). The primary outcome was postoperative pulmonary complications within 5 days. A time-weighted etCO2 analysis and propensity score matching were also performed to adjust for confounding. Results Of 2793 participants, 891 (29.4%; 52% female) had low etCO2, compared with 1972/2793 (70.6%; 65% female) participants with normal-high etCO2. Compared with participants with normal-high etCO2, higher minute volumes (normalised to body weight) were delivered in participants with low etCO2. Postoperative pulmonary complications developed in 278/821 (34%) participants with low etCO2, compared with 462/1972 (23%) participants who had normal-high etCO2 (adjusted hazard ratio, 1.3; 95% confidence interval, 1.1–1.6; P<0.001). The time-weighted analysis showed an inverse linear relationship between the mean etCO2 and postoperative pulmonary complications, which was also confirmed by propensity matching. Conclusions Low etCO2 occurs often during intraoperative ventilation and is associated with a higher rate of PPCs. The etCO2 level has an inverse dose-dependent relationship with postoperative pulmonary complications.