Perioperative cerebral oxygenation in patients undergoing aortic valve replacement
In cardiac surgery, the onset of cardiopulmonary bypass (CPB) is associated with a significant reduction in cerebral tissue oxygen saturation (SctO2),1,2 which may not always be fully reversible even after discontinuation of CPB.3 However, although intraoperative SctO2 monitoring with near-infrared reflectance spectroscopy (NIRS) has been used in adult cardiac surgery patients4 and a low SctO2 has been linked to neurocognitive dysfunction5 and other adverse outcomes,6,7 data on extended postoperative SctO2 values are scant to absent. Accordingly, we aimed to measure extended changes in SctO2 in adult patients undergoing single aortic valve replacement because the typical age of such patients, the nature of their condition and the features of the surgery and postoperative period may expose these patients to greater risk of prolonged cerebral oxygen desaturation.
From March 2013 to December 2014, we prospectively studied a convenience sample of adult (>18 years) patients in a single tertiary referral university hospital admitted for primary elective single aortic valve replacement. After induction of anaesthesia and before the start of CPB, we measured SctO2 by means of a NIRS monitoring device (Somanetics INVOS; Covidien, Boulder, Colorado, USA) with sensors placed on the right and/or left forehead. SctO2 was recorded continuously and average values were stored every 30 s until the patient was awake and extubated in the ICU. In the postoperative period, we recorded hourly mixed venous oxygen saturation (SvO2) and cardiac index (CI) derived from a pulmonary artery catheter (Edward Lifesciences Corp., Irvine, California, USA). The Austin Health Office for Research, Austin Hospital, Heidelberg, Victoria, Australia, granted permission (HREC/13/Austin/100) to use the data for formal analysis and publication on 27 August 2013, with a waiver for informed consent.
All analyses were performed using STATA version 11.2 (Stata Corp., College Station, Texas, USA). We averaged and analysed bilateral SctO2 readings over predefined time blocks (baseline levels 5 to 15 min before CPB start, during the first hour on CPB and hourly average after ICU admission up to the last 60 min before extubation). We defined significant decreases in SctO2 as a decrease of more than 20% from its baseline value for more than 30 s. Similarly, we defined significant absolute desaturation episodes as any absolute NIRS-derived SctO2 value less than 60% for more than 30 s. Continuous data are presented as medians [interquartile ranges (IQRs)]. Categorical data are reported as numbers (percentages). We used Spearman's rank correlation coefficient to assess correlations between SctO2 and SvO2, CI, lactate and haemoglobin (Hb) during the first hour in ICU. Changes over time for multiple measures of SctO2, CI, SvO2 and acid-base status were tested by repeated measures analysis of variance using time-blocks, including baseline, as the repeated measures variable. We used Bonferroni's correction for postestimation with pairwise comparison of margins. A two-sided P value less than 0.05 was considered significant.
We studied 17 patients [9 (53%) females]. Their median (IQR) age was 75 (66 to 79) years and the median APACHE III score was 53 (42 to 59). Duration of surgery and time on CPB were 237 (208 to 248) and 101 (86 to 136) min, respectively. The median duration of mechanical ventilation was 11 (7 to 16) h and the patients were discharged from ICU after a median of 42 (24 to 49) h. We analysed 21 588 NIRS measurements [1408 (860 to 1567) readings per patient] over a median (IQR) observation time of 6.7 (3.3 to 11.1) h.
Baseline median SctO2 before CPB was 74 (64 to 82)%. Within the first hour on CPB, there was a significant (P < 0.001) decrease in SctO2 to 65 (61 to 73)% for a median reduction of 10 (6 to 15)% despite a CPB pump blood flow maintained at 2.