Sex Differences in the Incidence of Postoperative Delirium after Cardiac Surgery: A Pooled Analyses of Clinical Trials

Postoperative delirium after cardiac surgery is associated with long-term cognitive deficits, prolonged hospitalization and institutionalization, and increased mortality. Although mechanisms to explain postoperative delirium are unclear, studies suggest that underlying patient vulnerabilities such as frailty, depression, and subclinical Alzheimer’s dementia contribute to delirium.1 In prospective observational studies of delirium after cardiac surgery, multiple studies2,3 reported a higher proportion of women with delirium. The data reported in these studies suggest that female subjects had a 1.36 (95% CI, 1.09 to 1.70) and 3.74 (1.42 to 9.82) times greater risk of developing postoperative delirium compared to their male counterparts, respectively.2,3 This finding contrasts with data from noncardiac surgeries,4 where men are frequently reported to be at a relatively higher risk of developing delirium. We note that studies included in reviews and meta-analyses of postoperative delirium after cardiac surgery were not designed to measure sex differences, and many do not report data to permit independent post hoc analyses. Thus, sex as a predisposing risk factor for postoperative delirium after cardiac surgery has been poorly understood.In this pooled-analyses study, we combined data from three published clinical trials (the DEXACET,5 Hyperoxia,6 and MINDDS7) across two institutions, each involving patients who underwent nonemergent cardiac surgeries. Patients were screened for postoperative delirium using the Confusion Assessment Method (for MINDDS7) or a combination of Confusion Assessment Method and its intensive care unit (ICU) equivalent (Confusion Assessment Method-ICU, for the DEXACET5 and Hyperoxia6 trials). In this report, postoperative delirium was defined as per the respective protocols of the trials included. Unadjusted relative risk, standardized risk ratio, and their corresponding 95% CIs were estimated by Wald’s methods of unconditional maximum likelihood and normal approximation. The Mantel-Haenszel method was used to estimate the pooled relative risk under the assumption of a fixed-effects model. Subcohort analyses were also performed to explore the possibility of confounding by treatment.From the three trials, the pooled cohort comprised 614 subjects, of which 141 (23%) were women. The median ages were 69.0 [interquartile range, 63 to 76] for the DEXACET, 70.5 [68 to 75.0] for the Hyperoxia, and 69.0 [64 to 74] for the MINDDS trials (table 1). We found no differences in baseline neurocognitive assessments between male and female subjects. Although the MINDDS trial involved two treatment arms (dexmedetomidine and placebo), the DEXACET trial used a factorial design to compare three treatments (dexmedetomidine, propofol, and acetaminophen) to placebo. The Hyperoxia trial, on the other hand, randomly assigned study participants to either a fraction of inspired oxygen (Fio2) of 100% (i.e., the hyperoxia arm) or Fio2 of at least 35% (the normoxia arm). Overall, the proportion of subjects in each treatment arm was not significantly different between the sexes for any of the included studies. Except for the participants of the MINDDS trial, who were assessed twice daily for delirium up to the first 3 days postoperatively, the remaining two trials involved once daily assessment for delirium until discharge. Regardless of the frequency and/or duration of delirium assessments, the average time to diagnosis was between the first and second postoperative days. In all three studies, the risk of developing delirium among female subjects was at least 1.54 times the risk among male subjects (pooled relative risk = 2.05 [95% CI, 1.56 to 2.68]; fig. 1, Supplemental Table 1, https://links.lww.com/ALN/D171). These risk associations were not confounded by the interventions administered during the included trials (Supplemental Table 2, https://links.lww.com/ALN/D172).Our study has two major limitations: (1) differences in the frequency of assessment of delirium between the trials, and (2) differences in the predefined diagnostic criteria for delirium, i.e., within the first three postoperative days (for the MINDDS7 trial), or until patients were discharged (for the DEXACET5 and Hyperoxia6 trials). Additionally, the included trials were performed in two institutions located in close geographical proximity, potentially resulting in a homogeneous sample that may limit the generalizability of our findings. Nonetheless, the randomized, prospective nature of the studies, the collectively larger sample size (n = 614), and the consistency in the increased risk of delirium among female subjects from all three trials lend credence to findings that demand formal, in-depth investigations. Taken together, our data suggest sex differences in the incidence of postoperative delirium after cardiac surgery in older patients. Additionally, our data indicate that women were a minority of participants in three prominent interventional trials of postoperative delirium prevention. Although men do represent the majority of patients undergoing cardiac surgery, adequate recruitment of both sexes into explanatory and interventional studies is needed to further our understanding of potential sex differences and to reduce the burden of postoperative delirium on patients, families, and the healthcare system.Funding for this work was supplied by the National Institutes of Health National Institute on Aging (Bethesda, Maryland; R01AG053582, to O. Akeju; R03 AG060179, to S. Shaefi), and by Mallinckrodt Pharmaceuticals (St. Louis, Missouri; #31864, to B. Subramanian). The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The DEXACET5 and Hyperoxia6 trials received institutional review board (IRB) approval from the Committee on Clinical Investigations at the Beth Israel Deaconess Medical Center (Boston, Massachusetts). The MINDDS7 trial was approved by the Mass General Brigham IRB (Boston, Massachusetts). Subjects in all three studies provided written informed consent. All trials were registered at ClinicalTrials.gov (NCT02546765, NCT02591589, and NCT02856594 for DEXACET,5 Hyperoxia6 and MINDDS7 trials, respectively).Ms. Mueller reports receiving funding from Roche Diagnostics (Indianapolis, Indiana) and the University of Chicago (Chicago, Illinois) for statistical consulting projects related to biomarkers in preeclampsia. Dr. Akeju is listed as an inventor on brain monitoring patents assigned to Massachusetts General Hospital (Boston, Massachusetts). The other authors declare no competing interests.Supplemental Table 1: Crude and Standardized Relative Risks (and 95% CI) from Studies Included, https://links.lww.com/ALN/D171Supplemental Table 2: Relative Risk Stratified by Treatment Groups, https://links.lww.com/ALN/D172