Predictors of postoperative visual and auditory hallucinations in hypertensive vs. non-hypertensive cardiac surgery patients: insights from the prospective multicenter VAACS study.

backgroundHallucinations are common yet under-recognized after cardiac surgery and contribute to postoperative morbidity. Hypertension (HTN), the most prevalent comorbidity in surgical candidates, may alter susceptibility through impaired cerebrovascular autoregulation. Whether HTN modifies predictors of postoperative visual and auditory hallucinations remains unclear.

methodsIn the prospective, multicenter VAACS study, 1,332 adults undergoing CABG or valve surgery were stratified by hypertension (HTN, n = 517; non-HTN, n = 815). Visual and auditory hallucinations within 7 days were assessed with the Questionnaire for Psychotic Experiences. Multivariable Cox regression was performed separately in the HTN and non-HTN cohorts.

resultsVisual hallucinations occurred in 12.0% of HTN and 11.3% of non-HTN patients; auditory hallucinations in 6.2% and 7.4%, respectively. In HTN patients, visual hallucinations were independently predicted by higher EuroSCORE II (adjusted HR, 1.041; 95% CI, 1.021-1.061; p < 0.001) and higher neutrophil counts (adjusted HR, 1.018; 95% CI, 1.008-1.028; p < 0.001), while longer postoperative ventilation was associated with a reduced risk (adjusted HR, 0.861; 95% CI, 0.780-0.950; p = 0.003). For auditory hallucinations in HTN patients, younger age (adjusted HR, 0.965; 95% CI, 0.932-0.998; p = 0.039), prior antibiotic use (adjusted HR, 2.464; 95% CI, 1.550-3.916; p < 0.001), higher neutrophil counts (adjusted HR, 1.016; 95% CI, 1.005-1.027; p = 0.004), and shorter ventilation duration (adjusted HR, 0.791; 95% CI, 0.709-0.883; p < 0.001) were independently associated with risk. In non-HTN patients, EuroSCORE II was independently associated with visual hallucinations (adjusted HR, 1.051; 95% CI, 1.043-1.060; p < 0.001). For auditory hallucinations in non-HTN patients, a higher EuroSCORE II increased the risk (adjusted HR 1.101, 95% CI 1.086-1.117, p < 0.001), while valve surgery versus CABG was protective (adjusted HR 0.463, 95% CI 0.221-0.970, p = 0.041). ROC analysis of a simple model (EuroSCORE II, ventilation, neutrophils) showed better discrimination in HTN (AUC 0.68) than in non-HTN (AUC 0.54; p ≈ 0.04-0.045).

conclusionsPredictors of postoperative hallucinations differ by HTN status. In HTN patients, inflammatory burden, age-related vulnerability, antibiotic exposure, and ventilatory course were the dominant determinants of risk. In non-HTN patients, overall operative risk as reflected by EuroSCORE II and surgery type played a more prominent role. CLINICAL TRIAL NUMBER: Not applicable.