Incidence of Post-operative Delirium with Use of Benzodiazepines in Adult Cardiac Surgery: A Comparative Study of On-Pump and Off-Pump Procedures

INTRODUCTION

Post-operative delirium (POD) is characterized by acute disturbances in attention, awareness, and cognition. It remains one of the most clinically significant neurologic complications following cardiac surgery, with global incidence rates between 10% and 30%.^[1-5] Its occurrence is associated with prolonged hospitalization, increased morbidity, greater institutional care requirements, and long-term cognitive decline.^[6-9]

Cardiac surgical patients are uniquely susceptible due to aging physiology, comorbidities, and the neuroinflammatory and embolic burden associated with cardiopulmonary bypass (CPB).^[10-13] While off-pump coronary artery bypass grafting (OPCAB) was developed partly to mitigate such risks, comparative data on POD between on- and off-pump surgery remain inconsistent.^[14-16]

Pharmacologic factors, particularly benzodiazepines, have been debated extensively. Although prolonged post-operative benzodiazepine sedation increases POD risk,^[17-20] the role of a single midazolam induction dose is less clear. Current evidence suggests that short-duration exposure does not reach the neurochemical thresholds required to trigger delirium through GABAergic dysregulation.^[1,2,21,22]

Given these uncertainties, we conducted a prospective comparison of POD incidence between on-pump and off-pump surgery, standardizing midazolam exposure to clarify the relative influence of CPB versus limited benzodiazepine use.

MATERIAL AND METHODS

This prospective observational cohort included 100 adults aged 18–75 undergoing elective CABG or valve surgery. Patients were divided equally into on-pump and off-pump groups. Induction was standardized with midazolam 0.05 mg/kg IV, fentanyl, and vecuronium. Maintenance consisted of volatile agents or propofol infusion. No postoperative benzodiazepines were administered. POD was assessed using confusion assessment method-intensive care unit at 6, 12, 24, and 48 h postoperatively, a validated tool for delirium detection in critically ill patients.^[23] Demographic, surgical, and perioperative data were recorded. POD incidence was compared using Chi-square analysis, and relative risk (RR) and odds ratio (OR) were calculated.

Exclusion criteria included pre-existing dementia or psychiatric illness, but we acknowledge that the absence of formal pre-operative cognitive testing, for example, mini-mental state examination (MMSE) and Montreal cognitive assessment (MoCA) is a methodological limitation given the known influence of subtle cognitive impairment on POD risk.^[6,24-26]

RESULTS

POD occurred in 16% of on-pump patients versus 12% in off-pump patients (RR 1.33; OR 1.39; P = 0.57). The difference, although numerically present, lacks clinical significance given the narrow margin and overlapping confidence intervals.

Baseline demographic characteristics were comparable. No perioperative variable independently predicted POD in this cohort.

DISCUSSION

The principal finding of this study is the absence of a significant difference in POD incidence between on-pump and off-pump patients when midazolam exposure was standardized. This highlights the multifactorial nature of POD, where inflammatory responses, cerebral perfusion, embolic load, sleep disturbance, metabolic fluctuations, and patient vulnerabilities interplay in complex ways.^{[3,7,21,22,27]}

Short-term midazolam use did not appear to increase POD risk. Mechanistically, benzodiazepines contribute to delirium through GABAergic overactivation, impaired cortical connectivity, and neurotransmitter imbalance, but these effects are dose- and duration-dependent.^[18,21,27] A single induction dose likely does not produce sustained neurochemical disruption sufficient to precipitate delirium^[1,2].

Although CPB can enhance inflammatory cytokine release and microembolic load,^[28] existing literature indicates that OPCAB does not consistently reduce POD or long-term cognitive dysfunction compared with on-pump surgery.^[14-16] These findings reinforce that POD risk is less dependent on isolated intraoperative techniques and more related to cumulative perioperative stress and cognitive reserve.^[29-31]

Several confounders were not captured in this study, including anesthetic depth, cerebral oximetry trends, transfusion thresholds, glycemic variability, temperature management, and CPB circuit characteristics. These are well-recognized contributors to postoperative neurocognitive dysfunction and are targets of multicomponent delirium-prevention strategies.^[32-38]

CONCLUSION

POD incidence was comparable in on-pump and off-pump cardiac surgery patients who received uniform midazolam induction. A single dose of midazolam did not significantly influence delirium risk. These findings support the concept that POD is driven by multifactorial interactions rather than isolated perioperative exposures. Larger, multicenter studies integrating objective cognitive assessments and granular intraoperative metrics are required.