General Anesthesia and Postoperative Neurocognitive Outcomes

Delirium is a common and serious postoperative complication that manifests as an acute, fluctuating failure of the brain to support normal arousal, attention, and organized thinking. Postoperative delirium has been associated with delayed recovery from surgery and persistent neurocognitive disorders, as well as with other adverse outcomes. There is a longstanding and widely held concern that general anesthetic and sedative agents are implicated in postoperative delirium and other neurocognitive disorders. Conclusions regarding the neurotoxicity of sedative-hypnotic anesthetic agents have been derived largely from basic science studies demonstrating neuropathology, neurocognitive impairment, or behavioral changes after general anesthesia in animals¹ as well as observational studies of anesthetized patients undergoing surgery.^2,3 However, the translational and causal relevance of these studies is best examined by a randomized controlled trial methodology in which surgical patients presenting for the same operation receive either general anesthesia or a regional technique in which no sedative-hypnotic drugs are administered.

In this issue of JAMA, Li and colleagues present the findings of such a trial titled, “Effect of regional versus general anesthesia on incidence of postoperative delirium in older patients undergoing hip fracture surgery: The RAGA randomized trial.”⁴

The investigators randomized 950 surgical patients 65 years or older to receive either general anesthesia (with inhaled or intravenous drugs) or awake regional anesthesia (spinal, epidural, or combined techniques) without any sedation. Patients across nine university hospitals who presented for surgical correction of traumatic hip fractures were included in the study. The primary outcome of the trial was delirium detected by the confusion assessment method in the first week after surgery. The percentage of patients with delirium was 5.1% (24/470) in the general anesthesia group and 6.2% (29/471) in the regional anesthesia group (difference = 1.1%, 95% confidence interval −1.7% to 3.8%). There was also no significant difference between groups in the percentage of patients with delirium symptoms (syndromal plus subsyndromal delirium) in the first postoperative week, with 8.7% (41/470) in the general anesthesia group and 8.9% (42/472) in the regional anesthesia group. These results are consistent with the conclusion that avoidance of sedative-hypnotic medications does not prevent postoperative delirium. Importantly, various secondary outcomes related to delirium provided corroborative support: the total number of postoperative adverse events, hospital length of stay, and 30-day mortality were not significantly different between the 2 groups.

There are several important caveats. First, there was a much lower percentage of patients with delirium than in other trials that have evaluated older adults after surgery, suggesting that delirium might have been missed in some patients. Second, 15% of those patients with postoperative delirium (8/53) already had preoperative delirium, and should potentially have been excluded from the trial. Third, the majority of randomized patients (~63%) were from a single medical center, which brings into question the generalizability of the findings. Fourth, all the patients were enrolled in China, and the results might not apply to other populations.

These data further challenge the common narrative that general anesthetics play a causal role in the etiology of postoperative delirium and adverse long-term neurocognitive outcomes. However, RAGA’s conclusions are actually consistent with recent studies, which have provided evidence that: (1) compared with general anesthesia, regional anesthesia in vulnerable older adults does not decrease postoperative delirium incidence;⁵ (2) the healthy human brain is resilient to even deep anesthesia;⁶ (3) decreasing general anesthetic concentration during surgery does not prevent delirium;⁷ and (4) general anesthesia for coronary artery bypass grafting leads to cognitive outcomes that are no different from percutaneous coronary intervention, with minimal or no sedation.⁸

The recent REGAIN trial (Regional vs General Anesthesia for Promoting Independence after Hip Fracture) answers the pragmatic question: does it matter to patient-centered outcomes which of the two common real-world anesthetics techniques is used for hip fracture repair? In this clinical trial that involved 1600 patients (mean age 78) who were undergoing surgery for hip fracture, there was no significant difference between spinal anesthesia compared with general anesthesia with respect to the primary composite outcome (death or inability to walk independently at 60 days), which occurred in 18.5% of patients in the spinal anesthesia group and 18.0% in the general anesthesia group, and there was no significant difference in the incidence of postoperative delirium, which occurred in 20.5% of patients in the spinal anesthesia group vs 19.7% in the general anesthesia group. Importantly both techniques included some use of anesthetic or sedative agents, albeit with much less exposure in the regional anesthetic group. Thus, the REGAIN trial does not exclude the possibility that complete avoidance of all sedative agents is needed to decrease postoperative delirium incidence.

The RAGA trial builds on the findings of the REGAIN trial, and by assiduously avoiding any sedation in the regional anesthetic arm, provides important scientific insights regarding the actions of anesthetic agents. Anesthetic- and sedative-induced disruptions to brain networks resulting from relatively brief administrations in the setting of a surgical procedure are themselves, apparently, too transient to provoke delirium.

Work conducted in healthy human volunteers⁶ for the ReCCognition (Reconstructing Consciousness and Cognition) study is consistent with this paradigm-challenging conclusion of the RAGA Delirium trial. In this multicenter experiment with 60 healthy volunteers, 30 participants were randomly assigned to receive propofol induction followed by deep isoflurane anesthesia—but without surgery—for three hours, followed by three hours of cognitive testing, and another three days of assessing sleep-wake activity.⁶ To control for learning confounds and circadian effects, 30 non-anesthetized volunteers were randomized to a similar cognitive testing protocol, but stayed awake throughout the experiment. Within three hours of emerging, the participants who had been anesthetized were able to perform cognitive tests with approximately the same accuracy and speed as those who had not been anesthetized. Furthermore, the sleep-wake patterns of those exposed to deep anesthesia were virtually indistinguishable from the non-anesthetized volunteers in the three days to follow, suggesting that there was no delayed perturbation to brain networks.

While the RAGA trial suggests that general anesthesia is not itself implicated in causing delirium, perhaps excessively deep anesthesia is harmful. This follows from observational research finding that the duration of electroencephalogram suppression, a biomarker of supratherapeutic anesthesia, was predictive of postoperative delirium.⁹ Thus, it would stand to reason that decreasing anesthetic depth during surgery might decrease postoperative delirium incidence. In the ENGAGES trial (Electroencephalography Guidance of Anesthesia to Alleviate Geriatric Syndromes), 1232 patients older than 60 undergoing major surgical procedures were randomized to receive electroencephalogram-guided anesthesia or usual care.⁷ In the guided group, the goal was to minimize epochs of electroencephalogram suppression and to decrease volatile anesthetic administration to prevent postoperative delirium. The anesthetic management goals were accomplished but there was no accompanying reduction in delirium incidence. It is important to sound a note of caution that it is still unclear if periods of deeper general anesthesia cause delirium, with randomized trials on this question having discordant findings.^7,10–12

The RAGA trial does not address intermediate or long-term cognitive outcomes, although 1-year neurocognitive measurements are planned. This is important to consider because influential studies have reported that patients who have had surgery with general anesthesia experience long-term neurocognitive disorders,(2,3) but generally, these studies have not included a comparison with appropriately matched controls who did not undergo a general anesthetic. In order to address this, a recent retrospective study, using regression modeling and inverse probability of treatment weighting, compared cognitive outcomes between 665 adults older than 65 who underwent coronary artery bypass grafting (which requires general anesthesia) and 1015 who had percutaneous coronary intervention, a strategy for coronary revascularization which avoids major surgery and general anesthetics.⁸ Although memory decline and dementia occurred commonly, there was no significant difference in the neurocognitive outcomes between groups. Thus, inferring that sedative-hypnotic medications set in motion a cascade of events causing long term neurocognitive disorders in older adults is likely a misattribution, with the more credible culprit being underlying comorbidities such as cardiovascular disease. This conclusion reinforces the findings of small randomized trials that have also shown no significant differences in neurocognitive outcomes following coronary artery bypass grafting or percutaneous coronary intervention.^13,14 The RAGA trial will provide additional valuable data on the role of general anesthesia and sedative-hypnotics on postoperative neurocognitive disorders when the one-year assessments are completed.

An expanding body of evidence suggests that surgery with general anesthesia is not implicated in persistent neurocognitive disorders at a population level,¹⁵ with the qualifier that there might still be individuals or patient sub-groups with specific vulnerability. The RAGA trial now adds an important piece to the puzzle by providing reassurance that general anesthesia or sedation may reasonably be chosen by patients and clinicians for hip fracture repair, and possibly other invasive procedures, without incurring an increased risk of early postoperative delirium.