Perioperative Brain Health in the Older Adult: A Patient Safety Imperative

Abstract

While people 65 and older represent 16% of the population in the United States, they account for more than 40% of surgical procedures performed each year. Maintaining brain health after anesthesia and surgery is not only important to our patients but is an increasingly important patient safety imperative for the specialty of anesthesiology.

Aging is a complex process that diminishes the reserve of every organ system and often results in a patient that is vulnerable to the stress of surgery. The brain is no exception, and many older patients present with preoperative cognitive impairment that is undiagnosed. As we age, a number of changes occur in the human brain, resulting in a patient that is less resilient to perioperative stress, making older adults more susceptible to the phenotypic expression of perioperative neurocognitive disorders.

This review summarizes the current scientific and clinical understanding of perioperative neurocognitive disorders, and recommends patient-centered, age-focused interventions that can better mitigate risk, prevent harm, and improve outcomes for our patients. Finally, it discusses the emerging topic of sleep and cognitive health, and other future frontiers of scientific inquiry that might inform clinical best practices.

Introduction

Medical systems have grown not only in sophistication but also complexity, and it should not be surprising that harm, from the systems of care intended to improve health, continues to be pervasive. Indeed the World Health Organization has declared that “adverse events due to unsafe care is likely one of the 10 leading causes of death and disability in the world”.¹ The specialty of anesthesiology has been a leader in improving patient safety and we have continued to innovate and develop new frontiers in quality and safety.² In alignment with the progressive complexity of medical systems the specialty has fundamentally changed. Advances in intraoperative monitoring, modern anesthetics, and improved techniques for securing the airway have provided an opportunity for a new era, and our specialty has embraced a more holistic vision of care that is focused on perioperative medicine with embedded processes of care that focus on returning patients to long-term functional, psychosocial, and cognitive health.

In the United States someone turns 65 years old every 8 seconds.^3,4 It is estimated that the number of people 65 years and older will double to 95 million by 2060. While this group represents 16% of the population in the United States, it accounts for more than 40% of surgical procedures performed each year.⁵ In other words, older adults, a cohort of patients vulnerable to preventable harm, undergo a disproportionate percentage of surgical procedures.^6,7

Aging is a complex process that diminishes the reserve of every organ system, changes the pharmacokinetics and pharmacodynamics of anesthetics/medications, and produces a pro-inflammatory state, resulting in increased risk of morbidity and mortality associated with surgery and anesthesia.⁸ Of particular concern to our patients is the risk of postoperative neurocognitive disorders which are common postoperative complications in older adults. A fundamental goal of patient-centric, age-friendly healthcare systems should be to prevent harm to cognitive health across all settings of care. The focus of this review is the key patient safety topic of cognitive health of the older adult during the perioperative period.

From The Patient’s Perspective

To use the term “patient-centric” with integrity we must address the patient’s perspective at the outset. Anecdotes do not define a problem but do start to paint a picture of the perceptions of our patients. In the author’s experience patients do come in with real world perceptions of the effect of surgery and anesthesia on brain health.

Anesthesiologists should be prepared to have discussions with patients and their families regarding brain health concerns and address their perceptions. While many older adults with known cognition issues are quick to express their concerns, most patients aren’t sure which questions to ask regarding perioperative cognitive changes. Older patients should be empowered to ask questions regarding how best to prepare for surgery, what optimization strategies are available before and after surgery, and what role caregivers can have in their recovery. Resources have been created for patients that explain what cognitive changes may occur after surgery with educational guides on maintaining brain health after surgery.^9,10 Families are important team members and can provide extra help with prescribed authority to activate a Rapid Response Team when concerned about the condition of the patient. Prompt reporting of symptoms can help reduce the burden of perioperative neurocognitive disorders on overall health.

Perioperative Neurocognitive Disorders

The Nomenclature Consensus Working Group has suggested using the overarching term Perioperative Neurocognitive Disorders (PND) defined as “cognitive impairment or change identified in the preoperative or postoperative period”.¹¹ Perioperative neurocognitive disorders are further characterized in the following domains:

1. Preoperative cognitive impairment

2. Postoperative delirium

3. Delayed neurocognitive recovery - impairment within the first 30-days of surgery

4. Postoperative neurocognitive disorder - occurring between 30-days and 12-months after surgery.

Postoperative delirium (POD) is the most common surgical complication in older adults and is defined as an acute fluctuating state of confusion, inattention, and level of consciousness that occurs within the first seven days after surgery.¹¹ The reported incidence of POD in the literature has a high degree of variability that is dependent on the study methodology, patient population, and surgical procedure and is generally in the range of 5–65%.¹² Further, it is estimated that new or worsening long-term cognitive dysfunction occurs in more than 10% of non-cardiac surgical patients over the age of 60 years, and has been reported in up to 50% of patients, depending on existing comorbidities or type of surgery.^13,14

Perioperative neurocognitive disorders are associated with prolonged hospital stays, more days on mechanical ventilation, higher risk of falls, increased healthcare costs, and functional decline. Even after discharge, patients who develop PND are at an increased risk of institutionalization, death, and dementia. Patient health care costs associated with PND can reach $60,000 during the first year after discharge.^13,15 The annual national healthcare costs for delirium alone have been estimated at $150 billion.^13,15

Given this clear patient safety issue, anesthesiologists have an opportunity to lead in improving safety through perioperative best practices. Recognizing the necessity for quality improvement, the American Geriatrics Society and American College of Surgeons assembled an expert panel on POD to provide a set of evidence-based recommendations for the optimal care of older adults.^16,17 In 2015, the American Society of Anesthesiologists launched the Perioperative Brain Health Initiative and published the Best Practices for Postoperative Brain Health,¹⁸ promoting attention to this critical issue. Several other evidence-based guidelines have been issued throughout the world.^19,20 Despite this call to action, and the fact that simple strategies are projected to reduce the incidence of PND by 40%, these disorders remain a far too frequent event for older adults.^16,21

Age as a risk factor for PND

Perioperative neurocognitive disorders most often arise from acute surgical stress in at-risk patients.²² Older patients are at higher risk as age-associated changes take place across multiple physiologic systems and result in diminished function, multiple co-morbidities, and increased vulnerability. Structural and functional alterations that are often present in the aging brain have a spectrum of phenotypic expression; from cognitive “super-agers” who defy normal aging, to a “normal” decline in cognitive ability which allows for a relatively high function and quality of life, to mild cognitive impairment, and finally dementia. Normal aging has its greatest effect on the cognitive domains of processing speed, memory, attention, reasoning, and executive function.

With aging there are widespread structural changes that occur in the brain with atrophy being universal. Atrophy is most prominent in the prefrontal cortex and hippocampus which are responsible for complex thought processes and memory. Atrophy is best explained by a reduction in synaptic density, neuronal size, and synaptic branching.²³ As we age there are also reductions in neurogenesis, impairment of the glymphatic system, neurotransmitter imbalances, deterioration of the microcirculation, and dysfunction of the blood-brain barrier. Of particular note is “inflammaging” or a low-level, systemic, chronic inflammatory state that can contribute to diseases of aging and may make the older patient susceptible to an additional inflammatory hit that often occurs with surgery.^24,25 There is no unitary theory that fully explains PND but as we age, our brain is less resilient to perioperative stress, making us more susceptible to the phenotypic expression of PND.^26,27

The Preoperative Period

Perioperative neurocognitive disorders result from a complex interaction between the patient’s baseline vulnerability, and the physiological stress and inflammation associated with the perioperative period.^24,28,29 Perioperative triggers, combined with a decreased physiologic reserve that is often present in the older patient, make older patients more likely to develop PND.^26,27 The ability to preoperatively identify and optimize modifiable risk factors presents an opportunity to employ interventions that might decrease the incidence and severity of PND.

Risk Assessment

The framework for risk assessment includes a combination of predisposing factors and precipitating factors (see Figure 1). Predisposing factors include the patient’s baseline medical condition while precipitating factors include factors related to perioperative care and hospitalization. Table 1 identifies predisposing risk as either modifiable or non-modifiable.

Figure 1.

Risk Factors for Perioperative Neurocognitive Disorders. There are three types of factors to consider. 1. Predisposing Factors which encompass the baseline medical condition of the patient. 2. Surgery and Anesthesia Factors which are associated with the type of surgery and anesthetic factors. 3. Postoperative Factors include additional characteristics that occur during the course of hospitalization.

Table 1.

Risk factors for Perioperative Neurocognitive Disorders

Modifiable Risk Factors	Non-Modifiable Risk Factors
Sleep disturbances Frailty Polypharmacy Pain Decreased vision/hearing Poorly controlled diabetes Infection Neuropsychiatric disorders Poor nutrition	Age > 65  Preoperative cognitive impairment / AD  Other Neurologic Disorders Controlled diabetes Severe vascular disease Surgery

AD = Alzheimer’s Disease

Just as we preoperatively evaluate for cardiovascular risk, identifying risk factors for developing PND allows for implementation of risk-reducing strategies. While several risk factors for PND are non-modifiable (e.g., age) and should be part of the decision matrix for informed consent; other modifiable risk factors (e.g., frailty, polypharmacy, sleep, pain, decreased vision/hearing, diabetes, infection, neuropsychiatric conditions, and poor nutrition) can be optimized perioperatively (Table 1). It is unlikely that one isolated intervention will dramatically change the course of PND, but a comprehensive bundle of interventions based on modifiable risk factors for brain health might positively change the course of this disease.

Although preoperative neurocognitive disorders are relatively common in older adults, they are often undiagnosed and untreated.^30,31 It has been reported that up to 24% of patients over 65 years of age have probable cognitive impairment when screened preoperatively.^31,32 These patients are more likely to develop PND and have worse outcomes.^30–34 In 2012, the American College of Surgeons and the American Geriatrics Society jointly published guidelines recommending routine preoperative neurocognitive assessments in an effort to better determine the risk of PND.³⁰ Yet nearly a decade later, routine preoperative neurocognitive assessments have not been widely implemented.

Cognitive Screening Tools

Various neurocognitive screening tools have been developed to estimate risk and guide clinical decision.³⁵ Screening tools like the Mini-Cog©, the Mini-Mental State Examination (MMSE) or the Montreal Cognitive Assessment (MoCA), are fast and applicable to the preoperative setting.³⁶ For example, the Mini-Cog is a brief test (3–5 minutes) that requires no specialized personnel or technology, has minimal education/cultural/language bias and has shown to be predictive of adverse outcomes including POD.³¹ The MoCA and the MMSE 2^ndedition™ have different examination forms that decrease practice effects in serial administration and provide further information that can be used when developing perioperative care pathways. Other brief cognitive screening tests such as the clock-drawing test and the verbal fluency test are options for very simple and quick assessments.¹⁸ Table 2 outlines commonly used screening tools that apply to the preoperative setting.

Table 2.

Examples of preoperative neurocognitive screening tools

Test	Test time	Notes
Mini-Cog	5 min	Fast to administer. High sensitivity and specificity
MMSE	10 min	Widely used. Tests time and spatial orientation, language ability, instant memory, delayed recall, attention, calculation, visual spatial ability, and executive function. Low sensitivity and ceiling effect when this scale is used to detect impairment of a single cognitive domain and mild changes. Possible remote administration.
MMSE 2nd Edition	20 min	Alternative to MMSE. More sensitive to subcortical dementia and to changes associated with aging. No ceiling effect. Different forms that decrease practice effects in serial administration.
MoCA	15 min	High test/retest reliability. Tests cognitive function from visuospatial, executive function, naming, memory, attention, language, abstraction, and orientation. Different forms that decrease practice effects in serial administration. Possible remote administration.
NIH Toolbox- cognitive battery	31 min	Simple and easy to implement. The battery consists of tests to assess: executive function, attention, episodic memory, working memory, language and processing speed. Stability among different age groups, races, genders, and education level. Good reliability and validity. Possible remote administration.
Clock-drawing test	< 2 min	Very fast and easy to administer.
RUDAS	5 min	Designed to minimize effect of cultural learning and language diversity
Verbal Fluency Test	< 5 min	Very fast to administer. Phonetic test not as sensitive as semantic test. Can be influenced by education

MMSE = Mini-Mental State Examination; MoCA = Montreal Cognitive Assessment; RUDAS = Rowland Universal Dementia Assessment Scale; NIH = National Institutes of Health; min = minutes

More comprehensive neurocognitive assessments are sensitive and can track deterioration in several domains of cognition over time, however, they require a high degree of training, and a significant amount of time to perform which are barriers to integration into clinical practice.¹⁸ When selecting a test, one must also keep in mind that cognitive screening tools do have limitations, including education, cultural, and language bias.^37–39 The international study of post-operative cognitive dysfunction²⁸ and a recent systematic review article⁴⁰ exploring the methodology of measuring PND, demonstrate the need for broadly testing several memory domains and motor skills through baseline and subsequent timed postoperative neurocognitive assessments.⁴⁰

A recent consensus statement recommended a practical and pragmatic preoperative assessment, like a brief cognitive test (table 2), for all surgical patients over 65 years of age that helps triage limited resource interventions for the high-risk patient.¹⁸ The response to an abnormal preoperative cognitive test should be three-fold: 1) Referral for a comprehensive evaluation that would diagnosis and chronically manage a potential cognitive deficit, 2) The information should be shared with the patient and family as part of the informed consent and shared-decision making process, and 3) Resource-intense targeted interventions should be considered as part of a care pathway specific to this high-risk patient.

Frailty

Frailty, is commonly defined as the decline in physical and cognitive reserve leading to increased vulnerability to stressors, and is associated with adverse postoperative outcomes.⁴¹ The presence of frailty either with normal aging or in those with preexisting cognitive impairment, increases the risk for developing PND.^42–44 The presence of frailty can affect performance on neurocognitive assessment tests,^45,46 and frailty and cognitive screening were recommended by the American College of Surgeons and the American Geriatrics Society in their 2012 jointly published guidelines as a means to risk stratify patients preoperatively.³⁰

Several frailty tests have been developed and described in the literature with the Fried Phenotype (or some derivation) being the most used and studied.⁴⁷ Screening for frailty and cognitive impairment has become a patient safety priority recognized by the perioperative community as we come to understand that both predict adverse outcomes.⁴⁸ Given that clinical practices for the perioperative community vary greatly and the frailty assessment may not be implemented in all practices, the American Society of Anesthesiologists has curated a Frailty Toolkit to provide education and resources including various frailty assessment tools such as Frailty Index, Clinical Frailty Scale, and Risk Analysis Index.⁴⁹

Interventions

Ideally, the decision for surgery should be a “teachable moment” to support lifestyle changes that will improve durable cognitive health for patients. Although not normally in the domain of the pre-operative clinic, ideally patients would be educated on modifiable risk factors for which personal lifestyle changes would improve their long-term cognitive health. The extent to which these interventions would improve short term cognitive resilience to the stress of the perioperative period is speculative. Accordingly, this education intervention would take place in advance of the procedure, with enough time to allow for lifestyle interventions before surgery and continuous postoperative management strategies that involve the family, caregivers, and communities of each patient. This is a fundamental paradigm shift in the way that preoperative clinics currently function, expanding the vision beyond immediate intraoperative risk to a more holistic and integrated approach in the best interest of durable patient health. The guidelines from the World Health Organization regarding “Risk Reduction for Cognitive Decline and Dementia” provide an framework to begin this work.⁵⁰ They are summarized in Table 3.

Table 3.

A summary of guidelines for risk reduction of cognitive decline and dementia by the World Health Organization.⁵⁰ The recommendations most plausibly applicable to the preoperative period are listed.

	Strength of Recommendation
1. Physical activity	Strong
2. Tobacco cessation	Strong
3. Nutritional interventions	Strong
4. Weight management	Conditional
5. Management of hypertension	Conditional
6. Management of diabetes mellitus	Conditional
7. Management of dyslipidemia	Conditional
8. Cognitive interventions	Conditional
9. Interventions for alcohol use disorder	Conditional

Although there is limited supporting evidence, the World Health Organization’s framework presents key strategies for a prehabilitation program. Increasing resilience through cognitive training, improved nutrition status, reduction of disease burden, strength and mobility training have all been shown to improve postoperative outcomes, including reductions in postoperative complications, POD, and cognitive decline.^42,44 There is data regarding prehabilitation programs consisting of interventions targeted to physical health, nutrition, frailty, and anemia leading to a modest decrease in the incidence of delirium.⁵¹ Recent evidence offers preliminary support for the use of cognitive prehabilitation to reduce the incidence of POD.⁵² Incorporating the fundamentals of prehabilitation into comprehensive systems that occur across the continuum of care are even better approaches to mitigate risk (e.g., Hospital Elder Life Program) and will be discussed later in this review.

Avoiding the use of medications such as those listed in the American Geriatrics Society Beers Criteria® list is recommended.^53,54 These include benzodiazepines, anticholinergics, antihistamines, antipsychotics and phenothiazines. There are scenarios where the benefit of the above medications may outweigh the risks. e.g., the patient with severe anxiety preoperatively or refractory post-operative nausea and vomiting unresponsive to medications with less risk.

Identifying sleep disturbances, such as obstructive sleep apnea, present the opportunity for interventions such as continuous positive airway pressure therapy to improve the quality of sleep. A final recommendation is the involvement of a physician with expertise in geriatric medicine throughout the perioperative period.

Consent

Informed consent includes a discussion of planned procedures and their risk, benefits and alternatives. While there is some controversy over what is the standard for risk disclosure, it is clear that patients want to know common and serious risks.⁵⁵ For older surgical patients, POD is the most common complication after surgery. Several expert publications endorse disclosure of the possibility of POD.^17,18 However, a recent survey suggests that only about a quarter of anesthesiologists regularly provide information about the risks of delirium.⁵⁶ To assure that consent for surgery is properly informed, risk of PND should be discussed with older patients and their families.^57,58 There has been some controversy regarding whether the etiology of PND defines whether it should be discussed by the anesthesiologist or the surgeon. However, it remains that many complications such as mortality or myocardial infarction are multifactorial, and that each provider should consider these as part of their discussion.

The Intraoperative Period

At face value, intraoperative variables seem like rich targets for interventions to impact the course of PND. Although, a plausible hypothesis, recent data suggests that the type of anesthetic, per se, does not influence PND. A recent single-center cohort study exposed healthy adult volunteers (40 to 80 years old) with no underlying cognitive dysfunction to 2-hours of general anesthesia, without a surgical procedure.⁵⁹ The primary hypothesis of the study was that time to recovery of cognitive function after general anesthesia increases with age. They did not observe an association between age and recovery to baseline, suggesting that sedative/hypnotics alone do not appear to be associated with an adverse cognitive recovery in healthy adults.⁵⁹ It should be noted that the investigators studied patients with normal cognition without significant co-morbidities, and the application of these results to patient cohorts with cognitive impairment and frailty is limited.⁵⁹ Moreover, in two large multicenter randomized trials involving older adults undergoing orthopedic surgery, neuraxial anesthesia was not superior to general anesthesia with respect to the incidence of POD.^60,61 In the study by Neuman, et al⁶⁰ it should be noted that 13.7% of patients in the spinal anesthesia group and 11.7% of patients in the general anesthesia group had pre-existing dementia. The study by Li, et al⁶¹ included approximately 40% of patients in each group with pre-existing dementia.

Given the disruptive effect of anesthetics on the brain, some have hypothesized that an anesthetic overdose, might contribute to PND. Specifically, that the use of a processed electroencephalogram (EEG) to minimize the anesthetic dose during surgery might be associated with a lower rate of postoperative neurocognitive decline. Sieber et al,⁶² randomized patients for hip surgery to receive a spinal anesthetic and either light or deep propofol sedation. They observed that use of light sedation decreased the incidence of post-operative delirium from 40% to 19%. A meta-analysis published in 2018,⁶³ concluded that the use of EEG guided anesthetic depth was associated with a decrease in POD. This was followed by two large randomized clinical trials.^64,65 In the ENGAGES trial, EEG-guided hypnotic administration, compared with usual care, did not decrease the incidence of POD. However, subsequently, Evered et al, demonstrated that titrating hypnotic dose to the EEG did indeed reduce the incidence of POD and improved cognitive function at one year. The prevailing hypothesis may be that there is a subset of cognitively frail patients, with preexisting cognitive impairment, who are uniquely susceptible to the adverse effects of a relative hypnotic overdose and might benefit from intraoperative EEG monitoring.⁶⁶ The use of intraoperative EEG in conjunction with other strategies, such as age-adjusted minimum alveolar concentration are part of the Best Practices for Postoperative Brain Health.¹⁸

Several studies have demonstrated an association between intraoperative hypotension and an increased incidence of POD.^67–70 There is limited data to support a causal relationship between intraoperative hypotension and POD.⁷¹ Maintenance of cerebral perfusion is of utmost importance in vulnerable populations.⁷² However, the lower limit of cerebral autoregulation varies widely between patients, and absent measuring brain perfusion, it is unknown what the ideal mean arterial pressure is in an individual patient.⁷² Moreover, many older adults have chronic hypertension which is known to shift the autoregulation curve to the right. The limited data cited above support a recommendation to keep the mean arterial pressure at a minimum of 65 mmHg, although a higher number may be necessary in some patients.

Other recommendations include monitoring of glucose levels and maintenance of normothermia.¹⁸ Tight glucose control has been linked to hypoglycemia leading to worse outcomes and an increased rate of POD and is no longer recommended.^73,74 Avoiding even mild hypoglicemia⁷⁴ may be equally as important as avoiding hyperglycemia⁷⁵ when it comes to preventing PND.

The Postoperative Period

Anesthesiologists often only briefly follow-up with their patients in the immediate postoperative period.¹⁸ These conversations mainly focus on pain and antiemetic strategies. New strategies must be adopted in our approach to cognitive recovery, one which engages multimodal tactics, including the return of visual and hearing aids, early mobilization, and resumption of a normal diet. Polysomnographic studies have revealed extreme sleep disruption during hospital stays.^76–79 Maintenance of sleep friendly hospital environments (noise reduction, appropriate light), and a decrease of iatrogenic factors such as frequent care-related interruptions was shown to improve outcomes and decrease the incidence of delirium. This is particularly important for the older patient for whom the restorative properties of natural sleep are another key part of their recovery. Importantly, family engagement and social support should be implemented early in the preoperative period and deployed in tandem with the involvement of an interdisciplinary health care team and incorporated in age friendly comprehensive systems of care.

Comprehensive systems of care that implement targeted, structured, multicomponent nonpharmacological approaches are even better approaches to mitigate risk. The Hospital Elder Life Program (HELP™) is an evidence-based intervention that supports this approach.⁸⁰ One example of the HELP program instituted three universal protocols (orientation, cognitive stimulation and mobilization) as well as targeted protocols dependent on specific risk factors for each patient. In patients over 70 years of age, POD decreased from 19.4% in the control group to 2.6% in the intervention group. Moreover, physical and cognitive function was improved in the intervention group as compared to the control group at 30 days after discharge.⁸¹ In a recent study, a novel non-pharmacological, multidisciplinary prevention program was applied to orthopedic, general, or cardiac surgery older patients. Risk factors for delirium and symptoms of delirium were assessed daily with interventions targeted to individual patient needs. Interventions included cognitive, motor, and sensory stimulation, meal companionship, stress relaxation and sleep promotion. Delirium incidence and duration were reduced in the intervention group.⁸²

To paraphrase an old adage “if it isn’t measured it can’t be improved” i.e., if you don’t assess the cognitive state of the patient, the physician has little hope of improving cognitive recovery after anesthesia. For POD, an accepted standard is the Confusion Assessment Method which is simple to use and requires only a few minutes to administer. Often, the first signs of delayed neurocognitive recovery and postoperative neurocognitive disorder are generally clinical signs such as memory impairment and inability to concentrate that may trigger a more sophisticated level of neuropsychometric testing.

A suggested care pathway is outlined in Figure 2. It should be noted that this represents the authors synthesis of current evidence and should be considered a starting point for clinicians to modify to local needs.

Figure 2.

A suggested template to consider when developing a perioperative care pathway for the older adult at risk for Perioperative Neurocognitive Disorders.

Future Directions and Emerging Issues

Much work remains to be accomplished to mature the science that inform best practices regarding PND. Emerging issues that we will highlight include the use of new technologies to define risk and provide decision support, the interaction of sleep and cognition, the cognitive super-ager, and the surgical burden of patients with mild cognitive impairment/Alzheimer’s disease (AD).

While current forms of cognitive screening prior to surgical care are useful for perioperative interventions, new technologies, such as natural language processing and automated speech analysis^83–85 as well as gait evaluation technology,⁸⁶ offer low-cost and pragmatic approaches to risk stratification. One of the advantages of speech analysis is that, by using voice-recognition software, risk prediction can be done virtually.

Current pragmatic and clinically-implementable assessments have moderate accuracy for risk predicition,^87,88 and by-in-large utilize hierarchical logistic regression models.⁸⁹ One technology with promise to advance the identification and treatment of PND is artificial intelligence (AI) and more specifically machine learning (ML). Machine learning is a subset of AI which uses big data and algorithms that imitate the human brain in a way to process large amounts of data to address complex problems with specific applications to define risk, and provide diagnostic and treatment solutions in an efficient manner. Recently, POD prediction models that use ML and AI technologies have been published.^90–92 One such model⁹² used 115 predictive features from electronic health record data with encouraging results as defined by the area under the receiver operating characteristic curve.

Cognitive testing and web-based products that improve long term cognitive health through coaching experiences for the population at large are in development.⁹³ Specific to the perioperative period tools have been developed with the goal of delivering end-to-end digital health platforms that are accessible to the perioperative care team. These tools^94–97 introduce in-depth neuropsychological tests in a time efficient manner with minimal barriers to implementation and are reported to be eligible for reimbursement by payors.

Sleep and Cognitive Health

Sleep and its coupling to circadian rhythms are integral to brain health (e.g., memory consolidation, metabolic waste clearance) and other systemic functions.^98–101 During normal aging, the duration of total sleep and slow-wave sleep is decreased and is associated with an increase in sleep fragmentation and difficulty of falling asleep.¹⁰² The prevalence of sleep disturbances at home, can affect up to 40% of older adults and is associated with impairment of spatial memory, verbal fluency, attention, processing speed, and executive function.^101,103 The presence of sleep disturbance before hospital admission is an independent risk factor for PND. Environmental and patient factors in the postoperative hospital setting (pain, light, environmental changes, monitoring systems, constant interruption) can contribute to the impairment of waste clearance, especially for older patients and could contribute to PND.⁷⁸ Thus, the prevalence of sleep disturbance in the older patient coupled with surgery and hospitalization may result in a nexus of precipitating factors in a patient vulnerable to PND.

The glial-lymphatic (glymphatic) pathway (see Figure 3), a recently theorized cerebral system, facilitates the homeostasis of various brain functions.^104–106 Like the peripheral lymphatic system, it clears and empties protein waste products, and is most active during sleep. Further investigation is needed to determine whether perioperative sleep disturbances alter the glymphatic system and if perioperative medications/anesthetics aid or impede glymphatic function.¹⁰⁵

Figure 3.

Three-dimensional representation of lymphatic transport and drainage concept highlighting the periarterial and the perivenous space, and the astrocytic end-feet with aquaporin 4 (AQP4) water channels and forming a sheath around the blood vessels. Cerebrospinal fluid is driven by convection through the periarterial space and is propelled across the astroglia end-feet to mix with interstitial fluid and waste products. From there, the waste and excess fluids are driven towards the perivenous space to ultimately be directed towards the lymphatic vessels and general circulation for breakdown and clearance. The black particles represent “waste” particles in the interstitial fluid (e.g., amyloid beta). SubA = subarachnoid space; Oli = oligodendrocyte. Reproduced with permission.¹⁰⁶

Anesthesia and surgery alter the circadian rhythm of sleep and delay secretion of melatonin that helps regulate sleep.¹⁰⁷ Anesthetic and sedative agents have variable action targets. Those which act by modulating the GABA_A receptor converge at the level of the hypothalamus while α2 adrenergic agonists, such as dexmedetomidine, converge on sleep pathways within the brainstem, thus behaving more like natural sleep.¹⁰⁸ Several clinical studies reported that the use of dexmedetomidine for sedation in the intensive care unit is associated with reductions in delirium, and improved outcomes when compared to benzodiazepines and/or opioid-based sedation strategies.^109–112 Brain waste clearance via the glymphatic system is more efficient during slow-wave sleep,¹¹³ which can be mimicked with drugs blocking central adrenergic transmission. Although highly speculative, one hypothesis would suggest that dexmedetomidine’s delirium-preventive effect is related to superior brain waste clearance during comfort sleep when compared to other hypnotic regimens in the intensive care unit setting.^105,114 Preclinical studies demonstrated that dexmedetomidine, by blocking norepinephrine release from the locus coeruleus, enhanced glymphatic system transport by 30% relative to other agents.¹¹⁴ Future studies are needed to further elucidate the benefits of dexmedetomidine, among other agents, as good hypnotic/sedative candidates that mirror the restorative functions of sleep.^78,115–118

Much can be learned from both ends of a wide spectrum of cognitive function of older patients to identify the complex interplay of factors when older adults present for surgery. On one end, recent studies have shown that there is a subset of patients, the “cognitive super-agers”, who defy the typical age-related decline in cognitive function. Do these patients have enhanced resilience to the stresses of the perioperative period? If so, are there variables that define resilience in these patients that may provide a foundation for interventions in the cohort of patients that are vulnerable to PND? Individuals who have reached extreme age with preserved cognitive health may present an opportunity to investigate factors that promote optimal cognitive function after surgery.¹¹⁹

At the other end of the spectrum, is AD. According to the Alzheimer’s Association over six million Americans live with AD, and that number is projected to grow to 12.7 million people by the year 2050.¹²⁰ Such numbers are daunting and present a number of perioperative challenges.¹²¹ Even more concerning is a likely large number of patients with undiagnosed mild cognitive impairment on the precipice of a steep decline in cognition. Does surgery and anesthesia accelerate cognitive deterioration in this vulnerable cohort of patients? Unfortunately, the delta between what we know, and a mature set of knowledge is substantial.

Population studies have demonstrated a bidirectional association between AD and PND.^122–125 Preexisting cognitive impairment is associated with PND,³⁴ and in patients with a high degree of AD-related cortical atrophy, POD is associated with an accelerated rate of long-term cognitive decline.^123–126 How perioperative management affects the acceleration of cognitive decline is unknown.¹²⁷ It has been shown that the apolipoprotein E4 gene was associated with an increased risk for early POD after controlling for known demographic and clinical risk factors.¹²⁸ Nonetheless, the evidence of the effect of surgery and anesthesia on amyloid deposition is conflicting. Sprung et al, did not find evidence of amyloid deposition after exposure to surgery and anesthesia.¹²⁹ Klinger et al found that amyloid deposition at six weeks in a surgical cohort was not different from that in AD patients not having surgery, but increased after one year at a rate greater than controls.¹³⁰ Avidan et al found that the rate of cognitive decline in patients with pre-existing dementia was not affected by the occurrence of a surgical event.¹³¹ A recent cohort study also provided some reassurance that the neurocognitive changes identified postoperatively are unlikely to accelerate amyloid or tau related neuropathological processes associated with AD.¹³²

Future studies on these two ends of the patient spectrum (cognitive super-ager and AD) could provide important information on the resilient brain, presenting an opportunity to prioritize perioperative interventions and improve the quality and quantity of life in older patient populations.

Summary

A half a century ago, surgery in older adults was confined to “unequivocally necessary cases”. Today, advances in surgery and anesthesia offer many options meant to improve safety, function, and quality of life. Though not proven causative, it appears that perioperative stresses can jeopardize brain health in vulnerable patients such as the older adult. The associated risks should be evaluated and reviewed within the patient/family-physician partnership in order to develop a perioperative care pathway for those most likely to suffer from PND.

Anesthesiologists, can directly impact this troublesome and growing issue, by coordinating perioperative care, partnering with surgical and other perioperative teams to integrate treatments into a postoperative discharge plan that extends beyond the hospitalization period.¹³³ We must make every effort to safeguard postoperative brain recovery for all our patients.

Glossary of Terms

PND

Perioperative Neurocognitive Disorders

POD

postoperative delirium

MMSE

Mini-Mental State Examination

MoCA

Montreal Cognitive Assessment

EEG

electroencephalogram

HELP

Hospital Elder Life Program

AD

Alzheimer’s disease

AI

artificial intelligence

ML

machine learning