1 INTRODUCTION

Postoperative neurological disorders have been attracting increasing attention in the world, with a vast amount of research conducted on these poorly understood disorders. Postoperative neurological disorders contain postoperative delirium (POD), postoperative cognitive dysfunction (POCD), postoperative covert ischemic stroke, and hemorrhagic stroke, causing cognition decline and poor long-term functional outcome in the elderly. Postoperative neurological disorders increase mortality and cause substantial financial burden on family and society. With a rapid increase in the number of elderly patients undergoing elective surgical procedures, postoperative neurological disorders need more concern and further investigation. This review aims to describe general features and the latest evidence-based knowledge of postoperative neurological disorders.

2 POSTOPERATIVE NEUROLOGICAL DISORDERS

Postoperative neurological disorders include neurological complications such as delirium, cognitive dysfunction, acute cerebral ischemic stroke, and hemorrhagic stroke that occur after surgery, especially in the elderly. As more elderly patients undergo surgery, the incidence of postoperative neurological disorders is rapidly increased. Although their exact etiology and pathogenesis remain elusive, several risk factors have been recognized.

2.1 POD

Postoperative delirium is defined as acute emergence of confusion, disorientation, perceptual disturbances, emotional dysregulation, or sleep disturbances, manifesting within a certain period of time. The prevalence of delirium ranged from 5% to 50% when assessed with Confusion Assessment Method for the Intensive Care Unit (CAM-ICU).1 POD can occur soon after general anesthesia and operation, for example, in the post-anesthesia care unit (PACU). In many cases, POD is frequently linked to anesthesia.2 Compared to dementia which chronically deteriorates brain function, POD is usually acute, transient, and presenting common causative factors.3 It contributes to prolonged hospitalization, increased mortality rate, and reduced long-term quality of life, which adds an additional burden to patients and families.

Besides DSM-5 listed in Table 1 and ICD-10 diagnostic criteria which supplements a disturbance in sleep-awake cycle including insomnia, reverse of sleep-awake cycle, various assessment tools have been developed to recognize POD.4 The CAM-ICU is a screening tool which consists of the assessment of four characteristics: 1) acute onset and fluctuating course of mental state, 2) inattention, 3) disorganized thinking, and 4) altered level of consciousness.5 Delirium is diagnosed when both characteristics 1) and 2) are satisfied with 3) or 4) electively satisfied.6

TABLE 1. DSM-5 Diagnostic criteria for delirium DSM−5 Diagnostic criteria for delirium A disturbance in attention (reduced ability to direct, focus, sustain, and shift attention) and awareness (reduced orientation to the environment). A disturbance that develops over a short period (usually hours to a few days) represents a change from baseline attention and awareness and fluctuates in severity during the day. An additional disturbance in cognition (memory deficit, disorientation, language. visuospatial ability, or perception). The disturbances in Criteria 1 and 3 are not better explained by another pre-existing established or evolving neurocognitive disorder and do not occur in a severely reduced level of arousal, such as coma. There is evidence from the history, physical examination, or laboratory findings that the disturbance is a direct physiological consequence of another medical condition, substance intoxication or withdrawal (due to a drug of abuse or to a medication), or exposure to a toxin, or is due to multiple etiologies. Abbreviation: DSM-5, the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition 2.2 POCD

Postoperative cognitive dysfunction, also defined as postoperative neurocognitive disorder (pNCD),7, 8 is characterized by cognitive decline persisting for more than 30 days but less than 12 months following surgery. Unlike POD (Table 2), consciousness, orientation, and attention are not obviously affected in POCD.9 However, patients can still manifest impairment in memory, perceptual function, and language.9, 10 The incidence increases among the elderly, especially those over 60 years old.11 For elderly patients, cognitive decline may result in prolonged hospitalization, reduced quality of life, even increased mortality, which has been neglected in the assessment of patient's prognosis, especially for those undergoing general anesthesia and surgery.12 The diagnosis criteria for POCD are more complex than for POD, as POCD requires a subjective impression of postoperative cognitive decline in neuropsychological test.13

TABLE 2. Differential diagnosis of POD and POCD POD POCD Epidemiology In all ages but more common in older people over 60 In all ages but more common in older people over 60 Manifestation Disturbance in attention and awareness, emotion, cognition, and fluctuating severity of consciousness. Cognitive deficits (impairment of memory, perceptual function, language, ability to combine tasks) Diagnostic tools Various delirium scale, for example, CAM-ICU Pre- and postoperative psychometric testing Timing Days to weeks Persisting for months Prognosis Reversible if underlying causative factors are treatable Reversible but with long-time impairment Abbreviations: POCD, postoperative cognitive dysfunction; POD, postoperative delirium. 2.3 Postoperative covert stroke

Cerebrovascular disease, a leading global cause of death and disability with approximately 6.2 million deaths due to stoke, is estimated to become the second leading cause of death by 2030.14 According to a systemic analysis for the Global Burden of Disease Study, the mortality rates caused by stroke range from 30.6% to 48.3%15 and are significantly related to operations.16 Covert stroke has been increasingly recognized over the years. It represents brain infarcts with silent and subtle manifestations that can be detected on brain imaging.17, 18 Covert stroke may contribute more to poor outcomes and prognosis in elderly patients presenting cognitive decline, as its subtle manifestation can lead to ignorance of cognitive symptoms.19

The incidence of postoperative covert stroke has gradually increased due to the aging population. So far, only a few studies have examined its mechanisms.20 Little is known about perioperative covert stroke except its association with substantially increased mortality.21 One multicenter prospective cohort study reported that postoperative covert stroke was found in 7% among 1114 participants over 65 years old who underwent inpatient, elective, noncardiac surgery, which were assessed with brain magnetic resonance imaging (MRI) after surgery and Montreal Cognitive Assessment (MoCA) on preoperative baseline and 1-year follow-up.22 Among the patients with a complete 1-year follow-up, cognitive decline after surgery occurred in 42% of participants who had postoperative covert stroke and 29% of participants who did not have postoperative covert stroke.22 In addition, another study suggested that covert stroke can increase the risk of POD, overt stroke, or transient ischemic attack (TIA) during one-year follow-up.22

2.4 Hemorrhagic stroke

Although hemorrhagic stroke comprises only 20% of all strokes, the perioperative hemorrhagic stroke could detrimentally deteriorate patients’ recovery and prognosis.23 According to the American Heart Association (AHA) and the American Stroke Association (ASA),24, 25 hemorrhagic stroke is divided into the following conditions: 1) focal bleeding in brain parenchyma and 2) in subarachnoid space or ventricular space due to rupture of blood vessel other than trauma. Perioperative hemorrhagic stroke may occur after cerebral hyper-perfusion due to a sudden surge of blood pressure (BP) for a certain amount of time.26 Hypertension is the most common risk factor for hemorrhagic stroke.27 Most anesthetics induce hypotension, so they are unlikely to provoke hemorrhagic stroke while under anesthesia.28 However, if postoperative hypertension persists for several hours, it can lead to certain condition linked to a sudden surge of cerebral perfusion.29

What discussed above is a brief introduction of four major types of postoperative neurological disorders with current understanding, which present disorientation, memory deficit, changes in awareness and attention from the baseline. Cognitive decline and poor prognosis of elderly patients after surgery have been an increasing concern around the world. With few studies on postoperative neurological disorders, the mechanisms and pathophysiology remain unknown,30, 31 especially for POD, POCD, and covert stroke. Besides, various manifestations of postoperative neurological disorders add more difficulties for research. As postoperative complications become increasing concerns, the mechanisms, prevention, and management are focus points that still need further research.

3 RISK FACTORS OF ACUTE POSTOPERATIVE NEUROLOGICAL DISORDERS

The exact mechanisms and pathophysiology of POD, POCD, and postoperative stroke are unclear. In the following paragraphs, we discuss potential risk factors of postoperative neurological disorders.

3.1 Risk factors of POD and POCD

It is commonly accepted that interactions between predisposing factors and precipitating factors play an important role in the occurrence of postoperative neurological disorders.32 The smaller vulnerability a patient has, the less occurrence of neurological disorders.32 For example, as advanced age is a predisposing factor, patients over 65 years old may present POD or POCD when exposed to only a few precipitating factors.33 On the contrary, younger patients exposed to the same precipitating factors may not experience POD or POCD. The recent evidence-based risk factors including some novel candidates for POD as listed in Table 3.32 POD and POCD share almost the same risk factors based on current limited researches.34 Moreover, SARS-CoV-2 (COVID-19) infection, a new uprising disease, has been found to be a potential novel risk factor of POD and POCD during the pandemic in the past two years, which is related to an accelerated onset with neurological manifestations35 and deterioration of cognitive decline.36

TABLE 3. Risk factors of POD and POCD in perioperative patients Risk factors Reference Preoperative factors Advanced age 32 Comorbidities (eg, cerebrovascular including stroke, cardiovascular, peripheral vascular diseases, diabetes, anemia, Parkinson's disease, depression, chronic pain, anxiety disorders, renal failure, and alcohol use disorders) 126-133 Preoperative fluid fasting and dehydration 32, 131 Preoperative blood transfusion 130 Hyponatremia and hypernatremia 128 Drugs with anticholinergic effects 32 Quinolone 130 Urine albumin to creatinine ratio (UACR) 134 Intraoperative factors Site of surgery (abdominal and cardiothoracic) 126, 135 Hybrid procedure 130 Intraoperative bleeding 136 Duration of surgery 126 Bispectral index (too low or too high) 127 Intraoperative electrolyte disturbance 128 Hyperthermia or hypothermia 137, 138 The rate of decline in intraoperative hemoglobin concentration 136 Postoperative factors Pain 139 Use of bed sensors 140 Nursing home residency 141 Delayed ambulation 129 Abbreviations: POD, postoperative delirium; POCD, postoperative cognitive dysfunction 3.2 Risk factors of postoperative stroke

Postoperative stroke remains one of the most serious complications after surgery. As for both ischemic stroke and hemorrhagic stroke, risk factors include conventional vascular risk factors, the type of surgery, and other perioperative events (Table 4). Vascular factors and nutritional state are important preoperative risk factors for stroke and postoperative stroke. Postoperative stroke happens more often in cardiovascular, general thoracic, and neurosurgery.37, 38 Among factors below, the BP and coagulation state are specific modifiable risk factors for postoperative stroke according to large-scale database studies, which will be further discussed.

TABLE 4. Risk factors of perioperative stroke Risk factors Reference Preoperative factors Vascular factors (eg, age, sex, history of stoke or TIA, arrhythmia, coagulopathy) 37, 41, 42, 142-144 Anemia 42 Malnutrition Preoperative central nervous system malperfusion 145 Cerebral diffusion-weighted imaging lesions 146 Renal dysfunction 147, 148 Intraoperative factors Type of surgery (cardiovascular, neurosurgery, left pneumonectomy other types of surgery) 37, 38, 42 Specific intraoperative events (arrhythmia, hypertension, hypotension) 44 Postoperative factors Adverse events (cardiac arrest, severe arrhythmia) 149 New-onset atrial fibrillation 150, 151 Abbreviation: TIA, transient ischemic attack 3.2.1 BP and stroke

Blood pressure fluctuation is an important risk factor for postoperative stroke.39 Emergency surgeries raise the incidence of neurological disorders and even affect long-term cognitive functions.40 As mentioned before, older patients are usually complicated with underlying diseases including hypertension, coronary artery disease, and arrhythmia, and thus, they have a greater risk of suffering from postoperative neurological disorders.37, 41, 42 It has been reported that patients developing postoperative stroke (ischemic or hemorrhagic) have higher average mean arterial blood pressure (MAP >80 mmHg).37 Moreover, a previous study suggests a strong relationship between elevated pulse pressure and stroke, which increases cerebral vulnerability to ischemic stroke.42 Additionally, evidence showed that intraoperative hypotension (IOH) is associated with the risk of major postoperative cardiac or cerebrovascular events.43 The short exposure to MAP of 55–65 mmHg is significantly associated with postoperative adverse cerebrovascular events, while maintaining systolic blood pressure (SBP) within 10% of the reference value may prevent postoperative adverse events compared with standard care (only treating if SBP <80 mmHg or <40% of the reference value).44, 45

3.2.2 Coagulation and stroke

For elderly patients, oral anticoagulants, including vitamin K antagonist (VKA) and non-vitamin K oral antagonists (NOACs),46 are common and effective therapies for the prevention of thromboembolism and stroke. However, they may lead to an inherent risk of bleeding. For elective and urgent surgery, reversal of anticoagulants is a necessary process for perioperative management.47 Traditional broader anticoagulants VKAs, such as warfarin, while effective, had multiple dietary and drug interactions and a great risk for intracranial hemorrhage, where 4-factor (II, VII, IX, and X) prothrombin complex concentrate (4F-PCC) should be used for VKAs reversal.48 Recently new types of anticoagulants targeted specific clotting factors (factors IIa and Xa), including dabigatran and rivaroxaban, have been approved for anticoagulation use with less adverse effect and less risk of hemorrhage.49 In addition, it is recommended to monitor the coagulation state of patients during the perioperative period.50

Postoperative neurological disorders occur by the interaction of predisposing factors and precipitating factors. Risk factors of POD, POCD, and perioperative stroke are listed above, and there may be other underlying risk factors unknown. Therefore, more researches are needed to understand possible risk factors and pathophysiology related to postoperative neurological disorders. The preventive strategies and protocols need to be established for known risk factors as well.

4 PREVENTION 4.1 Perioperative prevention of POD

Prevention strategies should be designed based on predisposing factors and parts of precipitating factors, which are the most effective measures against delirium.51 The Hospital Elder Life Program (HELP) released a multicomponent intervention guideline to prevent delirium52 (Figure 1).

Nonpharmacological and pharmacological interventions during preventive process of POD recommended in HELP and latest evidence-based advance research on prevention of POD, including interventions throughout the whole process of surgery and possible medications for treatment that needed further research. POD, postoperative delirium; HELP, Hospital Elder Life Program

4.1.1 Nonpharmacological interventions

Based on the studies of HELP, the multicomponent nonpharmacological intervention significantly reduces the incidence of delirium, including reorientation (using orientation calendar, clocks), early mobilization, promotion of sleep cycle, adequate hydration, visual and hearing aids, and increased supervision in hospital.53 If implemented by a skilled interdisciplinary team, these measures are effective against POD.54 Apart from common contents of HELP, a recent clinical trial has found that tailored, family-involved HELP could be beneficial for reducing POD, maintaining, or improving cognitive function, which may increase the implementation of this program.55 During the operation, however, the EEG (electroencephalography)-guided anesthetic administration, compared with usual care, failed to decrease the incidence of POD.56

4.1.2 Pharmacological interventions

Although several clinical trials have run various pharmacologic measures, there is a lack of strong evidence for effective prevention. One network meta-analysis demonstrated that haloperidol plus lorazepam might be the best treatment, while ramelteon may be the best preventive medicine for POD.57 Besides, a recent randomized clinical trial has evaluated the effectiveness of tropisetron and found that tropisetron could decrease the incidence of delirium after noncardiac procedures in adults.58 Another trial has found that postoperative scheduled intravenous (IV) acetaminophen, combined with IV propofol or dexmedetomidine, could reduce the incidence of POD in hospital versus placebo group, while patients receiving one single intravenous anesthetic had no significant improvement for POD.59 According to a Cochrane review that examined antipsychotic medications for preventing delirium in hospitalized, non-ICU patients, medications such as cholinesterase inhibitors, melatonin, and melatonin-receptor agonists have no clear effect in preventing delirium.60

Besides, measures against precipitating factors are mainly adopted in the perioperative period, including limiting fasting time, reducing the use of preoperative medications such as benzodiazepines, opioids, and anticholinergic medications. For elderly patients with risk factors, preoperative neurological assessment is essential for predicting and preventing of POD.61

4.2 Prevention of POCD 4.2.1 Preoperative interventions

The evaluation of patients’ baseline is important for the identification and prevention of POCD. Neuropsychological tests should be used before and after operations.62 Also, cognitive training and exercise63 have been proven to be beneficial for preventing POCD occurrence.64 Besides, Lu et al. found that pretreatment of parecoxib sodium combined with dexmedetomidine can decrease the incidence of POCD in patients undergoing arthroscopy by over 10%.65 These interventions should be particularly considered in high-risk patients.

4.2.2 Intraoperative interventions

In addition to preoperative interventions, there are several intraoperative measures to consider, including minimal exposure to anesthetics with careful monitoring. Regarding the anesthetic choice, Chen et al. found that the use of inhaled anesthetics66 in cardiac surgery generated higher postoperative scores in the Mini Mental State Exam (MMSE) compared with total intravenous anesthetics.67 Also, propofol may have a significant advantage in reducing POCD incidence compared with dexmedetomidine and midazolam sedation in elderly patients, in which midazolam has the highest inhibitory effects on cognitive functions.68

As for monitoring measures, one trial has shown a decreased incidence of short-term POCD with bispectral index (BIS)-guided deep anesthesia during the operation.69 Lastly, it has been suggested that postoperative management, including early identification and treatment of postoperative complications, may decrease the risk of POCD, which will be discussed in the treatment section. All interventions discussed above are listed in Figure 2.

Preventive measures of POCD based on latest research on it during preoperative, intraoperative, and postoperative process and possible effective medications that may contribute to the prevention. POCD, postoperative cognitive dysfunction. BIS, bispectral index. COX-2, cyclooxygenase-2

4.3 Prevention of stroke

As for perioperative ischemic (covert) stroke and hemorrhagic stroke, the control of modifiable risk factors is the most effective preventive strategy. These identified modifiable risk factors include hypertension, diabetes mellitus, hyperlipidemia, obesity, and smoking.70

4.3.1 Modifiable risk factor—hypertension

Hypertension is the most important modifiable factor for stroke,71 attributing to more than half of all stroke events worldwide.72 Antihypertensive medications are recommended for patients with BP over 140/90 mmHg.73 The most common medications include β-adrenergic agonists (β-blockers), calcium channel blockers (CCB), diuretics, angiotensin-converting enzyme inhibitors (ACEI), and angiotensin II receptors blockers (ARB), and the choice of therapy depends on individual comorbidities.73 Adequate BP control is important, while the goal adjustment should also be considered for older patients to avoid complications such as hypotension and dizziness. Although perioperative use of β-blockers might be beneficial in reducing heart rate and sympathetic activity and controlling BP, there is no association between β-blockers and perioperative outcomes.74, 75

4.3.2 Modifiable risk factor—hyperlipidemia

Hyperlipidemia is another remarkable risk factor, as several clinical trials and meta-analyses have reported decreased vascular events and mortality rates in patients with treatment for hyperlipidemia, especially lowering LDL-C (low-density lipoprotein-cholesterol).72 A randomized trial evaluated the benefits of statin as secondary prevention of stroke and found that atorvastatin can reduce the incidence of stroke in patients with recent stroke or TIA.76

Treatments of hypertension and hyperlipidemia have been proven to prevent stroke. However, the effectiveness of other factors such as weight and blood glucose control, or