Surgical context promotes the appearance of the main contributing factors for thrombosis and thromboembolic events, secondary to systemic inflammation and hypercoagulability. The incidence of these perioperative adverse events reported in by studies may vary according to the definition of ischaemic events. Major cardiac or vascular surgeries have higher incidence of perioperative stroke, mainly because manipulations of the aortic arch or heart are recognized as a source of embolic stroke. In noncardiac and non-vascular surgeries, ischaemic events are related to premorbid conditions and insufficient vascular reserve like in elderly with history of cerebral attack or coronary or renal disease. Bateman et al. identified comorbidities associated with perioperative stroke from the national database of hospital discharge available in the United States [4]. It includes age, gender, diabetes mellitus, atrial fibrillation, congestive heart failure, history of prior stroke, renal disease or cardiac valvular disease. When adjusted on most of these factors, perioperative ischaemic stroke was consistently associated with worse outcome [4]. The risk of perioperative thromboembolic event is increased in patients under anticoagulation or antiplatelet therapies, with a rebound of hypercoagulability, when therapies are stopped. However, the management of discontinuation for anticoagulation and antiplatelet therapy has not been clarified. Cardiac arrhythmia, especially atrial fibrillation, combined with the hypercoagulable state is a consistent risk of cardioembolism [8]. However, the benefit of β-blockade on cardiac events was counterbalanced by the higher incidence of stroke and worse outcome in patients who underwent a noncardiac surgery, and the relation to perioperative hypotension has been suggested [9].

Limiting inflammatory consequences of surgery should benefit to high-risk patients for thrombosis. In this context, a beneficial impact of statins was suggested but remains uncertain in patients without preoperative administration [10, 11].

In-hospital perioperative strokes, even if in-hospital, present a worse outcome than community-onset strokes, probably because of the delayed recognition of symptoms and then the delay of neuroimaging and low rate of thrombolysis [12]. The main difficulty is the evaluation of the neurologic deficit duration and establishment of “last time known well”. Following anaesthesia, the last normal neurologic exam is at anaesthesia induction. In all cases, the suspicion of stroke must activate emergent multidisciplinary discussion for the rapid answer to two key questions: (1) Is there any contraindication for thrombolysis? (2) Is the timing adequate for endovascular clot retrieval? Magnetic resonance imaging will help for locating the occlusion of the vessel and dating the event. A careful risk-benefit analysis should be conducted in order to balance the risk of haemorrhage after recent surgery. Recent intracranial and spinal surgeries remain well-recognized contraindications for thrombolysis. Other therapeutic options for surgical patients who may be at high risk of haemorrhage, such as intra-arterial selected thrombolysis, were suggested but are not fully evaluated. A clear procedure should be anticipated in each institution in order to ensure the best management when such postoperative deleterious event occurs.

Haemodynamic events are the other mechanism for occurrence of postoperative stroke. The circumstances are less characterized, but anaesthetists should consider factors such as surgical positioning, limited cerebral autoregulation or “vascular reserve” in some patients.

It is well admitted that the sitting surgical position may expose the patient to cerebral hypoperfusion because of the important hydrostatic gradient between the head and heart levels. In such position, an intensive monitoring and haemodynamic resuscitation should prevent deleterious events. The more frequent and less intensively monitored is the common “beach chair” position for shoulder, breast or skull base surgeries. Even if the orthostatic gradient is lower in beach chair than sitting position, it could challenge cerebral perfusion, especially in patients with stenosis on main cephalic arteries, or intracranial arteriolar diseases. Perioperative cerebral oxygen saturation can be monitored by near-infrared spectroscopy, and some desaturations have been detected during surgery, but no clear association with neurological outcome was established [13]. A strategy of prevention of ischaemic events should be preferred than detection, since no reliable method of detection can be firmly recommended. In such patients, baseline level of arterial blood pressure should be maintained along surgery. In some case, this objective of management with elevated blood pressure should be balanced with the risk of significant bleeding in the surgical focus. The difficulty to target the best arterial pressure arises from the difficulty to assess an effective autoregulation for each individual. Moreover, the classical concept of autoregulation may not be applicable in many patients [14, 15]. In this concept, the recommendations would be the tight control of blood pressure with the lower limit at 30% of baseline measured before anaesthesia. The perioperative management of blood pressure may also have an impact on postoperative occurrence of delirium [16].

Since the interactions between cerebral and systemic haemodynamics [17] are difficult to anticipate and could do more harm than good, anaesthetists must particularly pay attention to other cerebral blood flow-regulating processes, like the vasoconstriction in response to hypocarbia. A recent review depicted the integrated effects of carbon dioxide and perfusion pressure on autoregulation phenomenon [18]. Carbon dioxide partial pressure in the blood is easy to monitor with a non-invasive method by capnography and becomes a standard in anaesthesiology. Then the adjustment of carbon dioxide level must be cautiously achieved for patients with limited cerebral “vascular reserve”.

Positioning of the head is also of importance for adequate brain perfusion. Besides atherosclerosis, other imputable factors were identified like deviation from classical vessel configuration or compression. From study performed on cadavers, in order to avoid lowering of cerebral blood flow, the variations of the head from resting position should remain below 45° for flexion/extension, 45° for rotation and 30° for tilt [19].

Among patients with cerebral haemodynamic impairment, the patients with Moyamoya syndrome require particular strict perioperative anaesthetic management. Moyamoya disease (MMD) is characterized by chronic progressive stenosis at the termination of internal intracranial carotid [20]. Even if this syndrome remains rare in Europe in comparison with Asia (tenfold less), it affects young patients (third and fourth decades of life); the number of patients with MMD is increasing with significant progress in their medical management. With progression of carotid abnormalities, compensatory mechanisms will develop the enlargement of small arteriolar collaterals to bypass the occlusion but with impaired vasodilatory response. Such haemodynamic impairment may be compensated by the increase in oxygen extraction. We will address in this paragraph the management of Moyamoya patients for general surgery and not discuss the anaesthetic management for surgical cerebral revascularization (for review [21]). Before surgery, a detailed history of the disease is necessary for the preoperative evaluation. The anaesthetist will evaluate the severity of the arterial disease on cerebral magnetic resonance imaging (parenchymal lesions and vascularization) or conventional angiography. Cerebral blood flow studies such as transcranial Doppler ultrasonography, positron emission tomography and single-photon emission computed tomography with acetazolamide challenge could be helpful for the evaluation of the “cerebral blood flow reserve”. Considering haemodynamic in Moyamoya disease, maintenance of cerebral blood flow is extremely dependent on systemic conditions and deteriorates with hypotension, hypocarbia or hypercarbia [22]. Arterial blood pressure should be carefully monitored and kept close to patient’s baseline blood pressure. A haematocrit greater than 30% should avoid the risk of perioperative ischaemic events even if there is no specific recommendation. Because of the high risk of vascular events, postoperative care will preferably take place in the intensive care for the close monitoring of blood pressure and volemia. Antiplatelet agents should be started on the first postoperative day because some deterioration seems to occur as a consequence of emboli from microthrombus formation at sites of arterial stenosis.

In summary, although embolic and haemodynamic events with cerebral complications are expected in the context of cardiac and major vascular surgeries, their occurrence is rare but more insidious in general surgeries. Such observations force the clinician to identify the patients at high risk of perioperative stroke and the potential modifiable factors [1]. In such high-risk patients, anaesthesia procedure should be conducted according to haemodynamic objectives of blood pressure levels, stabilized haemodynamic and careful systemic carbon dioxide monitoring.

Despite high incidence and serious implications, postoperative delirium is frequently under-evaluated because of its variable clinical presentation (hypoactive or hyperactive) and despite the availability of reliable diagnostic tools like Confusion Assessment Method [23]. Elderly patients are at higher risk of postoperative delirium [24], and, 3 months after surgery, cognitive impairment is still detectable in older patients with consequences on their quality of life and increased mortality risk [25, 26]. In a prospective case-control study in patients without severe preoperative cognitive disease, postoperative delirium was higher after emergent (17.9%) than elective surgery (6.7%), and physiologic or psychological preoperative conditions including anxiety and depression were associated factors [25]. In this context, the perioperative management of the patients with limited vital function supports the concept of perioperative care for improving outcome. Notably, for patients with previous disabilities, it seems reasonable to delay nonurgent surgery beyond 9–12 months after ischaemic stroke, according to data from Danish registry [27].

Besides pre-existing medical conditions, anaesthetics and inflammatory response to surgery were suggested as potential precipitating factors for delirium. Similar to other cognitive impairments, the leading hypothesis for the pathomechanism incriminated in delirium was a central cholinergic deficiency related to drugs that impair cholinergic function [28–30]. Drugs such as atropine, antihistamines, corticosteroids or benzodiazepines should be minimized in vulnerable patients [28]. Alternative drugs might be proposed in order to limit or prevent delirium, like haloperidol, dexmedetomidine or subanaesthetic dose of ketamine, but should be further investigated before being firmly recommended [31].

Recent investigations have illustrated the debate about the impact of type of anaesthesia (general or local) or genetic factors [32, 33] on incidence of postoperative cognitive dysfunction. In 225 patients over 60 years of age who underwent a cardiac surgery, postoperative delirium was associated with cognitive decline after 1 year [26]. These patients with significant postoperative cognitive decline were significantly older, less educated and less likely to be men or white but had a history of stroke and higher comorbidity score. Accordingly, some studies reported that chronic diseases developed with age like cardiac diseases [34] or cancer [35] seemed predisposed the patients to cognitive decline, regardless of the surgical treatment. On the other hand, in some patients, preoperative cognitive disorders may be improved by surgery, especially when surgery decreases chronic pain and inflammation, improves cerebral blood flow or favours the functions in daily life. A common pathomechanism between postoperative delirium and dementia was suggested, but a recent study and meta-analysis rule out this association [36].

Alzheimer’s disease (AD) is the most common form of dementia in adults, affecting 35 million people worldwide. AD is characterized by progressive worsening of symptoms, including global cognitive decline in memory, orientation, judgement and reasoning. Because the incidence of AD is expected to increase continuously in the next future in the absence of significant therapeutic breakthrough, the anaesthetic management will have profound implication for anaesthesiology in elderly patients and patients already diagnosed with AD [37]. The experimental models for AD depicted a dysregulation of the homeostasis of tau protein and of processing of amyloid proteins. The imbalance in activities of enzymes that catalyses the phosphorylation/dephosphorylation of the tau protein results in its hyperphosphorylation and leads to neuronal cell death and degeneration. Besides, the alteration of amyloid protein metabolism results in its accumulation, then synaptic dysfunction and neuronal damage.

Despite in vitro studies and animal data, the possible relation between anaesthetic neurotoxicity, postoperative cognitive dysfunction and AD remains elusive. There is no rigorous clinical trial for either recommending or contraindicating anaesthetic procedures on the basis of neurotoxicity in the elderly. The impact of anaesthetics, especially inhaled anaesthetics, was investigated in experimental models and in patients with AD, because they might accelerate the progression of the disease [38]. The main hypotheses would be that inhaled anaesthetics (1) impact the processing and metabolism of β-amyloid protein, resulting in its accumulation instead of clearance from extracellular space into blood and cerebrospinal fluid, and (2) trigger hyperphosphorylation of tau protein and its aggregation. Among inhaled anaesthetics, isoflurane and sevoflurane were similarly described inducing apoptosis in experimental animal model and accumulation of β-amyloid metabolites in CSF in AD patients. Potential upstream mechanisms of inhaled anaesthetics are discussed also like the elevation of cytosolic calcium followed by the activation of apoptotic pathways, expression of inflammatory mediators or release of radial oxygen species and mitochondrial damage. By contrast, desflurane was not associated to aberrant processing of tau protein metabolites or postoperative cognitive decline. Most of these data were experimental but, if confirmed, might implicate changes in anaesthesiology of susceptible patients [39].

There is no absolute contraindication in AD for regional anaesthesia techniques. Pragmatically, the anaesthetist should consider the inability of some AD patients to understand their environment or to cooperate. Therefore, the lack of cooperation and unanticipated outbursts during the surgical procedure are arguments for general anaesthesia.

Multiple sclerosis (MS) is a chronic inflammatory disease characterized by demyelination of the white matter in the central nervous system. Young adults and women are predominantly affected. The severity of symptoms may be highly variable, resulting in bedridden at worst.

The most common anaesthetist’s concern is exacerbation of the pre-existing deficit [40, 41]. Nevertheless, anaesthetists should not automatically take the blame for postoperative worsened or new symptoms. Stressful condition, fever, infection, surgery and delivery may cause such exacerbations, making it very difficult to separate the effects of these factors and anaesthesia. The different anaesthetic options have to be discussed for every patient, regarding the progression of the disease, the respiratory function and the cardiac function.

A detailed history of the disease is necessary for the preoperative evaluation. Discontinuation of treatment for multiple sclerosis might be associated with disease recurrence, especially in patients with a highly active disease. Decision on discontinuation should be discussed with a neurologist, based on the presence of side effects and drug interactions.

Clinical assessment of respiratory dysfunction is a major issue for the preoperative evaluation. Respiratory dysfunction may appear in early stages of the disease, due to lack of respiratory muscle coordination caused by cerebellar impairment [42]. The preoperative assessment should at least include the ability to cough and clear respiratory secretions and the ability to exhale deeply. Obstructive sleep apnoea is another common sleep disorder in MS patients that should be identified in advance.

Cardiac evaluation is also important for anaesthetic management. MS treatment may cause cardiomyopathy and autonomic dysfunction is possible, resulting in an increased risk for haemodynamic instability during induction. Hypotension with reduced response to intravenous fluid or vasopressor therapy should be expected.

The choice for general or regional anaesthesia depends on the preoperative assessment and the surgical procedure. There are no specific precautions for inhaled and intravenous anaesthetic agents or opioids. For patients with significant motor impairment, succinylcholine might be avoided because of a potential hyperkalemic response. The response to nondepolarizing muscle relaxants is unpredictable. On one hand, a relative resistance to these agents has been described, due to an increased number of postjunctional receptors. On the other hand, lower doses of short-duration nondepolarizing relaxants should be used in MS patients with motor weakness. Baclofen, used to regulate spasticity, may cause muscle weakness, rendering the patient extremely sensitive to the action of nondepolarizing muscle relaxants. Despite this side effect, baclofen should not be stopped abruptly, because of the risk of delirium and convulsions. Because of the unpredictable response, neuromuscular monitoring is imperative.

Despite the lack of evidences, many anaesthetists, fearing forensic procedures, will consider the presence of pre-existing neurological disease, a contraindication for regional anaesthesia [43]. Above threshold concentrations, local anaesthetics are neurotoxic. The loss of the protective effect of the myelin in patients with MS results in the spinal cord and nerves being exposed to higher concentrations. As a consequence, the safety of regional anaesthesia cannot be guaranteed. Nevertheless, it may be advantageous because of a decrease stress response to surgery. Epidural anaesthesia has been studied extensively in the obstetric population and is considered safe. For subarachnoid anaesthesia, controversy still exists. When performing epidural or spinal anaesthesia, recommendations are to use shorter-acting agents and the minimum dose necessary, with or without epidural opioids.