1 Introduction

The global population is ageing, and by 2030, one in six people will be aged over 60 worldwide [1]. As a result, the number of older people requiring both routine and emergency anaesthesia for surgical procedures is expected to rise significantly [2]. This poses a number of challenges given that both the prevalence and degree of multi-morbidity, functional and cognitive impairment, and polypharmacy increase with age [2]. All of these factors lead to higher levels of risk, both in terms of surgery itself and the inherent risk of the accompanying anaesthetic [2]. As a result, there has been a rise in the number of perioperative services for older people, which aim to identify medical comorbidities, establish the level of risk, and pre-optimise older people undergoing both emergency and elective procedures [2].

Frailty describes the increased vulnerability to acute and everyday stressors that occurs with declining function and reserve of multiple organ systems, often associated with ageing [3]. Rockwood described the Clinical Frailty Scale (CFS), which quantifies the level of frailty on a scale of 1–9, based on the degree of functional and/or cognitive deficit [3]. Multiple definitions of frailty, and frailty sub-types exist, and conceptually, frailty can be difficult to operationalise.

Brain frailty is one such sub-type, and has been described in studies of stroke [4, 5], yet no formal definition exists. Brain frailty can be considered as a state of increased vulnerability of the brain to both acute and everyday stressors (e.g. infection, trauma, transient hypoperfusion, postural changes), due to reduced functional capacity or reserve (e.g. cognitive impairment, stroke, age-related changes, small vessel disease and atrophy). Studies have proposed a number of neuroimaging markers of brain frailty including: leukoariosis, microbleeds, prominent perivascular spaces, lacunar infarcts and intracerebral haemorrhage, which are related to poorer outcomes following acute stroke [5]. In addition to these structural markers, functional changes are seen with brain ageing, such as impaired cerebral autoregulation (CA), and altered coupling between neuronal activity and cerebral blood flow (neurovascular coupling) [6]. These processes are altered in a number of age-related disorders associated with increased frailty, including stroke [7] and dementia [8].

This brief review will outline the relevance of the concept of brain frailty to the practice of anaesthesia by discussing the effect of direct and indirect effects of anaesthesia on the brain, and how this may affect cerebrovascular morbidity and mortality in those with brain frailty. We will consider the effect of different anaesthetic techniques on cerebral haemodynamics, the importance of blood pressure (BP) control and the role of carbon dioxide (CO2) in this frail population.

2 Effect of anaesthesia on cerebral haemodynamics

CA describes the processes by which the brain regulates cerebral blood flow, to ensure adequate but not excessive perfusion, thus protecting against both cerebral ischaemia and haemorrhage. During surgery, the autoregulatory capacity of the brain is challenged by fluctuations in BP induced by a combination of noxious surgical stimuli, fluid loss, anaesthetic agents, patient positioning and pre-existing co-morbidities.

The cerebrovascular response to anaesthesia is heavily dependent on the anaesthetic agents utilised. Volatile agents (most commonly sevoflurane and isoflurane) have been demonstrated to induce impairment in CA in a dose dependent manner [9]. In contrast to this, CA remains intact during propofol based anaesthesia [10] up to a rate of 200 mcg/kg/min−1, which is well within the range of the infusion rates used in clinical practice.

Total intravenous anaesthesia (TIVA) is growing in popularity, partly driven by the perceived environmental benefit compared to the use of volatile agents. Remifentanil is used alongside propofol during TIVA to provide analgesia and to ablate spinal reflexes, and this too has been shown to have no significant impact on CA [9]. Therefore, in those requiring a general anaesthetic, TIVA perhaps preserves the cerebrovascular defences better than volatile agents and this should be considered when creating an anaesthetic plan.

Spinal anaesthesia is often utilised, where surgically suitable, in patients with multiple cardiorespiratory co-morbidities who may respond poorly to a general anaesthetic. Interestingly, relatively little has been written about the impact of spinal anaesthetic on cerebral blood flow and CA. One small study reported a statistically significant decrease in cerebral blood velocity following the administration of bupivacaine to older patients. However, no metrics of dynamic CA were collected and the clinical relevance of the observed difference remains unclear [11].

3 Blood pressure variability

As described earlier, BP can fluctuate significantly during surgery. Evidence suggests that increasing BP variability (BPV) is associated with worse outcomes in both cardiac [12] and non-cardiac surgery [13]. However, increasing BPV is less harmful than frank hypotension, which was found to have a much stronger association with mortality [13]. In the context of the frail brain, intraoperative control of BP is especially important due to the stress placed on already impaired autoregulatory processes [7] by both hypo- and hypertension.

The degree of variation of intraoperative BP, and therefore intraoperative cerebral perfusion pressure, can be minimised with careful practice. Intermittent hypertension in response to particularly painful surgical steps can be avoided with adequate analgesia and adequate depth of anaesthesia. Conversely, intraoperative hypotension can be avoided with peri-operative correction of hypovolaemia, careful titration of the induction agent, and avoiding excessively deep anaesthesia. For both hyper- and hypotension, prompt recognition allows timely correction, and in a frail population undergoing significant surgery, invasive BP monitoring facilitates this more reliably than intermittent non-invasive brachial BP monitoring. Continuous non-invasive beat-to-beat BP monitoring, using devices such as the Finometer [14], gives the benefit of an arterial line without the risks associated with the procedure. However, at present, they remain absent from most theatre complexes.

The ageing brain is more sensitive to hypotension as a result of a rightward shift of the autoregulatory curve, which renders the brain vulnerable to significantly reduced cerebral blood flow at a given BP, compared to a younger patient (Fig. 1, [15]). With regard to induction, in those with cardiovascular co-morbidities, there is an argument to use alternative induction agents such as ketamine or etomidate to avoid the profound hypotension that may accompany induction with propofol. However, these drugs themselves have side effects and so propofol remains commonplace due to its familiarity, its availability, its suitability in TIVA and its effect on ablating laryngeal reflexes to facilitate laryngeal mask and laryngoscope insertion. Hypotension on induction can be avoided with careful titration to effect, and where required, with concurrent vasopressor infusions in the frail, unwell patient.

Fig. 1
figure 1

Adapted from Beishon et. al with permission [15]. Schematic diagram demonstrating the change in the autoregulation curve with advancing age. Solid line; younger population. Hashed line; older population. Vertical lines demonstrate the lower bound of autoregulatory capacity for the respective populations

Full size image

With its constant and easily titratable delivery of opioid analgesia with remifentanil and individualised monitoring of anaesthetic depth with bi-spectral index, TIVA provides an effective method of controlling BP for all the reasons described above. Its drawback is that the induction bolus of propofol can be large, and therefore using a low target initially alongside a prophylactic metaraminol infusion may be considered.

4 Blood pressure targets

Improved understanding of peri-operative stroke risk, particularly in those with pre-existing cerebrovascular disease has helped inform our understanding of the susceptibilities of the aged or diseased cerebrovasculature to insults including anaesthesia and ventilatory changes [16].

There are a wealth of non-randomised data in cardiac and non-cardiac surgery demonstrating adverse outcomes associated with intraoperative hypo- and hypertension. Despite this, there remains a lack of consensus on targets for intraoperative BP. There is inherent variation due to the surgical procedure involved—however, there is a necessity in all instances to ensure BP related bleeding risk is weighed against the risk of organ ischaemia [17]. This is particularly pertinent to consider in those with atherosclerotic disease of key arterial supplies to major organs. A randomised study of older adults (mean 70 years, standard deviation 7 years) undergoing major abdominal surgery highlighted the need for individualised BP targets aimed at achieving a systolic BP (SBP) within 10% of their usual baseline SBP value [18].

Alterations in cerebral perfusion and autoregulation can occur during hypotension and during impaired cerebrovascular reactivity states prompted by hypocapnia or hypercapnia. These states can raise the intraoperative stroke risk, possibly explained by the “steal phenomenon” [19]. Whilst anaesthetised, a higher BP is related to greater CO2 reactivity during anaesthesia with either sevoflurane or propofol, which is an important consideration in older individuals [16].

A fall in BP can be seen during spinal anaesthesia, prompted by a fall in cardiac output and not by a fall in systemic vascular resistance [20]. Interestingly, postoperative delirium has an association with extremes of mean arterial BP in older hip fracture patients undergoing spinal anaesthesia. Indeed, one study of cardiac surgery patients demonstrated that pre-operative impairment in CA is associated with a greater risk of post-operative delirium [21], which suggests that uncontrolled fluctuations in cerebral perfusion may play a role. However, further hypothesis generating studies examining autoregulatory impairment and thresholds for degree of delirium and cognitive impairment are needed [22].

5 Conclusion

In summary, anaesthetic practice in the frail population needs to consider both the effect of anaesthesia on CA processes, and needs to minimise the ‘challenge’ to CA by avoiding large fluctuations in BP. Normocapnic ventilation is important to reduce the impact of increased cerebrovascular reactivity on the already frail brain, and TIVA provides a method of avoiding BP fluctuations while preserving CA. Further work comparing TIVA to volatile agents in the frail population, with end-points including cognitive dysfunction and cerebrovascular disease would help anaesthetists plan for the increasingly ageing and co-morbid population. Additionally, a consensus on what defines ‘brain frailty’ would enable this susceptible population to be identified and accounted for in anaesthetic practice.