The principles and conduct of anaesthesia for emergency surgery


C. Morris


In this second article we examine the principles underlying delivery of the components of anaesthesia. Topics considered include anaesthetic technique, management of the airway and lung ventilation, induction and maintenance of anaesthesia, patient monitoring including the place of cardiac output devices. We summarise recent research on the management of shock and sepsis syndromes including goal directed therapy and examine some controversies around intravenous fluid therapy. Finally, we discuss intra-operative awareness and challenges during emergence including peri-operative cognitive dysfunction.

Choice of anaesthetic technique

Optimal anaesthetic technique is dependent upon an individualised assessment of patient and surgical requirements. Prospective, randomised, controlled trials of rival anaesthetic techniques, e.g. the GALA trial, are rare [1]. Emergency anaesthesia lacks this evidence base and extrapolations from meta-analyses typically fail to reach clear recommendations for clinicians [2]. The nature of surgery, patient co-morbidities, and urgency will frequently decide the anaesthetic choice but clinicians should remain mindful of the options of local or regional anaesthesia and sedation in appropriate patients, especially when a ‘lesser’ procedure becomes feasible e.g. percutaneous drainage.

Management and protection of the airway including pulmonary aspiration

The fourth National Audit Project (NAP4) of the Royal College of Anaesthetists recorded the epidemiology and management of airway complications in UK practice and is essential reading [3]. It is emphasised that 25% of emergency airway management takes place outside operating theatres and since publication, several resources and recommendations have become available [4–6]. In the UK, out-of-theatre tracheal intubation frequently occurs in the absence of essential staff and equipment and capnography was not used in nearly a third of cases in a large observational study [7].

Both difficult laryngoscopy and difficult tracheal intubation are more common in emergency patients, often outside the theatre setting [8, 9]. One American study of non-theatre emergency tracheal intubation by experienced practitioners identified difficulties in over 10%, with over 4% suffering complications [10]. Predictors of complications included multiple attempts and poor views at laryngoscopy and intubation in a ward or emergency department setting. Human factors are increasingly recognised in airway complications and an unfamiliar environment during airway management is likely to be important [3].

Recommendations for prior planning for both airway management and airway disaster management form part of anaesthetists’ training in the UK [11, 12]. The distinction between an expansive management algorithm at the planning stage [13] and a ‘failed intubation drill’ with limited choices [14] is emphasised, yet ‘drill practice’ and simulation are typically rare in UK anaesthesia despite the need for establishing behavioural patterns in defined circumstances [15]. End-tidal capnography [16, 17] should support clinical examination for confirmation of airway placement and there were criticisms in NAP4 around interpretation of common capnograph recordings (e.g. cardiac arrest or airway obstruction) [3].

Emergency anaesthesia is often associated with delayed gastric emptying. Pulmonary aspiration of gastric contents was the commonest (over 50%) anaesthetic cause of death in NAP4, although the majority occurred during supraglottic airway anaesthesia, and only once during rapid sequence induction [3]. The incidence of aspiration syndromes varies by definition and patient population and has been reported as 1 per 400 000 elective anaesthetics or 1 per 900 emergency anaesthetics [18], or collectively, 1 in 7000 [19]. One might expect the increasing need for surgery in the sick and more elderly population to increase the incidence of aspiration, but some workers identify this as relatively constant [20].

The top five factors implicated in aspiration in the Australian Anaesthetic Incident Monitoring Study were: emergency cases; inadequate anaesthesia; abdominal pathology; obesity; and opioid medication [21]. It remains intuitive that a lower pH and larger volume increase the severity of aspiration syndromes, classically > 25 ml at pH < 2.5 [22]. Furthermore, obesity and gastro-oesophageal reflux, both significant risk factors for regurgitation of gastric contents, are common, and their incidence is increasing [23, 24]. Recent peri-operative fasting times and guidelines [25, 26] often need to be disregarded for emergency anaesthesia and other aspects of airway management (e.g. a low threshold for tracheal intubation) are probably more important in preventing aspiration [27]. The identification of a full stomach during left upper quadrant ultrasound is familiar to most FAST practitioners and this non-invasive test has been described as a precursor to anaesthesia, although this is a ‘rule in’ test with poor correlation with gastric volume [28].

Before elective surgery, histamine-2 receptor antagonists and proton pump inhibitors can significantly reduce the incidence and severity of aspiration syndromes [29]. Other interventions to reduce gastric contents volume and/or increase pH might intuitively reduce the risk and severity of pulmonary aspiration, including pre-surgery gastric tube drainage, use of post-pyloric feeding tubes and use of pro-kinetic therapy, although evidence is limited and extrapolated from other populations e.g. elective surgery, critical care or obstetrics [30–34]. The presence of a tracheal tube cannot guarantee airway protection and in cases of intestinal obstruction, intra-operative enterostomy and drainage may be preferable to ‘milking back’ large volumes of fluid that may bypass a nasogastric tube.

Cricoid pressure, described over 50 years ago, remains controversial in anaesthesia practice [35]. Although it forms part of mainstream practice in the UK, many countries do not use cricoid pressure and have similar reported incidences of aspiration [36, 37]. Conversely, some authorities suggest that cricoid pressure may increase the incidence of difficult laryngoscopy [38], failed intubation and regurgitation.

Emergence from anaesthesia and extubation are times during which the aspiration risk is higher, and clinicians should remain vigilant until patients are fully awake, with airway protective reflexes. The Difficult Airway Society (DAS) now recommends stratification of patients into ‘low-risk’ or ‘at-risk’ with distinct extubation algorithms to follow [39, 40]. Awake extubation is strongly recommended for both groups. Clinicians should have a low threshold for suspecting pulmonary aspiration, e.g. unexplained hypoxaemia, and a low threshold for planned admission to critical care facilities if this is suspected [41].

While beyond the scope of this article, the NAP4 report highlighted the high incidence of complications around tracheostomy management, often within the intensive care unit [3]. Guidelines and standards of care from the Intensive Care Society [42] and a number of resources are available alongside the NAP4 findings [5, 6].

The rapid sequence induction: evolution over time

Managing an anatomically difficult or obstructed airway in the emergency setting in a patient at high risk of gastric aspiration presents one of the most challenging scenarios an anaesthetist can face. While a ‘rapid sequence induction’ (RSI) is arguably the ‘gold standard’ technique to secure the airway and avoid aspiration, it carries significant risks if laryngoscopy and/ or intubation fail. Failure to consider alternative techniques to secure the airway, in particular awake fibreoptic intubation, was highlighted by the NAP4 report [3].

Recent surveys have identified, that a significant number of anaesthetists, possibly the majority, employ ‘modifications’ to the ‘classical’ RSI and these are considered in Table 1 [43–45]. Factors promoting clinicians to modify their technique included obesity, risk of reflux and trauma [46, 47].

Table 1. Components of a ‘classical’ and ‘modified’ rapid sequence induction (RSI).
Classical RSIModified RSIComment
Pre-oxygenation and avoidance of mask ventilation until trachea intubatedLittle justification to omit pre-oxygenation but many clinicians attempt mask ventilation [45]The role of neuromuscular blockade in facilitating difficult mask ventilation is controversial [48]; many cases are made easier following blockade, but spontaneous ventilation becomes impossible
Induction agent thiopental and avoidance of other agentsAlternatives used e.g. propofol, ketamine, opioid adjuncts [15]Factors promoting this may include availability, preparation time, familiarity, perceived cardiovascular stability, risk of allergic reaction
Neuromuscular blocking agent suxamethoniumCommonly rocuronium is used in doses of 0.75–1.5−1 [15]The largely equivalent time to intubation must be balanced against the risk of ‘can’t intubate, can’t ventilate’. The availability of sugammadex may have altered the risk benefit ratio (see below).
Cricoid pressureOmittedThe ability of cricoid pressure to prevent reflux and possibly worsen the view at laryngoscopy make it controversial. Most UK clinicians do use cricoid pressure [36–38]
Direct laryngoscopy with Macintosh bladeMany clinicians routinely utilise alternative blades e.g. McCoy or select a videolaryngoscope or modified device (e.g. AirtraqTM)Grade 3 or 4 views at laryngoscopy are associated with failure to intubate and attendant risk of hypoxaemia and aspiration of gastric contents. Furthermore, in the presence of suspected cervical spine injury, modifications may reduce cervical spine excursions [49, 50]
Confirmation of tracheal intubation using capnographyThe only argument against routine use of capnography is its absolute lack of timely availabilityCapnography was strongly recommended in NAP4 and publications before and since [3, 5, 17]

Management of ventilation

The requirement for tracheal intubation for airway protection and neuromuscular blockade to facilitate cavity surgery lead to the common use of intermittent positive pressure ventilation (IPPV) in emergency anaesthesia. Modern recommendations for IPPV have mostly derived from management during acute respiratory distress syndrome (ARDS). Many of these principles hold true during emergency anaesthesia as acute lung injury is the commonest cause of postoperative respiratory failure, and respiratory complications may exceed cardiac complications [51].

Modern ‘lung protective’ IPPV during ARDS suggests low tidal volumes of 6–8−1 (avoiding volutrauma) and limited inflation airway pressures < 30 cmH2O (avoiding barotrauma) [52–55]. More controversial is the selection of ‘high’ vs ‘low’ PEEP with limited evidence of benefit for higher values in critically ill patients [55–58], and little evidence to support clinicians during anaesthesia [59].

The deleterious effects of non-physiological IPPV have been reviewed previously [60]. In animal models, even 90 min of ‘non-injurious’ ventilation produces upregulation of pro-inflammatory genes and their products [61, 62]. It is of concern that tidal volumes > 10−1 may have damaging effects even after relatively short applications during human anaesthesia [63], and there are few circumstances where clinicians cannot limit tidal volumes to < 10−1 [64].

While permissive hypercapnia may be practised on intensive care [65], pre-existing acidosis, renal impairment, hyperkalaemia or elevated intracranial pressure may make this inappropriate during emergency surgery; the risk/benefits should be considered for each case.

The nature of the patient’s condition may carry significant implications for IPPV management. For example, a laparotomy and washout following a perforation may significantly reduce intra-abdominal pressure and aid IPPV, while conversely a polytrauma patient undergoing large volume resuscitation may develop intra-abdominal compartment syndrome, making IPPV more challenging. The management of intra-abdominal compartment syndrome is considered elsewhere [66], and simple adjuncts, e.g. a nasogastric tube, should be considered as one only needs to restore the abdomen to the compliant part of its pressure-volume relationship, rather than achieve complete decompression [67].

Induction of anaesthesia and pharmacology

UK practice for the majority of emergency cases will be to employ a form of RSI as discussed above. A national survey in 2001 found that thiopental and suxamethonium were the most widely used drugs for RSI. Propofol and rocuronium were used by more than a third of respondents and most (75%) routinely administered an opioid [15], and modifications seem to persist [47].

Historically etomidate, an imidazole ester, was popular for induction in cardiovascularly unstable, septic or shocked patients, due to its lesser effect as a circulatory depressant. Concerns regarding the safety of etomidate have been sustained [68] due to its effects on 11 β-hydroxylation in the steroid synthetic pathway; however, adrenal dysfunction in critical illness is a complex disorder [69]. Prolonged infusions were recognised to increase mortality [70], and even single bolus dose administration in the critically ill may be harmful [71, 72]. While the adrenal suppressant effect is largely accepted, the clinical implications of this are less clear, and support for the choice of etomidate in emergencies continues [73]. In retrospective analysis of septic shock patients, less cardiovascular depression was observed without significantly increased mortality or effect of steroid therapy following etomidate [74]. The controversy around etomidate is reinforced by equivocal findings following steroid replacement in critically ill patients with adrenal suppression [75, 76].

Ketamine, a phencyclidine derivative, has made a significant resurgence in emergency practice. Ketamine has many ‘ideal’ qualities for an intravenous induction agent, with analgesia and a sympathomimetic effect (and indirect haemodynamic stability) both considered positive. The risks of emergence phenomena, nausea and vomiting are minimised by prolonged postoperative sedation in intensive care. A French study demonstrated that adrenal insufficiency following even single dose etomidate was increased and that ketamine is a safe alternative to use [77]. It is certainly our routine practice to use ketamine, including in cases where raised intracranial pressure may be a risk or actually present, in all RSIs where cardiovascular instability is anticipated [78].

A variety of alternative selections and combinations are used in clinical practice and perhaps under-represented in the literature. Reduced dose thiopental [79], propofol, midazolam plus opioid combinations are frequently used in the UK during RSI [15]. Propofol is the most common induction agent for out-of-theatre tracheal intubation in the UK [7]. Induction agent choice has not been related to outcome for patients undergoing RSI in the emergency department who are subsequently admitted to intensive care units, although the risks of developing hypotension and receiving a vasopressor at induction were greatest for propofol, and etomidate may have increased mortality before patient matching [80]. Propofol is frequently considered as an appropriate agent for emergency anaesthesia but there are strong pharmacokinetic and dynamic considerations that argue against this, and some workers suggest that in the presence of haemorrhagic shock, propofol doses should be reduced to 10–20% of an ‘elective’ dose should [81]. Furthermore, the vagotonic effects of phenylpiperidene opioids may be especially poorly tolerated and patients may require an element of tachycardia to maintain cardiac output, especially with relative hypovolaemia. A recent study in non-critically ill patients (ASA 1-2, elective surgery) has demonstrated improved haemodynamic effects of combination ketamine 0.75−1 and propofol 1.5−1 named ‘Ketofol’ as an admixture administered from a 20 ml syringe for induction; this combination certainly requiring further study [82].

It is common practice in the elective setting to check if manual ventilation is possible before administering a neuromuscular blocking drug (NBD) over concerns of a ‘can’t intubate, can’t ventilate’ (CICV) scenario. This is a controversial area and some workers argue that ‘check ventilation’ lacks rationale and may increase patient risk [48]. It is interesting that NBDs during RSI are classically administered without ‘check ventilation’. Certainly NAP4 recommended that no patient should undergo salvage airway (e.g. cricothyroid access) during CICV without prior administration of NBDs [3].

The choice of rapid onset NBD has evolved from exclusive use of suxamethonium, with the advent of sugammadex offering the potential of rapid and complete reversal of aminosteroid NBDs e.g. rocuronium or vecuronium. A large UK trial demonstrated equivalence in intubating conditions at 50 s between rocuronium 1−1 and suxamethonium 1−1 [83], and a comparison of rocuronium 0.6−1 vs suxamethonium 1.0−1 showed no significant difference in the rate of acceptable intubating conditions in patients for emergency surgery [84]. When compared with rocuronium during RSI, suxamethonium was considered clinically superior by a Cochrane review panel as it had a shorter duration of action [85]. However, this must be considered cautiously for the patient that cannot be woken due to surgical condition or pre-induction physiology (e.g. aortic aneurysm rupture, apnoeic head injury etc.), or contraindications to suxamethonium (e.g. hyperkalaemia, muscle crush injury, malignant hyperthermia etc.).

Sugammadex is a cyclic gamma dextrin with the ability to reverse deep neuromuscular paralysis with rocuronium, and is of interest in this context where airway strategy entails ‘waking the patient up’ and restoring spontaneous ventilation [14]. This topic has undergone review previously [86]. Lee et al. have shown that reversal of neuromuscular blockade by sugammadex (16−1) is significantly faster than spontaneous recovery from low dose suxamethonium (1−1) when given exactly 3 min following rocuronium 1.2−1 [87]. Furthermore, the time to onset of spontaneous ventilation following placement of the tracheal tube was significantly faster following combination of rocuronium (1−1) and sugammadex (16−1) than with low dose suxamethonium (1−1). However, simulation evidence suggests that rocuronium followed by sugammadex ‘salvage’ is unrealistic in current anaesthetic practice [88]: the mean time from decision to administration in a CICV situation was 6.7 min (SD 1.5 min), following which a further 2.2 min would be required to achieve a train-of-four ratio of 0.9. In addition to this, task focus and distraction may lead to a failure to decide or use the sugammadex in a timely fashion, leaving a desaturating patient with profound neuromuscular blockade and no spontaneous ventilation.

The course of action in a rocuronium or suxamethonium ‘CICV’ scenario is the restoration of oxygenation and shortly thereafter ventilation [12]. If circumstances dictate that patients are not appropriate to ‘wake up’ following CICV, they should not undergo cricothyroid access until a NBD has been given, even if this was not the first line plan, as a proportion of these patients may become possible to ventilate. It should be emphasised that restoration of neuromuscular function does not ensure the patient will be able to ventilate, especially if airway obstruction exists [89]. Similarly, following the decision to wake a patient up, if deterioration continues prompt administration of a NBD should be considered. Retrospective data suggest that cricothyroid access using small bore cannula and high pressure ventilation are associated with unacceptable failure rates in delivering both oxygenation and ventilation [3]. Use of an incision and > 4 mm airway, either as small (6.0 mm) tracheal tube or large bore cannula (e.g. QuicktrachTM, Teleflex Medical, Rusch, Durham, NC, USA) enable oxygenation, exhalation and ventilation and can be delivered through conventional ventilation systems (< 60 cmH2O pressure).

Patient monitoring: not just the vital signs

Before emergency anaesthesia, equipment, medications and monitoring should be checked; a revised checklist for anaesthetic equipment is available from the Association of Anaesthetists of Great Britain and Ireland (AAGBI) [90]. Of particular note is that a ‘two-bag test’ should be performed and a self-inflating bag must be immediately available in any location where anaesthesia is conducted. The 2007 AAGBI monitoring guidelines are established and familiar to most anaesthetists (Table 2) [16].

Table 2. AAGBI essential monitoring devices for safe conduct of anaesthesia [16].
  1. NBD, neuromuscular blocking drug

Induction and maintenance of anaesthesia
  Pulse oximeter
  Non invasive blood pressure monitor
  Airway gases: oxygen, carbon dioxide and vapour
  Airway pressure
The following must also be available
  A nerve stimulator whenever a NBD is used
  A means of measuring a patient’s temperature

The UK National Reporting and Learning System (NRLS) data reports over 300 anaesthesia incidents per annum related to equipment and probably reflects under-reporting. Almost 40% of critical incidents concern monitoring, most commonly screen failure, whilst nearly 18% involve problems with IPPV [91]. The recent NAP4 report described a deficiency in the use and interpretation of monitors during anaesthesia outside of theatres [3].

‘Additional monitors’ in the AAGBI guidance may include invasive devices e.g. arterial and central venous pressure monitors, pulmonary artery catheters (PACs) and the range of cardiac output monitors currently available, urethral catheters, continuous temperature monitoring devices and occasionally intracranial or abdominal pressure monitors.

Cardiac output monitoring is increasingly used in emergency surgery, and is also recommended for use in high-risk elective cases; however, the uptake of such technology is slow [92] and clinicians may be guided by experience and anecdote rather than evidence and guidelines [93–95]. Much of the evidence around cardiac output monitoring in emergency anaesthesia has been extrapolated from studies in the critically ill using PACs. The landmark PACMan trial demonstrated no mortality benefit following insertion of PACs in ICU patients (compared with the 80% of patients with an alternative device) associated with a 10% complication rate [96]. A Cochrane systematic review of 12 randomised controlled studies could not demonstrate a beneficial effect of PACs on mortality, cost of care, or hospital length of stay in either patients on general ICUs or a subgroup of high-risk surgical patients [97], and a Health Technology Assessment concluded that ongoing routine costs could not be justified [98]. It is beyond the scope of this article to consider the PAC and cardiac output debate, which has largely been rendered obsolete by clinicians avoiding routine PAC placement. However, many of the failings of routine PAC use are being replicated with alternative ‘semi-’ or ‘non-invasive’ devices and the benefit of routine cardiac output monitoring in the critically ill remains controversial [99].

Oesophageal Doppler monitoring (ODM) for peri-operative fluid optimisation has been evaluated in elective bowel and cardiac surgery, and fractured neck of femur surgery, demonstrating reduced hospital stays [100–102]. It is recommended by NICE for use in patients undergoing major or high-risk surgery, or other surgical patients in whom a clinician would consider using invasive cardiac monitoring [103]. However, there are methodological concerns over its ability to direct care in the presence of vasoactive infusions and in terms of other outcomes beyond hospital stay in elective surgery, the evidence is limited. Similarly, a recent Health Technology Assessment highlighted the need for ongoing research [104]. It is a fallacy to expect any monitor to improve patient outcomes; rather it is the therapy it directs which holds this potential. With this in mind a recent review highlighted the difficulty in identifying the correct type of fluid to use in combination with the ODM and that harm with colloid strategies could not be excluded [105]. Indeed, more recent work could not demonstrate a benefit of ODM guided care and raised the possibility of hydroxyethyl starch related harm [106]. There remains controversy over the use of ODM in clinical practice [107].

Echocardiography is an imaging modality that allows diagnosis and haemodynamic monitoring and recent development of a UK training pathway and imminent inclusion in critical care training curriculum will see its use increase [108]. Echocardiography encompasses two complimentary approaches: transthoracic or (TTE) and transoesophageal (TOE) and has been considered previously [109]. TOE is routinely used during cardiac anaesthesia and lends itself to continuous patient monitoring where access may limit TTE e.g. during a laparotomy. Beyond cardiac output data, it also allows a direct view of cardiac physiology and a number of simplified algorithms to allow haemodynamic assessment and diagnosis exist with courses to support training [110–112]. TOE has been shown to influence diagnostic and management decisions and the range of indications, that include elements of emergency anaesthesia, is expanding [113, 114]. TOE has an expanding role for peri-operative monitoring in high-risk general surgical patients but requires extensive training, and equipment is both expensive to purchase and semi-invasive with finite complications. A second generation of smaller TOE devices, often with restricted scanning planes and limited to 2-D imaging, is becoming available and these are currently undergoing evaluation by NICE (personal communication, 13/09/2012). Currently, however, there is little evidence to confirm the presumed intuitive benefit of echocardiography, and the decision to use the device must be made on an individual basis.

An alternative approach to assessing cardiac output is to infer the ‘adequacy’ of oxygen delivery and identify markers of potential cell ischaemia. A number of devices are available, with increasing supportive evidence for their use, including near infrared spectroscopy [115–118], orthogonal polarised spectroscopy [119, 120], tissue perfusion monitors, sublingual capnography and gastric tonometry, microdialysis catheters [121], and continuous non-invasive blood/biochemical sampling. These devices are best regarded as research based, with very plausible theoretical groundings, but their place in routine management of emergency patients has not been defined. The identification of lactate as a prognostic variable and potential therapeutic target has been considered in our acompanying article. Recent evidence suggests that rapidity of normalisation is likely to infer further prognostic details [122].

Managing the circulation and shock

It is beyond the scope of this article to consider fully all aspects of shock and haemodynamic disturbance that could manifest during emergency anaesthesia, and that are largely specific to the condition e.g. severe sepsis vs traumatic haemorrhage vs peri-operative myocardial infarction. However, as early as the 1970s Shoemaker et al. identified convective oxygen delivery variables associated with outcome in the critically ill, so called ‘survivor values’ (cardiac index > 4.5 l.min−1.m−2, oxygen delivery index > 1000 ml.min−1.m−2 and oxygen consumption index > 170 ml.min−1.m−2) [123, 124]. It was hypothesised that if therapy could ‘drive’ patients to reach these values their outcomes might improve. The interest in strategies including ‘goal directed’, ‘optimisation’, ‘supranormal’ continues and remains controversial. There is consistent evidence that an optimisation strategy can improve outcomes as long as it is delivered before the onset of cell ischaemia and organ failure [125, 126]. Recent evidence for immunomodulating effects are conflicting [127, 128], and some workers have demonstrated other benefits e.g. on gastro-intestinal function [129]. There is equal evidence that such strategies delivered after the onset are largely unsuccessful, making extrapolations from the elective high-risk setting challenging [130, 131]. There remain many controversies within peri-operative optimisation [132–134] even in the elective setting, and studies are ongoing [107, 135], including studies in emergency patients [136].

The benefit of an early (< 6 h) goal-directed strategy in sepsis was suggested by Tuchschmidt et al. and confirmed by Rivers et al., and this was one of the founding interventions of the Surviving Sepsis Campaign resuscitation bundle [54, 137, 138]. This study demonstrated significant benefits through a strategy of optimising haemodynamic status and markers of tissue oxygenation but as a landmark single centre study it requires to be reproduced by other workers, and a number of trials are ongoing [139, 140].

Modern haemodynamic management encompasses contemporaneous pressure (e.g. mean arterial pressure (MAP)) and flow measurements on a global (cardiac output) and regional (e.g. splanchnic) basis. The optimal target for each parameter requires individualisation and rigid recommendations for practice are inappropriate. Most authorities would aim to maintain MAP > 65 mmHg [54, 141, 142], and a small study of critically ill patients targeting a MAP of 85 vs 65 mmHg did not improve clinical outcomes [143]. However, in theatre and in ICU, a MAP higher than 65 mmHg may be appropriate for those with chronic hypertension or compromised end organ perfusion e.g. raised intracranial pressure [144], and peri-operative hypotension has been adversely associated with outcomes in aortic vascular emergencies [145].

However, one of the principal goals of detailed haemodynamic monitoring is to allow individualised vasoactive and fluid therapy. While the principles of inotrope therapy are familiar to all anaesthetists, there is relatively limited evidence to guide this important class of drugs. An excellent review of vasopressor and inotrope use in septic shock has been published recently [146]. Catecholamine infusions remain the mainstay of vasoactive support for most forms of shock refractory to fluid therapy alone. More recent studies of vasoactive therapies are summarised in the Table 3.

Table 3. Summary of main findings in recent vasoactive therapy studies.
StudyComparisonMain findingsComments
Annane D et al. [147]Dual therapy inotrope dobutamine and noradrenaline vs adrenaline monotherapyNo difference outcomes, no difference metabolic conditionsAdrenaline monotherapy established as a treatment option
Russell JA et al. [148]
VASST study
Comparison of escalating doses of combined noradrenaline and vasopressin in refractory shockNo difference2012 meta-analysis of nine trials (includes Russell JA et al.) involving terlipressin or vasopressin with noradrenaline showed vasopressin to be safe and associated with a reduced mortality in adults and those with septic shock [149]
De Backer et al. [150]
Comparison of haemodynamic therapy noradrenaline vs dopamineNo significant difference mortalityHigher rates tachyarrythmias with dopamine and excess mortality in cardiogenic shock

Despite the above studies the optimal selection of inotropic agents remains challenging. In a retrospective series, no patient requiring four or more vasoactive agents survived [151]. Recent interest around dopexamine, with its proposed ability to preserve splanchnic perfusion, has perhaps declined, and its role in peri-operative optimisation remains controversial [152–154]. A major limitation of catecholamine inotropes is positive chronotropy, and elegant theoretical arguments propose that the negative chronotropic actions of beta blockade may be combined with post-adrenoceptor phosphodiesterase inhibitor stimulation of cAMP production and positive inotropy, but the impact on clinical outcomes remains obscure [155]. The landmark POISE trial evaluated the place of beta blockade in elective surgery, concluding that although peri-operative myocardial infarction was reduced, stroke and overall mortality was increased in those given beta blockers peri-operatively [156]. The results of the VASST trial [148] were unable to confirm a benefit to adding (non-catecholamine) vasopressin to escalating doses of noradrenaline in the critically ill, although more recent meta-analysis suggests some outcome benefits for vasopressin and terlipressin as alternative vasopressors [149].

An adjunct to inotropic support in the critically ill entails steroid replacement. Previous studies had established that adrenal insufficiency may be prognostic and supplementation produces faster resolution of shock [157]. The CORTICUS trial, however, established that the short synACTHEN test lacked prognostic value and steroid supplementation may reduce the duration of vasoactive support at the expense of secondary infections [76]. Furthermore, the components of the Surviving Sepsis Campaign management bundles require urgent re-evaluation as tight glycaemic control and drotrecogin alfa have been challenged by more recent studies [158–161], with the latter withdrawn by its manufacturer.

Intravenous fluid therapy – which fluid and how much?

Despite multiple international studies, evidence for the optimal intravenous fluid for peri-operative resuscitation remains elusive, but consensus guidelines do exist [162]. Evidence for restrictive fluid management during elective surgery is conflicting [163–168], and accepting the contrast to the elective setting [169], an individualised strategy is appropriate during emergency anaesthesia. There is an emerging consensus in trauma management that avoidance of excessive non-blood product administration (i.e. synthetic colloid and crystalloid) around a strategy of damage control surgery improves outcome [170, 171]; this has not been established in sepsis management where fluid volumes administered may be much higher, but the correct volume to administer remains ill-defined [172, 173]. However, many workers have suggested benefit to prompt resuscitation following trauma and so ‘permissive hypotension’ and delayed resuscitation must be based around early surgical control to allow more complete resuscitation [174].

The largest trial of crystalloid (0.9% saline) vs colloid (iso-oncotic albumin) was the SAFE study and it remains the definitive study awaiting the results of the CHEST study in late 2012 (J. Myburgh, personal communication) [175, 176]. SAFE failed to demonstrate an overall benefit of colloids but identified a trend to harm with colloid in trauma and benefit in sepsis. However, contrary to classical teaching, administered crystalloid volumes were only 1.4:1 of colloid for comparable clinical effects, questioning fundamental assumptions of fluid compartments, transvascular movement and fluid kinetics; this subject has recently been reviewed [177].

While awaiting the findings of CHEST there is little evidence of superiority over crystalloid and consistent concerns regarding the safety of hydroxyethyl starch suspensions and harm cannot be excluded [105, 159, 178, 179]. Furthermore, much of the HES work is being re-appraised following the retraction of prominent research by Boldt et al. (see

Many researchers and clinicians remain concerned by the entity of hyperchloraemic acidosis following large volume crystalloid resuscitation. The clinical impact of this phenomenon is unclear [180], as it is for many classes of metabolic acidosis [181], but the GIFTASUP guidelines recommend balanced salt solutions to avoid it [162].

The optimal haemoglobin concentration remains a controversial value and an individualised approach is recommended. The landmark TRICC trial suggesting that a restrictive transfusion trigger < 7.0 g.dl−1 was safe and beneficial in established and perhaps ‘milder’ critical illness with most patients having APACHE < 20 [182]; this study forms the basis of a recent Cochrane review [183]. Clinical guidelines for transfusion are available from 2009 for trauma and critical illness [184]. This is markedly different to the dynamic situation in theatres where ongoing haemorrhage, elderly patients and significant comorbidities may make this trigger too low. Indeed, within US critical care units, transfusion practice remained relatively unaffected by lower triggers after publication of the TRICC trial [185]. The recent FOCUS trial of peri-operative transfusion in fractured neck of femur patients at risk of cardiovascular events demonstrated no benefit in transfusion thresholds of 10 g.dl−1 vs 8 g.dl−1 [186]. Extrapolation to bleeding general surgical patients is difficult yet, where feasible, intra-operative cell salvage, anti-fibrinolytic therapy and prompt surgical haemostasis alongside an individualised transfusion trigger are guiding principles. The use of transfused blood stored for > 2 weeks remains UK practice and whether this affects outcomes is unclear, with suggestions of harm with both ‘older’ and ‘fresher’ blood products in differing analyses [187–189]. This remains an important clinical question that the Age of Blood Evaluation trial hopes to answer in due course [190]. The UK Defence Medical Services routinely use blood under this age [191] and while older blood may be perceived to have inferior oxygen transfer capabilities (e.g. depletion of 2,3-diphosphoglycerate) [192] a small study of patients suffering traumatic brain injury demonstrated that transfusion of UK-stored blood improved cerebral oxygenation markers within hours of transfusion [193].

Maintenance of anaesthesia and awareness

Awareness under anaesthesia is a subject that is challenging to consider and has been reviewed previously [194–196]. Concealed within a reported incidence of 0.1–0.2% it is well recognised that during emergency anaesthesia it may be especially prevalent. Factors increasing the risk of awareness include ASA 3-4 status, children, previous substance abuse, obstetrics, lack of benzodiazepine or neuromuscular blockade and ‘out of hours’ and emergency cases [194]. Estimates of awareness are further undermined by lack of routine post-operative screening. Reduced anaesthetic doses (in the face of haemodynamic instability) are probably the most common cause and of note, in 23% of cases no volatile or intravenous anaesthetic was administered, although this review found no association with NBDs. Most awareness was during maintenance, with only 4.5% reported at induction or with prolonged laryngoscopy [197].

No ‘ideal’ depth of anaesthesia monitor exists and it is perhaps troubling that most anaesthetists spend their working lives administering titrated drugs with no target organ monitoring. The various techniques and technologies available include the isolated forearm technique, and both electroencephalograph-derived and auditory evoked potential-derived signals; these have been reviewed previously [198]. Their effectiveness and role in preventing awareness are controversial and affected by the anaesthetic technique or agents employed. A recent study of electroencephalographic bispectral index monitoring to guide general anaesthesia in a population at high risk of awareness did not reduce this incidence [199] in direct contrast to the B-AWARE study [200]. Depth of anaesthesia monitors are subject of current guidance in development from NICE [201].

The use of end-tidal volatile anaesthetic concentrations has advantages and 0.7 minimum alveolar concentration (MAC50) is considered a level where awareness is highly unlikely [202], although the exact dose cannot be known [195] and some workers suggest 0.5 MAC [203]. In the study by Avidan et al., 75% of patients were managed at < 0.7 MAC with no reported awareness [199]. Nonetheless, regardless of haemodynamic instability, there is broad consensus not to reduce volatile anaesthetic < 0.7 MAC, and this would seem reasonable in emergency and non-emergency settings. Currently, the fifth UK National Audit Project NAP5 is collecting data on accidental awareness in association with general anaesthesia and its findings are eagerly awaited [204].

Postoperative care – the problematic recovery

There is increasing recognition of postoperative cognitive dysfunction (POCD) and delirium, especially amongst the elderly. These are important conditions that consistently worsen patient outcomes, but are, in practice, difficult both to avoid and to manage [205]. The impact of general anaesthesia on these conditions, relative to regional anaesthesia, is controversial, with conflicting evidence that general anaesthesia produces more short-term delirium and confusion, and little evidence that these changes are significantly different at six months [206–210]. Anaesthetists must be mindful of the high risk that emergency surgical patients have for developing delirium and POCD, and the negative impact this has on outcome, although many contributing factors, e.g. age, use of opioids etc. may be challenging to modify [211, 212].

Competing interests

No external funding or competing interests declared.