Patients presenting with vascular emergencies including acute aortic syndrome, ruptured thoracic or abdominal aortic aneurysms, thoracic aortic trauma and acute lower limb ischaemia have a high risk of peri-operative morbidity and mortality. Although anatomical suitability is not universal, endovascular surgery may improve mortality and the results of ongoing randomised controlled trials are awaited. Permissive hypotension pre-operatively should be the standard of care with the systolic blood pressure kept to 50–100 mmHg as long as consciousness is maintained. The benefit of local anaesthesia over general anaesthesia is not definitive and this decision should be tailored for a given patient and circumstance. Cerebrospinal fluid drainage for prevention of paraplegia is often impractical in the emergency setting and is not backed by strong evidence; however, it should be considered postoperatively if symptoms develop. We discuss the pertinent anaesthetic issues when a patient presents with a vascular emergency and the impact that endovascular repair has on anaesthetic management.
In-hospital mortality after open repair of ruptured abdominal aortic aneurysm (rAAA) approaches 40–50% with total mortality, including pre-hospital deaths, closer to 80% [1–3]. Traumatic thoracic aortic injury is often immediately fatal with a 30–40% mortality in those who arrive alive to hospital and an overall mortality of 90% . Thirty-day mortality rates are > 50% following ruptured thoracic aneurysm , 13% for acute type B dissection  and > 25% following embolectomy for acute limb ischaemia [7, 8].
We define a ‘vascular emergency’, for the purpose of this review, as an acute condition requiring intervention within hours, involving the descending aorta and its main lower limb branches. Specific pathologies meeting these criteria include acute aortic syndrome, ruptured thoracic and abdominal aneurysms, traumatic injuries of the thoracic aorta and acute lower limb ischaemia. Pathology involving the ascending aorta or aortic arch that is generally managed by cardiac surgeons is excluded.
This review aims to discuss the optimal anaesthetic management to minimise morbidity and mortality and the impact of a move towards endovascular repair. An English language electronic search using Medline and Embase databases identified relevant scientific literature from the last 10 years and was supplemented with a manual search of reference lists from reviewed articles.
Overview of vascular emergencies
Acute aortic syndrome
Acute aortic syndrome consists of three interrelated conditions: penetrating atherosclerotic ulcer; intramural haematoma; and aortic dissection. Penetrating atherosclerotic ulcers and intramural haematoma are treated as for type-B dissections with the usual caveats for surgical intervention . Aortic dissection occurs when an intimal tear allows blood under pressure to separate the layers of the aorta. The result is a ‘true lumen’ and ‘false lumen’ separated by a dissection flap. The Stanford classification defines dissection along treatment lines. Type A dissections (∼60%) involve the ascending aorta, proceed to emergency surgery and are not discussed in this review. Type B dissections (∼40%) are limited to the descending aorta or aortic arch and emergency surgery is considered for those with end-organ ischaemia, intractable pain, uncontrollable blood pressure or rupture .
The Crawford classification of rTAAA is based on the extent of aortic involvement (Fig. 1) that affects both the intra-operative and postoperative course. Emergency surgery for thoraco-abdominal aortic aneurysms is generally reserved for cases of rupture.
Thoracic aortic trauma
Thoracic aortic trauma often occurs during a motor vehicle accident, as sudden deceleration results in movement of the relatively mobile aortic arch on the descending aorta, in the peri-isthmic region (Fig. 2). Polytrauma in this setting adds controversy to the timing of repair – on the one hand expeditious surgery minimises the risk of rupture; however, systemic heparinisation, aortic cross clamp and lateral positioning can exacerbate co-existing injuries. Alternatively, delaying repair until other injuries are stable risks aortic rupture and controlled hypotension in the interim can have deleterious consequences, especially in patients with co-existing head injury. Growing interest in early endovascular treatment shows evidence of improved outcomes with this approach .
Ruptured abdominal aortic aneurysm
Most patients with rAAA who reach hospital alive have a contained retroperitoneal haematoma with tamponade effect. Immediate death usually results from free intra-peritoneal rupture [1, 11, 12]. Open operations on the abdominal aorta proceed through a midline laparotomy.
Acute lower limb ischaemia
Acute limb ischaemia is characterised by a pale, pulseless, paralysed, parasthetic and/or painful limb. Most commonly the cause is embolic or thrombotic [13, 14]. The Rutherford classification  (Fig. 1) provides a common language when discussing severity of acute and chronic limb ischaemia. Acute limb ischaemia is distinct from ‘critical limb ischaemia’, a severe form of chronic ischaemia that includes gangrene and rest pain. Restoration of flow in an acutely ischaemic limb can be attempted by surgical revascularisation (embolectomy, endarterectomy or bypass grafts) or pharmacological dissolution of thrombus [14, 16]. Fasciotomy (if there are signs of compartment syndrome) or even amputation may be necessary.
Endovascular aortic reconstruction (EVAR), first reported in 1991 , was quickly adopted for patients with leaking  and rAAAs . A completely percutaneous modification (avoiding arteriotomy) has been reported in the acute setting . An aorto-uni-iliac device is easier and quicker to deploy and requires only one suitable iliac artery ; however, it does necessitate a femoro-femoral bypass graft to provide contralateral lower extremity blood flow.
Thoracic endovascular aortic reconstruction (TEVAR) produces additional considerations. Stent coverage of the left subclavian artery, which supplies antegrade flow to the ipsilateral arm and left vertebral artery (Fig. 2), may be problematic. If the left subclavian artery is covered, the left arm receives retrograde flow from the left vertebral artery potentially resulting in a steal phenomenon  characterised by dizziness and visual changes. Patients with a left internal thoracic artery (LITA) to coronary artery bypass graft may suffer myocardial ischaemia as the LITA comes off the left subclavian artery. If the proximal landing zone necessitates coverage of the left subclavian artery, revascularisation should be considered using, for example, a carotid to left subclavian artery graft . Temporary apnoea and hypotension may be required to ensure precise positioning of the graft. Asystole induced by adenosine or rapid transvenous pacing were previously required, but modern self-expanding stents require only a modest drop in blood pressure, which can be achieved with a bolus of propofol, nitroglycerine or esmolol.
Rapid control techniques
Some authors strongly recommend a ‘clamp before you cut’ approach  using a transfemoral aortic occlusion balloon before laparotomy to reduce the haemodynamic disturbance following release of abdominal wall tamponade. In a small series, Philipsen et al. demonstrated that by transferring unstable patients with rAAA directly to the operating room and placing an aortic occlusion balloon under local anaesthesia, they could then assess suitability for EVAR using angiography. Overall 30-day mortality was only 17% despite significant pre-operative instability .
Suitability for EVAR
Approximately, 40–50% of patients with rAAA have suitable anatomy for endovascular repair [25, 26]. Adequate proximal and distal landing zones and appropriately sized groin access vessels are required. Although transporting a potentially unstable patient presents some risk, a pre-operative CT helps determine suitability. Lloyd et al. performed a ‘time to death’ study in patients with rAAA who reached hospital alive but did not undergo surgery for various reasons. They reported that 87.5% of patients survived more than two hours which is usually sufficient time to undertake a CT scan . Endovascular stenting is less desirable in patients with Marfan’s syndrome and other connective tissue disorders as long-term durability in this subset of patients is uncertain .
Is endovascular surgery the answer to reducing mortality?
In an elective setting, three randomised trials compared EVAR with open AAA repair, showing an early mortality benefit for EVAR [29–31] that was not sustained at 2 to 4-year follow-up [31–33]. Following aortic rupture, endovascular surgery has shown promise in reducing mortality. Mortality rates following EVAR for rAAA are less than 20% in several non-randomised studies [1, 21, 34–37] and 17–30.2% in other large meta-analyses [2, 38–41]. However, randomised controlled trial (RCT) data are currently lacking. The only prospective RCT to date was terminated early, published as a pilot study and showed no difference in 30-day mortality . Three RCTS are currently underway – the AJAX trial, ECAR trial  and the multicentre IMPROVE trial; all are yet to publish their results.
For thoracic aortic emergencies, an endovascular approach shows favourable outcomes in non-randomised studies [44–47]. For uncomplicated aortic dissection, medical management is traditionally preferred; however, long-term outcomes are poor, with 5-year mortality approaching 50% . This tradition may be challenged with improved outcomes for endovascular treatment. Xu et al. used TEVAR in 63 patients with acute type B dissections, and showed a 30-day mortality of 3.2% and 4-year survival of 89.4% .
A mortality benefit not withstanding, EVAR avoids aortic cross clamping and abdominal incision, allows proximal control without general anaesthesia , and shows reductions in procedure time, blood loss, intensive care unit (ICU) and hospital stay [35, 38, 50]. EVAR for thoracic emergencies avoids the need for a high aortic cross clamp and one lung ventilation.
There are certain complications specific to endovascular repair including graft malpositioning, iliac artery rupture and retrograde conversion of a type B dissection to a type A dissection [9, 48]. Endoleaks and postimplantation syndrome cause postoperative morbidity.
Anaesthetic management of vascular emergencies
One of the major challenges is the inadequate time to evaluate and optimise patients who present with a vascular emergency. Especially for rAAA, the aim is to minimise door to cross clamp time; however, where possible (i.e. the patient is conscious) a brief targeted assessment should address allergies, medications, cardiac history, routine blood tests, an arterial blood gas and ECG. It is clear that vascular patients are amongst those with the highest peri-operative risk. The American College of Cardiology/American Heart Association 2007 peri-operative cardiovascular evaluation and care for non-cardiac surgery guidelines  and 2009 update  rank major vascular surgery as high-risk even in the elective setting. Atherosclerosis affects vessels systemically and the anaesthetist should assume involvement of the coronary, cerebral and renal circulations.
Aortic dissections occur in a slightly different and younger population with connective tissue and congenital conditions playing a greater role. These conditions present anaesthetic implications beyond the vascular emergency (Table 1). Thoracic aortic trauma often occurs in young patients , often with other life-threatening injuries.
Table 1. Clinical features and anaesthetic implications of conditions that predispose to aortic dissection.
Careful positioning (risk of joint dislocation) Bag-mask ventilation may be difficult Longer laryngoscope blades may be required Failed spinal anaesthesia and increased risk dural puncture with epidural (dural ectasia) Restrictive lung function (chest wall abnormalities) Spontaneous pneumothorax possible
Triad: arterial aneurysms, hypertelorism & cleft palate or bifid uvula Also: Craniosynostosis Retrognathia Abnormal cervical spine Skeletal features similar to Marfan’s Developmental delay
Difficult airway management Development delay may preclude local anaesthesia alone
For aortic rupture, permissive hypotension should be a standard of care with the aim of keeping the systolic blood pressure 50–100 mmHg as long as consciousness is maintained [1, 36, 57]. Excessive fluid administration, in an attempt to normalise blood pressure, dilutes coagulation factors, disrupts thrombus and can cause expansion and rupture of a contained retroperitoneal haematoma . For dissection, aortic wall stress should be minimised by keeping the heart rate <60 beats.min−1 and systolic blood pressure 100–120 mmHg [9, 58, 59]. Intravenous beta-adrenoceptor antagonists are ideal initial treatment and if systolic blood pressure remains more than 120 mmHg after adequate heart rate control, vasodilators are then added. Adequate pain control is an important adjunct.
Preparation of equipment and operating theatre
The location where anaesthesia will take place (operating theatre or interventional radiology) should be prepared with a similar check-list to that in Table 2. The establishment of monitoring is in part dictated by the time available before surgery. Attachment of an ECG, pulse oximeter and non-invasive blood pressure (NIBP) cuff causes no delays. Pre-induction invasive arterial monitoring is optimal; however, if not rapidly achieved, it can be obtained after the aorta is cross clamped, with NIBP measurements taken every minute in the interim. The right arm is usually preferable since an aortic occlusion balloon placed via the left brachial artery or the left subclavian artery may be involved in a type B dissection or require coverage by an endograft. For most thoracic aortic emergencies, two arterial lines are ideal – one in the right arm and one ‘post-subclavian’ either in the left upper limb or a femoral artery.
Table 2. Suggested preparation and anaesthetic equipment for a vascular emergency.
Equipment & preparation
Standard monitors (NIBP, pulse oximeter, five lead ECG) Arterial, central ± pulmonary artery catheter, lines + transducers Temperature probe Urinary catheter
Alert blood bank, cross match minimum four units packed red blood cells Rapid infusor Fluid warmers (×2 ideal) Cell saver
Induction drugs (consider ketamine) Metaraminol or phenylephrine boluses Dilute adrenaline i.e. 4 microg.ml−1 for boluses Atropine Calcium chloride (1 g in 10 ml × 1–2) Sodium bicarbonate (50 ml 8.4% solution × 1–2) Insulin and dextrose (hyperkalaemia treatment) Nitroglycerin (to treat clamping related hypertension) Vasopressor and inotrope infusions (noradrenaline, adrenaline etc.)
Forced air warmer Double lumen tracheal tube (should be available for thoracic aortic surgery including TEVAR in case of rupture)
Transoesophageal echocardiography Depth of anaesthesia monitor (given the high risk of intra-operative awareness)
Use of central venous access is determined on a case-by-case basis. A pulmonary artery catheter may be considered and whilst RCT evidence suggested no added mortality resulting from pulmonary artery catheter guided treatment, equally no benefit was seen . Intra-operative transoesophageal echocardiography (TOE) improves detection of regional wall motion abnormalities, and helps guide management of volume resuscitation and haemodynamic changes during clamping and unclamping. TOE can be used to confirm left atrial cannula position during left heart bypass and assists endograft positioning; the true vs false dissection lumens can be identified, and detection of the endovascular guidewire within the false lumen using TOE  can avert disaster.
Single lung ventilation is required for open thoracic aortic repairs. Pathology of the thoracic aorta may distort or compress the left main bronchus, making placement of a left-sided double-lumen tracheal tube difficult or hazardous . A right-sided double-lumen tube or bronchial blocker can be used as an alternative. If postoperative ventilation in ICU is planned, conversion to a single-lumen tube is performed before transfer.
General vs local anaesthesia
Open emergency aortic operations are traditionally performed under general anaesthesia. However, for endovascular procedures, particularly those with percutaneous groin access, local anaesthesia ± sedation has become a viable option. Neuraxial techniques have been described for emergency procedures ; however, surgical urgency, limited patient co-operation and thrombolysis often preclude their use. If neuraxial techniques are used, timing in relation to anticoagulation and antiplatelet agents must be considered .
When a rAAA proceeds under general anaesthesia, it is usual to prep and drape the patient before induction to minimise time to clamp after loss of abdominal wall tone. Ketamine achieves rapid blood-brain equilibration and acts as a sympathomimetic, making it an ideal agent for a rapid sequence induction in a haemodynamically compromised patient . Embolectomy and other lower limb revascularisation procedures can proceed under local infiltration, unless fasciotomy is required. For EVAR or TEVAR, local anaesthesia can be used for the initial part of the procedure, followed by induction of general anaesthesia after the aorta is secured and before exposure of the contra-lateral femoral artery and femoro-femoral crossover, if required. Several studies have proven the feasibility of performing EVAR for rAAA under LA [36, 50, 66, 67]. Table 3 outlines the benefits vs disadvantages of each option.
Table 3. Benefits and disadvantages of local anaesthesia (LA) vs general anaesthesia (GA) for emergency vascular operations.
Extensive groin dissection, or femoro–femoral bypass precludes LA alone Patient co-operation may be limited Pain from retroperitoneal haematoma, tourniquet or lower limb ischaemia can be problematic Transoesophageal echocardiography cannot be used Acid-base derangements following reperfusion are challenging to manage in a spontaneously breathing patient
Apnoea, induced hypotension and patient immobility during stent deployment more easily achieved Transoesophageal echocardiography can be used Rapid conversion to open repair if required Optimal surgical conditions
Loss of tamponade effect of abdominal muscle tone Potential for GA-induced cardiovascular collapse
Is local anaesthesia a better choice?
For elective vascular surgery, several studies have shown a reduction in cardiac morbidity [68, 69] and a dramatic reduction in vascular graft occlusion [68, 70, 71] linked to regional anaesthesia. Elective EVAR using local anaesthesia has demonstrated benefits including reduced ICU and hospital stay [72–75] and decreased pulmonary morbidity . A meta-analysis of 39 articles and 2387 patients undergoing elective EVAR found that general anaesthesia was an independent risk factor for death, with an odds ratio of 5.1 (95% CI 1.9–13.3) .
In the emergency vascular literature, Karkos et al. reported in a small study that local anaesthesia for EVAR for rAAA was a significant predictor of survival . However, in another study, patients presenting with acute lower limb ischaemia who had their procedure under local anaesthesia (41%) had worse outcomes as an anaesthetist was not present and these two factors are likely to be linked . Once again, level one evidence is lacking. A given technique must be appropriately applied for the patient and circumstances. Meticulous attention to detail within the chosen technique is arguably more important than the technique itself.
Blood loss and coagulation management
If staffing allows, one anaesthetist should be assigned to manage fluids, blood products and blood samples to guide transfusion strategy. Most open operations on the aorta result in considerable blood loss secondary to aortic rupture, disruption of retroperitoneal vessels during dissection and coagulopathy. Coagulopathy develops due to hypothermia, dilution, acidosis and fibrinolysis induced by the aortic cross clamp. Lysine analogues (i.e. tranexamic acid) are sometimes used to minimise fibrinolysis based on the strong evidence seen in cardiac surgery. However, use in aortic surgery without cardiopulmonary bypass does not have the same body of evidence .
In a patient with ongoing bleeding, our current practice is to maintain their haemoglobin concentration > 9 g.dl−1, INR < 1.5, platelets > 100 × 109.l−1 and fibrinogen > 2 g.l−1. Laboratory-based coagulation tests (with inherent delays) or point-of-care tests such as thromboelastography (TEG) can be used. A study from Denmark compared a standardised transfusion strategy in patients having open repair of rAAA, with historical controls. A ‘transfusion package’ comprising packed red blood cells, fresh frozen plasma and platelet concentrates was combined with TEG analysis to guide transfusion therapy . There was a reduction in 30-day mortality from 56% (control group) to 34% (intervention group) (p = 0.02) at the expense of increased transfusion. Cell salvage should be considered when anticipated blood loss is > 1000 ml . A retrospective study of cell salvage during rAAA found that allogenic blood use was reduced by an average of three units per patient . A recent Cochrane review concluded that cell salvage reduced peri-operative allogenic blood transfusion although other clinical outcomes were not altered , but this was not specific for vascular surgery.
For open aortic repairs, the anaesthetist must be prepared to manage the haemodynamic effects of aortic cross clamping and unclamping as outlined in two excellent reviews [82, 83]. Venodilators (e.g. nitroglycerin) are useful for clamping-related hypertension to increase venous capacitance, reduce preload  and permit fluid loading in preparation for unclamping. Arteriolar dilators (e.g. nitroprusside) should be used with caution as the reduction in both distal and proximal aortic pressure can further affect perfusion distal to the clamp . An increased incidence of spinal cord injury attributed to nitroprusside is probably due to a combination of decreased distal aortic pressure and increased cerebrospinal fluid (CSF) pressure secondary to cerebral vasodilation [82, 83].
Preparation for unclamping includes volume replacement, aggressive treatment of acidosis and close communication between surgeon and anaesthetist. Venodilators should be discontinued and the anaesthetist should be prepared to treat with calcium, bicarbonate, hyperventilation, vasopressors and fluid. By gradually releasing the cross clamp and/or reperfusing the lower limbs one at a time, the washout of cardio-depressant mediators is slowed, re-oxygenation is delayed and the production of oxygen free radicals is decreased . If unclamping is poorly tolerated, reclamping can temporarily stabilise the patient. Reperfusion following an ischaemic insult (i.e. embolus) should be managed as for unclamping.
Normothermia should be maintained with warming beginning before induction. Warming fluid and blood products, increasing ambient temperature and forced air-warming blankets can minimise heat loss. Active warming of the legs during cross clamp must be avoided as limb ischaemia will be exacerbated. A retrospective study of patients with rAAA compared survivors with non-survivors, with survivors showing higher mean temperature on arrival to the operating theatre and higher intra-operative and end-of-procedure temperatures .
Strategies to reduce morbidity
The incidence of renal failure after rTAAA repair is reported to be between 4% and 40%  and 16–26% following rAAA repair [3, 85, 86]. Prolonged clamp time, hypotension, decreased cardiac output, anaemia, high contrast load in endovascular techniques and pre-existing renal impairment [86, 87] are important risk factors. Myoglobinuria from rhabdomyolysis, iatrogenic ureteric injuries and nephrotoxic drugs contribute to renal dysfunction.
There is no level-one evidence to support the use of N-acetyl-cysteine , dopamine [89, 90], loop diuretics or mannitol  to prevent renal impairment. The main protection goals are maintenance of the balance between renal oxygen supply and demand with normovolaemia, adequate cardiac output and haemoglobin concentration. Suprarenal clamping duration should be minimised and cold renal perfusion to reduce the metabolic demands of the kidney may be a useful adjunct [5, 92].
Spinal cord protection
Spinal cord injury occurs in 2–10%  of elective operations on the thoracic aorta with a higher incidence in the emergency setting . Prevention focuses on maintenance of spinal cord perfusion pressure (maintenance of mean arterial pressure, CSF drainage, left heart bypass [94, 95]) and increasing ischaemic tolerance (steroids, hypothermia.) Two RCTs have shown that CSF drainage reduces paraplegia in patients having elective TAAA repair [96, 97] and another failed to demonstrate a difference . A meta-analysis comprising these three RCT and 11 additional cohort studies supported CSF drainage as an adjunct to prevent paraplegia . However, a Cochrane review  of the RCTs alone concluded that there are limited supportive data.
Complications of CSF drainage include headache, persistent CSF leak, subdural or epidural haematoma and meningitis. Wynn et al. reported a 1% incidence of neurological deficit and 0.6% mortality related to CSF drainage . Dardik et al. reported the incidence of symptomatic subdural haematomas to be 3.5% with a mortality of 50% . These risks may be reduced by limiting the volume of CSF removed and avoiding postoperative drainage in asymptomatic patients [101, 102]. If a CSF drain has not been inserted pre-operatively, as is often the case in the emergency setting, it can still be placed postoperatively if symptoms develop, which may reverse paraplegia [103–106].
Most patients require ICU care postoperatively with monitoring for complications, further rewarming and ongoing treatment of coagulopathy. Persistent hypotension may be caused by adrenal insufficiency resulting from interruption of adrenal blood supply (suprarenal clamping), adrenal haemorrhage , or a systemic inflammatory response causing down-regulation of the hypothalamic-pituitary-adrenal axis. An unexpectedly high incidence was seen in one small study .
Abdominal compartment syndrome (intra-abdominal pressure > 20 mmHg with new organ dysfunction)  has a 4–12% incidence following open repair of rAAA . Some endovascular series have shown higher rates  due to an inability to evacuate retroperitoneal haematoma. Risk factors include hypotension, hypothermia, acidosis, aggressive fluid/blood product resuscitation, anaemia [1, 111, 112] and the need for an aortic occlusion balloon . Some authors recommend postoperative monitoring of intra-abdominal pressure via bladder pressure [1, 113, 114] or even surveillance colonoscopy in patients at increased risk [111, 115]. Others suggest delayed abdominal closure in patients with multiple risk factors .
Delirium is a well-recognised complication after vascular surgery that affects over 30% of patients [116, 117] and is associated with prolonged stay and higher morbidity and mortality [118, 119]. Independent predictors of delirium include older age, history of cerebrovascular accident and pre-operative beta-adrenoceptor antagonist administration . Endovascular repair of AAA was associated with significantly lower postoperative delirium when compared with an open approach, irrespective of age .
For patients who present with a vascular emergency, perioperative morbidity and mortality remains alarmingly high. The increasingly-used endovascular approach shows promise to reduce overall mortality, although, whatever the impact on mortality, endovascular surgery has created additional considerations for the anaesthetist including the potential to proceed under local anaesthesia alone.
The authors acknowledge the assistance of Marina Englesakis, UHN Health Sciences Library, for her assistance with the literature review.
No external funding or competing interests declared.