Acute respiratory distress syndrome (ARDS) is characterised by acute inflammation affecting the gas exchange surface of the lung, presenting clinically with acute hypoxaemia, in the presence of bilateral pulmonary infiltrates on chest radiography.
It has to be made clear from the outset that ARDS is not a disease. It is a state of dramatically diminished lung function of variable aetiology. A working definition of ARDS was established in 1994 by the American-European Consensus Conference. A more recent report recommends use of definitions of mild, moderate and severe ARDS, based on the degree of hypoxaemia (Table 44.1). Additionally, the exclusion of pulmonary oedema secondary to cardiac failure is now not mandated, positive end expiratory pressure (PEEP) is accounted for as an indicator of severity, and a known ARDS risk factor must be present within 7 days of onset.
|Mild ARDS||Moderate ARDS||Severe ARDS|
|PaO2:FiO2||201–300 mmHg||≤200 mmHg||≤100 mmHg|
|PEEP||PEEP ≥ 5 cmH20||PEEP ≥ 5 cmH20||PEEP ≥ 10 cmH20|
|Chest X-ray||Bilateral opacities||Bilateral opacities||Bilateral opacities|
PaO2/FiO2 arterial partial pressure of oxygen/inspired oxygen fraction.
PEEP positive end expiratory pressure.
ARDS develops after exposure to a wide variety of insults, and given the nature of the diagnostic criteria, should be considered a syndrome rather than a disease. Initiating insults can be divided into two groups: direct (e.g. pneumonia) or indirect (e.g. sepsis) as outlined in Table 44.2. Risk of progression to ARDS varies according to the type, number and severity of predisposing conditions, as well as the genetics and other patient characteristics including gender, body mass index, smoking status, and alcohol usage. Diagnosing and treating conditions that mimic, or are associated with, ARDS is the first principle of successful management; prediction scores such as the Lung Injury Protection Score (LIPS) are intended to facilitate early recognition and treatment.
|Indirect ARDS||Direct ARDS|
Incidence and Outcomes
Data from the late 1990s determined that the age adjusted incidence of ARDS was 86.2 per 100,000 person-years, with an in-hospital mortality rate of 38.5%. Other data suggest that the incidence of ARDS may have declined since the 1990s. The decline in incidence has been attributed to improvements in healthcare delivery and process, including adherence to low tidal volume ventilation, and early recognition and management of sepsis. More recent data bring the reported declines into question, and suggest that ARDS occurs in 10.4% of all ICU admissions, with a hospital mortality of 34.9–46.1%, although these data are not without controversy.
Once treatment of the precipitating condition has commenced, prognostic factors relate to the patient’s response to therapy. Severe arterial hypoxaemia (PaO2/FiO2 < 100 mmHg) and an increase in the pulmonary dead-space fraction (>0.60) are associated with increased mortality. Mortality also correlates with the number of organ system failures, increasing to 83% when three or more are present. Complications appearing during the course of ARDS, including circulatory shock, acute renal failure and liver dysfunction, allied with age over 60 years, are associated with a higher mortality. Lung function in most survivors returns to normal over 6–12 months, although the majority have persistent, abnormal exercise endurance, and neuromuscular and neurocognitive morbidity significantly impairs longer term health related quality of life.
The alveolar epithelium is composed of approximately equal numbers of flat type I cells (hAT1) and cuboidal type II cells (hAT2). hAT2 have several critical functions, including surfactant production, ion transport and functioning as progenitor cells for the regeneration of hAT1 cells after injury. Typical histological appearances of ARDS include extensive necrosis of hAT1 and the formation of protein-rich hyaline membranes on a denuded basement membrane. The extent of alveolar epithelial damage is a predictor of outcome.
Alveolar fluid clearance by hAT2 is primarily driven through sodium uptake on the apical membrane, followed by extrusion of sodium on the basolateral surface by Na+K+-ATPase. Loss of alveolar epithelial integrity and down regulation of sodium and chloride transporters results in the accumulation of protein-rich and highly cellular oedema fluid in the interstitium and alveoli. Loss of surfactant producing hAT2, allied with the effect of plasma proteins in the airspace, contributes to ARDS through atelectasis, increased oedema formation and impairment of local host defence. Clinically this presents with collapse of lung units and reduced pulmonary compliance. Epithelial cells also play key roles in regulating the inflammatory response through production of injury-driven pro-inflammatory cytokines and chemokines, the expression of leucocyte adhesion molecules, and cell–cell interactions with resident lung cells, particularly alveolar macrophages.
The pulmonary endothelium forms a continuous barrier of endothelial cells, which regulate fluid permeability as well as modulating host inflammation, vascular tone, angiogenesis and interactions with blood-borne cells. Loss of barrier integrity, characterised by the formation of reversible intercellular gaps between endothelial cells, is accepted as the ultrastructural basis for the pulmonary oedema observed in ARDS. Gap formation is induced by the binding of mediators, including thrombin and tumour necrosis factor-alpha (TNF), which induce cytoskeletal rearrangement and endothelial barrier disruption.
Similar to the alveolar epithelium, the lung endothelium orchestrates and propagates the inflammatory response. Endothelial cells release cytokines and chemokines, up regulate the expression of adhesion molecules, and shift from an antithrombotic to a pro-thrombotic activated state, resulting in capillary thrombosis and extravascular fibrin deposition, potentiating pulmonary inflammation and contributing to the increased dead-space fraction observed in ARDS. Inflammation in the vascular space also counteracts hypoxic pulmonary vasoconstriction, partly by causing dysregulation of the production of vasoactive mediators including prostanoids, endothelins and nitric oxide.
The recruitment of circulating inflammatory cells into the lung has long been recognised in ARDS. Neutrophils are central to the initiation and propagation of the inflammation observed, and neutrophilic alveolitis is a histological hallmark of ARDS (see Figure 44.1). The extent of neutrophilia present within the bronchoalveolar lavage fluid (BALF) of patients with ARDS has been reported to correlate with clinical outcome. However, the presence of neutrophils per se is not damaging, rather the priming/activation status of these cells is the major determinant of their subsequent injury-inducing behaviour. Recent data have shown that the healthy human pulmonary endothelium plays a role in host defence by trapping primed/activated neutrophils, facilitating their depriming, and later releasing them back into the systemic circulation in a quiescent state. Failure of this homeostatic depriming mechanism was observed in patients with ARDS.
Figure 44.1 Bronchoalveolar lavage fluid from a patient with ARDS. Photomicrograph of modified Wright’s stained cytospin. The white arrow denotes a hypersegmented neutrophil; the black arrow denotes a neutrophil which has been efferocytosed by a macrophage.
Monocytes appear to play a role in regulating neutrophil influx to the lung. However, whilst direct depletion of monocytes in mice consistently reduced LPS-induced blood and alveolar neutrophilia, as well as lung injury, peripheral mononuclear cell depletion in humans was unsuccessful in preventing the recruitment of neutrophils into the alveolar space.
Ventilator Associated Lung Injury
The application of mechanical ventilation exacerbates ARDS in a process called ventilator-associated lung injury (VALI). Mechanical forces applied during ventilation cause physical disruption of the alveolar-capillary membrane leading to pulmonary oedema, whilst cyclical stretch induces activation of cell- signalling pathways in epithelial, endothelial and inflammatory cells. In the context of the already injured lung, this results in pro-inflammatory and/or pro-fibrotic responses both locally and in the systemic circulation. Thus, in addition to local injury, VALI can drive systemic inflammation and extrapulmonary organ damage. This is reflected in the observation that the majority of patients with ARDS die from multisystem organ failure, rather than hypoxaemic respiratory failure.
Resolution of Inflammation and Repair of the Injured Lung
Most patients gradually recover normal physiology and lung function. In the alveolar epithelium, hAT2 cells proliferate in response to stimulation by epithelial growth factors. hAT2 cells are thought to act as progenitor cells for both daughter type II cells and type I cells. Local and bone marrow derived stem cells may also contribute to repair. Resolution of inflammation is macrophage and T-cell driven. Collectively, these mechanisms combine to reconstitute the epithelial barrier and restore lung function.
If epithelial injury is severe, or repair is impaired, a fibroproliferative phase of ARDS can ensue either following, or in parallel with, epithelial injury. During this phase, mesenchymal cells proliferate, neovascularisation occurs, and the alveolar space becomes filled with activated fibroblasts and myofibroblasts that synthesise excessive collagenous extracellular matrix. A small proportion of patients progress to a chronic phase of respiratory insufficiency characterised by widespread pulmonary fibrosis, with disordered lung architecture.
The main clinical priority is identification and treatment of the initiating pathology and any complications that may have ensued. It may be necessary, particularly in the immune compromised host, to undertake invasive diagnostic procedures, such as bronchoscopy, to establish a diagnosis. Further, it has been shown that intercurrent pulmonary infection occurs in up to 70% of patients with ARDS, necessitating vigilance for the subsequent development of infection.
The magnitude of the clinical burden of VALI was demonstrated by the ARDS Network study, in which patients with ARDS were randomised to receive either a high tidal volume (12 ml/kg predicted body weight (PBW)) with end inspiratory pressure limited to a plateau pressure Pplat ≤ 50 cmH2O, or a low tidal volume (6 ml/kg PBW) and Pplat ≤ 30 cmH2O. A PEEP ladder was used to determine the PEEP level administered, according to the fraction of inspired oxygen and the respiratory rate. The low tidal volume cohort demonstrated a 9% absolute reduction in mortality (40% to 31%). Accordingly, low tidal volume ventilation has become the standard of care in ARDS, and the goal of mechanical ventilation has shifted from normalisation of gas exchange parameters to minimising VALI with pragmatic acceptance of modest biochemical derangement.
High PEEP and Recruitment Manoeuvres
The application of PEEP has been used to mitigate the pulmonary oedema and atelectasis of ARDS. Several studies have failed to show clinical benefit of higher PEEP levels; nonetheless, a meta-analysis demonstrated improved survival in those with most severe physiological derangements (PaO2:FiO2 < 200 mmHg). Despite these data, consensus for this approach is lacking, particularly as response is heterogeneous and may be associated with lung hyperinflation and haemodynamic compromise. Determination of ‘optimal PEEP’ levels has thus far been elusive; calculation of driving pressure, estimation of transpulmonary pressures using an oesophageal probe and imaging techniques have shown promise but none have been assessed in prospective trials and cannot be recommended routinely.
Recruitment manoeuvres involving a transient increase in transpulmonary pressure are designed to promote reinflation of collapsed alveoli. A variety of techniques have been proposed, including graded incremental pressure increases, and a sustained high inflation pressure manoeuvre followed by a decremental reduction to an optimal PEEP level determined by dynamic analysis of flow-volume relationships, as well as image guided strategies. Whilst data support improvements in oxygenation, such manoeuvres are associated with complications, in particular haemodynamic compromise, barotrauma and exacerbation of existing air leaks; these appear to relate to the frequency of applied manoeuvres.
Despite inverse ratio ventilation (IRV) demonstrating no clear benefit when compared with conventional ventilator modes, the application of airway pressure release ventilation (APRV) in ARDS has been advocated by some centres. APRV utilises inverse ratio, pressure controlled, intermittent mandatory ventilation in patients with unrestricted spontaneous breathing to maintain alveolar recruitment and improve oxygenation, whilst limiting inflation pressures and sedation. Unlike IRV, APRV does not mandate paralysis and hence is an attractive mode. Nonetheless, APRV risks tidal hyper-inflation, increased transpulmonary pressures, and has shown no clear benefit over conventional modes. The use of high-frequency oscillatory ventilation (HFOV) to deliver small tidal volumes at high frequency with high mean airway pressures has been similarly tempered by two large multicentre trials demonstrating no benefit in one and increased mortality in the other. Accordingly, despite the theoretical mitigation of VALI, HFOV can no longer be recommended in ARDS, and conventional ventilator modes should suffice for the majority of patients. Where conventional modes fail to facilitate low tidal volume ventilation and acceptable physiology, extracorporeal gas exchange techniques should be considered in appropriate patients.
Extracorporeal Gas Exchange
Extracorporeal gas exchange (ECGE) is technology whereby blood is drained from a major vein, pumped through an artificial membrane to facilitate gas exchange, and returned to the venous or arterial system, depending on the physiological needs of the patient. Various extracorporeal circuit arrangements are available and are discussed in Chapter 24.
Prompt administration of appropriate antibiotics in sepsis, a reduction in iatrogenic injury through use of reduced tidal volume ventilation and stewardship for hospital-acquired aspiration/ventilator-associated pneumonia, coupled with lower blood transfusion thresholds and the removal of females from the plasma donor pool have combined to limit hospital-acquired ARDS. The use of sedation holds, spontaneous breathing trials and early mobilisation/rehabilitation programmes, where appropriate, are also likely to have contributed.
Non-hydrostatic pulmonary oedema in the context of increased pulmonary vascular permeability is pathogonomic of ARDS. Limitation of fluid administration is therefore an attractive strategy. Studies have shown an association between a persistent positive fluid balance and poor outcome in ARDS. The latter hypothesis is supported by a study assessing the safety and efficacy of ‘conservative’ versus ‘liberal’ fluid management strategies. Although the primary endpoint (death at 60days) did not differ between strategies, benefits in ventilator free and organ failure free days were observed in the ‘conservative’ group, without an increase in renal dysfunction.
Wherever possible, supplemental feeding should be administered via the enteral route. The use of pro-kinetic drugs where needed, and the avoidance of agents that delay gastric emptying, may limit the risk of aspiration pneumonitis.
The anti-inflammatory and antifibrotic properties of steroids present a rational therapy to dampen the dysregulated inflammation which propagates ARDS. Numerous studies have failed to demonstrate clinical benefit from a short course of high dose steroids for the prevention, or treatment, of early ARDS (within 72 hours of onset), with some suggesting steroids may be harmful. The role of steroid therapy in late or unresolving ARDS, i.e. ≥7 days, remains controversial, with data suggesting improvements in both mortality and liberation from mechanical ventilation, but post hoc analysis indicating an increased mortality in patients treated after day 13. In addition, a significant proportion of patients required re-ventilation following steroid weaning, and concerns persist about the relationship between steroid therapy and ICU-acquired weakness.
Inhaled vasodilators, including nitric oxide, prostacyclin and prostaglandin E1, improve ventilation-perfusion matching and pulmonary hypertension by inducing selective vasodilation in well-ventilated lung. Transient improvements in oxygenation are well established, but inhaled vasodilators are expensive, challenging to administer and associated with adverse events without concomitant mortality benefit. Their use should therefore be limited to those patients with refractory hypoxaemia, either as a bridge to extracorporeal oxygenation, or to provide time for ancillary therapies to take effect, or for those patients with reversible pulmonary hypertension in the setting of ARDS.
Neuromuscular blockade is used in over half of ARDS patients to prevent ventilator dysynchrony. A recent randomised controlled trial found the early use (<48 hours of ARDS onset) of cisatracurium improved 90-day survival in those patients with a PaO2:FiO2 ≤ 150mmHg. Of note, neuromuscular weakness was not more prevalent in the cisatracurium group. Further trials are required to clarify whether these effects are reproducible and to clarify the mechanisms of benefit.
Despite promising experimental data, recent large trials of a range of therapies, the majority of which were designed to reduce cellular and mediator driven inflammation, have failed to demonstrate improvements in clinical outcomes. Statins, beta-2-agonists, ketoconazole, vitamin D supplementation and antioxidants were all found not to confer benefit. This has highlighted the need to limit the heterogeneity of subjects recruited into clinical trials through improved understanding of the pathobiology.
Several clinical trials of mesenchymal stem cells (MSCs) for the treatment of ARDS are currently underway (ClinicalTrials.gov: NCT01775774, NCT02444455, NCT02215811, NCT02611609). However, engraftment within the lung does not seem to be the major therapeutic effect of MSCs, rather the effect derives from their capacity to secrete paracrine factors that modulate immune responses and alter the host responses to injury. Pre-clinical work has shown that cryopreserved allogeneic human MSCs are therapeutic in a human ex vivo lung model, but the antimicrobial effects of the MSCs could be largely duplicated by keratinocyte growth factor (KGF), a major paracrine product of MSCs. A clinical trial investigating the efficacy and safety of KGF in ARDS has been completed, but the results have not as yet been reported (ISRCTN95690673).
Mechanical ventilation in the prone position is frequently used in the management of refractory hypoxaemia in ARDS. Approximately 60% of patients demonstrate improvements in oxygenation, which are often sustained on return to the supine position. Proposed mechanisms of benefit include more homogenous distribution of ventilation, better ventilation-perfusion matching and a reduction in VALI. Recently, proning, combined with a protective ventilatory strategy, for at least 16 hours a day in patients with severe ARDS (PaO2/FiO2 < 150 mmHg with PEEP of at least 5 mmHg and FiO2 of ≥0.6) demonstrated a significant reduction in both 28-day and 90-day mortality. Whilst proning confers a risk of pressure ulcers, facial oedema and endotracheal tube obstruction/displacement, favourable haemodynamic effects have been observed with increasing cardiac index in those patients with preload reserve.
ARDS is a heterogeneous syndrome rather than a single disease, hence a variety of supportive and therapeutic approaches may be required for optimal management.
The mainstay of ARDS management is the identification and treatment of the predisposing condition, along with supportive care, which includes lung-protective ventilation.
Fluid restriction, after patients are appropriately resuscitated.
The early, and short-term, use of cisatracurium in more severe ARDS cases may improve outcome over and above lung-protective ventilation.
There are currently no licensed pharmacological therapies for ARDS.
1. ARDS is defined by the following except:
2. The following have demonstrated a mortality benefit in ARDS:
3. ARDS management includes:
Transfusion to >11 g/dl haemoglobin concentration to optimise tissue oxygen delivery
Administration of beta-2-agonists to improve alveolar fluid clearance
Early appropriate antibiotics in the management of sepsis related ARDS
Liberal administration of intravenous fluids to ensure adequate tissue perfusion
Daily recruitment manoeuvres
4. Regarding ECMO:
It has demonstrated a clear mortality benefit in multiple clinical trials in ARDS
Provides just oxygenation with no carbon dioxide removal
Has a complication rate <5%
Should be considered in patients who deteriorate despite optimal medical therapy and 4 weeks of supportive ventilation
Facilitates protective lung ventilator strategies
5. Evidence-based mechanical ventilation in ARDS includes:
Low tidal volume (6 ml/kg PBW) ventilation
High frequency oscillatory ventilation (HFOV)
Airway pressure release ventilation (APRV)
Inversed I:E ratio ventilation
The combination of the aging of the population and improved survival after myocardial infarction has increased the prevalence of heart failure. Most patients with advanced heart failure are admitted to hospital as a result of acute decompensation but some patients with new onset heart failure may present acutely in extremis. Patients with impaired ventricular function who undergo surgery may also present with low output states in the ICU.
The clinical syndrome of heart failure can result from any structural or functional impairment of ventricular filling or ejection of blood. Coronary artery disease remains the most common cause of heart failure. In younger patients (such as the population referred for cardiac transplantation), dilated cardiomyopathy (often of unknown aetiology) is the commonest cause. The true incidence of acute myocarditis in patients with a short history of heart failure is not known because of the difficulty in confirming the diagnosis.
Patients requiring admission to critical care because of severe heart failure usually have one of two clinical syndromes:
2. Cardiogenic shock: defined as tissue hypoperfusion induced by heart failure after correction of preload. It is usually characterised by hypotension (systolic BP <90 mmHg), oliguria (<0.5 ml/kg/hour) and evidence of end-organ dysfunction such as renal, hepatic and cognitive impairment, and elevated blood lactate level.
Bedside assessment may provide clues to the haemodynamic profile of a patient. It is important to distinguish between elevated and non-elevated cardiac filling pressures (‘wet’ or ‘dry’), and adequate or severely impaired tissue perfusion (‘warm’ or ‘cold’) (see Figure 45.1).
Figure 45.1 Assessment of haemodynamic profile using haemodynamic signs and symptoms of patients presenting with heart failure. JVP jugular venous pressure.
Elevated filling pressure can be diagnosed clinically by orthopnoea or by elevated jugular venous pressure. Blood pressure is only a guide to tissue perfusion; proportional pulse pressure (pulse pressure/systolic pressure) less than 25% has been reported to correlate with a cardiac index less than 2.2 l/min/m2 in the population of patients referred for cardiac transplantation.
Patients admitted to the critical care unit are most likely to have elevated filling pressures and inadequate end-organ perfusion (wet and cold). This group is associated with the highest mortality.
Both an acute deterioration and chronic impairment of cardiac function can lead to a progressive decline in renal function. The umbrella term cardiorenal syndrome is used to describe this. The pathophysiology varies depending on specific clinical circumstances; however it involves transrenal perfusion pressure, intrarenal haemodynamics and systemic neurohormonal factors. Alterations in the balance of vasoconstrictor and vasodilator hormones adversely affect renal function, and the combination of increased central venous pressure with low systemic pressure may lead to a severe compromise of net renal perfusion pressure.
In the acute setting, worsening renal function (acute kidney injury) frequently complicates hospital admissions with acute decompensated heart failure. Patients with worsening renal function have a higher mortality and morbidity and increased duration of hospitalisation.
In the more chronic state, renal dysfunction develops in patients who have chronic volume overload, prior renal dysfunction, right ventricular dysfunction and high baseline diuretic requirements. When filling pressures are measured they exceed the optimal levels required to maintain cardiac output. A stable clinical state may be maintained in some patients with high serum urea and creatinine levels but the prognosis is poor. Inotropic infusions may relieve the congestion and improve renal function but the problem often recurs when inotropes are withdrawn.
In addition to a detailed clinical history and physical examination a number of investigations are required:
1. Electrocardiogram to determine rhythm and aetiology of heart failure (e.g. acute coronary syndrome or myocarditis).
2. Chest radiograph for heart size, pulmonary congestion, lung consolidation or pleural effusions.
3. Echocardiography to assess regional and global left and right ventricular function, valve structure and function, pericardial effusion, and mechanical complications of myocardial infarction. The pulmonary artery systolic pressure may also be estimated from the tricuspid regurgitation jet and echocardiography.
4. Blood tests: full blood count, coagulation screen, C-reactive protein, creatinine and electrolytes, glucose and liver function tests in all patients. Troponin and plasma BNP (B-type natriuretic peptide) may also be indicated. For the assessment of a patient presenting with acute dyspnoea a low BNP has a high negative predictive value for heart failure as the aetiology. BNP may be less helpful in the critical care setting.
5. Coronary angiography if revascularisation is indicated.
1. Non-invasive monitoring – temperature, respiratory rate, blood pressure, continuous ECG monitoring and pulse oximetry are required for all patients.
2. Arterial line – essential in unstable patients for continuous arterial blood pressure monitoring and frequent analysis of blood gases.
3. Central venous line – monitoring right sided filling pressure is often essential in patients with advanced heart failure and a central venous line is required for the delivery of fluids and drugs. Estimation of superior vena caval or right atrial oxygen saturation can be a useful marker of oxygen transport. Central venous pressure may be significantly affected by positive end-expiratory pressure ventilation.
Pulmonary artery catheter (PAC) – PAC allows direct measurement of right atrial (RA), right ventricular (RV), pulmonary artery (PA), pulmonary capillary wedge pressure (PCWP) and calculation of pulmonary and systemic vascular resistance. Mixed venous oxygen saturation can also be monitored. This is particularly useful in the presence of severe tricuspid regurgitation when the cardiac output derived by thermodilution may be inaccurate.
In patients with heart failure, right atrial pressure does not correlate well with left sided filling pressure. In many situations, an estimate of left atrial pressure is invaluable. In patients with a high pulmonary vascular resistance (PVR) associated with heart failure, direct measurement of pulmonary pressure is important. In patients requiring inotropic or vasoconstrictor drugs, monitoring of cardiac output and estimation of systemic vascular resistance facilitates therapy based on pathophysiological principles.
Complications associated with the use of a PAC increase with duration of use and it should not be left in situ longer than necessary. In advanced heart failure, therapy tailored to haemodynamic goals as guided by PAC has been shown to result in sustained improvement in symptoms, stroke volume and cardiac output.
Cardiac power output (CPO) describes the relationship between flow and pressure in the circulation and is a powerful predictor of prognosis in cardiogenic shock. It is calculated as the product of simultaneous mean arterial pressure (MAP) and cardiac output corrected for a constant and expressed as watts.
Although PAC is the ‘gold standard’ there are a number of non-invasive cardiac output monitoring devices available that can derive haemodynamic data, but most have not been validated in low cardiac output states.
4. Echocardiography – there is an emerging role for both transthoracic and transoesophageal echocardiography in monitoring haemodynamics in critically ill patients; however, interpretation of data requires specific training and expertise.
The treatment of chronic heart failure has been the subject of several large randomised clinical trials and evidence-based guidelines are available.
Critically ill patients with acute heart failure are a heterogeneous group with respect to aetiology, haemodynamic abnormalities and comorbidities, and are therefore difficult to subject to randomised trials. Treatment strategies should be based on the underlying pathophysiology with the aim of reversing haemodynamic abnormalities. If an underlying treatable cause is identified, clinical condition should be optimised so that definitive treatment can be carried out with the minimum of risk. Immediate therapy should focus on relieving symptoms. Reducing congestion is often the most effective way of achieving this (see Table 45.1).
General management involves a multidisciplinary team due to the complexity of the pathophysiology in the critically ill heart failure patient. In addition to the specific treatments described below, management should include investigation and correction of anaemia, electrolyte abnormalities, thyroid and adrenal function and optimal glucose control. Infection is common; therefore standard infection control measures and evidence-based antimicrobial treatment are essential. Patients should receive appropriate nutrition, mobilisation and physiotherapy. Provision of psychological support and involvement of palliative care teams are also important considerations.
Achieving an adequate level of oxygenation at the cellular level is important to prevent end-organ dysfunction. Treatment should aim to achieve optimal rather than supraphysiological arterial oxygenation, i.e. arterial oxygen saturation above 95%. Respiratory muscle fatigue often results from hypoxaemia and low cardiac output.
Either continuous positive airway pressure (CPAP) or non-invasive ventilation (NIV) can be used to reduce the work of breathing. Both CPAP and NIV result in pulmonary recruitment and an increase in functional residual capacity and a reduction in pulmonary oedema. Clinical trials comparing CPAP with standard therapy have shown a decreased need for endotracheal intubation.
The major goals of medical therapy in the heart failure patient in the ICU are (i) reducing venous congestion (optimising preload), (ii) optimising afterload with vasodilators, in the absence of severe hypotension, and (iii) inotropic support.
The reduction of elevated filling pressures is the most effective way to relieve symptoms of heart failure. Patients with acute decompensation of chronic heart failure are likely to be on diuretic therapy when admitted. Data are lacking on the relative efficacy and tolerability of different diuretics. In the acute setting, a loop diuretic is administered intravenously with dose titration to produce optimal urine output. A large bolus of diuretic may also lead to reflex renal vasoconstriction and a higher risk of ototoxicity. An intravenous infusion of furosemide at 5–10 mg/hour is sufficient in most patients once steps have been taken to increase the cardiac output. Fluid restriction (usually to 1.5 l/day) is an important adjunct to diuretic therapy in severely fluid-overloaded patients. Using a ‘fluid challenge’ in such patients with obvious fluid retention is irrational and has no place in treatment of heart failure patients in the cardiothoracic ICU; inadequate urine output in these patients is usually related to a low cardiac output and treating this often requires inotropic therapy. Once filling pressures have been reduced to normal, the dose of diuretic should be reduced; the dose required to maintain euvolaemia is usually less than that required to achieve it.
The combination of a thiazide (e.g. metolazone or bendroflumethiazide) with a loop diuretic can augment the diuresis achieved in patients with chronic heart failure and is of use in the acute setting. Heart failure patients are often hyponatraemic in the ICU and care needs to be taken not to exacerbate this with combination diuretic therapy. Serum potassium should be monitored as hypokalaemia may predispose to arrhythmias. Combining loop diuretics with a mineralocorticoid receptor antagonist like spironolactone or eplerenone may be effective provided the serum potassium is <5 mmol/l and serum creatinine <200 µmol/l.
In the absence of severe hypotension, vasodilators are indicated in most patients with acute heart failure. Decreasing preload relieves congestion and decreasing afterload is usually beneficial as most patients with heart failure are vasoconstricted (Table 45.2). When administering vasodilators or positive inotropic drugs, the following equation is useful in manipulating the circulation:
MAP − CVP = CO × SVR
where MAP is the mean arterial pressure, CVP is the central venous pressure, CO is the cardiac output and SVR is the systemic vascular resistance.
|Enoximone||0.25–0.75 mg/kg||1.25–7.5 μg/kg/min|
|Milrinone||25–75 μg/kg||0.375–0.75 μg/kg/min|
|Levosimendan||12–24 μg/kg||0.05–0.2 μg/kg/min|
In low doses, nitrates are venodilators but high doses can also cause arterial dilatation. They are particularly useful in acute coronary syndromes associated with heart failure. Oral nitrates in combination with hydralazine have been shown to be beneficial in chronic heart failure and at least two randomised controlled trials have shown that intravenous nitrate in combination with furosemide is superior to furosemide alone. Tolerance to nitrate can develop within 24 hours of commencing an infusion.
Nesiritide (recombinant brain natriuretic peptide) is licensed in the USA for the treatment of acute heart failure. It relaxes smooth muscle leading to arterial and venous dilatation and leads to an increase in cardiac output without direct positive inotropic effect. Compared with nitroglycerin, nesiritide has been shown to produce faster relief of dyspnoea and a more pronounced decrease in pulmonary capillary wedge pressure. Nesiritide has both natriuretic and diuretic effects although up to half of patients with advanced heart failure are reported to be resistant to its natriuretic effect. There is no conclusive evidence that nesiritide improves renal function; clinical studies have not demonstrated better clinical outcomes and it may increase risk of adverse cardiovascular events. The role of nesiritide in the management of heart failure remains unclear.
Hydralazine is a potent arteriolar vasodilator. A combination of hydralazine and nitrates has been shown to be beneficial in patients with chronic heart failure. In patients who cannot tolerate an angiotensin converting enzyme inhibitor (ACEI) because of hyperkalaemia or worsening renal function, it is reasonable to use this combination orally or intravenously.
Inotropic agents are indicated in the presence of tissue hypoperfusion and pump failure often manifested by worsening renal function or fluid retention (peripheral or pulmonary oedema) refractory to treatment with diuretics and vasodilators. A common clinical scenario is a volume-overloaded patient with hypotension, hyponatraemia, and a rising serum urea and creatinine on intravenous diuretic therapy. Continuing such therapy is likely to exacerbate metabolic abnormalities and is unlikely to induce a significant diuresis. Intravenous inotropic therapy may be necessary to improve the patient’s haemodynamic state until some form of definitive therapy or long-term palliation can be considered.
Although inotropic agents in heart failure can result in short-term beneficial haemodynamic effects, they increase myocardial oxygen consumption and carry a risk of inducing life-threatening arrhythmia. Rational use of inotropic therapy in a critically ill population requires invasive haemodynamic monitoring. The lowest effective dose should be used; patients receiving beta-blockers may require higher doses. Very few trials have been conducted in patients with advanced heart failure and there is no definite evidence of the superiority of one agent compared to any other.
Dopamine has complex effects, which vary according to dose. Low doses of dopamine (<2 μg/kg/min) are thought to act predominantly on peripheral dopaminergic (D1) receptors leading to vasodilatation. Although controversial, low dose dopamine has been shown to increase renal blood flow. In addition it is possible that in some patients it may also have an inotropic effect.
A higher dose of dopamine (>5 μg/kg/min) has β-adrenergic effects increasing cardiac output and α-adrenergic effects increasing peripheral vascular resistance and arterial pressure.
Dobutamine acts through stimulation of β1 and β2 receptors and is positively inotropic and chronotropic; it may also induce mild arterial vasodilatation. High doses of dobutamine (above 10 μg/kg/min) cause vasoconstriction but the exact effect at any given dose varies between patients. Heart rate is generally thought to increase less than with other catecholamines, however atrioventricular conduction is facilitated. The commonly used dose range is 2 to 10 μg/kg/min.
This drug is used to increase systemic vascular resistance (SVR) because of its affinity for α-receptors. The lowest dose required to increase the SVR (and hence the blood pressure), and to maintain perfusion of vital organs, should be used. Septic shock is a common indication for its use; the occasional patient after acute myocardial infarction will present with a low SVR due to cytokine release and will benefit from noradrenaline. Excessive vasoconstriction is associated with a reduction in cardiac output.
Type III Phosphodiesterase (PDE) Inhibitors
PDE inhibitors block the breakdown of cyclic-AMP, which modulates intracellular calcium handling. Enoximone and milrinone are the two agents used in clinical practice. They cause marked peripheral vasodilatation and have positive inotropic effects and are therefore useful in patients with advanced heart failure who have an elevated SVR and a low cardiac output. Because of their powerful vasodilating effect, haemodynamic monitoring is recommended whenever they are used. Both agents have a long elimination half-life and tend to accumulate if the patient is oliguric. As their site of action is distal to the β-adrenergic receptor, PDE inhibitors maintain their effect in patients who have been treated with beta-blocking drugs. In patients with atrial fibrillation, they may increase ventricular rate less than dobutamine.
This drug has two main mechanisms of action: Ca2+ ion sensitisation of the contractile proteins (positive inotropic effect) and smooth muscle K+ channel opening (peripheral vasodilating effect). There is also a suggestion that levosimendan has a PDE inhibiting effect. Intravenous infusions of levosimendan are usually maintained for 24 hours, but the haemodynamic effects persist beyond this, probably because of the long half-life of its metabolite. Levosimendan infusions in patients with heart failure have been associated with a dose-dependent increase in stroke volume and cardiac output, a decline in the pulmonary capillary wedge pressure, a decrease in SVR and PVR, a slight decrease in blood pressure and a slight increase in heart rate. An improvement in symptoms of dyspnoea and fatigue has been shown in trials comparing levosimendan with dobutamine. The haemodynamic effects were seen even in the presence of beta-blocker therapy. Tachycardia and hypotension are side effects associated with the use of levosimendan and it is not recommended in patients with a systolic blood pressure below ~85 mmHg.
Angiotensin Converting Enzyme (ACE) Inhibitors and Angiotensin Receptor Blockers (ARB)
ACE inhibitors were the first drug class shown to improve outcome in severe chronic heart failure. ARBs are used as alternatives in patients intolerant of ACEI. They have no role in the early management of unstable heart failure patients but should be introduced as soon as the patient is haemodynamically stable and has acceptable perfusion and renal function. ACE inhibitors decrease renal vascular resistance, increase renal blood flow and promote sodium and water excretion. However, in patients with a very low cardiac output, they may significantly decrease glomerular filtration rate. If patients with acute decompensation of chronic heart failure are admitted to the ICU, it may be necessary to discontinue them temporarily.
The role of beta-blockers in the management of chronic heart failure is well established following several large trials involving many thousands of patients. In volume-overloaded patients, beta-blockers are likely to increase the severity of heart failure and are best avoided. There is no consensus on the management of a patient receiving beta-blockers for chronic heart failure admitted to hospital with acute decompensation. Most will require at least a decrease in the dose of the drug but in patients requiring (beta-agonist) inotropic therapy it is logical to discontinue beta-blockers altogether.
Patients should receive standard prophylaxis for prevention of venous thromboembolism and anticoagulation for established indications, for example acute coronary syndromes, pulmonary embolism and atrial fibrillation. Patients with evidence of ventricular thrombus on echocardiography should also receive anticoagulation. Care is needed as concomitant liver dysfunction may lead to a prolonged prothrombin time. In patients with a creatinine clearance below 30 ml/min, low molecular weight heparin in therapeutic doses should be used cautiously, probably with monitoring of factor Xa level.
Patients with gross fluid retention and hyponatraemia present a difficult clinical problem. Diuretics often worsen hyponatraemia and sometimes features of the cardiorenal syndrome become apparent. Inotropic drugs may help in this situation but if therapy needs to be prolonged, the risk of arrhythmia needs to be considered. Continuous venovenous haemofiltration (CVVH) is effective in removing fluid and the rate of fluid removal can be tailored to the patient’s needs. CVVH may also remove cytokines with myocardial depressant properties (e.g. tumour necrosis factor) as macromolecules up to 20,000 Da can pass through the ultrafiltration membrane. If necessary, large volumes of fluid can be removed in a relatively short time to get the patient ready for a definitive procedure (heart transplantation, mechanical circulatory support).
CVVH usually requires large-bore central venous access although there are devices to allow ultrafiltration via cannulae in peripheral arm veins. Although the maximum rate of fluid removal is less than that attainable by central CVVH, an adequate rate is achieved for the most common clinical situations.
Compared to high dose diuretic therapy, ultrafiltration has been reported to induce less neurohormonal activation and vasoconstriction.
Intra-aortic Balloon Pump (IABP)
IABP use may reduce afterload, thereby decreasing left ventricular stroke work and myocardial oxygen consumption, as well as augmenting diastolic blood flow in the coronary and systemic circulation. Functional mitral regurgitation, a common problem in a patient with a dilated left ventricle, decreases with the use of the IABP.
The IABP is extremely useful in critically ill patients with heart failure who can be stabilised until definitive therapy can be carried out. In patients requiring support beyond that provided by IABP, consideration should be given to the use of mechanical circulatory support. There is no evidence of improved survival with any form of short-term mechanical support, including IABP, in patients with cardiogenic shock. However these therapies may improve haemodynamics and get the patient fit for definitive treatment (implantable device or heart transplantation).
Advanced heart failure in patients on critical care units carries a high mortality. Optimal management requires close cooperation between a cardiologist with an interest in heart failure, an intensive care physician and a cardiac surgeon. With appropriate therapy, many critically ill patients can be resuscitated and returned to a productive life.
Heart failure is most commonly secondary to coronary artery disease and carries a poor overall prognosis.
Admission to critical care is usually due to pulmonary oedema or cardiogenic shock.
The major goals of medical therapy for the heart failure patient in the ICU are (i) optimising preload, (ii) optimising afterload and (iii) increasing cardiac output with inotropic support when required.
Aggressive diuresis may be needed, and renal replacement therapy may be necessary to remove an adequate amount of fluid.
Inotropes are often started to enhance cardiac output, but may lead to other complications; close haemodynamic monitoring is essential to direct therapy.
1. Clinical presentation of acute heart failure may include the following symptoms and signs:
2. Investigations and monitoring for the patient with acute heart failure may include:
3. First line drug therapy in the treatment of chronic heart failure includes the following:
4. Regarding inotropic drugs used for the treatment of patients with acute heart failure:
5. Regarding the use of an intra-aortic balloon pump which of the following statements is true?
Both systolic and diastolic blood pressure are increased
Haemodynamic effects include an increase in afterload and increase in cardiac output
Severe mitral regurgitation is a contraindication
Aortic regurgitation is a contraindication
Cardiac surgery’s profound intervention in systemic blood pressure and circulation carries a high potential to affect the brain in a way that may offset the benefits of successful cardiac procedures. Astonishing progress has been achieved in safety and effectiveness, which has enabled cardiac surgery to become feasible in ever older and sicker patients. However, these are at higher risk of stroke, which in turn increases mortality fivefold, and neurological complications prolong intensive care treatment and rehabilitation. Conditions treated by emergency cardiac surgery, such as aortic dissection and acute cardiac failure, carry a high risk of neurological complications even without surgery, which may only come to light in the postoperative period.
Cardiac surgeons closely monitor effectiveness and adverse effects, so the complications of cardiac surgery have been well identified. Prevention of neurological complications includes thorough preoperative screening and perioperative monitoring, and this has deepened our understanding of potential risks. This chapter will provide an overview of the most significant neurological problems encountered around cardiac surgery, their diagnosis and management, and an outline of preventive options.
Neurological Considerations in Patients Undergoing Cardiac Surgery
Changing patient characteristics indicate a growing potential for neurological complications: the age of patients in cardiac surgery has increased, and the case mix accepted for CABG has also changed through the advent of PCI as an alternative procedure. Competent preoperative neurological assessment including a detailed cerebrovascular history may help identify patients at risk of developing neurological complications. In CABG patients generally, the risk of severe carotid stenosis may be higher than in the general population, and carotid ultrasound before elective CABG benefits carefully selected patients the most.
Cardiopulmonary bypass has been a prerequisite for modern cardiac surgery, and in the past 60 years this has become a safe, routine procedure. Venous blood from the systemic circulation is drained to a reservoir, then pumped through a filter and membrane oxygenator system to the ascending aorta, which is cannulated distal to the aortic cross-clamp. In a second venous-to-arterial circuit originating from the bypass, some of the blood is diverted back to the heart together with cardioplegia solution, and from the heart to the venous reservoir after being purified from embolic material in a cardiotomy reservoir.
Neurological complications related to the cardiac bypass are fortunately rare. One main source of problems is related to embolism, when during preparation of the cardiopulmonary bypass circuits, cross-clamping, and turbulent or high-velocity blood flow can dislodge atheromatous material from the aortic wall. In addition, bypass circulation requires heparinisation and, together with cardioplegia solution and other admixtures causing haemodilution, alters the flow and oxygenation qualities of the blood perfusing the brain. Furthermore, contact between the blood and non-biological filter membranes and bypass surfaces induces a systemic inflammatory response with potential neurological significance which is not yet well understood.
A number of technical modifications may help reduce the risk of neurological complications. Preoperative transoesophageal or intraoperative epiaortic ultrasound may help identify patients particularly at risk for embolism related to aortic cross-clamp and potentially may allow modification of the surgical strategy. Most procedures with cardiopulmonary bypass are performed using mild hypothermia. The importance of temperature management for brain protection in hypoxic conditions has been long recognised, both for cardiac surgery and for treatment after cardiac arrest. However, a Cochrane review of studies comparing neurological outcomes after hypothermic and after normothermic bypass surgery did not establish a clear benefit.
Deep hypothermia is used to prolong toleration of hypoxia to the brain in surgery with prolonged cerebral circulatory arrest, for instance aortic arch surgery or pulmonary thromboendarterectomy. Although the neuroprotective effect of deep hypothermia is beyond doubt, the precise temperature required to achieve maximum benefit, the parameters to which hypothermia should be implemented and how rewarming should take place remain unclear. Rapid rewarming after therapeutic hypothermia can trigger epileptic seizures. Other potentially neuroprotective strategies include the use of anterograde or even retrograde selective brain perfusion. The neurological benefits have not been established, although some evidence suggests that selective anterograde perfusion may permit procedures to take place in moderate rather than deep hypothermia. The use of Alpha-stat metabolic management or pH stat management, in which CPB gas sweep rate is adjusted at lower temperatures to compensate for the hypothermic alkaline drift, remains controversial. Proponents of pH stat management believe that it improves oxygen delivery, increases cerebral blood flow and thus allows more effective cooling of the brain. Alpha-stat management allowing an alkaline pH during hypothermia, on the other hand, may improve enzyme activity and protein function, preserve cerebral autoregulation, and produce less risk of embolism through reduced cerebral blood flow. Which method is more beneficial may depend on multiple factors, including the age of the patient, but the topic remains controversial.
It still remains to be proven whether ‘off-pump’ procedures are significantly safer than conventional cardiac bypass. The risk of postoperative delirium, which may prolong intensive care, is also claimed to be lower. It is unclear which systemic blood pressure is optimal for brain protection during surgery. Trials comparing a mean arterial blood pressure of 50 mmHg with 70 mmHg failed to demonstrate a significant benefit for either approach. Either target would mean a cerebral perfusion pressure (CPP) lower than the CPP of 70 mmHg currently advised for brain protection in acute traumatic brain injury; whether this has clinical significance has so far not been investigated.
Acute Neurological Complications of Cardiac Surgery
Central Nervous System
Stroke is the commonest neurological complication after cardiac surgery, occurring through embolism or hypoperfusion. The incidence of clinically relevant stroke has decreased to under 2% after CABG and under 4% after single valve replacement, despite the demographic increase in the proportion of elderly at-risk patients. Postoperative diffusion-weighted MRI scans demonstrate clinically silent new lesions in up to 18%. Valve replacement and combined CABG and valvular surgery procedures carry a higher risk of causing embolic territorial and branch infarcts, whereas prolonged cardiopulmonary bypass increases the risk of watershed infarcts related to hypoperfusion. In most cases currently, there is no acute active treatment as major surgery is a contraindication for thrombolysis, but the increasing availability of emergency clot retrieval may change the situation. Stroke associated with left ventricular assist devices has recently been shown to benefit from endovascular treatment, if recognised early. One likely consequence will be that rapid diagnosis in patients evidencing a neurological focal deficit on awakening becomes necessary more often in the near future, increasing the need for imaging. Ischaemic and haemorrhagic stroke cannot be differentiated clinically, so the different management mandates CT scanning of the brain, if emergency MRI is not feasible. Specific locations of ischaemic stroke are associated with particular risks, such as ‘malignant’ complete MCA territory stroke with the risk of life-threatening hemispheric swelling, or cerebellar stroke with the risk of hydrocephalus and brain stem compression; both these forms of stroke have high fatality if diagnosed late, but often have good outcomes if treated with prompt surgical decompression.
Intracranial haemorrhage is infrequent in cardiac surgery, and is related to a combination of intentional treatment and effects of bypass on platelet function and clotting factors, especially in patients with previously unrecognised predisposing conditions (Figure 46.1). The location of haemorrhage can be intraparenchymal, subdural, epidural and subarachnoid, and is often atypical. Intracranial haemorrhage warrants neurosurgical consultation and some cases may need decompression.
Figure 46.1 A 39 year old female admitted for redo pulmonary endarterectomy. Postoperatively she was heparinised, but also had low platelets (66 × 109). Failure to awaken, and right sided weakness prompted imaging. CT showed subdural haemorrhage and bleeding into two pre-existing, unrecognised arachnoid cysts. (a)–(e) Initial CT scans; (d) tonsillar herniation into the foramen magnum; (e) upward transtentorial herniation; (f), (g) CT post hemicraniectomy and cyst decompression shows improvement of midline shift.
Encephalopathy is an umbrella term for diffuse brain dysfunction that can be due to multiple different aetiologies. In the acute phase after surgery, encephalopathy manifests as failure to awaken, and global hypoxia or ischaemia of the brain needs to be differentiated from sedative overhang. Elderly patients or those with cerebrovascular disease are at greater risk of encephalopathy. Patients scheduled for CABG are a population with risk factors for cerebrovascular disease, with a high prevalence of silent infarcts on preoperative MRI of the brain. A clinical diagnosis of global encephalopathy demands the absence of a focal neurological deficit, which makes a careful examination of oculomotor and other cranial nerve functions important. Dysconjugate eye movements imply the presence of a focal brain stem lesion, for instance stroke in the basilar distribution, which may go undetected in routine CT of the brain. Multiple simultaneous emboli can simulate global brain dysfunction, making imaging of the brain mandatory as well as extensive metabolic and infectious screening in prolonged encephalopathy. Negative CT but persistent failure to awaken will mean an indication for MRI of the brain and/or electrophysiological studies. Lumbar puncture may be needed after imaging, in the early postoperative period, to exclude subarachnoid haemorrhage, and in the later postoperative period to exclude infection.
Encephalopathic patients, when they awaken, may have persistent confusion or delirium, a fluctuating conscious state, or recurrent agitation and hallucinations. Repetitive movements such as tremor, asterixis, choreoathetosis or myoclonus are frequent, and are often suspected to be epileptic seizures. The clinical diagnosis or exclusion of epileptic seizures versus movement disorders, stereotypies or paroxysmal dysautonomia is notoriously unreliable. EEG is mostly needed to confirm the suspicion and make the indication for anticonvulsant drugs. Even in the absence of abnormal movements or stereotypes, a prolonged EEG may be necessary to exclude non-convulsive status epilepticus (NCSE). This condition has only recently been recognised and its incidence in cardiosurgical patients is unclear, but it has been shown to be prevalent in neurotrauma and in sepsis patients. A fluctuating dyscognitive state may be the only clinical sign of NCSE, and the features may be so unspecific that only EEG may differentiate it from other causes of delirium.
Posterior reversible encephalopathy is a syndrome of vasogenic oedema affecting predominantly white matter, more often the posterior region of the brain, and is diagnosed by MRI. It is occasionally seen in the immediate postoperative phase where it can occur as a form of ‘reperfusion syndrome’, for example after repair of aortic stenosis, after carotid endarterectomy or after transplantation when there is a rapid shift of arterial pressures. More often, in patients who have undergone transplantation, PRES is related to toxicity of calcineurin inhibitors, in particular tacrolimus, where a pronounced tremor is often a feature, or to many other drugs (Figure 46.2).
Figure 46.2 A 29 year old patient post double lung transplant. Agitation and visual hallucinations developed shortly after initialising the full dose of tacrolimus. CT scan was unremarkable. MRI shows PRES with symmetric high posterior white matter signal in FLAIR image sequence.
Optic neuropathies are uncommon but well-recognised complications of cardiac surgery. Infarction of the optic nerve results in permanent monocular loss of visual acuity and visual field defects, with optic disc swelling (anterior optic neuropathy) or without (posterior optic neuropathy). A monocular disturbance of vision needs to be differentiated from incongruous hemianopia due to infarction affecting the optic tracts, which does not affect the pupillary reaction and where the visual field defect in each eye differs in size. Infarction of the visual cortex causes homonymous hemianopia with an identical visual field defect in both eyes.
Brachial plexus injury causes denervation in the distribution of multiple peripheral nerves, and after median sternotomy with sternal retraction a lower brachial plexus injury of varying severity occurs in up to 5% of patients. The majority are mild and transient, with sensory and motor symptoms resembling ulnar nerve injury, although findings on careful examination may show a Horner’s syndrome, or sensory deficits and reflex abnormalities exceeding the distribution of a single peripheral nerve. Persistent and severe deficits warrant neurophysiological examination, which may help prognostication by estimating the degree of axonal injury versus demyelination. Intraoperative monitoring of sensory nerve conduction may help to avoid brachial plexus injury, and identify factors which make brachial plexus injury less likely, such as minimising the opening of the sternal retractor, caudal positioning of the retractor and reducing asymmetric traction. Transient injury to sympathetic fibres through ipsilateral jugular venous cannulation causing Horner’s sign, but with no abnormalities in the arm, is a differential diagnosis to partial lower brachial plexopathy.
The phrenic nerve passes through the mediastinum adjacent to the pericardium and is therefore vulnerable to injury during surgery. The frequency of unilateral diaphragmatic weakness due to phrenic nerve injury may reach 10% after cardiac surgery. Diaphragmatic weakness causes weakness of inspiration and atelectasis, and predisposes to postoperative respiratory complications, especially in the presence of pre-existing pulmonary disease. Most cases recover within months, but weakness may persist if there is significant axonal degeneration, as with nerve transection. Rare bilateral phrenic nerve palsy causes failure to wean from the ventilator.
Unilateral recurrent laryngeal nerve injury causes dysphonia; the left side is more often affected due to its longer intrathoracic course. Bilateral laryngeal palsy leads to severe stridor and aspiration. Several mechanisms can be suspected, from central venous catheterisation to traction along the nerve’s intrathoracic course, for instance traction on the oesophagus, subclavian arteries, or on the heart. Other mono‑ neuropathies have been described related to nerve compression or trauma during prolonged surgery, including ulnar, radial, long thoracic, spinal accessory, peroneal, lateral femoral cutaneous, and facial nerves. Most of these are related to typical nerve compression sites, and are seldom persistent.
Intensive care unit acquired weakness (ICU-AW) is a complication which was originally named critical illness polyneuropathy, but is now recognised to affect primarily muscle at least as often, causing an acute or subacute myopathy. The exact causation remains unclear, but ICU-AW is most often associated with prolonged ventilation, sepsis and use of steroids and muscle relaxants. Muscle biopsy findings are variable, with some cases showing only Type 2 fibre atrophy, which is an unspecific abnormality, others showing necrotising myopathy or acute myosin loss. The clue to diagnosis is a patient who after prolonged ventilation is tetraplegic, fails to wean, but has relatively preserved eye and often facial movements. Nerve conduction studies are often normal, electromyography may be myopathic but is often unspecific, and muscle biopsy may be needed for confirmation. Crucially, CNS pathology, including spinal cord damage, has to be conclusively excluded.
Guillain–Barré syndrome (GBS), an autoimmune acquired demyelinating neuropathy, may rarely occur after cardiac or other surgery. In the chronic postoperative phase GBS has been described as a consequence of CMV reactivation after cardiac transplantation or as a symptom of graft-versus-host disease. Neuropathy after transplantation can be caused by tacrolimus toxicity, or by other commonly used neurotoxic medications such as amiodarone.
Delayed Neurological Complications
Immediate postoperative neuropsychological impairment in the absence of a structural defect is common but reversible over the course of the days following surgery. A more persistent cognitive impairment, however, wears off over weeks or months; it is considered comparable to cognitive problems after non-cardiac surgery, possibly due to brain injury at a cellular level. Better cognitive performance is seen at 3 months after CABG surgery performed off pump compared with patients undergoing bypass, with a risk of cognitive decline reduced from 29% to 21%, but the difference becomes insignificant after 12 months (30.8% versus 33.6%) and currently no surgical technique has been shown to be protective.
Delayed cognitive decline occurring over the 5 years after CABG surgery has been described by various authors. However, comparative studies have demonstrated a similar change in patients treated with angioplasty and no difference between neuropsychological performance in patients treated with bypass CABG and with off-pump CABG. Therefore, the theory that the delayed decline is causally linked to CABG surgery remains unproven, although it seems that cognitive decline may be linked to an accumulation of vascular risk factors, and to the patients’ preoperative conditions. The best preventative measure is therefore meticulous attention to the control of vascular risk factors postoperatively, and no decline has been demonstrated where this was the case. Microembolic brain injuries have been implicated in the genesis of delayed Alzheimer dementia after cardiac surgery, but there is currently no evidence that patients who undergo cardiac surgery are really at increased risk for dementia or Alzheimer’s disease, and the significance of microemboli is not yet established.
The survival rate today of modern cardiac transplantation is impressive, exceeding 80% at 1 year, 70% at 5 years and 50% at 10 years. Neurological complications constitute a significant proportion of mortality and postoperative morbidity; whether modifications of the transplantation procedure such as preoperative use of ventricular assist devices, or of combining heart and lung transplantation, have an impact on morbidity is still under investigation. Interestingly, an increasing number of patients undergo cardiac transplantation with a known genetic neuromuscular disease causing cardiomyopathy, but also limb muscle weakness, and the outcomes for these patients are apparently no worse than for patients without underlying neurological disease.
Patients accepted for cardiac transplantation are likely to have heart failure and a high risk of embolism prior to surgery, and some may have needed prolonged preoperative treatment with ventricular assist devices or ECMO, which themselves carry a substantial risk of cerebral embolism (Figure 46.3). Until recently, the only treatment option for acute stroke was thrombolysis, unfeasible for patients who had just undergone heart surgery, but the situation may change with increasing availability of clot retrieval.
Figure 46.3 A 26 year old female after fulminant heart failure due to sarcoid cardiomyopathy, with emergency transfer under CPR for ECMO and ventricular assist device. Imaging took place after encephalopathy and agitation were noted. Initial CT shows a left basal ganglia infarct (a) and a right occipital cortical infarct (b). MRI shows both watershed infarcts (c) and a left occipital pole infarct (d). There was good functional recovery after cardiac transplant.
The early neurological complications related to transplantation surgery itself are similar to those found after other cardiac surgery, but they are more frequent. In one series, the frequency of perioperative CNS complications in heart transplantation was 19.8% versus 3.1% in routine and 10.3% in emergency CABG surgery. The incidence varies widely between series. Perioperative neurological complications occurred in 23% in the Mayo Clinic series, and in other series early complications were even more frequent. Perioperative stroke occurs in up to 11%. Delirium and encephalopathy may occur in 10–20% of cases, including posterior reversible leucoencephalopathy.
Seizures are reported in 2% to 20% of patients, often as part of a generalised encephalopathy. Once embolic stroke has been excluded, the presumed aetiologies are rapid shifts of brain perfusion and drug effects, particularly of calcineurin inhibitors. As these patients are on multiple drugs, the possibility of unexpected interactions always needs to be considered. Levetiracetam is today the anticonvulsant of choice, due to its advantages of availability in intravenous preparation, lack of depression of conscious state and lack of enzyme induction or effects on cyclosporine, tacrolimus or warfarin levels. CSF studies are necessary in most cases, to exclude early onset of a CNS infection, although infections are more common beyond the immediate perioperative phase.
In this early period complications related to the peripheral nervous system are reported in between 4% and 11% of cases, and include both the nerve compression and nerve and plexus traction problems outlined earlier in this chapter, as well as immunological complications such as Guillain–Barré syndrome or early drug toxicity.
Later after transplantation, neurological complications include many conditions not seen in other types of cardiac surgery, broadening the spectrum of possible differential diagnoses, which emphasises the importance of early accurate identification. In the Mayo Clinic series reported by van den Beek and coworkers, the risk for neurological complications over a period of 18 years was 81%, including sleeping disorders in 32%, polyneuropathy in 26% and cumulative risk of stroke in 14%. The cause of death was neurological in 13%. Severe late complications are dominated by the effects of stroke and of problems related to immunosuppression, including direct side effects of immunosuppressive drugs, infection, transplant rejection, or post-transplantation lymphoproliferative disorder and other malignancies. Post-transplantation lymphoproliferative disorder (PCNS-PTLD) is an uncommon Epstein–Barr virus driven malignancy that may cause systemic uncontrolled B-cell proliferation affecting many organs, or primarily affecting the brain and central nervous system, according to a recent series, in 0.18% of heart and/or lung transplant recipients. It may regress with reduction of immunosuppressive treatment, or may require radiotherapy or chemotherapy.
Pharmacological side effects of immunosuppressive drugs are dominated by the effects of calcineurin inhibitors, and today in the first line by tacrolimus related side effects. These can be both direct toxic effects as well as immune reactions triggered by tacrolimus. Calcineurin toxicity affects both the CNS with tremor in early stages, and in more severe cases seizures, encephalopathy, cortical blindness and psychosis. Patients may have very variable MRI findings of the brain, and commonly posterior reversible encephalopathy syndrome (PRES) and vasogenic oedema, but imaging may be completely normal despite CNS symptoms. Calcineurin inhibitor toxicity may also affect the PNS, causing demyelinating neuropathy or inflammatory myopathy. Toxicity is not always closely related to the tacrolimus levels, but generally improves when the dose is reduced. Combination with other immunosuppressants (steroids, mycophenolate, sirolimus) may be helpful to minimise the dose, but newer immunosuppressants can demonstrate similar side effects. OKT3, another less frequently used immunosuppressant, has been seen to cause aseptic meningitis.
Infections of the CNS after cardiac transplantation occur in 2–3% of patients of recent larger series including some 600 patients, with a broad range of potential infectious agents; the incidence of any specific agent is too low to narrow the range of differential diagnoses. Aspergillus species, Cryptococcus neoformans, Listeria monocytogenes, Herpes zoster and JC virus occurred more than once. Aspergillus is a particular diagnostic concern. Other infectious agents that have been described include Streptococcus pneumoniae, West Nile Virus, HHV6, Toxoplasma gondii, Nocardia species, Candida species, Cryptococcus, Balamuthia mandrillaris, Bipolaris spicifera, Tuberculosis, Acanthamoeba, Scedosporium and syphilis. ‘Acute flaccid paralysis’ due to West Nile Virus has not been described after cardiac transplantation, but clearly, due to increasing prevalence, it needs to be included in the differential diagnosis of neuromuscular weakness.
The best strategy to prevent neurological complications in cardiac surgery remains management of risk factors and thorough assessment of the patient’s risk profile. The risk of stroke increases with age, reaching 5% in patients over 80, although results have improved in newer studies. Other characteristics elevating a patient’s risk profile for stroke and neurological complications are the presence of hypertension, diabetes mellitus, smoking, a history of cerebrovascular disease, peripheral vascular disease and aortic atheroma, and the type of surgery required. Intraoperative factors associated with increased risk can be duration of the cardiopulmonary bypass and its haemodynamic performance. A postoperative risk factor for stroke is the presence of atrial fibrillation.
The 2011 ACC/AHA guidelines include recommendations on prevention of neurological complications. A number of potential strategies are aimed at reducing complications related to cardiopulmonary bypass. Shann et al. recommend the following evidence-based strategies:
1. Reducing emboli through arterial filtration, intraoperative aortic imaging, minimising direct reinfusion of pericardial suction blood, filtration of suction blood before reinfusion;
2. Minimising the prothrombotic response through reducing blood contact with non-biocompatible surfaces in CPB;
3. Ensuring perioperative normoglycaemia and avoidance of hyperglycaemia;
4. Avoiding low haematocrit and excessive haemodilution;
5. Managing rewarming rate and limiting arterial line temperature to 37 °C during rewarming;
6. Optimising metabolic management (pH stat versus alpha stat management remains a matter of controversy).
The ACC/AHA guidelines for CABG recommend perioperative administration of a beta-blocker drug to reduce the risk of atrial fibrillation, as well as perioperative statins and aspirin. Epiaortic ultrasound is advised to identify atherosclerosis of the aorta, and to enable measures to be taken that might help avoid embolism, such as choice of off-pump. Simultaneous carotid endarterectomy is recommended with CABG for symptomatic high grade (70–99%) internal carotid stenosis, because it can reduce the risk of embolism and stroke. The benefit of simultaneous CABG and CEA is not established in asymptomatic high grade stenosis, but the ACC/AHA guidelines and other authors suggest simultaneous endarterectomy and CABG in patients with high grade stenosis on one side, and severe stenosis or occlusion on the other.
Failure to awaken after cardiac surgery is a worrying neurological scenario encountered in the early postoperative phase, and may signify anything from transient encephalopathy to a catastrophic focal event. Patients may be systemically severely unwell, and ongoing procedures such as ECMO and pacing wires may limit the feasibility of standard neurological investigations. On the other hand, new treatment options such as clot retrieval increase the potential benefit of accurate early diagnosis.
In the comatose patient, diagnostic evaluation includes bedside clinical assessment, specialist review, appropriate use of neuromonitoring and neurophysiological techniques and timely choice of imaging. Any test is only as good as its availability at the right time, and a multimodal approach builds on good clinical evaluation.
Clinical assessment is primarily performed by the intensive care doctors and nursing staff, and their first responsibility is to recognise there is a potential neurological problem as early as possible, whether or not they have specialist neurological expertise. Standardised clinical scores are helpful, as long as their limitations are recognised and interpretation is consistent. The Glasgow Coma Score (GCS) is the most widespread tool, but it has significant limitations such as having inadequate validation in non-trauma patient groups, including largely irrelevant criteria such as the verbal response, and providing inadequate information on brain stem function or focal deficits. The clinical evaluation must reliably check for focal dysfunction and brain stem abnormalities, as either is unusual in generalised encephalopathy. The Full Outline of UnResponsiveness (FOUR) score provides better neurological information and in the author’s experience more validity in serial examination. Performance in neurological scores informs a more comprehensive assessment by the neurocritical care specialist, as early as possible. This should ideally set the priorities for imaging and electrophysiological tests.
Neurophysiological investigations extend the clinical assessment; they are tailored to the clinical situation and are best done after clinical review. EEG is the only method to diagnose non-convulsive seizures. Routine bedside neuromonitoring promises to provide continuous assessment of brain function, but is still in its infancy. Continuous electroencephalography (CEEG) allows real time detection of seizures, and may alert to worsening encephalopathy or (less well) to evolving large focal lesions. It is gaining use in neurocritical care units, but the amount of data generated is difficult to handle, requiring experienced personnel with specialist knowledge and significant time. EEG trending with specialised analysis software may help overcome the problem but still requires ‘raw’ EEG for interpretation. Telemedicine-based EEG services today enable competent EEG assessment even in remote hospitals. Simplified EEG-based monitoring systems have not proved their value and are handicapped by the lack of spatial resolution offered by 2 to 4 channels. Somatosensory evoked potentials (SSEPs) are used intraoperatively, and in addition the evolution of serial SSEPs after hypoxia gives robust prognostic information. Monitoring of brain perfusion and oxygenation based on transcranial Doppler ultrasound and near infrared spectroscopy (NIRS) is under development, but their clinical value is not yet clear.
Neuroimaging is indicated if either clinical or neuromonitoring data suggest a focal or evolving lesion, as well as in persistent reduced conscious levels. Computerised tomography (CT) is rapid and sensitive in detecting haemorrhage and supratentorial stroke, and feasible in the presence of metal implants and electronic devices. Mobile bedside CT scanners allow imaging in the ICU. CT angiography is rapid and has largely superceded diagnostic catheter angiography, and CT perfusion studies can detect vasospasm. Magnetic resonance imaging (MRI) has exquisite sensitivity in the brain stem and cervical spinal cord, which are largely inaccessible to CT, but it is more cumbersome, slower and is not suitable for unstable patients undergoing ECMO. With the exception of magnetic material implants, however, most disadvantages can be overcome through using modern imaging sequences and some specialised equipment. MRI is much more sensitive for white matter lesions, and has much more capability to date lesion onset and provide biochemical and functional imaging than CT.
Cardiosurgical neurology is a unique field in which good cooperation between cardiac surgeons, cardiac intensivists and intensive care neurologists is significantly improving clinical outcomes. We now have powerful diagnostic tools applicable to the perioperative situation. There are still major obstacles to overcome, including improving neurological training in the CT ICU, improving neurologists’ understanding of a complex field extending far beyond conventional neurology, and in making sophisticated neurological investigations available in a challenging environment within a very tight time frame. Currently, the capabilities of technology are still limited by personnel and training constraints, but the clear benefits they offer are rapidly driving change.
Neurological complications are major determinants of outcome after cardiac surgery.
Risk assessment prior to surgery takes into account age, cerebrovascular risk factors and comorbidity, type of surgery, and a detailed neurological assessment to identify patients at higher risk of neurological complications.
The postoperative neurological assessment must screen for focal neurological deficits in order to diagnose vascular deficits early, in time to consider revascularisation in the brain.
Focal neurological deficits in a patient with postoperative encephalopathy suggest a structural abnormality rather than drug effects of generalised hypoxia, and necessitate urgent brain imaging.
A fluctuating impairment of the level of consciousness may suggest non-convulsive status epilepticus even if abnormal movements or stereotypies are absent.
EEG is mandatory to diagnose and adequately treat status epilepticus.
Lower brachial plexus injuries are the commonest mechanical peripheral nerve injuries after sternal retraction in cardiac surgery. Peripheral nerve injuries need to be clinically differentiated from complications related to the central nervous system, as neurophysiological tests are often unrevealing in the acute phase.
Calcineurin inhibitor (tacrolimus, cyclosporin A) toxicity can cause multiple peripheral and central neurological symptoms. Toxicity does not closely correlate with the individual’s blood levels, but improves with dose reduction.
The 2011 AHA/ACC guidelines recommend strategies to reduce the risk of neurological complications.
1. The following factors are important for assessing the risk of neurological complications: