The prevalence of heart failure (HF) is increasing and is a co-morbidity in up to 20 per cent of patients undergoing major non-cardiac surgery.
A thorough preoperative assessment is required to adequately risk stratify patients.
Cardiac biomarkers such as brain natriuretic peptide (BNP) are not routinely required, but may be useful in symptomatic patients.
Transthoracic echocardiogram (TTE) should be considered in patients with suspected heart failure to determine the severity and aetiology.
HF therapy should ideally be optimised for 3 months prior to surgery.
Specialist heart failure advice should be sought early in the preoperative process for complex cases.
Heart Failure (HF) is a heterogeneous condition that affects more than 900,000 patients living in the United Kingdom, and more than 25,000 are newly diagnosed each year (National Heart Failure Audit, April 2011–March 2012). Of the major cardiovascular disorders, HF is the only condition with a steady increase in prevalence (Braunwald, 2013). This prevalence continues to rise with age, with around 10 per cent of men and 8 per cent of women in the United States over the age of 60 having the condition (Roger et al., 2012). HF is associated with substantial morbidity and mortality with a 5-year survival of 50 per cent, worse than many cancers (Askoxylakis et al., 2010).
Up to 20 per cent of patients undergoing major non-cardiac surgery will have a diagnosis of HF (Hammill et al., 2008). Patients with HF have long been known to have significantly higher perioperative cardiac mortality and morbidity compared to patients without HF (Hernandez et al., 2004). The operative mortality can be up to 63 per cent higher in patients with HF, and the risk of 30-day rehospitalisation is 51 per cent higher than patients without heart failure or coronary artery disease (Hammill et al., 2008).
Currently available preoperative risk stratification models include HF as a risk indice. As far back as 1977, Goldman and colleagues observed that the presence of a raised jugular venous pressure and a third heart sound on auscultation were associated with an increase in perioperative HF (Goldman et al., 1977). More recently, the Lee index or revised cardiac risk index incorporated a prior history of HF as one of six preoperative variables to predict post-operative myocardial infarction, pulmonary oedema, ventricular fibrillation, cardiac arrest and complete heart block (Lee et al., 1999). Please refer also to chapter 24 , page 282.
Definition, Diagnosis and Pathophysiology
HF can be defined as a disorder of cardiac structure or function resulting in failure of the heart to maintain adequate tissue perfusion. HF is defined clinically as a syndrome and the diagnosis of HF is dependent on meeting three criteria: symptoms (e.g. breathlessness, peripheral swelling), signs (e.g. raised jugular venous pressure and pulmonary crackles) and measurement of ejection fraction (EF). EF is the stroke volume divided by the end-diastolic volume, and is considered an important prognostic indicator in HF. The lower the EF, the worse the survival (Curtis et al., 2003). The major HF trials enrolled patients with EF ≤35 per cent, a group defined as systolic HF or HF reduced EF (HF-REF), and it is in this group that clinically effective therapies have been demonstrated. If the EF is preserved (EF ≥50%) the phrase HF with a preserved EF (HF-PEF) is used. Usually this is a diagnosis of exclusion, and most patients have evidence of diastolic impairment. Patients tend to be older, female, hypertensive, obese and less likely to have had a prior myocardial infarction. Prognosis is better in patients with HF-PEF than those with HF-REF (Yancy et al., 2006). At present, perioperative management of HF-REF and HF-PEF should remain similar pending further randomised controlled trials (McMurray et al., 2012).
The identification of the underlying aetiology is crucial to allow for the correct therapeutic intervention. Myocardial pathology causing systolic ventricular dysfunction is the commonest cause; however, abnormalities involving the valves, pericardium, cardiac arrhythmias or conduction and ventricular diastolic dysfunction can also cause HF (Felker et al., 2000; McMurray et al., 2012).
How each of these pathologies causes HF differs, but the overall result is a reduced cardiac output state. For example, in patients with myocardial infarction and associated LV systolic impairment, the death of cardiac myocytes leads to adverse remodelling and the ventricle dilates with a corresponding decline in the contractile function. If left untreated, there is progressive enlargement of the ventricle and worsening cardiac function, due to a cascade of physiological and neurohumoral responses. Mean arterial pressure is a function of cardiac output and peripheral vascular resistance. Therefore, as the cardiac output falls, the mean arterial pressure also falls. In response, the sympathetic nervous system is activated and catecholamines are released. Catecholamines cause an increase in heart rate (which increases cardiac output) and an increase in peripheral vascular resistance, thereby preserving the mean arterial pressure and completing the negative feedback loop. It also results in the activation of the renin angiotensin aldosterone system. The kidneys secrete renin when they are hypoperfused. Renin breaks down angiotensinogen to angiotensin I, which is further broken down by angiotensin, converting enzyme to angiotensin II. This acts on angiotensin receptors, causing vasoconstriction and stimulation of aldosterone and antidiuretic hormone, thus promoting the conservation of salt and water. In the short term, this preserves the mean arterial pressure and cardiac output, but in the long term, the net result of the sympathetic and neurohumoral activation is to cause a pathophysiological vicious cycle, leading to continued pathological remodelling of the ventricle and decline in cardiac output. Inhibition of these two mechanisms is the basis for current HF therapy.
A thorough clinical history and examination are a prerequisite to risk stratifying patients preoperatively. Further enquiry into a patient’s co-morbidities is essential, and will provide important prognostic information. Clinical risk factors, such as age (Turrentine et al., 2006), aetiology, New York Heart Association class (NYHA class) (Argenziano et al., 1999; Wechsler and Junod, 1989), EF (McEnroe et al., 1990), plasma natriuretic peptide concentration (Dernellis and Panaretou, 2006) and key comorbidities including renal dysfunction, diabetes, sleep-disordered breathing and anaemia, are associated with an increased operative risk (Eagle et al., 1989; Goldman et al., 1977; Lee et al., 1999).
Functional capacity of a patient is a simple way to assess a patient’s operative risk. NYHA functional classification dates back to 1928 and has been used to select and assess treatment effect in almost all randomised therapeutic HF trials, and thus remains an important prognostic marker. It places patients in one of four categories: NYHA class I represents patients with no symptoms and NYHA classes II, III, and IV, mild, moderate and severe symptoms, respectively. For example, in patients undergoing coronary artery bypass grafting, those with symptomatic HF have been shown to have a higher operative risk than those who are asymptomatic (Argenziano et al., 1999; Wechsler and Junod, 1989).
Despite the clear relationship with symptom severity and survival, problems do exist in applying NYHA assessment, including the subjectivity and thus accuracy of differentiating between the different classes. Exercise testing (6-minute walk test and cardiopulmonary exercise test (CPET)) may represent a more objective measure of exercise capacity and symptoms in patients with HF; however, the data are limited (Bagur et al., 2011; Sinclair et al., 2011). Another measure of functional capacity is the metabolic equivalent (MET). 1 MET is the metabolic demand at rest. 4 METs would be equivalent to walking up two flights of stairs or a hill and greater than 10 METs would be required for intense sporting activity (Fletcher et al., 2001). Please see also chapter 1, page 4.
CPET combines standard exercise testing with the measurement of ventilator gas exchange. Patients with HF and a reduced peak VO2 (the highest oxygen uptake during exercise) or VE/VCO2 slope at peak exercise (relationship between minute ventilation and CO2 production) have a worse prognosis (Guazzi et al., 2012). Please refer also to chapter 6, page 70. Angiotensin converting enzyme (ACE) inhibition and beta blockade have been shown to improve performance on CPET (Guazzi and Arena, 2009). HF patients with a peak VO2 of <11 mL O2/kg/min are likely to be at higher risk for surgery (Guazzi et al., 2012). Overall, the benefit of preoperative CPET has not been definitively evaluated. There are no randomised controlled trials, and the current trials are small and non-blinded (Young et al., 2012). However, the peak VO2 may have a future use in the preoperative setting.
B-type natriuretic peptide (BNP) and N-terminal pro-BNP are produced in response to left ventricular wall stress and remain a powerful prognostic biomarker in HF (Karthikeyan et al., 2009). Similarly, preoperative BNP confers a risk for non-cardiac surgery. In a study of 1638 patients undergoing non-cardiac surgery, those with raised plasma BNP had a higher rate of cardiac deaths, non-fatal myocardial infarction, acute pulmonary oedema and ventricular tachycardia post-operatively. Raised BNP was superior to the Goldmann multifactorial clinical index in this group (Dernellis and Panaretou, 2006). There is also a strong interaction between BNP and cardiovascular events in the immediate post-operative period (Karthikeyan et al., 2009). Measuring the BNP post-operatively in addition to the preoperative measurement adds further predictive value for mortality and non-fatal myocardial infarction (Rodseth et al., 2014).
If the NT-proBNP is normal (<300ng/L), HF diagnosis is unlikely. However, if the NT-proBNP is significantly elevated (see Table 4.1), acute HF is likely and should be confirmed by echocardiography if not already documented. Routine testing for BNP in asymptomatic patients is not recommended; however, for those with a poor functional capacity (e.g. NYHA III or IV or <4 METs), a BNP may be considered to aid in risk stratification (Kristensen et al., 2014).
TTE remains the most available and versatile tool in evaluating cardiac function in the preoperative assessment of patients with known or suspected HF. Cardiac function can be assessed in a variety of ways with TTE. For further details please refer to chapter 5A. Parameters routinely measured include left ventricular (LV) function (systolic and diastolic), valve function, inferior vena cava (IVC) diameter and deformation imaging. A reduced EF has been associated with an increased incidence of cardiac complications following surgery (Cowie, 2012; Rohde et al., 2001). The diastolic function of the LV can be assessed with mitral inflow pulsed wave Doppler, measuring the ratio of the E and A waves in patients with sinus rhythm or Doppler tissue imaging (DTI), using deformation and strain. In addition, TTE also provides useful assessment of either primary organic or secondary functional valve dysfunction. The imaging of the IVC remains a useful adjunct to the clinical assessment of fluid overload. The IVC should collapse by at least 50 per cent during inspiration. An increased IVC diameter with reduced respiratory variation implies an increase in the right atrial pressures.
In summary, TTE can improve risk stratification prior to non-cardiac surgery. In patients defined as high-risk, TTE should be considered prior to their operations (Kristensen et al., 2014). Cardiovascular magnetic resonance imaging remains an alternative in patients with poor echocardiographic windows.
Heart Failure Management
Two decades of randomised trials have established a variety of treatment options for systolic HF. HF management involves a combination of non-pharmacological (e.g. lifestyle modification/education including fluid restriction and low salt diet, smoking cessation) and pharmacological approaches. Current international guidelines (McMurray et al., 2012) recommend an initial triple treatment strategy with optimal dose ACE inhibitors (or angiotensin 2 receptor blockers (ARBs) in patients with ACE intolerance), beta-blockers and mineralocorticoid receptor antagonists (MRAs). Diuretics are recommended in HF patients with symptoms and signs of fluid overload. Thereafter, if specific criteria are met, drugs, including ivabradine (If channel sinus node inhibitor), digoxin and a combination of hydralazine/isosorbide dinitrate, can also be considered as treatment options (see later in this chapter). Furthermore, patients with HF-REF demonstrating ventricular conduction delay as evidenced by bundle branch block on ECG should also be evaluated for cardiac resynchronisation therapy (CRT) prior to major surgery.
In contrast, no treatment has yet to been shown to reduce mortality in patients with HF-PEF. Most patients with HF-PEF have diastolic dysfunction and LV wall stiffness and are at risk of pulmonary oedema and fluid overload in the perioperative period. Therefore careful attention to control of afterload and adequate diuretic therapy are important considerations.
In patients with a new diagnosis of HF-REF, it is recommended that intermediate or high-risk surgery be deferred for at least 3 months after initiation of HF treatment, so as to allow for treatment uptitration and possible reverse remodelling of LV function. The rapid preoperative initiation of high doses of beta-blockers and/or ACE inhibitors is contraindicated.
All HF patients should also be clinically euvolaemic with stable blood pressure prior to proceeding to elective surgery. Furthermore, it is imperative the following drugs are not prescribed/or at least avoided in patients with systolic HF. These include: the anti-diabetic thiazolidinediones (glitazones), which cause worsening HF and increase the risk of HF readmission; NSAIDs and COX-2 inhibitors, which can exacerbate fluid retention and worsening renal function and thus worsening HF; most negatively inotropic calcium channel blockers with the exception of amlodipine and felodipine and finally, the addition of an ARB to the combination of ACE inhibitors and MRA, due to the risk of renal dysfunction and hyperkalaemia.
Lifestyle modifications may be of benefit in the treatment of HF and should be encouraged in the preoperative setting. Regular exercise has the greatest amount of evidence in improving prognosis. One large randomised controlled trial showed an adjusted 11 per cent reduction in mortality or hospitalisation with 3 months of supervised training followed by home exercise in patients with severe LV impairment (<35%) and NYHA class II and III symptoms (O’Connor et al., 2009). Routine fluid restriction is not advised in those with mild to moderate symptoms, but may be considered in severe heart failure. The current guidelines do not recommend salt restriction due to equivocal evidence (McMurray et al., 2012; Paterna et al., 2008) and further clinical trials are required.
ACE inhibitors were the first to be shown to reduce mortality in HF-REF (The SOLVD Investigators, 1991). They prevent the cleavage of angiotensin I to angiotensin II. The effect of this is to reduce salt and water retention and attenuate the compensatory vasoconstriction, which in the long term is detrimental to heart function (as outlined earlier in this chapter). Various signalling pathways including via angiotensin II mediate ventricular remodelling in response to volume and pressure overload, and preventing this is important in maintaining cardiac function (Opie et al., 2006). Unwanted side effects include renal dysfunction, hyperkalaemia, cough and, rarely, angioedema. If a patient cannot tolerate an ACE inhibitor, an ARB should be considered.
An observed problem with ACE inhibition in the immediate preoperative setting is the potential for increased hypotension on induction of anaesthesia (Colson et al., 1992). This is usually easily remedied with vasopressors or intravenous crystalloid replacement. Some groups advocate holding the ACE inhibitor on the day of procedure or giving it on the evening before, and restarting post-operatively. It is imperative in the post-operative period that patients already taking ACE inhibitor continue to do so.
Beta-blockers have also been shown to reduce mortality in patients with HF (CIBIS-II, 1999; Packer et al., 2002). Through the inhibition of the sympathetic nervous system on beta receptors, they suppress arrhythmias and have been shown in the long term to reduce adverse LV remodelling and improve LV function. The long-term effect of this is a reduction in adverse ventricular remodelling and an improvement in LV function. Concern has also been raised about the use of beta-blockers in the preoperative setting following the controversial Dutch Echo Studies (Dunkelgrun et al., 2009). Furthermore, a recent study has shown that loading with high-dose metoprolol acutely prior to surgery in patients at high risk of cardiovascular disease was associated with a higher risk of stroke and all cause mortality compared to placebo (POISE Study Group, 2008), and thus current recommendations are not to initiate high-dose beta-blockade prior to non-cardiac surgery. However, beta-blockers should be continued in HF patients throughout the perioperative period (Kristensen et al., 2014). For details please refer to chapter 1, page 11.
Mineralocorticoid antagonists (e.g. spironolactone and eplerenone) inhibit the effect of aldosterone on the collecting duct in the kidneys and therefore prevent the reabsorption of water and sodium and have a potassium-sparing effect. When given in addition to ACE inhibitors and beta-blockers, they reduce mortality by 30 per cent relative to placebo (Pitt et al., 1999).
Although there is no prognostic benefit, diuretic therapy remains the mainstay of treatment in fluid-overloaded HF patients. Loop diuretics act on the ascending limb of the loop of Henle and the thiazide diuretics on the distal convulted tubules. By competing for binding sites on the chloride component of the Na+/K+/2Cl- co-transporter, loop diuretics increase the urinary excretion of sodium, chloride, potassium, magnesium and calcium. Thiazide diuretics block the Na+/Cl- transporter, reducing the reabsorption of sodium and chloride and therefore increasing their excretion along with water. Important side effects include electrolyte disturbances such as hypokalaemia, and therefore serum electrolytes should be monitored closely and corrected if necessary. A thorough assessment of fluid balance should be undertaken preoperatively. If there is evidence of volume overload, the diuretic may need to be increased. However, if the patient is hypovolemic, the diuretic dose should be reduced or stopped and intravenous fluids may be required. There is a high risk of iatrogenic fluid overload post-operatively, and therefore careful fluid balance assessment is required, and it is likely diuretics will be needed.
Ivabradine acts on the If channel in the sino-atrial node and slows the heart rate of patients in sinus rhythm. This negative chronotropic effect has been shown to significantly reduce composite cardiovascular death or HF hospitalisation compared to standard care (Swedberg et al., 2010). Ivabradine should be considered in patients with symptomatic systolic HF with an EF <35 per cent and a heart rate >70 bpm despite optimal beta-blocker therapy.
Digoxin, a cardiac glycoside, can also be considered in patients with symptomatic HF. It is particularly useful in patients in atrial fibrillation. Moreover, analogous to the beneficial effect observed with ivabradine, digoxin may have a role in patients in sinus rhythm (Hood et al., 2004). However, more recently there has been concern that the use of digoxin may in fact be harmful. A recent meta-analysis of more than 300,000 patients receiving digoxin demonstrated a 14 per cent increase in mortality in those patients with HF (Vamos et al., 2015).
Combination therapy with hydralazine and isosorbide dinitrate has been shown to reduce morbidity and mortality particularly, in the African American population (Taylor et al., 2004). Current international guidelines recommend both drugs in patients intolerant to ACE inhibitor/ARBs.
Cardiac Resynchronisation Therapy (CRT) – atrial-synchronised biventricular pacing has been found to be a safe and effective treatment for patients with severe systolic dysfunction and evidence of intraventricular conduction delay and should be considered prior to major surgery (Bristow et al., 2004; Cleland et al., 2005). It was the COMPANION (Bristow et al., 2004) (n = 1520 patients) and CARE-HF (Cleland et al., 2005) (n = 804 patients) trials, designed to investigate clinical end points, that firmly established the role of CRT in mainstream HF therapy by demonstrating a significant reduction in combined all-cause mortality and all-cause hospital admissions. On this basis, the recent European Society of Cardiology (ESC) Heart Failure Guidelines 2012 provide a class I recommendation with a level of evidence A for implantation of a cardiac resynchronisation device with or without a defibrillator function (CRT-D/CRT-P), for patients with LV systolic dysfunction (EF 35%), symptomatic HF despite optimal medical therapy and a QRS duration of >120 ms who are expected to survive with good functional status for >1 year. In patients with mild symptoms (NYHA class II), CRT is recommended with a QRS duration of either ≥150 ms irrespective of QRS morphology, or ≥130ms plus an LBBB pattern and an EF ≤30 per cent (McMurray et al., 2012). Please see also chapter 3, page 28.
A large proportion of preoperative patients has HF, and this is likely to further increase. A variety of treatment options is available for patients, including pharmacological and device therapy, and these should be optimised, if possible, prior to surgery. It is also important that HF specialist input is sought early and continued throughout the patient journey, including the post-operative period.