Management of severe heart failure

Past paper SAQs concerned with heart failure have been unsurprisingly common in the past papers, considering the prevalence of this in our daily workload. What is surprising is that they have focued extensively on right heart failure. In fact, the second paper of 2018 contained two questions on this topic. As no massive breakthroughs or upheavals have occurred in the literature, one might surmise that recently one of the examiners has completed a PhD on the subject, or otherwise the topic of right ventricular dysfunction forms a special interest area for them. Whatever the reason, by using past papers as a shadow curriculum the CICM candidate might come to the concusion that for the purposes of revision the left ventricle is a useless vestigial organ which has zero importance in critical care.

The past paper questions on heart failure have consisted of the following SAQs:

• Question 5 from the second paper of 2018 (right heart failure investigations)
• Question 27 from the second paper of 2018 (right heart failure management)
• Question 12 from the second paper of 2017 (multiple problems, among them being right heart failure)
• Question 2 from the second paper of 2015 (right heart failure)
• Question 17 from the second paper of 2009 (right heart failure)
• Question 12 from the second paper of 2000 (cardioselective beta blockers)
• Question 14 from the first paper of 2000 (medical management of CCF)

In spite of the dextrocentricity (Right-wingedness? Left hemineglect?) of the past papers, this chapter is an effort to address all heart failure in a way which lends a structure to one's approach to a generic CCF question.  It is of course difficult to do so in a manner which is comprehensive, given that heart failure management is a massive daunting topic. In order to approach this in a systematic manner, I have separated the problems of heart failure into the variables which govern cardiac output, and the means of manipulating those variables. The approach resembles that of the similar chapter from the section on cardiothoracic intensive care. 

The definitive resource for this is probably this 2022 guideline statement by the AHA/ACC and the 2023 ESC focused update on heart failure, an update of their 2021 guideline. Unless otherwise reported, these were the source for most of the information below. For the sane exam candidate who may be disinclined to read the entire fifty-page document,  of particular interest may be Section 7.3.2,  Pharmacological Treatment for Stage C HFrEF. For Question 17 from the second paper of 2009, the single most useful article on right heart failure was the JACC review by Lahm et al (2010). From Oh's Manual, one needs to be familiar with  Chapter 24 (Acute  heart  failure) by Nicholas  Ioannou,  Pratik  Sinha  and  David  Treacher (7th ed)

In brief:

Strategies for the Management of Heart Failure

Management of preload	Management of afterload	Management of contractility	Cheating
Diuretics Fluid restriction Venodilators Aldosterone antagonists Beta-blockers Maintenance of sinus rhythm and atrial systolic contribution Pacing to maintain AV synchrony	Left ventricle Vasodilators ACE-inhibitors Beta-blockers Right ventricle Normoxia Normocapnea Avoidance of excessive postive respiratory pressures Pulmonary vasodilators	Inotropes Digoxin Dobutamine Milrinone Levosimendan Cardiac resychronisation Supportive hormones and micronutrients (cortisol, insulin, calcium, glucagon, thyroxine, thiamine etc) Revascularisation	Increase cardiac output by unnatural means: LVAD ECMO Increase the pacemaker rate Decrease the organism's demand for cardiac output Hypothermia Paralysis/sedation

Definition and classification of heart failure

"Failure" implies the heart stops performing some duty which it is usually expected to faithfully execute, which suggests a mainly functional definition. Of these, the best known is the NYHA classification:

Class I	No limitation of physical activity. Ordinary physical activity does not cause undue breathlessness, fatigue, or palpitations.
Class II	Slight limitation of physical activity. Comfortable at rest, but ordinary physical activity results in undue breathlessness, fatigue, or palpitations.
Class III	Marked limitation of physical activity. Comfortable at rest, but less than ordinary activity results undue breathlessness, fatigue, or palpitations.
Class IV	Unable to carry on any physical activity without discomfort. Symptoms at rest can be present. If any physical activity is undertaken, discomfort is increased.

The reason they are known to all are not that these are the best possible classification categories, but because they are so simple that anyone can apply them, and so impossibly old that everyone has heard of them.  They originate from the 1928 booklet, "Diseases of the heart and blood vessels: nomenclature and criteria for diagnosis", of which only the more recent editions seem to be available to the casual reader. This classification system precedes all others and is held up as "a landmark in diagnostic medicine" because it was among the first to connect aetiology and functional/clinical features into a set of diagnostic criteria and risk stratification tools. It is also ancient. Of the first editions, one criticism was that extensive discussion of postmortem diagnosis supported the text, eg. where syphilis or endocarditis could be attributed only following culture acquired by autopsy. Because modern society is more supportive of antemortem diagnosis, we now have somewhat better criteria, such as these Universal Definitions from Bozkurt et al (2021):

"HF is defined as a clinical syndrome with symptoms and/or signs caused by structural and/or functional cardiac abnormality and corroborated by elevated natriuretic peptides and/or objective evidence of pulmonary or systemic congestion"

So, we require:

• Symptoms and/or signs caused by structural/functional cardiac abnormalities 

And

• at least 1 of: 
  • Elevated natriuretic peptides; 
  • Objective evidence of cardiogenic pulmonary or systemic congestion. 
  • for HFpEF, an LVEF of 50% or greater

This definition, one might observe, still permits symptoms, but also involves physical examination findings and the results of modern investigations. Additionally, a classification system based on LVEF is now well entrenched, based on the LV ejection fraction:

Type of HF According to LVEF	Criteria
HFrEF (HF with reduced EF)	LVEF ≤40%
HFimpEF (HF with improved EF)	Previous LVEF ≤40% and a follow-up measurement of LVEF >40%
HFmrEF (HF with mildly reduced EF)	LVEF 41%–49% Evidence of spontaneous or provokable increased LV filling pressures (eg, elevated natriuretic peptide, noninvasive and invasive haemodynamic measurement)
HFpEF (HF with preserved EF)	LVEF ≥50% Evidence of spontaneous or provokable increased LV filling pressures (eg, elevated natriuretic peptide, noninvasive and invasive haemodynamic measurement)

The 40% cutoff was based on the use of this threshold value for enrolment into influential trials (eg. PARADIGM‐HF, EPHESUS‐HF, MERIT-HF, DAPA-HF, DAPA‐HF, TRED‐HF)

The main disadvantage of these, of course, is the linguistic contortions required to pronounce the classification types. One often hears otherwise serious people mouth things like "heff reff" or "huff puff", which erodes their dignified authority somewhat. Additionally, other concerns can be raised:

• The system says nothing of the competence of the valves, which means rather severe failure due to mitral disease might be overlooked by the finding of an ostensibly normal EF
• The system is blind to the right side of the heart
• There is no mention of any patient-centric variables, eg. exercise tolerance or symptoms
• No attempt to unite this with the rate of onset or progression

On the other hand:

• Hard to argue with objective measurements
• these are the category criteria that separated the groups in heart failure trials, so there must be some meaning to them
• Yes, these are spot measurements, but you can still use them to patch together a trajectory

These are probably here to stay, and like the KDIGO criteria, they are therefore a ripe fruit for the CICM examiners to pick and serve.

Epidemiology of heart failure in the ICU

American and European societies must differ from the Australian society substantially in terms of cardiac failure risk factors, insofar as the epidemiological data from one continent is never quoted in the guidelines of the others.

• The Americans have a prevalence of 1.9% to 2.8% in adults (Bozkurt et al , 2023)
• The Europeans quote a prevalence of 1-2% of adults
• In Australia, the prevalence seems to be something like 0.6%, which seems unusually low

Considering the hospital acts as a highly sophisticated system to identify and concentrate sick patients, and considering how inconsistently the definitions are applied, the prevalence of heart failure in the ICU is probably underestimated, and when  Dar & Cowie (2008) listed large scale studies, the reported % of ICU admissions with heart failure ranged from 15% to 100%. The reduced EF variant seems to be the most prevalent (50% of them), and the others are somewhat less common.

Aetiology of heart failure in the ICU

For those practicing in the morally bankrupt West, where luxury and vice has softened all the ventricles, the dominant cause of heart failure is ischaemic heart disease exacerbated by hypertension and diabetes. The prevalence of this cause is said to be two thirds of all cases (Montone et al, 2024). But, as historically the aetiology questions from CICM exams have generally been of the "list the causes" format, trainees should develop a system for generating a list of differentials. What follows is only one such possible system. 

Causes of Acute Heart Failure

Vascular Coronary artery disease Aortic dissection Arrhythmia Pulmonary emboli Infectious Systemic sepsis Endocarditis Myocarditis Neoplastic Infiltration by extracardiac tumour Atrial myxoma Drug-induced Beta-blockers and calcium channel blockers Doxorubicin Alcohol Cocaine

Idiopathic and infiltrative Amyloidosis Peripartum cardiomyopathy Congenital Neuromuscular disorders, eg. Duchenne's Congenital heart disease Autoimmune Vasculitis of the coronary vessels Autoimmune endocarditis/myocarditis Sarcoidosis Traumatic Contusio cordis - cardiac contusion Endocrine and environmental Hypothyroidism or thyrotoxicosis Hypothermia Acidosis and alkalosis Electrolyte derangement

Those interested in more detail than this are offered a 70-page analysis of the global data from 2017, with the caveat that this will probably do nothing to improve their exam success.

Natural history and prognosis of heart failure in the ICU

It's bad. It has improved, but remains bad. Mortality from the Framingham data was 67% over five years from 2000 to 2010. According to Taylor et al (2017), it has improved slightly, with 10-year survival increasing from 19.8% in 2000 to 26.2% in 2007. Those enrolled in trials seem to do better than the general population. 

Management of ventricular preload

There are several methods of managing preload in heart failure. Typically, with LV failure the LV preload needs to be  on the lower side, because high LV preload tends to lead to pulmonary oedema. In contrast, with RV failure the preload needs to be high (but not too high) because the useless RV is essentially a passive conduit into the pulmonary circulation.

Preload has all sorts of precise and imprecise definitions  created by intelligent people - but it is essentially the volume of blood which manages to make its way back to the heart during diastole.  Generally speaking, in manipulating preload we have only two major choices: we can either change the volume, or the venous compliance. Volume can be changed by adding or removing fluids of some sort. Venous compliance can range from 30 to 300ml/mmHg, and can be manipulated by vasoactive substances. 

Manipulation of venous return in left heart failure

Diuretics: If there is evidence of fluid overload, one normally reaches for the diuretics. The general tendency of heart failure patients to become boggy with fluid has resulted in a Class I recommendation in favour of diuretic use. Loop diuretics are recommended as a start. In order to increase the fluid loss, one may add a thiazide (the increase in the delivery of sodium to the distal tubule will improve the activity of frusemide).

Fluid restriction begins to play more of a role once the patient is extubated and begins to complain of thirst to teir attending relatives. Bottles and jugs are brought. Oedema ensues. At the early intubated stage, it is much easier to have fine control over fluid balance.

Aldosterone receptor antagonists (such as spironolactone) are recommended not only for their diuretic effect (which is minor), but also because they seem to improve mortality by another effect, perhaps related tot he role of the renin-angiotensin-aldosterone system in the development of CCF. The RALES trial demonstrated a 30% improvement in mortality for patients receiving spironolactone.

Venodilators such as hydralazine and nitrates are useful in decreasing preload in patients whose renal function does not permit the use of diuretics (eg. if they produce no urine).

Manipulation of venous return in right heart failure

Question 17 from the second paper of 2009 has asked specifically about the management of right heart failure. General principles of heart failure management are preserved in this context, but the approach required to optimise RV preload is dramatically different. The RV requires a higher filling pressure, and you can never guess precisely how high it has to be without experimenting. Traditional means of assessing fluid responsiveness are ineffective in  RV failure.

Increase of diastolic filling time

Beta blockers: Question 12 from the second paper of 2000 discusses the role of cardioselective beta blockers in the management of heart failure. This role seems to be largely related to the improvement in cardiac contractility associated with the increase in diastolic time. During a nice long diastole your useless floppy ventricle can fill with blood to a satisfactory point (i.e. where its Frank-Starling relationship allows it to contract with a satisfactory force). Another beneficial aspect of beta blockade is the limits it puts on cardiac conduction and automaticity. The beta-blocked patient is less likely to develop a lethal MI-inducing tachycardia while mowing their lawn, or to develop VT. Cardiac oxygen consumption is decreased, and because of decreased contractility the subendocardial blood flow is better. This mortality benefit has been demonstrated both in cardioselective and non-selective beta blockers. Bisoprolol, carvedilol and metoprolol all decrease mortality, but carvedilol is not cardioselective. 

The main disadvantage of beta-blockade in the context of critical illness is the fixed cardiac output. With beta blockade, the ICU patient is unable to reflexively increase their contractility or heart rate in response to stress. Sepsis and trauma rapidly demonstrate the advantages of a working sympathetic nervous system among non-beta-blocked people. Less important disadvantages include lethargy, nightmares and depression.

Atrial systolic contribution

Pacing may improve A-V synchrony, and result in an improvement of ventricular preload. Unfortunately, frequently the atrial conducting system is fairly shot by the time you get your dual-chamber pacemaker. The atria may be very dilated by this stage, and their contractility may be impaired by fibrosis. A specific situation where atrial pacing is very beneficial is functional mitral regurgitation, where the dilated noncontractile atrium stretches open the mitral valve annulus and increases the severity of MR: in this context, maintaining a sinus rhythm could produce some benefit by decreasing or reversing the MR, thereby maintaining haemodynamic stability (Al Qahtani et al, 2019). 

Enhancement of ventricular compliance

Realistically, you can do little about this variable. Ventricular compliance in the standard heart failure patient is not amenable to any specific therapy.  A stupid-level answer to this question would be "keep them from infarcting"; an ischaemic LV is much less compliant than a healthy one (Bardet et al, 1977). Obviously, sometimes there are easily reversible compliance problems (for instance, a pericardial effusion which could be drained) but this is not something you can routinely rely upon. In the context of recent cardiothoracic surgery you could even open the chest and allow the pericardial sack to remain open, which is the ultimate compliance-improving trick. 

Enhancement of lusitropy

In the presence of severe diastolic heart failure, one may be able to improve diastolic function (or at least prevent a deterioration in diastolic function) by manipulating the lusitropy, i.e. the relaxation of the myocardium. This is an active process which is mediated by the removal of calcium from the cell, and depends on SERCA proteins. Generally, positive lusitropy is viewed as a property of catecholamine inotropes,  which phosphorylate phosphholamban (thus stimulating the removal of calcium from the cytosol and into the sarcoplasmic reticulum by SERCA), and troponin-I, (thus inhibiting calcium interaction with troponin-C). It is hard to track down precisely which papers this comes from, but it is reproduced in Kam (Principles of Physiology for the Anaesthetist) and in Achilles Pappano and Withrow Weir's Cardiovascular Physiology.  It is a property observed with β-agonists such as dobutamine; for example Parker et al (1991) were able to demonstrate "significant acceleration of isovolumetric relaxation" in patients both with normal LV function and with CCF. Conversely,  β-blockers tend to inhibit lusitropy.

Synergy of diuretics with inotropes

The fluid overload state in heart failure is a thing which exists mainly because of poor cardiac output. The fluid is accumulated because of the relentless cycle of defensive sodium-hoarding that results from decreased renal perfusion, driven largely by the effects of the angiotensin and aldosterone. And in response to fluid accumulating, the ventricles dilate like idiots, their valves becoming leakier and their Starling relationships becoming more tense and distant. The resulting reduced cardiac output decreases renal perfusion yet more, reinforcing bad neurohormonal behaviours, i.e. angiotensin and aldosterone continue to continue to squeeze and hoard.  

So:

• To introduce a loop diuretic into this situation will usually deplete some small amount of volume, which reduces the circulating volume slightly. Unfortunately with a heart like this, and with the peripheral circulation already maximally squeezed (squose?), the normal baroreceptor response cannot maintain the cardiac output, and it decreases even more, perfusing the kidneys even less, and leading to even more neurohormonal silliness. When one gives even heroic doses of frusemide to these patients, one is often rewarded with little urine, and a substantial change in their EGFR.
• To introduce an inotrope into this situation will usually produce a modest improvement in urine output by increasing renal perfusion and reversing some of the abovementioned mechanisms, as well as mechanically (the annuli of valves are muscular and the valve leaflets come together better in systole when their annulus can be persuaded to contract harder, leading to decreased regurgitation). The effect is often unsatisfyingly small, and would lead to a very gradual return to a normal volume state. If one were presently experiencing lifethreatening pulmonary oedema, one would have no patience for this strategy.
• Ergo, the combination of a diuretic and an inotrope can yield a faster return to a more normal ventricular volume at a better position along the Starling curve

Management of afterload

Manipulation of systemic vascular resistance to unload the left ventricle

Afterload can be reduced by arterial vasodilators such as ACE-inhibitors, nitroprusside, hydralazine, and numerous others. the whole point is to achieve an increase in cardiac output without any additional myocardial workload. Unfortunately, there is only so far you can push this: it is important to maintain some blood pressure, so as to maintain a gradient for flow into capillary beds. It is also important to maintain a certain diastolic pressure to ensure coronary arterial perfusion

ACE-inhibitors and angiotensin II receptor blockers (ARBS) earn a Class I recommendation from the AHA. This recommendation rides on the back of many trials. For instance, The CONSENSUS study (1988) demonstrated an improvement in both mortality and NYHA grade with the use of enalapril, which is one of the drugs asked about in Question 14 from the first paper of 2000. These drugs have been placed in the "afterload reduction" basket, but in actual fact their mortality benefit may be due to their influence on cardiac remodelling. Other drugs with purely afterload-reducing effects would include prazocin and hydralazine. An ancient article (Mason, 1978) reports that the effect of these drugs is to increase cardiac output and reduce the symptoms of fatigue.

This strategy has benefits in the context of mitral regurgitation, where forward flow is improved by decreasing systemic vascular resistance (Stevenson et al, 1990).  It is effective even in aortic stenosis (mild and moderate AS will still benefit, according to a 1980s paper by Greenberg et al). In severe AS, no amount of fiddling with peripheral resistance will be of any benefit: the limiting resistance is at the level of the valve. Well, perhaps that absolute statement is not entirely fair, as a patient with severe AS could still theoretically develop an afterload even higher than their stenosis conditions, and could therefore benefit from vasodilators (or, alternatively, the LV could fail spectacularly, and the afterload improvement afforded by vasodilators could become beneficial, as in this series) 

Manipulation of pulmonary vascular resistance to unload the right ventricle

In order to improve forward flow in right heart failure, a pulmonary vasodilator is sometimes used. The modern arsenal consists of nitric oxide, prostacycline, bosentan and sildenafil. On a slightly different shelf one may place milrinone and levosimendan, which will dilate the pulmonary circulation along with the systemic.

Oxygen: The pursuit of normoxia is of some importance. Hypoxic pulmonary vasoconstriction is to be avoided; oxygen is a cheap and easy way to optimise pulmonary vascular resistance. Lahn et al (2010) recommend aiming for an SpO₂ in excess of 92%.

Avoidance of excessive PEEP: all positive respiratory pressure is transmitted to the pulmonary circulation and is counterproductive in right heart failure. It adds to afterload and is to be avoided. Typically, these people do not have any indications for high pressure mechanical ventilation (unless ARDS is their primary pathology). Ask yousrelf: am I doing this person any favours by continuing to ventilate them? Would they be better off try it themselves on some high-flow nasal prongs?

Avoidance of hypercapnea:  In the weird backwards-world of the pulmonary circulation, everything that normally causes vessel dilation has the opposite effect. CO₂ increases pulmonary arterial pressure and RV afterload. After 4 hours of hypercapnea (created experimentally with an FiCO2 of 10mmHg) healthy volunteers had an increase in PA pressure from 22mmHg to 32mmHg (Balanos et al, 2003).

Pulmonary vasodilators: Of these, the ones in common use are inhaled prostacycline, inhaled nitric oxide and systemic sildenafil. Of nitric oxide, a 2015 review by Benedetto et al writes "this remarkable story has experienced a roller-coaster ride with high hopes and nearly universal demonstration of physiological benefits but disappointing translation of these benefits to harder clinical outcomes". Inhaled nitric oxide can sure dilate some porcine pulmonary arteries, but it doesn't seem to save lives. Presently, the only guidelines supporting its use are feeble expert recommendations from Europe.

Management of contractility

Enhancement of contractility with inotropes

Digoxin may improve contractility, and may control the rate of AF with atrial systolic benefits. In the ICU, digoxin is frequently inappropriate as a rate control agent for AF because of its ineffectiveness in the presence of increased sympathetic tone. However, it may have a role to play as a positive inotrope irrespective of the underlying rhythm. In the community, digoxin does appear to improve the symptoms of heart failure, even in people who are in sinus rhythm, though it does not seem to decrease mortality. Specifically, hospitalisation rates were improved. Digoxin is probably the only orally bioavailable inotrope in the current repertoire. All others (eg. the ill-fated attempt to feed milrinone to outpatients) have failed hideously. Inotropes for cardiogenic shock are discussed elsewhere.

Actual proper inotropes  in cardiac failure are generally viewed as measures of last resort, mostly because of their side-effect profile. The patient with severe decompensated heart failure will not appreciate AF, and there is a serious risk of this happening, which is why inotropes are not on offer to every single patient with a reduced ejection fraction. A 2023 statement from the ESC describes some of the indications of inotropes in heart failure, which include:

• Where heart failure overlaps into cardiogenic shock
• Where the patient is resistant to diuretics, and needs an inotrope to help push blood through the kidneys
• Where a test of cardiac reserve is required prior to some kind of intervention
• As a bridge to transplant or VAD

Coordination of depolarisation with biventricular pacing

The AHA recommend cardiac resynchronisation therapy (CRT) for patients who have an LVEF less than 35% and a LBBB with a QRS duration of greater than 150 msec. CRT in this group may improve ventricular contractile function, diminish secondary mitral regurgitation, and reverse ventricular remodeling.

Support of cardiac metabolism

This broad topic includes such weirdness as the hormonal contribution to cardiac function (eg. from insulin, glucagon, thyroxine) as well as the all-important maintenance of the various electrolyte gradients. The AHA makes statements to scoff at diet modification or supplements, but the situation is very different at the ICU level. Critically ill patients have enough reasons to have impaired cardiac metabolism and their endocrinology is deranged. They may be septic and their cardiac mitochondria may have been disabled by cytokines and endotoxin. They may have a relative adrenal insufficiency. Thiamine deficiency or hypothyroidism may be occult in this population. In short, anything is worth a try.

The following list of correctable negatively inotropic metabolic variables is offered in Oh's Manual. To the canonic list, the author has added a few italicised variables.

• Acidosis (pH <7.20)
• Hypoxia (PaO₂ < 60mmHg)
• Hyperkalemia (K⁺ > 5.5)
• Hypomagnesaemia (Mg⁺⁺ <0.9)
• Hypocalcaemia (iCa⁺⁺ <1.0)
• Hypophosphataemia (PO₄^- <0.8)
• Thiamine deficiency
• Cortisol deficiency
• Thyroxine deficiency
• Alkalosis

Cheating

Cheating with the pacemaker rate

Let us say you have a "fixed" cardiac output. That cardiac output will increase by 30% if we increase your heart rate from 70 to 90. The temporary pacemaker is an ideal device for this, because it does not have the side-effects of chronotropic drugs. Similarly, controlling an unnaturally high rate (eg. the 150 bpms of 2:1 conducted atrial flutter) is beneficial as it will improve diastolic filling.

Creative management of demand

Heart failure is by definition the failure fo the heart to produce a cardiac output which satisfies the body's needs. If one can do nothing about the cardiac output, it is therefore logical to make changes to the demand side of the equation. Reducing the demand for cardiac output comes in several flavours, and includes beta-blockade, sedation, paralysis, hypothermia, and the use of mechanical ventilation.

Cheating with mechanical devices

LVAD and ECMO would be considered cheating, and are discussed elsewhere.