Heart failure learning module 4: invasive treatment of HF

The mechanisms underlying the benefit of CRT are not clear. There is no convincing evidence that any one of several measures of mechanical dyssynchrony is associated with the beneficial effects of CRT, and the latest (2014) National Institute for Health and Clinical Excellence (NICE) guidance on CRT has removed any reference to measuring mechanical dyssynchrony on echocardiography as part of patient selection.³

QRS morphology, rather than duration, may be the important factor for predicting response and survival following CRT device implantation; LBBB is associated with improved survival following CRT device implantation compared to patients with non-LBBB QRS morphology.⁴

However, it may be that a major part of the benefit of CRT is simply to shorten the AV delay. Old studies in patients with HF and a prolonged PR interval showed that sequential AV pacing reduced mitral regurgitation and could improve some aspects of myocardial function. It may be that CRT (by pacing both ventricles) shortens the AV delay, whilst avoiding the potentially harmful effects of RV apical pacing alone.⁵

A common misconception is that insertion of a CRT device is similar to that of a pacemaker used to treat bradycardia. Whilst some similarities exist, there are significant and important differences in terms of the procedure itself and potential adverse effects.

A CRT device is commonly implanted under local anaesthetic and takes one to three hours. The patient may be in hospital overnight.

CRT devices consist of the following:

• small battery-powered generator (about the size of a match box) containing circuitry and a battery
• leads (wires) that are placed in the heart using X-ray guidance.

The leads stay in contact with the myocardium at one end, while the other end is connected to the generator.

Insertion of the right atrial and RV leads is performed in the same way as for a pacemaker used to treat bradycardia.

The significant difference for a CRT device is placement of a lead to allow pacing of the LV. In general, a lead cannot be positioned within the LV due to the thromboembolic risk associated with a foreign body inside the systemic circulation.

The LV lead is typically placed transvenously via the coronary sinus (CS) to lie on the epicardial (outside) surface of the left heart.

The CS lies in the left AV groove, collects blood from the myocardium and drains into the right atrium through the Thebesian valve.

Step 1 Pre-shaped guide catheters are used to access the CS (figure 2). This can prove challenging if there is significant cardiomegaly with associated tricuspid regurgitation. Furthermore, unlike the arterial circulation, the venous anatomy varies considerably from patient to patient.	Key: ICD = implantable cardioverter defibrillator; RA = right atrial; RV = right ventricular
Step 2 After CS cannulation, contrast is injected to visualise the vessel with or without the aid of an occlusive balloon (figure 3).	Key: Os = ostium

Step 3
LV pacing requires placement of a pacing lead down a suitable tributary of the CS, preferably in a posterolateral location. The lead may then be advanced over an angioplasty guidewire into the appropriate position (figures 4 and 5).

Key: ICD = implantable cardioverter defibrillator; LV = left ventricular; RV = right ventricular

Key: ICD = implantable cardioverter defibrillator; LV = left ventricular; RV = right ventricular

A common problem with CRT is inadvertent phrenic nerve stimulation, which results in diaphragmatic contraction at the programmed pacing rate – a particularly unpleasant side effect. This has been overcome, to a certain degree, by the development of LV pacing leads with multiple pacing points (electrodes) that can be switched on or off electronically without the need to reposition the lead. Step 4 Once the LV lead is satisfactorily positioned, the delivery catheter needs to be removed by means of splitting whilst leaving the lead in situ (figure 6).

Key: CRT-D = cardiac resynchronisation therapy with defibrillation; ICD = implantable cardioverter defibrillator; LV = left ventricular; RA = right atrial; RV = right ventricular

The majority of adverse events during CRT device implantation relate to LV lead placement and include:

• failure to implant LV lead (4.5–8.5%)
• coronary sinus dissection (0.5–2.1%)
• cardiac perforation/tamponade (0.3–2.1%)
• extracardiac stimulation/phrenic nerve stimulation (0.8–4%)
• LV lead dislodgement (0.4%)
• death (0.01–0.03%).

Following successful implantation, the CRT device needs to be programmed to ensure as near to 100% pacing as possible. The P wave-to-pacing stimulus interval has to be short to prevent normal LV activation via the AV node.

The device should be checked in the pacing clinic for levels of biventricular pacing (aiming for 100%), arrhythmia detection and pacing variables at set intervals.

It is important to note that the major CRT RCTs have recruited patients fulfilling select inclusion criteria (LVEF ≤35%, mainly in NYHA class II–III, in sinus rhythm with LBBB and a QRS duration >150 ms). Evidence is therefore strongest in this patient group.

The COMPANION (Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure) trial (n=1,520) showed a relative risk reduction (RRR) in death of 24% with a CRT-P (p=0.059) and 36% with CRT-D (p=0.003) compared to medical therapy.⁹

In a similar patient population, the CARE-HF (Cardiac Resynchronisation — Heart Failure) trial (n=813) compared CRT-P to medical therapy.¹⁰ The primary end point of time to death from any cause or hospitalisation for a major cardiovascular event occurred in 39% of patients receiving CRT-P compared to 55% of patients receiving medical therapy alone, with a significant reduction in mortality in the CRT arm versus medical therapy alone arm (20% vs. 30%, respectively; p<0.002).¹⁰

Both the MADIT-CRT (Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronisation Therapy) (n=1,820; NYHA I, 15%; NYHA II, 85%)¹¹ and RAFT (Resynchronisation–Defibrillation for Ambulatory Heart Failure Trial) (n=1,798; NYHA class II, 80%, NYHA III, 20%)¹² trials confirmed a similar reduction in HF events and survival benefit for CRT in patients with mild symptoms.

In both trials, more than half of the patients had HF due to ischaemic heart disease.^11,12 The benefit of CRT was seen mainly in patients with LBBB and a QRS duration >150 ms. The number needed to treat (NNT) with CRT to gain one life year over a period of three years is <7 and the NNT with CRT to reduce one hospitalisation is <4.¹³

CRT can be considered for patients with LVEF greater than 35% and symptoms of HF who have another indication for pacing.⁷ The BLOCK-HF (Biventricular versus Right Ventricular Pacing in Heart Failure Patients with Atrioventricular Block) study randomised patients who had high-degree AV block to receive either CRT or conventional RV pacing (n=691; NYHA I or II; LVEF, 35–50%). The primary end point of all-cause mortality, urgent HF visit or increase in LV end systolic volume index occurred in 56% of patients who received conventional RV pacing, compared to 46% of patients treated with a CRT device, demonstrating superiority of CRT over conventional RV pacing.¹⁴

There are several other groups of patients for whom CRT may be useful, but the evidence is less certain.

CRT may be a useful treatment for patients with advanced HF; one study published from a single centre found that CRT device implantation in patients with severe LVSD and broad QRS complexes who are dependent on intravenous (IV) inotropic support can enable LV recovery to the point that patients are weaned off support and survive to discharge.¹⁵ It is important to note that guidelines recommend against the use of implantable defibrillators in patients with NYHA class IV symptoms.^3,7

There is no evidence from RCTs that patients with atrial fibrillation (AF) benefit from CRT device implantation. The only trial recruiting a significant number of patients with AF – RAFT (Resynchronisation–Defibrillation for Ambulatory Heart Failure Trial) – showed no benefit in the AF subgroup.¹²

The subsequent CERTIFY study (Cardiac Resynchronisation Therapy in Atrial Fibrillation Patients Multinational Registry) compared patients in permanent AF undergoing CRT combined with either AV nodal ablation or rate-limiting medications to patients in sinus rhythm undergoing CRT (AF and AV node ablation [n=443]; AF and rate-limiting medication [n=895]; sinus rhythm [n=6,046]).¹⁶ Patients with AF and CRT who had undergone AV node ablation had a similar mortality rate to patients with CRT in sinus rhythm and lower mortality than patients with AF and CRT who were taking rate-limiting medications.¹⁶

Only one patient in RAFT had undergone AV junction ablation and it is possible that the neutral results were due to insufficient biventricular pacing in patients with AF.¹²

For every 1% increase in biventricular pacing percentage, the risk of mortality decreases by 6–10% among patients in sinus rhythm with CRT-D.¹⁷ Biventricular pacing percentage was found to be inversely associated with mortality in patients, regardless of underlying rhythm, in a large cohort study (n=36,935).¹⁸

Case series of patients with AF and LBBB suggests that CRT might be beneficial, particularly when strategies are used to maximise biventricular pacing rates, either pharmacologically or with AV node ablation.¹⁹

However, case series (whether collected together into a ‘meta-analysis’ or not) are no substitutes for clinical trials. Proper RCTs of different strategies for patients with LVSD, AF and LBBB are needed to prevent patients from being exposed to anecdote-based medicine.

The Echo-CRT (Echocardiography Guided Cardiac Resynchronization Therapy) study (n=1,679) recruited patients with narrow QRS complexes and echocardiographic evidence of mechanical dyssynchrony.²⁰ CRT did not confer any benefit to patients with narrow QRS complexes and was stopped early for futility.²⁰ However, the trial was not included in the NICE Assessment Group’s systematic review, hence the inclusion of patients with QRS >120 ms (rather than >130 ms as in ESC guidance) in the NICE recommendations for CRT.³

Two individual patient data meta-analyses from the landmark trials of CRT have suggested the prognostic benefit of CRT was less certain with a QRS duration 130–149 ms, with a signal towards potential harm in patients with a QRS <130 ms (table 1).^21,22

As a result, the most recent ESC HF guidelines have downgraded the recommendation for CRT in patients with LBBB and QRS duration 130–149 ms from class I (‘recommended’) to class IIa (‘should be considered’).⁷

Interestingly, this change was made between the 2016 and 2021 iterations of the guidelines, despite the absence of new data. Evidence, perhaps, of the problems implementing evidence from subgroup analyses into practice – can we base practice on subgroup findings within a trial or should we focus on patients who meet the trial eligibility criteria only? Subgroup analyses may be statistically underpowered. However, ignoring the potential benefits seen in subgroups may mean that patients who could benefit from a treatment miss out. The change in the guidelines recognises the weaknesses of subgroup analyses while acknowledging that there may be some patients who might benefit from a treatment, even though they do not necessarily meet trial eligibility criteria.

As with assessing efficacy of a new pharmacological intervention, such as an angiotensin-converting-enzyme inhibitor, initial trials concentrated on selective high-risk and/or highly symptomatic patient groups. This is also true of CRT. To date, no RCT has been published on the efficacy of CRT in patients with HF with a preserved ejection fraction, or mild or moderate LVSD. Studies are currently ongoing.²⁴

An individual patient meta-analysis of all randomised controlled data of CRT (n=3,872) found that the reduction in mortality or hospitalisation with HF with CRT compared to medical treatment alone was only statistically significant in those with LBBB morphology (HR 0.66; 95% CI, 0.55–0.78) but not in patients with right bundle branch block (RBBB) (HR 0.74; 95% CI, 0.44–1.23) or non-specific interventricular conduction delay (HR 0.82; 95% CI, 0.48–1.41). However, the same analysis found that patients with longer QRS durations derived greatest benefit from CRT device implantation, regardless of QRS morphology. Additionally, there was a signal that CRT may be associated with greater risk of mortality or hospitalisation with HF in patients with QRS duration <130 ms.

The ESC and NICE subsequently attempted to distil these findings into usable clinical guidelines; the ESC guidelines recommend CRT for symptomatic patients with LBBB and QRS ≥130 ms (class I recommendation – ‘indicated’), but only advise consideration of CRT in patients with RBBB and QRS ≥130 ms (class IIb recommendation – ‘may be considered’). CRT is contraindicated in patients with QRS <130 ms.⁷ NICE recommend CRT for symptomatic patients with LBBB and QRS >120 ms with CRT only indicated in patients with RBBB if they have a QRS >120 ms and NYHA IV symptoms or QRS >150 ms.³

ICDs consist of a ventricular lead positioned in the RV apex or septum and the generator. The ventricular lead can act as a standard pacing lead, but it also contains a shock coil. To deliver a shock, the device generates and discharges a large current across the heart from the tip of the ventricular lead to the generator. In some patients, a lead is also positioned in the right atrium if atrial pacing is required.

The ICD has two modes of action on ventricular arrhythmias – anti-tachycardia pacing (ATP) and shock. In the event that a patient with an ICD goes into ventricular tachycardia (VT), the device will attempt to ‘pace’ the heart out of the abnormal rhythm; the ventricular lead paces the RV at a rate greater than that of the ventricular rate of the arrhythmia. Programming within the generator then reduces the paced rate to that of a normal heart for a period of seconds thus terminating the arrhythmia. In the case of recurrent VT not amenable to ATP or VF, a shock is delivered by the ICD from the RV lead tip to the generator. Patients have variable awareness of ATP but will always know when the ICD has shocked them.

All patients should be counselled regarding procedural complications, inappropriate shocks, implications for driving and planning deactivation of the device.⁷

Inappropriate shocks used to be a common complication of ICDs (up to 40%), usually triggered by AF, other supraventricular tachycardias or artefact.³⁵ Rates are far lower (~5%) with modern programming approaches.³⁶

ICDs should be deactivated in patients with deteriorating HF, terminal illness or those with a ‘do not attempt cardiopulmonary resuscitation (DNACPR)’ order in place. Generator changes should only be undertaken after careful reassessment of treatment aims.^4,7

Once a common cause of mortality amongst patients with HF, the rate of sudden cardiac death has fallen dramatically in recent years (figure 7).³⁴ Advances in the treatment of risk factors for HF mean that patients are diagnosed at a much later stage in life and advances in the treatment of HF mean that, once diagnosed, patients live longer with the disease. As a result, there are many more competing causes of morbidity and mortality for patients with HF. Although ICDs remain in guidelines, the additional prognostic benefit alongside ‘quadruple therapy’ is unknown.

Data obtained from Shen et al., 201734 and McMurray et al., 201939 Key: BEST = Beta-Blocker Evaluation of Survival Trial; CHARM = Candesartan in Heart Failure – Assessment of Reduction in Mortality and Morbidity; DAPA-HF = Dapagliflozin in Patients with Heart Failure and Reduced Ejection Fraction; EMPHASIS-HF = Eplerenone in Mild Patients Hospitalization and Survival Study in Heart Failure; GISSI-HF = Gruppo Italiano per lo Studio della Sopravvivenza nella Insufficienza Cardiaca‐Heart Failure; MERIT-HF = Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure; PARADIGM-HF = Prospective comparison of ARNI with an Angiotensin-Converting Enzyme Inhibitor to Determine Impact on Global Mortality and Morbidity in Heart Failure; RALES = Randomized Aldactone Evaluation Study; SCD = sudden cardiac death; SCD-HeFT =-Sudden Cardiac Death in Heart Failure Trial; Val-HeFT = Valsartan Heart Failure Trial

CardioMEMS™ device

The CHAMPION trial (CardioMEMS™ Heart Sensor Allows Monitoring of Pressure to Improve Outcomes in NYHA Class III Heart Failure Patients) (2011) of the CardioMEMS™ device (which measures pulmonary artery pressure) is the largest RCT of such devices to date.⁴¹ The investigators enrolled 550 patients (average age, 61 years; NYHA class III with an admission to hospital with HF within the preceding 12 months; ~20% of whom had an LVEF ≥40%) all of whom had the CardioMEMS™ device implanted.⁴¹ Patients were then assigned to either a ‘treatment’ group (in which investigators were aware of the pulmonary artery pressure) or a control group (in which investigators were ‘masked’ to the pulmonary artery pressures).⁴¹

HF medications and diuretics were titrated based on pulmonary artery pressure (where known) and patient symptoms and signs assessed over the telephone at pre-determined intervals.⁴¹ There was a 28% reduction in the primary end point of the rate HF hospitalisation after six months in the treatment group compared to the control group.⁴¹

Longer-term outcomes were similarly better in the treatment group compared to the control group with reductions in HF hospitalisation rate (HR 0.67; 95% CI, 0.55–0.80; p<0.0001), all-cause hospitalisation rate (HR 0.84; 95% CI, 0.75–0.95; p=0.0003), and combined end point of death or HF hospitalisation rate (HR 0.77; 95% CI, 0.60–0.98; p=0.03) during the 18-month randomised-access period.⁴² However, there was no mortality benefit with continuous pulmonary artery monitoring (19% vs. 23%; p=0.23).⁴²

The GUIDE-HF trial (Haemodynamic-Guided Management of Heart Failure) (2021) of the CardioMEMS™ device enrolled 1,000 patients (average age, 70 years; 94% NYHA class II/III; median LVEF, 38%; 45% of whom had an LVEF >40%; median NT-proBNP, 1,480 ng/L).⁴³ All patients had the device implanted and were then randomised to standard of care or care guided by haemodynamic monitoring from the device.⁴³ The primary end point was total HF events (hospitalisation plus urgent care visits) or all-cause mortality measured at 12 months. There was no difference in the primary end point after 12 months.⁴³

However, the authors reported a ‘COVID-19 (coronavirus disease 2019) sensitivity analysis’ in which they excluded all events that occurred after the onset of the COVID-19 pandemic.⁴³ The authors reported a 19% reduction in the primary end point, driven entirely by a reduction in HF hospitalisations with no effect on either urgent hospital visits for HF, or death.⁴³ The finding is counterintuitive and cannot be used to justify the use of the device beyond the findings of the whole trial.⁴³ The COVID-19 pandemic led to national lockdowns and a reduction in patients having contact with healthcare providers. The CardioMEMS™ device, which allows for remote detection of early decompensation and titration of medications, is perfect for such a situation. However, all benefit from the device was observed in the weeks and months before lockdowns came into effect. The device had no impact on outcomes during the time it was designed to have the greatest impact.

It remains unknown whether the costs involved in implanting, monitoring and maintaining the devices are outweighed by the clinical benefit.

The MultiSENSE trial (Multisensor Chronic Evaluation in Ambulatory Heart Failure Patients) (n=900; average age, 66 years; 94% NYHA II–III; median NT-proBNP, 754 ng/L) found that an algorithm using variables measured by pre-implanted CRT-D devices could identify patients at greater risk of HF-related events (death, hospitalisation with HF, increased diuretic treatment without hospitalisation).⁴⁴ Remote identification of patients at greater risk may allow earlier intervention that reduces the rate of adverse outcomes, although this was not tested in the study.⁴⁴

HeartPOD™ device

The HOMEOSTASIS trial (Hemodynamically Guided Home Self-Therapy in Severe Heart Failure Patients) (n=45; average age, 66 years; average LVEF, 32%; NYHA II–III) established the safety and potential outcome benefits of titration of HF medications based on readings from an implantable pressure monitor in the left atrium (HeartPOD™).⁴⁵ Medication changes depending on readings had been pre-programmed into the device by investigators with the broad aim of reducing or withdrawing diuretics from patients with low left atrial pressures or increasing diuretic in those with high pressures.⁴⁵ These changes could be implemented by patients based on instructions from the device without involvement of the medical team.⁴⁵

The LAPTOP-HF study (Left Atrial Pressure Monitoring to Optimize Heart Failure Therapy Study) of the HeartPOD™ device in patients with HFrEF and NYHA class III symptoms was stopped early due to a high rate of procedure-related complications upon device insertion.⁴⁶ Amongst the 486 patients enrolled (out of a planned 730) titration of treatment based on left atrial pressure was associated with a 41% reduction in the risk of admission with HF compared to the control group (p=0.005).⁴⁶

One problem with the above trials, and indeed with home telemonitoring in general, is that each device produces a large volume of data, much of which is difficult to interpret remotely from the clinical context. Another issue is that of the time taken to review and act upon the data generated; it is quite possible that by the time a doctor or nurse has had a chance to review the data and make a clinical decision, the adverse event predicted (e.g. hospitalisation with HF) has already occurred.

Patient-directed titration of HF medication (such as in HOMEOSTASIS and LAPTOP-HF) would ultimately bypass this step and may prove beneficial in the future. Interest in remote monitoring remains high and the advent of ‘wearable technology’ may lead to devices that can direct patients to manage their condition without the time-consuming involvement of their overstretched medical team or the risks associated with implantable devices.

A cardiac contractility modulation (CCM) device is an implantable device that delivers a low-voltage, biphasic current to the myocardium after the detection of the QRS complex during the ventricular refractory period.⁴⁷ The result is to increase the strength of LV contraction without triggering a new action potential. Implantation is similar to that of a simple pacemaker with right atrial and ventricular leads.

Early studies suggest that CCM devices can improve exercise tolerance and QoL in patients with HFrEF, NYHA class II–IV symptoms and narrow QRS complex (<120 ms). However, despite being available for over a decade, the development in the technology has progressed slowly and there is no evidence of benefit on mortality or morbidity.⁴⁸ More work is needed.

Recent years have seen a variety of implantable devices developed that are able to increase vagal tone by electrical stimulation of different nerve fibres. Increased vagal tone (activation of the parasympathetic nervous system) may be potentially beneficial for patients with HF in that it may counteract some of the deleterious effects of sympathetic activation.

Sleep disordered breathing (SDB) is a description of abnormal breathing patterns during sleep, affecting up to half of patients with HF. The condition is associated with a poor prognosis.⁵³

There are two major types of SDB:

• obstructive (upper airway collapses but respiratory effort continues)
• central (respiratory drive ceases).

Ultrafiltration is an extracorporeal process that attempts to mimic filtration in the Bowman’s capsule. Venous blood is removed from the body and passed through a circuit with a membrane permeable to smaller molecules, such as water and electrolytes. Blood is maintained at a pressure greater than the transmembrane pressure; hydrostatic pressure forces water and solutes to cross the membrane and out of the blood. The intravascular space is then refilled with fluid from the interstitium, thus reducing congestion.

It differs from dialysis in that the latter is usually intermittent, allowing solutes – which can be larger-sized molecules (e.g. toxins, lactate) – to diffuse through a semi-permeable membrane down the solute’s concentration gradient into the dialysate fluid.

To date, there are three key trials of ultrafiltration in patients with severe fluid retention: UNLOAD (Ultrafiltration Versus Intravenous Diuretics for Patients Hospitalised for Acutely Decompensated Heart Failure), CARRESS-HF (Effectiveness of Ultrafiltration in Treating People with Acute Decompensated Heart Failure and Cardiorenal Syndrome) and AVOID-HF (Aquapheresis Versus Intravenous Diuretics and Hospitalizations for Heart Failure) studies.^61–63

In the UNLOAD trial, investigators enrolled 188 patients admitted with HF (average age, 62 years; median BNP 1,256 ng/L in the ultrafiltration arm).⁶¹ Patients were randomised to either ultrafiltration or loop diuretic therapy within 24 hours of hospitalisation.⁶¹ The co-primary end points were weight loss and dyspnoea relief at 48 hours with net fluid loss and rehospitalisation with HF after 90 days among the pre-specified secondary end points.⁶¹ The ultrafiltration group had greater weight and fluid loss (5.0 kg vs. 3.1 kg, p<0.001 and 4.6 L vs. 3.3 L, p=0.001 respectively), but there was no difference in patient-reported dyspnoea.⁶¹ Rehospitalisation rate was lower in the ultrafiltration arm, but the trial was not powered to detect a significant difference. There was significantly less hypokalaemia and also no difference in serum creatinine with ultrafiltration compared with diuretic treatment.⁶¹

In the CARRESS-HF trial, investigators enrolled 188 patients admitted with HF and deterioration in renal function from baseline (median age, 69 years; median LVEF, 30%; median NT-proBNP, 5,013 ng/L in the ultrafiltration arm).⁶² Patients were randomised to titration of IV diuretic treatments (plus treatment with a thiazide diuretic, vasopressors or inotropes) based on signs and symptoms, or ultrafiltration at a rate of 200 ml/hr.⁶² Treatment was continued until the signs and symptoms of congestion were reduced at the judgement of the treating doctor. The co-primary end point was change in serum creatinine and change in weight at 96 hours.⁶² Both groups lost a similar amount of weight, but serum creatinine was significantly higher in the ultrafiltration treatment group compared to the pharmacological treatment group (20.3 µmol/L vs. −3.5 µmol/L; p=0.003).⁶²

It is difficult to know what the correct interpretation of the CARRESS-HF trial is. Factors suggesting that ultrafiltration might not be harmful (as is commonly interpreted) include:

• there was substantial dropout and cross-over of patients from the ultrafiltration to the pharmacological treatment arm in the first 24 hours⁶²
• although 90% of patients enrolled in the trial had residual congestion at the time of evaluation of the primary end point (96 hours), only 32% of patients in the ultrafiltration arm were still receiving their assigned treatment compared to 80% of patients in the pharmacological arm⁶²
• a recent per-protocol analysis of the CARRESS-HF data found that ultrafiltration was associated with greater cumulative (p=0.003) and net (p=0.001) fluid loss, and greater weight loss relative to baseline (p=0.02) compared to pharmacological therapy⁶²
• rather than being an indicator of disease progression or harmful treatment, a transient increase in serum creatinine during decongestion may actually indicate more effective fluid loss and is associated with better post-discharge outcomes than in patients with no change or a decrease in serum creatinine levels.⁶⁴

In the AVOID-HF trial, investigators aimed to enrol ~800 patients admitted with HF who would be randomised to ultrafiltration or conventional therapy. Unfortunately, the trial was terminated early by the sponsor with only 224 patients enrolled, so although the primary end point of HF event at 90 days favoured ultrafiltration, the effect was not statistically significant.⁶³

The underlying trigger for the subsequent changes in chronic HF caused by the stress signalling pathway is, in part, chronic stimulation of beta-adrenergic receptors (β-ARs) leading to alterations in gene expression by cyclic adenosine monophosphate (cAMP)-dependent signalling pathways.⁶⁶

Some of the changes in gene expression are currently being exploited as biomarkers in HF: for example, up-regulation of the human natriuretic peptide B (NPPB) gene for NT-proBNP. However, many of the changes to gene expression in HF affect functional proteins and are damaging to the heart.

Gene therapy is ‘the introduction or alteration of genetic material within a cell or organism with the intention of curing or treating a disease’,^65,67 and may allow direct modification of the damaging changes to gene expression that are associated with HF.

In therapeutic terms, this either increases the production of a deficient protein − commonly by delivering a new copy of the relevant gene − or inhibits production of a deleterious protein, for example by interfering with RNA.

Although there is much enthusiasm for gene therapy and much promising laboratory data, there are no convincing clinical trial data in humans as yet.

Potential targets for gene therapy⁶⁶

A plethora of changes occur in gene expression in HF affecting proteins involved in the following:

• contractile proteins (α- and β-myosin heavy chains)
• cytoskeleton (titin and tubulin)
• pro-inflammatory cytokines (tumour necrosis factor alpha [TNF-ɑ] and interleukin [IL]-6)
• receptors (β-ARs and ryanodine receptors)
• transporters (sarcoendoplasmic reticulum calcium adenosine triphosphatase [SERCA] and Na⁺/K⁺ -adenosine triphosphatase)
• metabolism switches from free fatty acids to carbohydrate (glucose transporter [GLUT], pyruvate dehydrogenase kinase [PDK])
• ion channels (Nav1.5 and K_ir)
• calcium binding proteins (calmodulin and phospholamban)
• exchangers (sodium-calcium exchanger [NCX] and sodium/proton exchangers [NHX])
• kinases (CaMKII).

Vectors and delivery systems for gene therapy⁶⁸

Once the therapeutic gene has been selected, an appropriate vector and delivery method is required. There are various techniques used in laboratory research (table 2) but in the small numbers of clinical trials that have been carried out, the most common vector is the adeno-associated virus and the delivery method is the antegrade intracoronary infusion without balloon occlusion.

Clinical experience of gene therapy

Numerous animal studies have targeted various genes and used a variety of vectors and delivery methods. The genes which have been successfully targeted encode for proteins involved predominantly in either calcium-handling (sarcoplasmic [endoplasmic] reticulum calcium adenosine triphosphatase (SERCA2a), phospholamban (PLB), protein phosphatase 1, inhibitor-1, S100A1 and small ubiquitin-related modifier 1) or the β-adrenergic system (upregulation of β-ARs, inhibition of G-protein-coupled receptor kinases, activation of adenyl cyclase).⁶⁸

Many of these laboratory studies have shown promise, but the only area which has advanced to clinical trials targets the SERCA2a gene using an adeno-associated virus as the vector, delivered during an intracoronary infusion. The function of SERCA2a in the normal physiology versus HF state.⁶⁹

Typical cycling of Ca²⁺ within cardiac cells commences with the depolarization of the cardiac action potential, leading to a minor influx of Ca²⁺ into the cytoplasm via L-type Ca²⁺ channels.⁶⁹ Subsequently, this initiates a larger release of Ca²⁺ from the sarcoplasmic reticulum (SR) reservoir through the calcium release channel (CaRC), with Calstabin (FKB12.6) playing a role in stabilising CaRC function.⁶⁹ This calcium-induced release further stimulates contraction by binding with troponin C in the cardiac myofilaments. During relaxation (diastole), Ca²⁺ is reabsorbed into SERCA2a and expelled from the cell via the NCX, with the function of SERCA2a regulated by PLB.⁶⁹

In HF, abnormalities in Ca²⁺ regulation occur, including diminished expression of SERCA2a and inhibition of its function by PLB, resulting in compromised Ca²⁺ reuptake during diastole, heightened diastolic tension, and impaired relaxation, leading to reduced availability of Ca²⁺ for contraction and weakened contractions.⁶⁹ Although PLB levels remain unchanged in HF, reduced phosphorylation of PLB intensifies its inhibition of SERCA2a, leading to an accumulation of cytosolic Ca²⁺, which is offset by an increase in NCX expression. Additionally, hyperphosphorylation of CaRCs by protein kinase A or CaMKII disrupts FKB12.6 association, destabilising CaRC and promoting Ca²⁺ leakage during diastole. This combination of increased Ca²⁺ extrusion via NCX and elevated leakage from CaRC during diastole fosters delayed after depolarisations and triggered activity, which may precipitate ventricular arrhythmias.⁶⁹

However, results have been disappointing; phase II trials, CUPID 2 (Calcium Up-Regulation by Percutaneous Administration of Gene Therapy in Cardiac Disease Phase 2b) and AGENT-HF (AAV1-CMV-SERCA2a Gene Therapy Trial in Heart Failure), failed to show an improvement in ventricular dimensions or outcome with gene therapy that restored function of SERCA2a in patients with HFrEF.^70,71

One frequently cited reason for the failure of both CUPID 2 and AGENT-HF was the relatively low myocardial uptake of the viral vectors compared to the pre-clinical animal studies.⁶⁸ Advancements in more ‘cardiotropic’ vectors mean that work to find an effective genetic therapy for patients with HF continues.

Genetic therapies have been proposed as a potentially curative treatment for heritable causes of HF such as muscular dystrophy but translating promising results in animal models into clinical trials is proving problematic.⁷²

Finally, there is ongoing interest in ‘gene-silencing’ therapies for transthyretin (ATTR) amyloidosis, an increasingly recognised cause of HF with normal ejection fraction in elderly patients. Production of transthyretin occurs in the liver and can be interrupted by RNAs specifically engineered to inhibit translation of the gene. Three such so-called ‘small interfering RNAs’ exist; revusiran, inotersen and patisiran.⁷³

Treatment with inotersen or patisiran can improve various neurological end points in patients with hereditary ATTR and polyneuropathy compared to placebo.^74,75 However, none of these studies had HF-related end points and enrolment into the only study involving patients with cardiac amyloidosis, the ENDEAVOUR study (Study to Evaluate the Efficacy and Safety of AZD4831 in Participants with Heart Failure with Left Ventricular Ejection Fraction >40%) of revusiran, was halted due to safety concerns arising from the phase II extension study in which more than half of patients receiving revusiran developed peripheral neuropathy.⁷⁶

On the cellular level, HF is characterised by the loss of functioning myocytes. Regenerating myocardial tissue with cell therapies is a potential future treatment for HF. Types of cells under investigation include bone marrow mononuclear and mesenchymal stem cells, skeletal myoblasts and cardiac progenitor cells.⁶⁶

The field is littered with many small studies conducted in single centres with no control groups; the promising preclinical data is yet to translate into convincing benefit in phase II and III trials. Insufficient retention of the cells (delivered by injection or infusion) is a significant problem and generation of ‘new’ myocardium is yet to be demonstrated.^78,79 Ventricular arrhythmia is a common adverse event in clinical trials, and the trials have only involved a very select group of patients. The safety of stem cell therapy has not yet been confirmed. Additionally, ethical issues surrounding the use of human embryonic stem cells may limit further work in the field.