2010 Focused Update of ESC Guidelines on Device Therapy in Heart Failure

The Committee for Practice Guidelines (CPG) of the European Society of Cardiology recognizes that new evidence from clinical research trials may impact on current recommendations. The current heart failure (HF) guidelines1 were published in 2008 and the cardiac pacing guidelines in 2007.2 In order to keep these guidelines up to date, it would be appropriate to modify the recommendations and levels of evidence according to the most recent clinical trial evidence. This Focused Update on the use of devices in heart failure 2010 is the first publication of its kind from the CPG. Practice Guideline recommendations should represent evidence-based medicine. Traditionally, these recommendations are based on the outcomes in the cohort of patients described by the inclusion criteria in the protocols of randomized clinical trials (RCTs). More recently, based on the fact that the characteristics of the patients actually included in a trial may differ substantially from the eligibility criteria, Guideline Task Force members frequently favour restricting the applicability of these recommendations to the clinical profile and outcomes of the enrolled cohort, representing a more accurate interpretation of the evidence provided by a trial's result. In contrast to previous guidelines, this focused update considers the characteristics of the patients included in the trials and contains several examples. In MADIT-CRT, although the protocol permitted inclusion of patients in both New York Heart Association (NYHA) I and II function class, only 15% of the patients included in this trial were classified as NYHA I, many of whom had been previously symptomatic. Similarly, although the inclusion criteria permitted randomization of patients with a QRS width of ≥130 m, the favourable effect on the primary endpoint was limited to patients with a QRS width of ≥150 ms, a prospective, pre-specified cut-off. The text accompanying these recommendations explains and justifies the decisions to diverge from a traditional recommendation based strictly on the protocol inclusion criteria. The Task Force hopes that the users of the Guidelines will appreciate that this adjustment provides a more realistic application of the trial evidence to daily clinical practice. The management of patients with HF represents a substantial economic burden and hospitalization is responsible for >50% of this expense.3 The initial expense of device implantation must be weighed against measures of short- and long-term efficacy with regard to survival, morbidity, and quality of life.4 The effective use of limited health care resources necessitates identification of the characteristics of the patient population most likely to benefit from cardiac resynchronization therapy (CRT) and treatment strategy should target these patients for device implantation. The clinical effects of long-term CRT have been evaluated in a large number of randomized multi-centre trials with crossover or parallel treatment assignment,5–11 using CRT pacemakers (CRT-P) or CRT-implantable cardioverter defibrillator (ICD) devices (CRT-D). Practice with regard to the choice of the CRT device varies widely between countries.4 Meta-analyses were also published,12–15 suggesting that the most efficacious option in patients with HF and low left ventricular ejection fraction (LVEF) would be a CRT-D. The usual study enrolment criteria were: NYHA function class III or IV despite optimal pharmacological treatment, LVEF ≤35%, sinus rhythm (SR), left ventricular (LV) dilatation but with varying definitions, and QRS duration ≥120/ ≥ 130 ms. All RCTs have confirmed a significant alleviation of symptoms and increase in exercise capacity conferred by CRT. On average, NYHA function class decreased by 0.5–0.8 points, the 6 min walk distance increased by 20%, and peak oxygen consumption increased by 10–15%. The functional benefits and quality of life improvements were sustained.11,16,17 In the COMPANION trial, CRT with or without an ICD, lowered the combined endpoint of all-cause mortality and rehospitalization for HF by 35–40%, mainly driven by the 76% lower rate of hospitalizations.10 In CARE-HF, CRT-P lowered the proportion of unplanned hospitalizations for worsening HF by 52%, and the number of unplanned hospitalizations for major cardiovascular events by 39%.11 CARE-HF and COMPANION were trials powered to examine the effects of CRT on combined primary endpoints of morbidity and mortality.10,11 In COMPANION, CRT-D was associated with a significant decrease in all-cause mortality (relative risk reduction: 36%; P = 0.003), while the 24% relative risk reduction in mortality associated with CRT-P was nearly statistically significant (P = 0.059). A limitation of COMPANION was the absence of pre-specified analysis to compare CRT-D and CRT-P, precluding demonstration of the superiority of one CRT strategy over the other. In CARE-HF, where only CRT-P was assessed, a 36% relative reduction in the risk of death (P 6 months. Post hoc analysis found that time to all-cause mortality or first all-cause hospitalization was significantly improved by both CRT-P and CRT-D as compared with optimal medical treatment. No significant benefit was observed on all-cause mortality. The 2-year mortality rates were 55% and 45% with CRT-D and CRT-P, respectively, compared with 62% in the control group. A significant functional improvement was also documented. These data support the use of CRT to improve morbidity (but not mortality) in ambulatory class IV patients. Favourable outcome was defined as freedom from death or major cardiovascular event in CARE-HF.18 A baseline typical left bundle branch block (LBBB) pattern predicted a favourable outcome. By multivariable analysis, prolonged PR interval and right bundle branch block (RBBB) were the only predictors of non-favourable outcome. The 5% of patients with RBBB had a particularly high event rate. One large study, MIRACLE ICD9 and one large meta-analysis15 support the choice of a CRT-D in patients in NYHA class III/IV, with LVEF of ≤35%, QRS width of ≥120 ms with a conventional indication for an ICD. The role played by CRT in patients presenting with no or only mild manifestations of HF, a depressed LVEF and a wide QRS complex, has been addressed in three trials. The MIRACLE ICD II9 trial enrolled 186 candidates for ICD, who presented in NYHA function class II and in SR, and whose LVEF was ≤35%, QRS duration ≥130 ms, and LV end-diastolic diameter ≥55 mm. All patients received a CRT-D, and CRT was randomly activated in 85 patients. Despite the development of significant reverse LV remodelling, their exercise capacity was not increased. The large MADIT-CRT20 and REVERSE21 randomized trials evaluated the incremental benefit conferred by CRT in medically optimally treated patients. MADIT CRT enrolled 1820 patients in NYHA function class I (15%) of ischaemic aetiology or II (84%) of any aetiology and SR, whose LVEF was ≤30% and QRS duration ≥130 ms.20 Using a 2:3 randomization scheme, 731 patients were assigned to receive an ICD and 1089 received a CRT-D. The primary endpoint was a composite of death from any cause and non-fatal HF-related adverse events. During a mean follow-up of 2.4 years, the relative risk of sustaining a primary endpoint was reduced by 34% in the CRT-D-treated group, a benefit attributable primarily to a 41% decrease in HF-related adverse events. The ∼3% annual mortality was similar in both study groups. MADIT-CRT was stopped prematurely by the Data Safety Monitoring Board when a rigorous, pre-specified, stopping boundary was crossed, ultimately, at the P 35% (the threshold value for ICD indication in HF) after 6–12 months of CRT. Third, CRT-D seems to be associated with a higher risk of device-related complications as compared with CRT-P.26 The relative risk–benefit advantage of CRT-D over CRT-P remains unclear, especially in this population with milder symptoms. Randomized studies of CRT to date have been almost exclusively restricted to patients in SR. This contrasts with the high prevalence of CRT use in routine practice as indicated by the recent ESC CRT survey,27 thus indicating a need for prospective controlled trials. Approximately one-fifth of patients receiving CRTs in Europe have permanent atrial fibrillation (AF). The prevalence of AF in patients with HF is linked to the severity of the disease: 5% in NYHA I as compared with 25–50% in NYHA III/IV patients.28,29 Patients suffering from AF and ventricular dysschrony are typically older, and have a higher prevalence of comorbidity and a worse prognosis than patients in SR.27,30–32 It should be emphasized that patients with symptomatic HF, AF, and an LVEF of ≤35% may satisfy the criteria for ICD implantation. The presence of QRS prolongation would favour implantation of a CRT-D in these patients. In that the evidence is limited in AF and most of the patients included in trials had a very wide QRS width, we restrict our recommendation for CRT-P/CRT-D to QRS ≥130 ms. Some patients with permanent AF may resume SR during long-term treatment or following successful left atrial ablation.33,34 No comparative data exist on the efficacy of rhythm versus rate control strategy in patients with either paroxysmal/persistent or permanent AF, HF, and QRS duration ≥120 ms. Current knowledge restricts us to the use of rate control strategy in the subgroup of patients with permanent AF. In this latter group of patients outcomes are more difficult to measure, since both heart rate control and CRT may contribute to the observed changes in clinical status.35 An adequate trial with pharmacologically induced rate control is advisable. However, there is consensus that essentially complete ventricular capture is mandatory in order to maximize clinical benefit and improve the prognosis of patients with permanent AF.36 This often requires creation of complete heart block by ablation of the AV junction given the frequently inadequate efficacy of pharmacological treatment of ventricular rate control at rest and during exercise. Frequent pacing is defined as ≥95% pacemaker dependency.37 Since the publication of the previous versions of guidelines on CRT, mortality data from a large prospectively designed registry including AF patients30 and several small observational studies38,39 in addition to a meta-analysis have been published.40 The majority of patients in this meta-analysis had undergone AV nodal ablation. A large, prospective, observational registry33 showed that, during long-term follow-up, hybrid therapy combining CRT with AV ablation (resulting in 100% effective biventricular stimulation) conferred improvements in LV function and exercise capacity comparable to those achieved in patients with SR. In the same cohort,28 the authors provided evidence that patients with HF and AF treated with CRT received the same survival benefit as those achieved in patients with SR only when AV ablation was performed shortly after CRT implantation. These observational data need to be confirmed in randomized controlled studies in the cohort of patients with HF and permanent AF. Although prospective randomized controlled studies specifically addressing the issue of CRT in patients with a narrow QRS complex are currently lacking, there are several retrospective observational series or small prospective trials demonstrating a clinical benefit of upgrading to biventricular pacing with long-standing right ventricular pacing, severe ventricular dysfunction, NYHA function class III symptoms, regardless of QRS duration.41–46 This may indirectly indicate that preservation and/or restoration of an intrinsic, near-normal activation sequence by biventricular pacing should be pursued regardless of rhythm. It is important to distinguish which part of the clinical picture may be secondary to the underlying bradyarrhythmia rather than LV dysfunction. Once severe reduction of functional capacity as well as LV dysfunction have been confirmed, then it is reasonable to consider biventricular pacing for the improvement of symptoms. Conversely, the detrimental effects of right ventricular pacing on symptoms and LV function in patients with HF of ischaemic origin and preserved LVEF have been demonstrated.47 The underlying rationale of recommending biventricular pacing should therefore aim to avoid chronic right ventricular pacing in HF patients who already have LV dysfunction.48 Initiation and up-titration of β-blocker treatment, indicated in patients with symptomatic HF, may reduce heart rate and increase pacemaker dependency. Patients with a CRT-P/CRT-D will better tolerate increased pacing time. This may permit initiation of β-blocking treatment or dosage increase in those patients who are already on therapy, confirming a frequently reported clinical observation of dosage up-titration in HF patients treated with CRT. Patients with end-stage HF have a poor quality of life, a very high mortality rate, and are potential candidates for implantation of a left ventricular assist device (LVAD). Although cardiac transplantation (CTX) is associated with high 1- and 10-year survival rates, organ supply is limited. The technical improvements and proven success of implantable LVADs have made it a reasonable treatment option in these patients, either as a bridge to CTX or as destination therapy. Patient selection for LVAD is crucial. Most patients are on continuous inotropic support. Patients with severe renal, pulmonary, or hepatic dysfunction as well as patients with active infection or cardiogenic shock should not be considered as candidates.49 One recent study was conducted in 200 patients as destination therapy, who were randomized in a 2:1 ratio to a continuous-flow device (HeartMate II) or a pulsatile device.50 Patients were in NYHA function class IIIB/IV with an LVEF of ≤25%. A peak VO2 of ≤14 mL/kg/min was an inclusion criterion in HeartMate II but gas-exchange data during exercise is not routinely available in clinical practice and may be inconclusive. The primary composite endpoint was, at 2 years, freedom from disabling or to or the endpoints included mean of the patients was years, and the mean LVEF was The primary endpoint was achieved in more patients with the continuous-flow device P < and survival at 2 was higher P = recent study patients in whom the continuous device was implanted as bridge to After 18 months, patients underwent LVAD for cardiac or required LVAD The an of that in practice of patients receiving an LVAD are not considered candidates for CTX at the time of