Cardiac contractility modulation in patients with heart failure refractory to drug treatment

Abstract

Cardiac contractility modulation in patients with heart failure refractory to drug treatment aims to strengthen myocardial activity through transmission of nonexcitatory impulses to the heart. The present study reviewed a total of 251 patients from four studies; the results, however, showed a low level of evidence. Crossover analysis showed a strong placebo effect for patients with implants that were switched on after three months. The technology is still in the development stage and further studies are needed.

Heart failure is a common condition of debilitated heart muscle, resulting in an insufficient blood supply to the body, primarily to the vital organs and peripheral extremities. The main symptom is dyspnea, caused by a significant decrease in the heart’s pumping capacity and, in many cases, is associated with pulmonary congestion. In many cases, patients with heart failure are rehospitalized after diagnosis because of recurring health impairments (1,2). Several studies have shown the effectiveness of medical therapy with angiotensin-converting enzyme inhibitors, beta-blockers and aldosterone antagonists. For patients with a persistent symptomatic condition despite optimal medical therapy, several treatments with electronic or mechanical medical devices, such as cardiac resynchronization therapy (CRT), cardiac contractility modulation (CCM) or heart-supporting systems, were developed (2–5). The present review (originally conducted for the Austrian Ministry of Health) investigates the effectiveness and safety of CCM.

Description of CCM

In CCM, nonexcitatory impulses are transmitted to the heart within the refractory phase, without initiating heart activity (compared with the pacemaker, which does increase heart activity). However, it increases contractility of the heart muscle during the following systole. After implantation and programming of the device, the patient is examined and monitored under stationary conditions, before being discharged into the domestic environment.

Constant monitoring and surveillance by the division of cardiology is necessary (6). This is similar to pacemaker surveillance; however, the patients are more frequently monitored.

Indication and therapeutic aim

Only patients with a normal, wide QRS complex in the electrocardiogram are eligible for CCM. Patients with a wider QRS complex are examined for CRT by a special ultrasonic technique. The aim of the therapy is to improve discomfort and restrictions on daily activities caused by cardiac insufficiency, which significantly reduces quality of life. In some cases, CCM could even help avoid or, at least, delay cardiac transplantation.

Estimated scope of services and costs

According to the modification and amendment proposal submitted to the Austrian Ministry of Health, the service was executed 25 times in 2008. Costs per application added up to €17.278.

METHODS

In 2008, the Austrian Ministry of Health applied for reimbursement of CCM for patients with cardiac insufficiency refractory to drug treatment.

The present systematic review sought to answer the following research question: In comparison with no intervention, is CCM effective and safe for patients with heart failure, normal QRS complex and for those who are refractory to drug treatment?

A systematic literature search was conducted on January 18, 2008, and was updated on February 1, 2010, using the following databases: Medline, Embase, Evidence-Based Medicine Reviews via Ovid, the Health Technology Assessments database of the Centre for Reviews and Dissemination (United Kingdom), the NHS Economic Evaluation Database (NHS EED) of the Centre for Reviews and Dissemination, and the International Network of Agencies for Health Technology Assessment (INAHTA) database. The Medline search was restricted to the English and German literature. After deduplication, a total of 38 bibliographical citations were identified. The exact search strategy can be obtained at Ludwig Boltzmann Institut – Health Technology Assessment (LBI-HTA; Vienna, Austria). The authors requested literature and information from Impulse Dynamics, Germany, and, among other material, information from a randomized trial was sent; however, it had already been identified within the systematic literature search. Therefore, the information sent by the company had no additional impact on the literature sources on which the assessment was based on. In addition, one study was identified through a manual search.

In total, 39 studies were available. Literature selection and assessment of internal validity were performed by two independent researchers. Differing opinions were resolved by discussions or through third-person opinion. A detailed list of criteria regarding the use of assessing internal validity of single-type studies is available in the internal manual of LBI-HTA (7).

Two controlled studies (8,9) with standard drug treatment and sham treatment (device implantation without emission of signals) interventions were available. In addition, two uncontrolled pre-post studies (4,10) gave limited evidence on the general effectiveness and safety. Study characteristics and results are shown in Table 1. Only patients with symptomatic heart failure were included. Exclusion criteria were atrial fibrillation and recent myocardial infarction. Patients eligible for CRT were also excluded from the study population. At implantation, individual response to left ventricular pressure on stimulation was tested directly. Individuals were included in the study analysis when the critical value was exceeded.

TABLE 1.

Study results

Author, year (Ref)	Borggrefe et al, 2008 (8)	Neelagaru et al, 2006 (9)	Stix et al, 2004 (10)	Pappone et al, 2004 (4)	Abraham et al, 2008 (17)
Country	International	United States	International	International	United States
Sponsor	Industry	Industry	Industry	Industry	Industry
Design	Randomized crossover study; double blind	Randomized, controlled trial; double blind	Uncontrolled pre-post study	Uncontrolled pre-post study	Randomized, controlled, open
Patients	n=164 (informed consent of 181, implantation procedure in 178, 164 randomized); 25 F, 139 M; 24% NYHA class* II, 76% NYHA class III	n=49 (informed consent of 107, implantation procedure in 52, 49 randomized); 15 F, 34 M; NYHA class III or IV	n=25; 2 F, 23 M; all NYHA class III	n=13; 0 F, 13 M; all NYHA class III	n=428; 119 F, 309 M
Age, years	59	56 (intervention) and 60 (control)	62	63	58
Intervention/control intervention	Group 1: n=80, 3 months CCM, 3 months sham treatment† Group 2: n=84, 3 months sham treatment†, 3 months CCM	CCM/sham treatment†	No control intervention	No control intervention	CCM + optimal medical therapy/optimal medical therapy alone
Follow-up, weeks	24	24	8	8+24	50
VO2max, mL/kg/min‡	Improvement of approximately 0.4 mL/kg/min in both groups after 3 months, impairment in sham group after 6 months of 0.9 mL/kg/min, improvement in CCM group after 6 months of 0.2 mL/kg/min	Improvement of 0.8 mL/kg/min – 1.2 mL/kg/min in both groups (no significant disparity between groups)	N/A	Improvement of 1.2 mL/kg/min (8 weeks) vs 2.5 mL/kg/min (24 weeks)	N/A
Quality of life (MLWHFQ)	Improvement of 10–12 scores in both groups after 3 months (CCM slightly better), impairment of 5 scores in sham group after 6 months, improvement of 1 score in CCM group after 6 months	Improvement of 16–18 scores in both groups, no significant disparity between groups	Improvement of 18 scores	Improvement of 18 (8 weeks) vs 29 scores (24 weeks)	N/A
6 min walk test	Improvement of 11–17 m in both groups after 3 months (CCM slightly better), impairment of 6 m in sham group after 6 months, improvement of 20 m in CCM group after 6 months	Improvement of 50 m in CCM group, of 35 m in sham group (no significant disparity between groups)	Improvement of 54 m	Improvement of 59 m (8 weeks) vs 80 m (24 weeks)	N/A
Ejection fraction increase	N/A	Minimal improvement of 1.3% – 1.8% in both groups (no significant disparity)	Improvement of 6%	Improvement of 6% (8 weeks) vs 14, 2% (24 weeks)	N/A
Overall complications	52% (93/178 patients with implantation procedure)	40% (21/52 patient with implantation procedure)	48% (12/25)	31% (4/13)	N/A
Implant infections	3.0% (5/166 implants)	4.1% (2/49 implants)	0.0% (0/25)	7.7% (1/13)	N/A
Mortality	3.4% (6/178 patients with implantation procedure)	4.0% (2/52 patients with implantation procedure)	8.0% (2/25)	0.0% (0/13)	N/A
Dropout rate	9/164 randomized, controlled trial	0/49	0/25	0/13

^*New York Heart Association (NYHA) class: Patients’ performance measure (the lower the class, the higher the performance);

^†Sham treatment: Implantation without impulse emission;

^‡VO₂max Maximal oxygen absorption (measure of potential endurance performance). CCM Cardiac contractility modulation; F Female; M Male; MLWHFQ Minnesota Living with Heart Failure Questionnaire; N/A No data or information available; Ref Reference; vs Versus

RESULTS

Clinical effectiveness

Two double-blind, randomized controlled trials (RCTs) (8,9) were available for assessing CCM therapy. In both studies, patient-relevant outcomes were measured. Apart from the small sample size, both trials showed good internal validity (9). In the study by Neelagaru et al (9), results of the 6 min walk test and maximal oxygen absorption with CCM improved; however, it was not significant compared with the control group (9). In other parameters (quality of life and ejection fraction [EF]), because both groups showed the same improvements, it was indicative of a placebo effect (11). In the study by Borggrefe et al (8), which had a crossover design, a placebo effect could also be observed because during the first three months, the values of the sham group also improved (8). Nevertheless, the following crossover period showed that the treatment resulted in a statistically significant improvement of oxygen intake and of the patients’ quality of life. Considering these data, the results of the pre-post studies (4,10) (Table 1) must be put into perspective. Even if patient-relevant effects are higher than in both RCTs, this is probably caused by a lack of control and blinding, as well as a regression to the mean. The question to answer is whether the effects observed in RCTs are clinically relevant. While maximum oxygen intake among treatment groups improved by 1.2 mL/kg/min, in other add-on treatments to standard therapy for heart failure patients, the maximum oxygen intake was 3.7 mL/kg/min (12), and in studies with angiotensin-converting enzyme inhibitors, the maximum oxygen intake was 2.1 mL/kg/min (13). Generally, differences of approximately 1 mL/kg/min among groups are reported to be clinically relevant (14). Such differences were not observed in the two RCTs. In the Minnesota Living with Heart Failure Questionnaire, which measures quality of life, a disparity of 5 points is reported as being clinically significant (15) – a group difference that was observed in only one of the RCTs. In other studies of patients suffering with heart failure, differences of 9 points (2) or 14 points (16) were reported among groups.

Safety

The complication rate in randomized studies was between 40% (9) and 52% (8); in noncontrolled studies, the complication rate was between 31% (4) and 48% (10). Implant infections occurred in 0% to 7.7% of patients. In the RCT by Borggrefe et al (8), 3.4% (six of 178) of the patients who underwent implant procedure died – two of them during the placebo (‘off’) phase and two during the CCM phase. The overall six-month mortality rate was 0% to 8%. There were no long-term data available.

Additional evidence: The FIX-HF-5 trial

In various studies involving clinical trials, the strength of evidence was low. In 2005, the FIX-HF-5 study (17) – a randomized, controlled, nonblinded trial – was initiated. The results were not published by the cut-off date of the literature search; however, they were presented at the 2009 annual meeting of the American College of Cardiology in Orlando, Florida (USA), by the authors of the FIX-HF-5 (18) study.

Until June 2007, 428 patients with heart failure (left ventricular EF of 35% or lower) and not eligible for CRT were randomly assigned to CCM (n=215) or drug therapy (n=213). The responder status regarding the anaerobe threshold was defined as the primary outcome of effectiveness. Patients with an improvement of at least 20% were defined as responders. The outcome was chosen because of its independence of the placebo effect. Intention-to-treat analysis showed 13.2% (28 of 213) of responders in the control group and 17.7% (38 of 215) in the CCM group. Twenty-seven per cent of data were not sufficient for evaluation. The main reason was that patients did not reach the anaerobe threshold. The analysis of the available data showed a response rate of 11.7% (18 of 154) in the control group and 17.6% (28 of 159) in the CCM group (P=0.093). The noninferiority analysis of the primary safety outcomes (mortality and hospitalization for any reason) after one year resulted in a levelled profile of 48% (193 of 213) in the control group and 52% (112 of 215) in the CCM group. In secondary effectiveness outcomes, maximal oxygen absorption, New York Heart Association score and quality of life (Minnesota Living with Heart Failure Questionnaire) were better in the CCM group than in the control group. There was no significant difference observed in the 6 min walk test.

In a subgroup analysis of patients with an EF of 25% or greater and New York Heart Association class III, an improvement in the CCM group was found. Nevertheless, because it is only applicable for a subgroup, this analysis had to be classified as ‘hypothesis generating’. No improvements in the prespecified primary outcome, the anaerobe threshold, were shown in the FIX-HF-5 study (18).

DISCUSSION

The grading tool of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group was used to assess the strength of evidence (7). The application of this GRADE scheme for the actual research question is presented in Table 2, including in published study results. The overall strength of evidence for effectiveness and safety of CCM in patients suffering heart failure refractory to drug treatment is low. Public reimbursement is currently not recommended. Existing evidence is not sufficient to assess the net benefit of the intervention in evaluation. The results of the actual multicentre RCT (FIX-CHF-5) presented at the 2009 annual meeting of the American College of Cardiology in Florida and the fact that we could not find other ongoing trials strengthens this appraisal.

TABLE 2.

Evidence profile – clinical effectiveness and safety of cardiac contractility modulation

Number of studies/patients	Design	Methodological quality	Consistency of results	Directness	Effect size*	Other modifying factors†	Strength of overall evidence
Comparative effectiveness: Increase of oxygen absorption capacity (VO2max)
2/213	Randomized, double blinded	Well	Yes	Yes	+0.6 until +1.2 mL/kg/min versus −0.5 until +0.8 L/kg/min; barely clinically relevant	None	Low
Effectiveness: Increase of oxygen absorption capacity (VO2max)
1/13	Uncontrolled pre-post study	Limited; no control group	Yes	Yes	+2.5 mL/kg/min	None	Very low
Comparative effectiveness: Improvement in quality of life (MLWHFQ)
2/213	Randomized, double blinded	Well	Yes	Yes	(1) +13 versus +5 scores; clinically relevant disparity (2) +18 versus +16 scores; barely clinically relevant	None	Low
Effectiveness: Improvement in quality of life (MLWHFQ)
2/38	Uncontrolled pre-post study	Limited; no control group	Yes	Yes	+18 until +29 scores	None	Very low
Safety: Infection of implants
4/251	2 RCTs, 2 uncontrolled pre-post studies	Well	Yes	Yes	2.8%	None	Low

^*Improvements are marked with +, impairments with –;

^†Low event rate or imprecise data, strong or very strong association, high risk of reporting bias, dosage-effect gradient, residual confounding plausible. MLWHFQ Minnesota Living with Heart Failure Questionnaire; RCTs Randomized controlled trials; VO₂max Maximal oxygen absorption