Keywords
INTRODUCTION
About 30% of patients with heart failure have conduction disorders that cause asynchronous ventricular contraction. This negatively influences their prognosis.1,2 Cardiac resynchronization is accepted as a complement to pharmacological treatment in persistent heart failure,3 and randomized clinical trials have shown that it can improve functional capacity and quality of life.4-7 Furthermore, the meta-analysis performed by Bradley et al8 clearly shows that cardiac resynchronization therapy (CRT) improves survival in heart failure patients,8 although about one third show no improvement.5,6 This has been related to the location of the electrode and to certain echocardiographic variables,9,10 but no clinical variables prior to implantation have been identified that might predict a failure to respond. The aim of this work was to prospectively analyze the variables that might predict such a lack of response to CRT.
PATIENTS AND METHODS
Patients
The patients selected for inclusion all suffered chronic heart failure, and all fell into New York Heart Association (NYHA) functional classes II-IV despite optimized pharmacological treatment (all had taken stable doses in the previous 2 months). All patients showed a left ventricular ejection fraction (LVEF) of ≤40%, a left ventricular end-diastolic diameter (LVEDD) of >55 mm, left bundle branch block with a QRS interval of ≥130 ms, and could only manage a distance of <500m in the 6-minute walking test. The study protocol was approved by our hospital's ethics committee, and all patients gave their signed consent to be included.
Methods
This study was a prospective, observational study in which patients were included in a consecutive fashion. Those that took part underwent a baseline examination that included determining their NYHA functional class, a 6-minute walking test,11 The patients also completed a quality of life questionnaire (the Minnesota questionnaire for heart failure;12 the greater the score, the worse the quality of life), echocardiogram (ventricular diameter, semiquantitative assessment13 of the degree of mitral failure, and calculation of the ejection fraction using the Simpson biplanar method), and radioventriculography.
After this initial evaluation, all patients were implanted with a cardiac resynchronization device (either a Guidant ContakHF®, a Contak-Renewal,® or a Renewal II®; the last of these was used only when a defibrillator was indicated according to accepted criteria), placing one electrode in the right atrium (if the patient showed sinus rhythm), another at the apex of the right ventricle, and the specially designed Easy Track®, Guidant® electrode in a distal cardiac vein. The coronary sinus was catheterized using a guiding catheter.
All patients underwent a complete check-up at 6 months, although some required non-programmed examinations. Patients were deemed to have shown a positive CRT response on the basis of a composite clinical variable including no cardiac death or heart transplant and a >10% improvement in the distance covered in the 6-minute walking test. Although this test only correlates moderately well with peak oxygen consumption in patients with moderate or severe heart disease,14 intrapatient reproducibility in the short term is very good.15 Previous studies have shown that a minimum variation of 10% is required in order to confirm (99% confidence interval) a real effect of therapy.14
Programming the Resynchronization Devices
The resynchronization devices were programmed in DDD stimulation mode in patients with sinus rhythm and in VVIR in those with atrial fibrillation. Attempts were made to ensure a high percentage of biventricular stimulation in these latter patients, either with drugs or through the ablation of the atrioventricular node.
For patients with sinus rhythm, the AV interval was programmed for 140 ms, and the PV interval for 120 ms, in agreement with the results obtained in earlier studies with this kind of patient.16-18 Biventricular stimulation was synchronous in all patients (VV interval=0).
Statistical Analysis
Qualitative variables are expressed as a number and percentage, whereas quantitative variables are expressed as means ± standard deviation (SD). Qualitative variables were compared using the χ2 test, and the Student t test (for independent samples) used to compare quantitative variables. Improvement at follow-up was analyzed using the McNemar test (for qualitative variables) and the Student t test for paired data (quantitative variables). Multivariate analysis was performed using stepwise logistic regression. Significance was set at P<.05. All data were analyzed using SPSS v10.0 software.
RESULTS
Sixty eight patients were selected for the implantation of a biventricular stimulation device between January 2001 and January 2003. Attempts were made to locate the left ventricular stimulation electrode in the coronary venous system, but this was not possible in 5 patients (7.4%), and in another the possibility of left ventricular stimulation was lost due to diaphragm stimulation or an increase in the capture threshold. A left ventricular stimulating electrode was successfully implanted epicardially in 4 patients.
Of the 64 patients who received CRT, 33 (51.7%) were given a pacemaker providing biventricular stimulation, and 31 (48.8%) received a defibrillator providing the same. One patient died of lung cancer 4 months after receiving the implant and was excluded from the analysis.
Tables 1 and 2 show the baseline characteristics of the 63 patients whose results were analyzed. Mean age was 68.3±8 years; 51 patients (81%) were male. Fifty patients (79.4%) fell into NYHA functional classes III or IV. The mean duration of the QRS interval was 177±25 ms and the mean LVEF recorded by echocardiography was 22.4±6%. Twenty nine patients (46%) had a history of ischemic heart disease (either coronary stenosis of >70% as demonstrated by coronary angiography, or at least one documented heart attack). Twenty one of these (72.4%) had suffered at least one heart attack (anterior localization in 16 patients [55.25%]).
Thirteen patients (20.6%) fell into NYHA functional class II at the time of inclusion. All had an LVEF of <30% and indication for a definitive pacemaker. A resynchronization device was implanted in each to prevent any clinical worsening of their condition that the inherent asynchrony of right ventricular stimulation might cause. Two additional patients had an LVEF of 35% and 40% but fell into NYHA functional class III despite optimized drug treatment.
Eighteen patients (28.6%) had experienced clinically documented sustained monomorphic ventricular tachycardia (SMVT). Prior SMVT was not significantly associated with having suffered a documented heart attack (23.5% in patients with no previous heart attack and 35.7% in those who had suffered one; P=.3), nor with age, functional class, the etiology of disease, or the dimensions or systolic function of the left ventricle.
Response at 6 Months Follow-up
Compared to baseline results, the group of 63 patients as a whole showed significant improvement at 6 months in terms of NYHA functional class, quality of life questionnaire score (41±20 compared to 30±18; P<.001), the distance covered in the 6-minute walking test (267±134 m compared to 398±150 m; P<.001), LVEDD (75±9 mm compared to 73±9 mm; P=.001), LVESD (61±12 mm compared to 57±9 mm; P=.007), LVEF (as determined by echocardiography) (23±6% compared to 28±9%; P<.001), and radioventriculography results (23±9% compared to 28±14%; P=.005).
With respect to the composite variable defined above for determining a positive response to CRT after 6 months, 44 patients (69.8%) showed satisfactory results (Table 3). During the follow-up period, no significant differences were seen between patients who responded positively and those who responded negatively with respect to the medication they used.
The patients who responded positively to CRT showed a significant improvement in their functional class, quality of life score (13±17 points), 6-minute walking test results (170±125 m) and LVEF (6±8%), and showed significant reductions in their LVEDD (3±5 mm) and LVESD (5±8 mm).
Nineteen patients (30.2%) experienced no clinical improvement in terms of the composite CRT response variable. Of these, 4 (6.3%) died of heart disease and 3 (4.8%) required a transplant during the 6 month follow-up period. Of the 12 remaining, no significant changes were seen in functional class, quality of life questionnaire score, LVEF, LVEDD, or LVESD (Table 3).
Predictors of Response to Treatment
The results were examined to determine which baseline clinical, electrocardiographic or echocardiographic variables were associated with CRT response (Table 4).
A lack of improvement in the mid term was associated with ischemic heart disease, the existence of clinically documented SMVT prior to implantation, and moderate (at least) mitral regurgitation (grades II-IV [scale=0-IV]). No other variables were associated with a lack of response.
These 3 predictors were included in regression analysis and showed independent predictive power of an absence of improvement: SMVT (OR=8.7; 95% CI, 1.8-41.3; P=.007), moderate (at least) mitral regurgitation (OR=8.03; 95% CI, 1.7-37.1; P=.008), and ischemic heart disease (OR=4.8; 95% CI, 1.2-18.3; P=.023).
If the patients who died or who received a transplant are excluded, and a subanalysis performed using only the data of those who covered >110% their original distance in the 6-minute walking test at 6 months (n=56), only 2 variables appear as predictors of an absence of improvement: prior SMVT (OR=8.8; 95% CI, 1.7-45.8; P=.009) and ischemic heart disease (OR=9.6; 95% CI, 1.6-59.5; P=.015).
DISCUSSION
This study reports the prospective results for our first 63 patients treated with CRT, and they agree with those of earlier studies with respect to the number of patients who responded to resynchronization therapy (69.8%) for heart failure with intraventricular conduction disorders.5,6
The main novelty of this work is that it identifies pre-implantation clinical and echocardiographic variables that help identify patients with less chance of clinically benefiting from resynchronization therapy. Improvement was defined in terms of a composite variable that takes into account those patients who are lost to follow-up through cardiac death or heart transplant (these patients experience a change in their clinical status far more important than a change in functional class, quality of life or tolerance to exercise). The available literature does not include death or heart transplant as a lack of response to CRT, variables that studies on the response to pharmacological treatment for heart failure have included.19 Variation in quality of life scores (which include psychosocial elements) is here understood to be more subjective and likely to be greater: it was therefore not included in the composite variable. Ischemic heart disease, the existence of documented SMVT at some point prior to implantation, and at least moderate mitral regurgitation were found to be independent predictors of a lack of response to CRT therapy.
Among those patients who improved clinically in terms of the composite variable at 6 months, the degree of improvement was greater than that reported in earlier studies comparing the effects of resynchronization and placebo therapy, and similar to that obtained with drugs widely prescribed in the treatment of heart failure. This may be because the groups of patients who truly improved were not analyzed independently in these other studies. Furthermore, as in the subanalysis of the MIRACLE20 and MUSTIC21 trials, it is here shown that CRT leads to inverse remodeling (the LVEDD and LVESD of the left ventricle become smaller and its systolic function increases). The data for the present patients show that this remodeling is significant in patients who improve from a clinical point of view. This suggests that, to a large degree, the improvement provided by resynchronization therapy could be due to the induction of inverse remodeling in the mid term. The subanalyses of the MIRACLE and MUSTIC trials mentioned above show that ischemic patients experience an improvement in their LVEF and a reduction in their LVEDD and LVESD with CRT at 6-9 months, although significantly less than that experienced by non-ischemic patients.
The patients who did not respond to CRT showed a notable worsening in the distance they covered in the 6-minute walking test, in functional class, in their quality of life questionnaire scores, in LVEDD and LVESD, and in left ventricular systolic function. Together, these results validate the composite clinical variable used to define those patients who improve with CRT.
The study performed by Reuter et al22 is one of the few in which patients have been classified on the basis of their CRT response (as determined by the same composite clinical variable); the results of this work also identify ischemic heart disease as a predictor of a lack of response. This might be due to akinetic or dyskinetic areas--areas of fibrosis with no viable myocytes--that are impossible to resynchronize despite stimulation at the point of greatest electrical delay.
The present results are also supported by those of Gasparini et al.23 These authors show that in patients with ischemic heart disease, CRT does lead to an improvement, but not so great as that achieved in non-ischemic patients.
Unlike the present study, which showed mitral regurgitation (at least moderate) to be a significant predictor of a lack of improvement, Reuter et al22 found no such association.
It may be that clinical SMVT predicted a failure to improve with CRT in the present patients since they had a greater probability of dying because of arrhythmia. However, all of them had a defibrillator, and when those who died or received a transplant were removed from the analysis, SMVT continued to be a predictor of a failure to improve.
Limitations
The main limitation of this work is the reduced sample size; this is why the confidence intervals for the relationships between the predictors and the likelihood of a lack of response are so wide. Similarly, this reduced sample size diminishes the statistical power to identify other possible predictors of a lack of response. However, this approximation in the identification of variables that might negatively influence the response to CRT--a laborious and costly form of treatment--is important. Further studies with larger samples should minimize these limitations.
The aim of this work was to identify predictors of a lack of response that can be easily recorded by the clinician, not the identification of special echocardiographic predictors. The echocardiographic measurements made were therefore not directed towards the evaluation of intraventricular or interventricular asynchrony, but focused on LVEDD and LVESD, left ventricular systolic function, and the degree of mitral regurgitation. These were then correlated with the clinical course of the patients. Since the goal was to analyze the influence of different variables on the response to CRT before implantation of the resynchronization device, factors such as the effect of the position of the left ventricular electrode were not analyzed.
The large number of patients with ventricular arrhythmias in our sample could be a selection bias; for many patients, the indication of CRT was based on the recommendations of the arrhythmia department. However, the percentage of patients that did not respond is similar to that seen in other studies.
CONCLUSIONS
Resynchronization therapy for moderate-severe heart failure with a wide QRS complex caused by left bundle branch block leads to an improvement in nearly 70% of patients. The remaining 30%, however, do not improve or their condition worsens. Those who do respond do so clearly and show substantial clinical improvement. Ischemic heart disease, clinically documented prior SMVT, and at least moderate mitral regurgitation are pre-implantation factors that predict a negative response to CRT. This should be borne in mind before embarking on such therapy.
Funding: This work was supported by a grant from the Fundación Española del Corazón for basic and clinical research in cardiology (2001). Ernesto Díaz-Infante was awarded the Dr Gilbert Queraltó grant by the SEC in 2003.
Correspondence: Dr. L. Mont Girbau.
Unidad de Arritmias. Instituto de Enfermedades Cardiovasculares. Hospital Clínic.
Villarroel, 170. 08036 Barcelona. España.
E-mail: lmont@clinic.ub.es