Paced Heart Rate Acceleration for Cardiac Conditioning

A clinical trial of exercise-similar heart rate acceleration delivered via cardiac pacing vs. sham intervention in subjects at rest will be performed. The study population comprises subjects with guideline-directed medically managed left ventricular dysfunction due to ischemic or non-ischemic cardiomyopathy and an existing implantable cardioverter defibrillator or biventricular implantable cardioverter defibrillator. The purpose of the study is to understand how the heart rate pattern of exercise contributes to the considerable cardiac conditioning effects of exercise and estimate whether the pacing approach may have translational clinical applicability. Fifty-two subjects will be randomized, single-blinded, to either the pacing intervention or a sham intervention which they will receive once daily, 3 days/week for 6 weeks. Baseline symptoms and clinical test results will be compared to the same measures at 2 weeks, 4 weeks and 6 weeks of intervention/sham and at 3 months and one-year post-intervention. The primary endpoint will be the change in left ventricular ejection fraction from baseline in intervention vs. sham groups (mixed effects linear regression with time and treatment arm as fixed effects and pre-specified covariates of sex and cardiomyopathy type as random effects). Secondary endpoints will include changes in quality of life, 6-minute walk distance, cardiopulmonary exercise test (CPET) measures, daily activity and major adverse cardiac events (MACE) at 3 and 12 months between pacing and sham groups. A "dose-response" analysis of outcomes at 2, 4, and 6 weeks of the intervention vs. sham compared with baseline will be performed.

Subjects and protocol: This is a phase II randomized, controlled, prospective clinical trial in subjects with systolic heart failure. It is designed to assess the effect on symptoms, functional, structural, and medical outcomes in response to an atrial paced reproduction of the heart rate exercise pattern delivered regularly over a 6-week period. A single center (University of Iowa) will execute the intervention/sham and subject testing. A separate center (Cleveland Clinic) will perform statistical analysis of deidentified data. Fifty-two subjects will be randomized 1:1 by computer generated code to either the intervention (pacing) arm or the sham arm. The subject, but not researcher, will be blinded to randomization status (single-blind). Analysis of data will be performed in a blinded fashion. Within 2 weeks prior to initiation of pacing vs. sham interventions, baseline testing will be performed including collection and storage of plasma, testing for blood urea nitrogen (BUN), creatinine, N-terminal pro-b-type natriuretic peptide (NT-proBNP), echocardiographic left ventricular (LV) dimensions, left ventricular ejection fraction (LVEF), and quality of life/symptoms by both the Minnesota Living with Heart Failure (MLHF) and the Kansas City Cardiomyopathy Questionnaire (KCCQ), designed for assessment of symptoms over 1 month and 2 week windows, respectively, distance on 6-minute walk, 1 week of pacing device accelerometer-derived activity measures, wearable activity monitor measures, and cardiopulmonary exercise test (CPET) for determination of metabolic equivalents (METs), maximum oxygen consumption (VO2max) and anaerobic threshold (AT) estimated noninvasively, using the rapid incremental exercise test. In addition, starting 1 week prior and continuing through 1 month after the pacing vs. sham intervention, each subject will be provided a wearable activity monitor and their daily activity analyzed as a cofactor. At the end of 2, 4 and 6 weeks of pacing/sham, as well as 3 and 12 months after the intervention, the same outcome measures as at baseline will be collected as well as major adverse cardiovascular events (MACE), weight, medications, and arrhythmias recorded on the subject's implantable cardioverter defibrillator (ICD) or biventricular ICD (BiV/ICD). At the end of the pace/sham period, subject blinding will be assessed by asking subjects if they believe they were in pace or sham group. Collected plasma will be saved for metabolomic testing in year 5 to correlate with findings in mouse models. Outcomes: All tests result analyses will be done by 2 independent clinicians, qualified in specific areas, blinded to the randomization. A reviewer who analyzes a testing parameter will analyze the same parameter for all subjects. Variability between reviewers will be analyzed. The primary endpoint will be change in left ventricular ejection fraction (LVEF) from baseline. Secondary endpoints will include quality of life (QOL) by MLHF score, 6-min walk distance, CPET (including VO2max, metabolic equivalents (METs) achieved, peak power output, total exercise time), daily activity change from baseline and MACE at 3 and 12 months between pacing and sham groups. A "dose-response" analysis of outcomes at 2, 4, and 6 weeks of the intervention vs. sham compared with baseline will be performed. Durability analysis of response at 3 and 12 months will be performed. The data from all subjects completing at least one pacing vs. sham session will undergo analysis for measures of tolerability (symptoms, maximum achieved pacing rate, appearance of arrhythmias, changes in vital signs and bioimpedance cardiac output) with respect to age, sex, degree, type of cardiomyopathy (ischemic vs. non-ischemic), left ventricular function, NT-proBNP, NHYA class, baseline oxygen consumption (VO2), baseline exertional tolerance, baseline physical activity level as assessed by the ICD of BiV/ICD accelerometer and wearable activity watch, autonomic function as defined by heart rate variability derived from electrocardiogram (ECG) and stored ICD or BiV/ICD data. Hospitalizations, deaths, or other adverse outcomes will similarly be analyzed with respect to baseline data and procedural data such as maximum achieved pacing rate. For those subjects completing the intervention/sham and follow up visits, the primary endpoint of change in left ventricular ejection fraction will be analyzed. Secondary endpoints will include changes in functional capacity, particularly 6-minute walk, and quality of life indicators as well as MACE with respect to the above study variables. Descriptive statistics will be used to summarize participant demographic and clinical characteristics at baseline. Categorical variables will be compared using Chi squared tests. For continuous variables, to relax the strict requirement of normality, non-parametric methods for unadjusted comparison will be used (Wilcoxon rank-sum test, also known as the Mann-Whitney two-sample statistic, for comparison of two independent samples. Wilcoxon matched pairs signed-rank test will be used for paired comparisons). For the primary endpoint analysis, a mixed-effects linear regression model will be used to compare the two study arms on the primary outcome of LVEF, with time (2, 4, and 6 wks) and treatment arm (intervention vs. sham) included as fixed effects, and pre-specified covariates of sex (binary male/female) and cardiomyopathy type (binary ischemic/non-ischemic) as random effects. Both stratified randomization by sex and cardiomyopathy type, as well as adjustment for these covariates in the analysis, will be performed. Although often only one or the other technique is sufficient, both are performed here given the well-known confounding effects of sex and cardiomyopathy types in heart failure trials and a desire to be highly rigorous in this regard. An intercept for time of treatment or sham (2, 4, 6 wks) will be included. The significance criteria for the model coefficient corresponding to treatment will be .025 with a one-sided 97.5% confidence interval. Analyses will be conducted on an "intention-to-treat" basis. Estimates and confidence intervals of effect sizes and standard deviation of outcome measures will be presented. Pre-specified subgroup analyses will probe for treatment effect heterogeneity in outcomes based on sex and ischemic cardiomyopathy vs. nonischemic cardiomyopathy. Secondary endpoints will be analyzed in a fashion similar to the primary endpoint analysis. Due to the controlled environment, missing data are expected to be very rare. If such exists, patterns will be evaluated, and multiple imputation performed.