CLINICAL STUDIES:

Harbinger Medical has conducted clinical studies to document the diagnostic potential of the Harbinger system in risk stratification of SCD. The two major clinical studies are outlined below.

Wedensky Modulation Study:

The performance of the Harbinger system was demonstrated in a clinical study performed at the Mayo Clinic in Rochester, MN, the University of Minnesota in Minneapolis, MN and St. Olav Hospital in Trondheim, Norway. The result of this study, which included nearly 300 subjects, has been reported in over 40 papers and presentations at the scientific sessions of industry leading organizations. The results of this clinical study were used to determine and select optimal technical specifications. The knowledge obtained was used to design the HIP study.

The HIP Study:

The HIP study was a prospective observational study of post-MI patients with implanted ICDs designed to test the hypothesis that the Harbinger tests could stratify patients into high risk and low risk categories, and the rate of life-threatening cardiac events for the high risk group would be significantly higher than for that of the low risk group. As determined by the Wedensky Modulation Study, a high-risk patient was identified using Wedensky Modulation analysis of the R-wave and was defined as one with a WMI > 0.5, with a low-risk patient being defined as one with a WMI ≤ 0.5. The study was powered to detect a hazard ratio of 1.5 or greater.

There were 334 patients enrolled in the study from September 2003 to August 2007, from four sites in the U.S. and four sites in Europe. 329 patients had complete Harbinger test records, and 268 patients had completed at least one 1-year follow-up. Of those 268 patients, 137 patients were in the low-risk group (WMI ≤ 0.50), and 131 patients were in the high-risk group (WMI > 0.5). For purposes of the study, a life-threatening cardiac event was defined to be the occurrence of a ventricular arrhythmia that led to a patient’s death or an event terminated by ICD therapy as determined by an attending physician following a review of the ICD data and the patient's associated clinical history.

The study was designed as a group sequential procedure adapted to survival analysis to detect the difference in event rates over a 24-month follow-up period. At a predefined interim analysis point, the 12-month hazard ratio for the cumulative event rates between the high- and low-risk groups was significant, and it was determined that the study sample size provided sufficient power. Under the established rules for the study, enrollment was therefore halted.

ABSTRACTS RELATING TO THE HIP STUDY:

1. Wedensky Modulation Index and Ejection Fraction Combined Provide Better Risk Stratification of Post-MI Patients

   Author Block: Peter A. Brady, MD, FRCP, Paul Erne, MD, Jesus E. Val-Mejias, MD, Joerg O. Schwab, MD, Rainer Schimpf, MD, Michael V. Orlov, MD, PhD, Tom Mattioni, MD, Marek Malik, MD, PhD and Jan P. Amlie, MD, PhD. Mayo Clinic, Rochester, MN, Kantonsspital Luzern, Luzern, Switzerland, Galichia Heart Hospital, Wichita, KS, Dept. of Medicine – Cardiology, University of Bonn, Bonn, Germany, 1st Department of Medicine-Cardiology, University Hospital Mannheim, Mannheim, Germany, Caritas St. Elizabeth’s Hospital, Boston, MA, Arizona Arrhythmia Consultants, Scottsdale, AZ, St. George’s Hospital, London, United Kingdom, Rikshospital, Oslo, Norway

   Abstract:
   Introduction: Risk stratification based on ejection fraction (EF) alone is limited. Wedensky Modulation Index (WMI), based on subthreshold transchest electrical stimulation delivered to every other QRS complex, is a novel noninvasive assessment of myocardial vulnerability. We hypothesized that a WMI may have predictive value in patients late after myocardial infarction and provide incremental (electrical) risk stratification beyond EF to predict best candidates for ICD therapy. Methods: A WMI was derived and prospectively evaluated along with EF using patient event outcomes. The combination of the WMI and EF was evaluated using patient event outcomes derived from a prospective observational study at 8 international centers that included 268 post-myocardial infarction patients with ICD who completed at least one 6-month follow-up over a 12 month period. Patients were placed into two groups: NEG (WMI ≤ 0.5 AND EF > 20%, n=119) and POS (WMI > 0.5 OR EF ≤20%, n=149). Cumulative ICD-treated arrhythmia event rates for the two patient groups were compared using Kaplan-Meier estimates and Cox regression analysis. Results: There was more than a two-fold increase in the number of events for the POS group compared to the NEG group (28% versus 13%, odds ratio = 2.53, p = 0.004). Cox regression analysis determined a > 30% absolute difference in event-free survival over 12 month follow-up between patients with WMI ≤ 0.5 and EF > 20% compared to patients with WMI > 0.5 and an EF ≤ 20% (p = 0.009). Patients with EF > 20% comprised 84% of population; however, when combined with a high WMI, these patients had a 26% event rate and > 15% difference in event-free survival over the 12 months (p = 0.011). Importantly, these results continued to hold when the patient population was restricted to patients with EF ≤ 35% (n = N). Conclusions: WMI adds important incremental prognostic data beyond ejection fraction in patients late after myocardial infarction. High WMI is associated with highest risk of arrhythmic event especially when combined with lower EF and is a non-invasive tool for distinguishing patients at highest arrhythmic event risk who may benefit from ICD therapy.

2. Non-invasive Risk Stratification Using Wedensky Modulation to Determine Cardiac Electrical Vulnerability Late After Myocardial Infarction

   Author Block: Peter A. Brady, MD, FRCP, Paul Erne, MD, Jesus E. Val-Mejias, MD, Joerg O. Schwab, MD, Rainer Schimpf, MD, Michael V. Orlov, MD, Tom Mattioni, MD, Marek Malik, MD, PhD and Jan P. Amlie, MD, PhD. Mayo Clinic, Rochester, MN, Kantonsspital Luzern, Luzern, Switzerland, Galichia Heart Hospital, Wichita, KS, Dept. of Medicine – Cardiology, University of Bonn, Bonn, Germany, 1st Department of Medicine-Cardiology, University Hospital Mannheim, Mannheim, Germany, Caritas St. Elizabeth’s Hospital, Boston, MA, Arizona Arrhythmia Consultants, Scottsdale, MN, St. George’s Hospital, London, United Kingdom, Rikshospital, Oslo, Norway

   Abstract:
   Introduction: Risk stratification for sudden cardiac death (SCD) remains problematic with reliance on left ventricular ejection fraction (LVEF) which predicts total mortality rather than arrhythmic risk. A novel non-invasive method developed by Harbinger Medical, Inc. that uses Wedensky Modulation (WM) evoked by sub-threshold transthoracic electrical stimulation delivered to every other QRS complex may predict risk of SCD by direct measurement of myocardial electrical vulnerability. This study sought to determine the utility of WM to predict arrhythmic events in patients late after myocardial infarction. Methods: The study was an international multi-centre prospective observational study of post-myocardial infarction patients with ICD implantation. A WM Index (WMI) was computed from differences in the spectro-temporal analysis (frequency/energy) of stimulated vs. non-stimulated beats. Patients were assigned to a WMI-L group (WMI ≤ 0.5, n=137) or WMI-H group (WMI > 0.5, n=131). Data were analyzed at 12 months and cumulative ICD-treated arrhythmia event rates for the two WMI groups were compared using Kaplan-Meier estimates. Results: A total of 268 pts were included with 21 events in the first year for the WMI-L group compared to 37 events in the same time period for the WMI-H group (log-rank p < 0.01). Comparing WMI-L to WMI-H, the hazard ratio for event rates was 2.1 at one year (95% CI of 1.2 to 3.6, Wald p < 0.01). When stratified both by WMI and LVEF, proportional hazards showed significant differences for cumulative event rates between WMI-L and WMI-H when adjusted for LVEF. However, LVEF did little to separate patients regarding events, either separately or when stratified by WMI. Conclusions: Wedensky modulation index is a new and important non-invasive test that identifies patients at risk of SCD following myocardial infarction independent of LVEF.

3. Wedensky Modulation of T-Wave Accurately Predicts Arrhythmic Events Author Block: Peter A. Brady, MD, FRCP, Paul Erne, MD, Jesus E. Val-Mejias, MD, Joerg O. Schwab, MD, Rainer Schimpf, MD, Michael V. Orlov, MD, PhD, Tom Mattioni, MD, Marek Malik, MD, PhD and Jan P. Amlie, MD, PhD. Mayo Clinic, Rochester, MN, Kantonsspital Luzern, Luzern, Switzerland, Galichia Heart Hospital, Wichita, KS, Dept. of Medicine – Cardiology, University of Bonn, Bonn, Germany, 1st Department of Medicine-Cardiology, University Hospital Mannheim, Mannheim, Germany, Caritas St. Elizabeth’s Hospital, Boston, MA, Arizona Arrhythmia Consultants, Scottsdale, AZ, St. George’s Hospital, London, United Kingdom, Rikshospital, Oslo, Norway

   Abstract:
   Introduction: Wedensky Modulation (WM) is based on subthreshold transthoracic electrical stimulation delivered to every other QRS complex. The WM Index (WMI) is computed from electrocardiographic (ECG) differences between stimulated and non-stimulated QRS complexes and accurately predicts arrhythmic events. We hypothesized that WM effects measured within T-waves would also accurately predict arrhythmic events. Methods: The WM T-wave index was prospectively evaluated post-hoc using patient ECG data derived from a recently completed study (prospective observational study, 8 global centers). There were 268 post-myocardial infarction (post-MI) patients with ICD implantation with at least one 6 month follow-up completed. Patients were placed into the TMI-L group (TMI = 0, T-wave index ≤ 0.5, n = 202) or the TMI-H group (TMI = 1, T-wave index > 0.5, n = 66). Cumulative ICD-treated arrhythmia event rates for the two TMI groups were compared using Kaplan-Meier estimates. Results: There were 35 events in the first year for the TMI-L group compared to 23 events in the same time period for the TMI-H group (log-rank p < 0.01). Comparing TMI-L to TMI-H, the hazard ratio for event rates was > 2 at one year (95% CI of 1.2 to 3.5, Wald p < 0.01). TMI is significantly different from WMI regarding future cardiac-related events prediction (p < 0.0001 regarding discordant pairs). When combined with WMI study results to form an R-wave T-wave index the event rate predicted by this combined index significantly improves the separate predictive power of the two indices, raising the positive predictive value (PPV) from 64% (WMI) and 40% (TMI) to 81%. The event rate for the RTI-L group (RTI = 0, n = 106) was 10% compared to the event rate of 29% for the RTI-H group (RTI = 1, n = 162). Conclusions: Post-MI infarction patients with a high T-wave modulation index may have a significantly increased risk of a life-threatening arrhythmia when compared to patients with a low index. When combining the R-wave index with the T-wave index, the predictive power of Wedensky modulation testing for future cardiac-related events accurately predicted more than 80% of events occurring in the first 12 months following myocardial infarction.