Boston Scientific CRMN11906 Implantable Defibrillator User Manual Cognis Part 4 Manual

Boston Scientific Corporation Implantable Defibrillator Cognis Part 4 Manual

UserManual.wiki >

Boston Scientific >

CRMN11906 User Manual >

Cognis Part 4 Manual

CLINICAL STUDY - COMPANION B-15Sub-study Primary Endpoint and Additional Tertiary EndpointsExercise performance––the co-primary endpoint, which consists of Peak VO2and Six-Minute Walk, is designed to demonstrate improvement in exerciseperformance with CRT (CONTAK TR and CONTAK CD pooled data) comparedto OPT at six months post-baseline.Additional tertiary endpoints included Quality of Life as measured by theMinnesota Living with Heart Failure Questionnaire®and NYHA Class.FOLLOW-UP SCHEDULEThe follow-up schedule included the following:• Enrollment––initial assessment of patient eligibility; taking of patient history• Baseline screening––special testing (included a Symptom-Limited TreadmillTest with measurement of oxygen uptake (Peak VO2), a Six-Minute Walk,Quality of Life [QOL] questionnaire and NYHA Classiﬁcation)• Randomization––randomization status (OPT, CRT-P, or CRT-D) wasassigned• Implant (CRT-P or CRT-D arm)––implant of investigational devices andacute device testing for those randomized to a CRT therapy arm• Routine follow-up––routine evaluation of device function and patientcondition at pre-discharge, one week, and one month post-implant• Three- and six-month visits––evaluation of randomized therapy with specialtesting and device function at three and six months after the Post-RecoveryVisit• Quarterly Visits––after the six-month visit, patients were seen for routineevaluation of device function and patient conditionDEMOGRAPHIC DATAAll baseline patient characteristics are presented in Table B-5 on page B-16.- DRAFT -

B-16 CLINICAL STUDY - COMPANIONTable B-5. Patient population characteristics for COMPANION (OPT and CRT-D)Characteristic OPT(N = 308)CRT-D(N = 595)P-valueAge (years) Mean ± SD 66.7 ± 10.7 65.6 ± 11.2 0.14Female 97 (31.4) 194 (32.6) 0.73Gender [N (%)]Male 211 (68.5) 401 (67.3) 0.73Class III 253 (82.1) 512 (86.1) 0.12NYHA Classiﬁcation [N (%)]Class IV 55 (17.8) 83 (13.9) 0.12Ischemic 58.7 54.6 0.13Ischemic Etiology (%)Non-ischemic 41.3 45.4 0.13LVEF (%) Mean ± SD 22.8 ± 7.2 22.5 ± 6.8 0.47Resting Heart Rate (bpm) Mean ± SD 72 ± 12 73 ± 13 0.37QRS Width (ms) Mean ± SD 156 ± 24 159 ± 24 0.09LBBB 69.8 72.9 0.21Non-speciﬁc21.4 16.8 0.21Conduction Abnormality (%)RBBB 8.77 10.2 0.21Duration of Heart Failure (years) Mean ± SD 4.86 ± 4.41 4.44 ± 3.83 0.43Diuretic 94.4 96.6 0.12ACE inhibitor or ARB 88.6 89.6 0.66Beta Blockers 66.2 67.6 0.69Aldosterone Antagonist 54.8 55.1 0.94Heart Failure Medications [(%)]Digoxin 67.2 70.9 0.25PATIENT ACCOUNTABILITY AND FOLLOW-UP DURATIONThe COMPANION study enrolled 1638 patients, with 1520 patients randomizedto a treatment group and one hundred eighteen patients (118) not randomizeddue to changes in patient condition or consent between time of enrollment andtime of randomization, such that the inclusion criteria were no longer satisﬁed.Of the 1520 patients, 595 were randomized to CRT-D with a mean follow-up of1.3 years and 308 were randomized to OPT with a mean follow-up of 1.1 years.Figure B-1 on page B-18 provides an overview of patient enrollment.Table B-6 on page B-17 gives a summary (by treatment group) of patientdisposition over time through 12 months after randomization. This does not- DRAFT -

CLINICAL STUDY - COMPANION B-17account for patients that had a hospitalization or death event that contributed tothe primary endpoint or secondary endpoint of all-cause mortality.Table B-6. Patient follow-up disposition 12 months post-randomizationCRT-D OPT#ofWithdrawnPatients#ofDeceasedPatients(N =595) #Reachedend ofstudy(Nov. 30,2002)#ofActivePatientsat endof timeinterval#ofWithdrawnPatients#ofDeceasedPatients(N = 308)#Reachedend ofstudy(Nov. 30,2002)#ofActivePatientsat endof timeinterval1Day-7Days4 3 0 588 6 0 0 3027Days-1 Month435576 10 3 1 2881Month-3 Months4 15 6 551 11 11 1 2653 Months-9 Months12 28 49 462 26 22 29 1889 Months-12Months1 12 35 414 11 11 19 147- DRAFT -

B-18 CLINICAL STUDY - COMPANIONCOMPANION Enrolled n = 1638 Fall out prior to Randomization n = 118 Randomized n = 1520 Randomized to OPT n = 308 Randomized to CRT-D n = 595 Randomized to CRT-P n = 617 CRT-D Implant Intents n = 7 CRT-D Implant Attempts n = 47 Implanted CRT-D Devices n = 541 System Safety Endpoint n = 541 All-cause mortality endpoint through 12/1/02 for OPT and CRT-D n = 903 All adverse events through 11/26/03 for OPT and CRT-D n = 903 All-cause mortality or first hospitalization and cardiac morbidity endpoints through 12/1/02 for OPT and CRT-D n = 903 Figure B-1. Study patient enrollment and randomization for CRT-D and OPT- DRAFT -

CLINICAL STUDY - COMPANION B-19Event Contributing to Primary Endpoint and Secondary Endpoint ofAll-cause MortalityA total of 903 COMPANION patients in the CRT-D (595) and OPT (308) groupswere eligible for the primary endpoint. Figure B-2 on page B-19 provides patientrandomization and status for the primary endpoint and Figure B-3 on page B-19provides patient randomization and status for the secondary mortality endpoint.Randomized n = 903 of 1520 OPT n = 308 CRT-P n = 595 Hospitalization n = 196 Death n = 20 Unknown n = 15 No event n = 77 Hospitalization n = 372 Death n = 18 Unknown n = 5 No event n = 200 No event n = 92 Event n = 216 No event n = 205 Event n = 390 Figure B-2. CRT-D and OPT patient randomization for primary endpointRandomized n = 903 Alive n = 218 Unknown n = 15 Dead n = 77 Alive n = 484 Unknown n = 6 Dead n = 105 OPT n = 308 CRT-D n = 595 Figure B-3. CRT-D and OPT patient randomization for mortality endpoint- DRAFT -

B-20 CLINICAL STUDY - COMPANIONDATA ANALYSIS AND RESULTS FOR PRIMARY ENDPOINT ANDSECONDARY ALL-CAUSE MORTALITY ENDPOINTSequential MonitoringThe COMPANION DSMB met approximately every six months to review thetrial’s progress and to review the safety and effectiveness data collected. An“O’Brien-Fleming” type boundary as implemented by Lan and DeMets wasused in monitoring the trial. The Group sequential procedure ensured thatthe total alpha spent across repeated analyses did not exceed the total type Ierror, in this case a=0.03.On November 18, 2002 the DSMB reviewed the study progress for the ﬁnaltime. The CRT-D arm of the Study had reached the target number of eventsfor both the combined mortality and hospitalization endpoint as well as theall-cause mortality endpoint prompting the DSMB to recommend to the SteeringCommittee that enrollment be stopped. All effectiveness follow-ups endedon December 1, 2002.ResultsPrimary Endpoint: All-cause Mortality or First HospitalizationThe Kaplan-Meier curves illustrate the time to all-cause mortality or ﬁrsthospitalization (Figure B-4 on page B-21). There were 216 primary endpointevents observed in the OPT arm and 390 in the CRT-D arm (p = 0.010; p =0.011 after adjustment for interim analyses). The median time to ﬁrst event was209 days in the OPT group and 269 days in the CRT-D group. The annualevent rates for OPT and CRT-D, respectively, were 68.0% and 55.9%, with ahazard ratio of 0.80; 95% CI (0.68, 0.95). This result demonstrated that CRT-Dsigniﬁcantly reduced the relative risk of all-cause mortality or ﬁrst hospitalizationby 20% when compared to OPT alone.- DRAFT -

CLINICAL STUDY - COMPANION B-211080 960 840 720 600 480 360 240 120 0 100 80 60 40 20 0 % of Patients Event-Free Days from Randomization All-cause Mortality or First Hospitalization CRT-D vs. OPT: p = 0.010 Number of Events OPT: 216 CRT-D: 390 OPT CRT-D 308 595 Number of Patients at Risk OPT CRT-D 176 385 115 283 72 217 46 128 24 61 16 25 6 8 1 0 HR = 0.80. 95% CI (0.68, 0.95) Figure B-4. Primary Endpoint: All-cause mortality or ﬁrst hospitalizationIn addition to the hazard ratio, point estimates of risk reduction were alsocalculated (Table B-7 on page B-21). These estimates will vary with time fromthe true treatment effect, and thus should be interpreted with caution.Table B-7. Risk reduction point estimates% Failure Absolute RiskReductionRelative RiskReductionOPT CRT-D6months 44.9%(38.9%, 50.3%)42.9%(38.7%, 46.7%)2.0% 4.5%12 months 68.0%(61.7%, 73.2%)55.9%(51.6%, 59.8%)12.1% 17.8%18 months 77.8%(71.6%, 82.7%)69.0%(64.5%, 73.1%)8.8% 11.3%Secondary EndpointsAll-cause Mortality––deaths from any cause were reported in 77 patientsrandomized to OPT and 105 patients randomized to CRT-D (p = 0.003, p= 0.004 after adjusting for interim analyses). The Kaplan-Meier curves are- DRAFT -

B-22 CLINICAL STUDY - COMPANIONdepicted in Figure B-5 on page B-22. These numbers correspond to an annualmortality rate of 19% in the OPT arm and 12% in the CRT-D arm, with a hazardratio of 0.64, 95% CI (0.48, 0.86). These results demonstrated that CRT-Dwas associated with a 36% reduction in the risk of all-cause mortality whencompared to OPT alone.1080 720 630 540 450 360 270 180 90 0 100 90 80 70 60 50 % of Patients Event-Free Days from Randomization All-cause Mortality CRT-D vs. OPT: p = 0.003 Number of Events OPT: 77 CRT-D: 105 OPT CRT-D 308 595 Number of Patients at Risk OPT CRT-D 284 555 255 517 217 470 186 420 141 331 94 219 57 148 45 95 HR = 0.64. 95% CI (0.48, 0.86) 25 47 4 21 2 1 810 900 990 Figure B-5. Secondary Endpoint: All-cause mortalityIn addition to the hazard ratio, point estimates of risk reduction were alsocalculated (Table B-8 on page B-22). These estimates will vary with time fromthe true treatment effect, and thus should be interpreted with caution.Table B-8. Mortality endpoint risk reduction point estimates% Failure Absolute RiskReductionRelative RiskReductionOPT CRT-D6months 9.0%(5.7%, 12.2%)7.3%(5.1%, 9.3%)1.7% 18.9%- DRAFT -

CLINICAL STUDY - COMPANION B-23Table B-8. Mortality endpoint risk reduction point estimates (continued)% Failure Absolute RiskReductionRelative RiskReductionOPT CRT-D12 months 18.9%(14.1%, 23.5%)12.1%(9.3%, 14.8%)6.8% 36.0%18 months 28.4%(22.3%, 34.1%)18.0%(14.4%, 21.5%)10.4% 36.6%Results for Secondary Cardiac Morbidity EndpointDuring a hospitalization more than one of the pre-speciﬁed cardiac morbidevents could occur. The Anderson-Gill extension to the Cox proportionalhazard model was used to analyze time to multiple cardiac morbid events.Caution must be used in interpreting p-values in this analysis because thisanalysis does not account for the competing risk of death.In Figure B-6 on page B-24, the frequency and duration of cardiac morbidevents are illustrated. CRT-D was associated with a 36% reduction (p <0.0001) in the proportion of patients with at least one event, a 52% reduction(p < 0.0001) in events on an annual basis, and a 41% reduction (p < 0.0001)in the hospital duration on an annual basis. These reductions are primarilydue to the reduction of hospitalizations for acute decompensation of heartfailure, worsening heart failure resulting in IV inotrope or vasoactive therapy> 4 hours (during an inpatient hospitalization) and cardiac surgery (includingpercutaneous intervention) (Figure B-7 on page B-24).- DRAFT -

B-24 CLINICAL STUDY - COMPANION55% 1.08 7.660504020103001.21.00.80.40.20.60.08642035%Proportion of Patients with ≥ 1 Event (%)0.52Event Rate (Events/Patient-Year)4.5Duration (Days/Patient-Year)Frequency (per Patient)36% reductionp < 0.0001Duration41% reductionp < 0.0001Frequency (per Patient Year)52% reductionp < 0.0001OPT (n = 308) CRT-D (n = 595)Figure B-6. Secondary Endpoint: cardiac morbidityCaution must be used in interpreting p-values; analysis does not account forcompeting risk of death.0.680.80.60.40.20.00.42Events per Patient Year0.570.310.380.06 0.060.08 0.04 0.050.02 0.02OPTCRT-DHospitalization for acute decompensation of HFWorsening HF resulting in IV inotrope or vasoactive > 4 hoursCardiac surgery, including percutaneous interventionHospitalization resulting in death from cardiac causesMechanical respiratory or cardiac supportResuscitated cardiac arrest or sustained VT requiring external intervention38% reductionp < 0.000146% reductionp < 0.000184% reductionp < 0.000150% reductionp = 0.01617% reductionp = 0.54No changep = 0.55Figure B-7. Cardiac morbidity by major component- DRAFT -

CLINICAL STUDY - COMPANION B-25For a given cardiac hospitalization, patients may have events in more thanone category, and if there are multiple occurrences in a single category, thenonly the ﬁrst occurrence was counted.ADDITIONAL STUDY DATAImplant DispositionTable B-9 on page B-25 identiﬁes the number of initial and subsequent implantprocedures attempted in patients randomized to CRT-D and the rate of successfor each additional implant procedure. There were 81 CRT-D patients thathad an unsuccessful initial implant for the CRT-D system. Fifty (50) of thesepatients had a second implant procedure, of which 33 were successful and 17were unsuccessful. Three patients had a third implant procedure, of which onewas successful. Therefore, there were 541 patients implanted with the CRT-Dsystem.Table B-9. CRT-D system implant dispositionAttemptsuccessfulFailed implant Reattempt notdone after thisprocedureInitial implants 588 (98.8%) 507 (85.0%) 81 (14.0%) 31 (5.2%)First reattempt 50 (8.4%) 33 (5.5%) 17 (2.9%) 14 (2.3%)Second reattempt 3 (0.5%) 1 (0.2%) 2 (0.3%) 2 (0.34%)ADDITIONAL OUTCOME MEASURESFirst Heart Failure HospitalizationsAn additional outcome that was not pre-speciﬁed in the protocol providesfurther insight into the results observed in the composite primary endpoint.This post-hoc analysis was conducted using cause-speciﬁc hospitalizations asadjudicated by the morbidity and mortality committee and therefore shouldbe interpreted with caution.The outcome of all-cause mortality or ﬁrst heart failure hospitalization wasanalyzed on an intention-to-treat basis and time to ﬁrst event.Hospitalizations were deﬁned per any of the following:• Care provided at a hospital for any reason in which the duration isassociated with a date change- DRAFT -

B-26 CLINICAL STUDY - COMPANION• Use of intravenous inotropes and/or vasoactive drugs for a duration > 4hours (inpatient or outpatient)NOTE: Hospitalizations associated with a device implant attempt orre-attempt are excluded.Those contributing to the heart failure hospitalization outcome were requiredby the Morbidity and Mortality committee to meet at least one of the followingadditional criteria:•IVdiuretics• IV inotrope/vasoactive therapy• Other parenteral therapy for the treatment of heart failure•Signiﬁcant alterations in oral therapy for the treatment of heart failureAll-cause Mortality or First Heart Failure HospitalizationThe Kaplan-Meier curves for all-cause mortality or ﬁrst heart failurehospitalization is shown in Figure B-8 on page B-27. OPT and CRT-D hadannual event rates of 45% and 29%, respectively with a hazard ratio of 0.60,95% CI (0.49-0.75), p < 0.001. Therefore, CRT-D was associated with a40% relative reduction in the risk of all-cause mortality or ﬁrst heart-failurehospitalization when compared to OPT alone.- DRAFT -

CLINICAL STUDY - COMPANION B-2710809608407206004803602401200100806040200% of Patients Event-FreeDays from RandomizationAll-cause Mortality or First Heart Failure HospitalizationCRT-D vs. OPT: p < 0.001Number of EventsOPT: 145CRT-D: 212OPTCRT-D308595Number of Patients at RiskOPTCRT-D216497161411118470762283913128711127HR = 0.60. 95% CI (0.49, 0.75)25Figure B-8. All-cause mortality or ﬁrst heart failure hospitalizationDisposition of HospitalizationImplantation of the CRT-D system generally requires hospitalization. Todifferentiate between the hospitalization required to implant the system andthose hospitalizations that occurred after the system was implanted, thefollowing terms are used:• Implant hospitalization––the elective hospitalization associated with eitherthe implant procedure or a repeat implant procedure if the initial procedurewas unsuccessful.• All other hospitalizations––patients who required a revision for an implantedsystem (e.g., lead dislodgment or infection) were included in this categoryas were hospitalizations for non-elective device related implants.The hospitalizations analysis (Figure B-9 on page B-28) and hospitalizationdays analysis (Figure B-10 on page B-28) depicts hospitalization datastratiﬁed by implant and non-elective hospitalizations. This analysis was onan intention-to-treat basis and includes patients who underwent an attemptedimplant procedure. Patients randomized to CRT-D had a follow-up durationapproximately 30% longer than OPT patients. Thus, hospitalization data- DRAFT -

B-28 CLINICAL STUDY - COMPANIONare normalized per patient-year of follow-up. An additional comparison ofhospitalization days for heart failure hospitalizations is shown in Figure B-11on page B-29.NOTE: CRT-D was associated with a reduction in all-cause mortality andtherefore there is a competing risk for hospitalizations. This data should beinterpreted with caution.2.52.01.00.51.50.0Initial Implant 0.07Non-elective 1.551.62 Initial Implant 0.83Non-elective 1.182.01OPT(n = 308)CRT-D(n = 595)Hospitalizations (In-patient) per Patient-Year of Follow-upFigure B-9. Hospitalizations per patient year1284260Initial Implant 0.2Non-elective 10.911.2Initial Implant 2.4Non-elective 8.611.1OPT(n = 308)CRT-D(n = 595)All-cause Hospitalization (In-patient) Days per Patient-Year of Follow-up10Figure B-10. Hospitalization days per patient year- DRAFT -

CLINICAL STUDY - COMPANION B-296421305.93.8OPT(n = 308)CRT-D(n = 595)Heart Failure Hospitalization Days per Patient-Year of Follow-up5Figure B-11. Heart failure hospitalization days per patient yearDATA ANALYSIS AND RESULTS - CRT-D SYSTEM SAFETYThe system-related complication-free rate analysis was not a predeﬁnedendpoint in the protocol. The intent of this analysis is to provide reasonableassurance of safety of the CONTAK CD system in this patient population.The system-related complication-free rate was deﬁned over a six-monthfollow-up period as the proportion of patients who are free of complicationsattributed to:• Any implanted component (e.g, pulse generator, coronary venous lead,right atrial pace/sense lead, cardioversion/deﬁbrillation lead)• The surgical procedure required to implant the CRT-D systemIn the COMPANION study, this analysis was performed on an intention-to-treatbasis and also extends to those patients who underwent an implant procedurebut did not ultimately receive a device. Of the 595 patients analyzed, 522(87.7%) were free of system-related complications.Of the 73 (12.3%) patients who experienced a system-related complication,the most common were loss of left ventricular capture (25 patients, 4.2%),loss of right atrial capture (9 patients, 1.5%), and phrenic nerve/diaphragmaticstimulation (8 patients, 1.3%).When analyzed on a time-to-event basis, the system-related complication-freerate was 87.7%. The safety performance of the CONTAK CD system comparesfavorably with the safety performance observed in the prior CONTAK CD study(P010012, May 2, 2002).- DRAFT -

B-30 CLINICAL STUDY - COMPANIONDATA ANALYSIS AND RESULTS - COMPANION SUB-STUDYThe Exercise Performance Sub-study consisted of the following components.CRT EffectivenessPrimary Co-primary endpoint consisting of Peak VO2derived from asymptom-limited exercise test and Six-Minute Walk, with CRT results pooledfrom the CONTAK TR and CONTAK CD arms.Effectiveness was determined by assessing both Peak VO2and Six-MinuteWalk distance improvements with CRT compared to OPT.Prospectively, success was deﬁned as occurring if either of the followingoccurred:•PeakVO2improved ≥0.7 ml/kg/min (p < 0.05) and 6 MWD improvementresulted in p < 0.10•PeakVO2improved ≥0.5 ml/kg/min (p < 0.10) and 6 MWD improvementresulted in p < 0.05Additional: Quality of Life as measured by the Minnesota Living with HeartFailure Questionnaire®‚ and NYHA Class- DRAFT -

CLINICAL STUDY - COMPANION B-31Patient Accountability (Figure B-12 on page B-31)OPTBaseline Visit87Peak VO2Assessable for effectiveness466MWAssessable for effectiveness57Patient relatedCrossoverWithdrawnDeathOther11641010654105Peak VO2Assessable for effectiveness2316MWAssessable for effectiveness261Patient relatedMode ChangeWithdrawnDeathOther28201622124121623CRTBaseline Visit318Randomized448Met Exclusion Criteria 6Intent/Attempt 37Figure B-12. Enrollment and follow-up of randomized patientsBaseline Characteristics— (Table B-10 on page B-31)Table B-10. Patient population characteristicsCharacteristic CRT(N = 318)OPT(N = 87)P-valueaAge (years) Mean ± SD 62.1 ± 11.8 63.1 ± 10.6 0.48Range 32.0–86.0 27.0–85.0Gender [N (%)] Female 109 (34.3) 24 (27.6) 0.24Male 209 (65.7) 63 (72.4)NYHA Classiﬁcation [N (%)] III 294 (92.5) 79 (90.8) 0.61IV 24 (7.5) 8 (9.2)- DRAFT -

B-32 CLINICAL STUDY - COMPANIONTable B-10. Patient population characteristics (continued)Characteristic CRT(N = 318)OPT(N = 87)P-valueaIschemic Etiology Ischemic 141 (44.3) 42 (48.3) 0.51Non-ischemic 177 (55.7) 45 (51.7)LVEF (%) Mean ± SD 22.5 ± 6.9 22.2 ± 8.0 0.79Range 5.0–35.0 5.0–35.0Resting Heart Rate (bpm) Mean ± SD 73.1 ± 12.8 73.5 ± 11.5 0.78Range 46.0–122.0 54.0–103.0QRS Width (ms) Mean ± SD 159.2 ± 25.0 155.7 ± 25.8 0.26Range 120.0–276.0 120.0–224.0LBBB/NSIVCD (%) LBBB 230 (72.3) 62 (71.3) 0.60Nonspeciﬁc 54 (17.0) 18 (20.7)RBBB 34 (10.7) 7 (8.0)Peak VO2(ml/kg/min) Mean ± SD 12.7 ± 3.3 12.4 ± 3.3 0.42Range 3.0–21.2 4.8–21.5Six-MInute Walk Distance (m) Mean ± SD 292.4 ± 65.5 291.6 ± 70.5 0.92Range 152.0–411.5 162.4–414.0Quality of Life Score (points) Mean ± SD 59.8 ± 23.1 55.4 ± 23.3 0.12Range 0.0–105.0 0.0–97.0Heart Failure Medications [N (%)] Diuretic 300 (94.3) 82 (94.3) 0.98ACE Inhibitor or ARB 286 (89.9) 82 (94.3) 0.22Beta Blockers 240 (75.5) 60 (69.0) 0.22Aldosterone Antagonist 178 (56.0) 51 (58.6) 0.66Digoxin 239 (75.2) 65 (74.7) 0.93a. Continuous data were analyzed using a two-tailed t-test procedure, and categorical data were analyzed using a two-tailedchi-square procedure. A p-value < 0.05 is considered signiﬁcant.CRT EffectivenessPeak VO2—Peak VO2was determined from a standardized protocol forexercise testing as a means of measuring a patient’s capacity for performingphysical activity (Figure B-13 on page B-33, Table B-11 on page B-33).- DRAFT -

CLINICAL STUDY - COMPANION B-332.0 1.0 0.5 1.5 0.0 Time (months) ∆ Peak VO2 (ml/kg/min) 3 6 CRT OPT ∆ = 0.8 p = 0.026 ∆ = 0.6 p = 0.074 (Critical boundary = 0.10) Mean ± SE Figure B-13. Maximal Oxygen Consumption ResultsTable B-11. Maximal Oxygen Consumption ResultsCRT OPTPeak VO2(ml/kg/min)N Mean ±S.E.N Mean ±S.E.P-valueaChange at 3 months 247 1.1 ± 0.2 52 0.3 ± 0.4 0.026Change at 6 months 230 1.2 ± 0.2 46 0.6 ± 0.4 0.074a. P-values obtained using one-tailed longitudinal analysis methods.The longitudinal analysis was performed on all available data. The percentagesof missing data at the six-month endpoints for Peak VO2and Six-Minute Walkwere 36 percent and 28 percent for the CRT arm and 47 percent and 34 percentfor the OPT arm. The longitudinal analysis performed is most appropriate whenmissing data occurs at the percentages found in this trial.Six-Minute Walk––the Six-Minute Walk test is a measure of a patient’s abilityto sustain exercise during an activity similar to that which a patient may typicallyperform on a daily basis. For this test, patients are instructed to walk as far aspossible in 6 minutes in a level corridor (Figure B-14 on page B-34, Table B-12on page B-34).- DRAFT -

B-34 CLINICAL STUDY - COMPANION50 20 10 40 0 Time (months) ∆ Six Minute Walk Distance (m) 3 6 CRT OPT ∆ = 24 p = 0.016 ∆ = 24 p = 0.017 (Critical boundary = 0.05) Mean ± SE 30 Figure B-14. Change in six-minute walkTable B-12. Change in six-minute walkCRT OPTSix-Minute Walk (m)N Mean ±S.E.NMean±S.E.P-valueaChange at 3 months 274 37 ± 5 63 13 ± 10 0.016Change at 6 months 260 41 ± 5 57 17 ± 10 0.017a. P-values obtained using one-tailed longitudinal analysis methods.NYHA Class––the determination for New York Heart Association (NYHA) Classis based on mutual assessment, by the patient and physician, of the patient’sheart failure symptoms both at rest and while performing ordinary physicalactivity(FigureB-15onpageB-35,TableB-13onpageB-35).- DRAFT -

CLINICAL STUDY - COMPANION B-35100804020600CRT OPT36NYHA Class (%)CRT OPTImprove 2 ClassesImprove 1 ClassNo ChangeWorsen 1 ClassTime (months)Figure B-15. Change in NYHATable B-13. Change in NYHACRT OPTNYHAClassiﬁcationChangeN Patients N PatientsP-valuea3 months Improve 2 Classes 294 22 (7.5%) 69 3 (4.4%) <0.01Improve 1 Class 142(48.3%)13 (18.8%)No Change 122(41.5%)48 (69.6%)Worsen 1 Class 8 (2.7%) 5 (7.3%)6 months Improve 2 Classes 291 20 (6.9%) 65 2 (3.1%) 0.032Improve 1 Class 149(51.2%)28 (43.1%)No Change 118 (40.6%) 34 (52.3%)Worsen 1 Class 4 (1.4%) 1 (1.5%)a. P-values are not adjusted for multiplicity and were obtained using a one-tailed Mantel-Haenszel chi-square method.Quality of Life (QOL)––Quality of Life was assessed using the 21-questionMinnesota Living with Heart Failure questionnaire. Each question, answeredby the patient, is ranked on a scale ranging from 0 to 5. A lower total scoreindicatesanimprovedqualityoflife(FigureB-16onpageB-36,TableB-14onpage B-36).- DRAFT -

B-36 CLINICAL STUDY - COMPANION25 10 5 20 0 Time (months) ∆ Quality of Life Score (points) 3 6 CRT OPT ∆ = 17 ∆ = 13 Mean ± SE 15 30 Figure B-16. Quality of Life scoreTableB-14. QualityofLifescoreCRT OPTQuality of Life(points) NMean ± S.E. (95% CI) NMean ± S.E. (95% CI)P-valueaChange at 3 months 289 23 ± 1 (20.1, 25.7) 72 6±3(0.6,11.3) < 0.001Change at 6 months 279 23 ± 1 (19.7, 25.4) 66 10 ± 3 (4.2, 15.2) < 0.001a. P-values are not adjusted for multiplicity and were obtained using one-tailed longitudinal analysis methods.Additional Functional Capacity DataIn addition to the Exercise Performance sub-study, functional capacity wasevaluated by means of NYHA Class, six-minute walk distance, and MinnesotaLiving with Heart Failure Questionnaire QOL for the all patients randomizedto OPT and CRT-D through 6-months of follow up.NYHA Class, six-minute walk distance, and QOL scores were signiﬁcantlyimproved in the CRT-D group compared to the OPT group at 3 and 6 months(Table B-15 on page B-36). These ﬁndings are similar to those presented inthe exercise performance sub-study and previous cardiac resynchronizationtherapy trials.Table B-15. Changes in six-minute walk, QOL, and NYHACRT-D OPT P-valueaSix Minute Walk Distance NMean ± SD NMean ± SDChange at 3 months 420 42 ± 98 172 8 ± 82 < 0.0001- DRAFT -

CLINICAL STUDY - COMPANION B-37Table B-15. Changes in six-minute walk, QOL, and NYHA (continued)CRT-D OPT P-valueaChange at 6 months 377 45 ± 98 141 2 ± 92 < 0.0001QOL NMean ± SD NMean ± SDChange at 3 months 514 -24 ± 28 243 -8 ± 21 < 0.0001Change at 6 months 479 -23 ± 28 207 -12 ± 23 < 0.0001NYHA N % Improved N% ImprovedChange at 3 months 543 55 242 24 < 0.0001Change at 6 months 498 57 199 38 < 0.0001a. P-values are not adjusted for multiplicity and were obtained using t-tests for continuous data and chi-square for categorical data.- DRAFT -

B-38 CLINICAL STUDY - COMPANION- DRAFT -

C-1CLINICAL STUDY - DECREASE HFAPPENDIX CSUMMARYThe DECREASE-HF study was designed to determine if LV-CRT and LV Offsetare safe and effective as compared to the control treatment (BiV-CRT) inpatients with heart failure and an indication for an implantable cardioverterdeﬁbrillator (ICD). The primary effectiveness endpoint was a composite ofpeak VO2and left ventricular end systolic diameter (LVESD). The primarysafety endpoints were heart failure related adverse event free rate and systemrelated adverse event free rate. The LV Offset arm supports the safety andeffectiveness of the LV Offset feature.The DECREASE-HF Study design has been previously described in medicalliterature.1STUDY DESIGNPatients were randomized (1:1:1) to receive one of these three therapies.Patients who could not be randomized due to their inability to complete baselinetesting or because Expert Ease recommended BiV-CRT were followed forsafety data only in a separate “safety arm.” Available data for all patients wereanalyzed by randomization group assignment, regardless of actual therapyreceived (i.e., intent to treat).The DECREASE-HF clinical investigation used CONTAK RENEWAL 2/4/4HEdevices to study the LV Offset feature as well as other features that are notavailable in the CONTAK RENEWAL 1/3/3HE devices. The 2/4/4HE devicesare physically and mechanically identical to the 1/3/3HE devices and they bothcontain the LV Offset feature. As such, the data from the DECREASE-HFclinical study regarding the LV Offset feature, studied by using the 2/4/4HEdevices, applies to the CONTAK RENEWAL 1/3/3HE devices. The LV Offsetarm supports the safety and effectiveness of the LV Offset feature.1.De Lurgio D, Foster E, Higginbotham M, Larntz K, Saxon L. A Comparison of cardiacresynchronization by sequential biventricular pacing and left ventricular pacing to simultaneousbiventricular pacing: Rationale and design of the DECREASE-HF clinical trial. J Card Fail.2005;11(3):233-239- DRAFT -

$C-2 CLINICAL STUDY - DECREASE HFFOLLOW-UP SCHEDULEThe follow-up schedule for the DECREASE HF study is detailed below(Table C-1 on page C-2).Table C-1. DECREASE HF follow-up scheduleFollow-up period Follow-up schedulePre-implant Initial assessment of patient eligibility; taking of patient history.Administration of baseline Quality of Life (QOL) questionnaire.Implant Implant of investigational devices and acute device testing.Two-week visit Physical assessment, including NYHA assessment, and device evaluation.Special Testingato establish the patient’s baseline condition, after which the randomizationassignment was assigned.Three- and six-monthvisitEvaluation of randomized therapy with Special Testing and device functionb.Quarterly visits After the six-month visit, patients were seen for routine evaluation of device functionand patient condition.a. Special Testing included a Symptom-Limited Treadmill Test with measurement of oxygen uptake (Peak VO2), Echocardiography,QOL questionnaire.b. Holter monitor recordings were taken at the three-month visit for patients in the Holter Substudy.INCLUSION/EXCLUSION CRITERIAPatients enrolled in the investigation were required to meet the followinginclusion criteria:• Must meet the general indications for a CRT-D implant• Moderate or severe heart failure, deﬁned as NYHA Class III-IV despiteoptimal pharmacological heart failure therapy• A 12-lead electrocardiogram (ECG) obtained no more than 90 days prior toenrollment documenting a sinus rate > 50 bpm, QRS duration ≥150 ms,PR interval ≤320 ms measured from any two leads and a P-wave duration< 150 ms measured from lead V1• Creatinine ≤2.5 mg/dL obtained no more than 14 days prior to enrollment• Left ventricular ejection fraction ≤35% [measured by echo, multiple gatedacquisition (MUGA) scan, cardiac catheterization, etc.] no more than 14days prior to enrollment- DRAFT -$

CLINICAL STUDY - DECREASE HF C-3• Willing and capable of undergoing a device implant and participating in alltesting associated with this clinical investigation• Have a life expectancy of more than 180 days, per physician discretion• Age 18 or above, or of legal age to give informed consent speciﬁc to stateand national lawPatients were excluded from the investigation if they met any one of thefollowing exclusion criteria:• Right bundle branch block morphology (per World Health OrganizationGuidelines), on a 12-lead ECG obtained no more than 90 days prior toenrollment• Had previous cardiac resynchronization therapy, a previous coronaryvenous lead, or met the general indications for antibradycardia pacing• Had a neuromuscular, orthopedic, or other non-cardiac condition thatprevented normal, unsupported walking• Had an atrial tachyarrhythmia that was permanent (i.e., did not terminatespontaneously and could not be terminated with medical intervention) orpersistent (i.e., could be terminated with medical intervention, but did notterminate spontaneously) within 180 days prior to enrollment• Had a hypersensitivity to a 0.7 mg dose of dexamethasone acetate• Had surgically uncorrected primary valvular heart disease• Required dialysis at the time of enrollment• Had chronic obstructive pulmonary disease (COPD), deﬁned as FEV1/FVC<60%• Had a myocardial infarct, unstable angina, percutaneous coronaryintervention, or coronary artery bypass graft during the 30 days prior toenrollment• Had hypertrophic obstructive cardiomyopathy or inﬁltrative cardiomyopathy(e.g., amyloidosis, sarcoidosis)• Had a mechanical tricuspid prosthesis- DRAFT -

C-4 CLINICAL STUDY - DECREASE HF• Were enrolled in any concurrent study, without Guidant written approval,that may confound the results of this studyDEMOGRAPHIC DATAPatient enrollment (Figure C-1 on page C-4) and baseline characteristics(Table C-2 on page C-4) are detailed below.Enrolled 360 Implant 342 Attempt 16 Death 1 Intent 2 Death 3 Received Randomization Assignment 325 Withdrawal 1 Incomplete Baseline Testing 13 Expert Ease BiV Suggestion 19 Randomized 306 LV-CRT 101 LV Offset 104 BiV-CRT 101 Safety Arm 32 Figure C-1. DECREASE-HF Study Patient Enrollment and RandomizationTable C-2. DECREASE-HF study patient characteristicsCharacteristic Measurement LV-CRT(N=101)LV Offset(N=104)BiV-CRT(N=101)P-valueaAge at Implant (years) N 101 104 101Mean ± SD 67.4 ± 9.6 66.6 ± 10.5 66.2 ± 10.6 0.69Range 45.4 - 87.3 32.4 - 85.6 40.6 - 86.2Gender [N (%)] Male 66 (65) 70 (67) 69 (68) 0.90Female 35 (35) 34 (33) 32 (32)NYHA Class [N (%)] III 98 (97) 100 (96) 101 (100) 0.16IV 3(3) 4(4) 0(0)- DRAFT -

CLINICAL STUDY - DECREASE HF C-5Table C-2. DECREASE-HF study patient characteristics (continued)Characteristic Measurement LV-CRT(N=101)LV Offset(N=104)BiV-CRT(N=101)P-valueaLVEF (%) N 101 104 100Mean ± SD 22.6 ± 6.6 22.4 ± 6.7 23.2 ± 7.1 0.67Range 8.0 - 35.0 9.0 - 35.0 5.0 - 35.0QRS Duration (ms) N 101 104 101Mean ± SD 165 ± 15 167 ± 16 168 ± 15 0.29Range 150 - 220 150 - 220 150 - 218PR Interval (ms) N 101 104 101Mean ± SD 195 ± 42 195 ± 42 194 ± 39 0.98Range 120 - 318 100 - 320 88 - 320P-Wave Duration (ms) N 101 104 101Mean ± SD 91 ± 22 96 ± 22 95 ± 24 0.21Range 39 - 140 40 - 140 40 - 145Concomitant Medicationsb[N(%)]ACE Inhibitor/ARB 88 (87) 88 (85) 91 (90) 0.50Beta Blocker 84 (83) 84 (81) 82 (81) 0.89Digoxin 47 (47) 55 (53) 46 (46) 0.52Diuretic 89 (88) 93 (89) 82 (81) 0.19Loop Diuretic 87 (86) 91 (88) 80 (79) 0.22Nonloop Diuretic 8(8) 8(8) 8(8) 1.00Aldosterone Antagonist 40 (40) 37 (36) 40 (40) 0.79Antiarrhythmic 21 (21) 14 (13) 13 (13) 0.22Etiology [N (%)] Ischemic 67 (66) 70 (67) 58 (57) 0.27Nonischemic 34 (34) 34 (33) 43 (43)Conduction Disorder [N (%)] Left Bundle BranchBlock94 (93) 95 (91) 97 (96) 0.68NonspeciﬁcIntraventricularConduction6(6) 8(8) 4(4)Right Bundle BranchBlock1(1) 1(1) 0(0)a. P-values for continuous variables were calculated from a Student’s t-test; p-values for discrete variables were calculated froma Chi-squared test.b. Patients may appear in more than one category.- DRAFT -

C-6 CLINICAL STUDY - DECREASE HFSTUDY RESULTSTherapy EffectivenessPrimary EndpointComposite Score––Effectiveness of LV Offset was measured using aComposite Score that combines six-month changes in Peak VO2and LVESD(FigureC-2onpageC-6,TableC-3onpageC-7).Basedontheseestimatesofclinically meaningful improvement (1 ml/kg/min and -5 mm, for Peak VO2andLVESD, respectively), a scaling factor of 5 was chosen to give each componentapproximately equal weight, as follows: Composite Score = (5 x change inpeak VO2) - (change in LVESD).To evaluate the effectiveness of LV Offset, the Composite Score was comparedto the control arm using a longitudinal analysis. The null hypothesis was to berejected if the upper one-sided conﬁdence bound of the difference were lessthan 10 points.The observed mean differences from the BiV-CRT control arm was 3.6 ± 2.4in the LV Offset arm, with upper one-sided conﬁdence bound of 8.2 showingstatistical equivalence to BiV-CRT.200481216LV Offset BiV-CRT6-month Composite Score200481216LV OffsetDifference from BiV-CRT in a 6-month ScoreAcceptable BoundaryAll patients with peak VO2 and LVESD data at a minimum of one visit, N=189Figure C-2. Composite Score Equivalence to BiV-CRT at Six Months- DRAFT -

CLINICAL STUDY - DECREASE HF C-7Table C-3. Composite score equivalence to BiV-CRT at six monthsStatistic LV Offset BiV-CRTNaEstimate ± SE NaEstimate ± SE3-Month Composite Score 71 12.4 ± 1.5 70 16.0 ± 1.56-Month Composite Score 70 12.8 ± 1.7 76 16.4 ± 1.7Difference at 6 Months(BiV-CRT - LV Offset)3.6 ± 2.4Conﬁdence Interval Upper Bound 8.2a. N refers to the number of patients with paired data.Secondary EndpointsPeak VO2––A patient’s capacity for performing physical activity was assessedusing six-month change in Peak VO2achieved during CPX testing. Theendpoint analysis includes only CPX tests that are representative of maximalpatient effort, deﬁned as achievement of a Borg RPE ≥16 or RER ≥1.1.As deﬁned in the Protocol, patients with a baseline Peak VO2greater than20 ml/kg/min were excluded from the analysis. A longitudinal analysis thatincluded all patients with data at a minimum of one visit was performed toestimate six-month change from baseline in each group.As shown in (Figure C-3 on page C-8) and Table C-4 on page C-8, bothtreatment arms showed a statistically signiﬁcant and clinically meaningful (≥1.0 ml/kg/min) improvement in Peak VO2, an endpoint considered clinicallymeaningful in previous randomized controlled trials of CRT. The null hypothesiswas to be rejected if the lower one-sided 95% conﬁdence bound were greaterthan zero. The observed lower one-sided conﬁdence bound for LV Offset is1.1 ml/kg/min.- DRAFT -

C-8 CLINICAL STUDY - DECREASE HF15 13 12 Peak VO2 (ml/kg/min) 14 Baseline 3-Month 6-Month LV Offset BiV-CRT All patients with data at a minimum of one visit, N=189 Figure C-3. Improvement in Peak V02 at Six MonthsTable C-4. Improvement in Peak VO2 at six monthsStatistic LV Offset BiV-CRTNaEstimate ± SE NaEstimate ± SEBaseline 89 12.7 ± 0.2 88 12.7 ± 0.23 Months 72 14.2 ± 0.3 71 14.5 ± 0.36 Months 71 14.3 ± 0.3 76 14.2 ± 0.3Improvement at 6 Months 1.6 ± 0.3 1.5 ± 0.3Conﬁdence Interval Lower Bound 1.1 1.0a. N refers to the number of patients with paired data.Left Ventricular End Systolic Diameter (LVESD)––TheeffectofLVOffset was also assessed using six-month change in LVESD. A recordedechocardiographic examination was performed at the randomization visit (priorto CRT initiation) and subsequently at the three-month and six-month visits. Alongitudinal analysis that included all patients with data at a minimum of onevisit was performed to estimate six-month change from baseline in each group.Both arms showed a statistically signiﬁcant and clinically meaningfulimprovement (≤-5 mm) in LVESD (Figure C-4 on page C-9, Table C-5 onpage C-9). The null hypothesis was to be rejected if the upper one-sided 95%conﬁdence bound were less than zero. The observed upper one-sided boundfor LV Offset is -4.2 mm.- DRAFT -

CLINICAL STUDY - DECREASE HF C-946 54 58 LVESD (mm) 50 Baseline 3-Month 6-Month LV Offset BiV-CRT All patients with data at a minimum of one visit, N=205 56 52 48 Figure C-4. Improvement in LVESD at Six MonthsTable C-5. Improvement in LVESD at six monthsStatistic LV Offset BiV-CRTNaEstimate ± SE NaEstimate ± SEBaseline 104 55.7 ± 0.5 100 55.7 ± 0.53 Months 97 50.5 ± 0.9 97 48.9 ± 0.96 Months 92 50.3 ± 0.9 91 47.1 ± 0.9Improvement at 6 Months -5.4 ± 0.7 -8.7 ± 0.7Conﬁdence Interval Upper Bound -4.2 -7.5a. N refers to the number of patients with paired data.Quality of Life (QOL)––The effect of CRT on the patient’s perceived qualityof life was assessed using six-month change in QOL score. The MinnesotaLiving with Heart Failure Questionnaire‚ was administered prior to implant andsubsequently at the three-month and six-month visits. A longitudinal analysisthat included all patients with data at a minimum of one visit was performed toestimate six-month change from baseline in each group.Both arms showed a statistically signiﬁcant and clinically meaningfulimprovement (≤-10 points) in QOL, an endpoint considered clinicallymeaningful in previous randomized controlled trials of CRT (Figure C-5 on pageC-10, Table C-6 on page C-10). The null hypothesis was to be rejected if the- DRAFT -

C-10 CLINICAL STUDY - DECREASE HFupper one-sided 95% conﬁdence bound were less than zero. The observedupper one-sided conﬁdence bound for LV Offset was -19.4 points.20 QOL Score Baseline 3-Month 6-Month LV Offset BiV-CRT All patients with data at a minimum of one visit, N=205 30 40 50 60 Figure C-5. Improvement in Quality of Life at Six monthsTable C-6. Improvement in QOL at six months.Statistic LV Offset BiV-CRTNaEstimate ± SE NaEstimate ± SEBaseline 100 54.6 ± 1.4 98 54.6 ± 1.43 Months 94 33.5 ± 2.4 95 34.0 ± 2.36 Months 88 31.3 ± 2.4 91 32.5 ± 2.4Improvement at 6 Months -23.4 ± 2.4 -22.1 ± 2.4Conﬁdence Interval Upper Bound -19.4 -18.1a. N refers to the number of patients with paired data.NYHA Class––The effect of CRT on the patient’s heart failure relatedsymptoms (as measured by NYHA Class) was assessed prior to implant andsubsequently at the three-month and six-month visits. The analysis of NYHAClass included all patients with data at enrollment and six months.As shown in Figure C-6 on page C-11 and Table C-7 on page C-11, both armsshowed a statistically signiﬁcant percentage of patients who improved at leastone NYHA Class, an endpoint considered clinically meaningful in previousrandomized controlled trials of CRT. The null hypothesis was to be rejected- DRAFT -

CLINICAL STUDY - DECREASE HF C-11if the lower one-sided 95% conﬁdence bound of the percentage of patientsimproving one or more NYHA Class were greater than zero. The observedlower one-sided conﬁdence bound for LV Offset was 47.9%.100 Percent of Patients Improved LV Offset BiV-CRT All patients with data at a minimum of one visit, N=187 60 40 20 0 80 Figure C-6. Improvement in NYHA Class at Six MonthsTable C-7. Improvement in NYHA Class at six monthsStatistic LV Offset BiV-CRTTotal Patients 95 92Number (Percent) of Patients Improved 54 (56.8%) 54 (58.7%)Lower Bound of One-Sided Exact 95% Conﬁdence Interval 47.9% 49.6%Table C-8 on page C-11 provides additional detail, showing the percent ofpatients who improved two or three classes, as well as the percent of thosewho had no change or worsened.Table C-8. Six-month change in NYHA by treatment group6-Month Change in NYHAaLV Offset (N=95) BiV-CRT (N=92) Total (N=187)Improved 3 Classes 1 (1.1) 0 (0.0) 1 (0.5)Improved 2 Classes 16 (16.8) 8 (8.7) 24 (12.8)Improved 1 Class 37 (38.9) 46 (50.0) 83 (44.4)- DRAFT -

C-12 CLINICAL STUDY - DECREASE HFTable C-8. Six-month change in NYHA by treatment group (continued)6-Month Change in NYHAaLV Offset (N=95) BiV-CRT (N=92) Total (N=187)No Change 35 (36.8) 36 (39.1) 71 (38.0)Worsened 1 Class 6 (6.3) 2 (2.2) 8 (4.3)a. All patients with paired data; N=187.Device EffectivenessPrimary EndpointVentricular Tachycardia/Fibrillation Detection Time––The objective of thisendpoint was to demonstrate that CRT does not affect the ability to detectVT/VF. The results for VT/VF detection time are shown in Table C-9 on pageC-12.Table C-9. VT/VF detection timeNumber of PatientsaMean SD Upper Bound of One-Sided95% Conﬁdence Interval338 2.46 0.58 2.50a. All patients implanted with non-missing data; N=338.The null hypothesis was to be rejected if the upper one-sided 95% conﬁdencebound for mean VF detection time were less than 6 seconds. The observedupper one-sided 95% conﬁdence bound for VF detection time was 2.50seconds. These data demonstrate device effectiveness in the detection ofVT/VF.Therapy SafetyPrimary EndpointHeart Failure-Related Adverse Event Free Rate––Therapy safety wasassessed by the heart failure related adverse event free rate observedthrough six months of therapy delivery (randomization visit through six monthspost-randomization). The heart failure related adverse event free rate isdeﬁned as the number of patients who do not experience a heart failure relatedadverse event divided by the total number of patients implanted and activeat the randomization visit. All patients who were successfully implanted andremained active at the randomization visit were included in the analysis.- DRAFT -

CLINICAL STUDY - DECREASE HF C-13Table C-10 on page C-13 summarizes the heart failure related adverse eventrates through the six-month visit. A Kaplan-Meier analysis is also presented inFigureC-7onpageC-14toshowtimetoevents.Table C-10. Heart failure-related adverse event free rate at six monthsAdverse EventaNumberofEventsNumber ofPatientsHeart FailureAdverseEvent Free RateLower One-Sided 95%Conﬁdence BoundMultiple heart failure symptoms 38 29 91.4 88.5Dyspnea - Heart failure 13 13 96.2 94.0Heart failure symptoms -Unspeciﬁed13 12 96.4 94.3Hypotension - Heart failure 10 9 97.3 95.4Weight gain - Heart failure 54 98.8 97.3Fatigue - Heart failure 4 4 98.8 97.3Pulmonary edema - Heart failure 4 4 98.8 97.3Renal insufﬁciency - Heart failure 4 3 99.1 97.7Gastrointestinal - Heart failure 3 3 99.1 97.7Multi-system failure - Heart failure 2 2 99.4 98.1Peripheral edema - Heart failure 2 2 99.4 98.1Chest pain - Heart failure 1 1 99.7 98.6Dehydration - Heart failure 1 1 99.7 98.6Dizziness - Heart failure 1 1 99.7 98.6Elevated BNP - Heart Failure 1 1 99.7 98.6Total 102 67 80.2 76.3a. All patients implanted and active at the randomization visit; N=338.The null hypothesis was to be rejected if the lower one-sided 95% conﬁdencebound for heart failure adverse event free rate through six months post-implantwere greater than 50%. The heart failure related adverse event free rate at sixmonths was 80.2% with a lower one-sided 95% conﬁdence bound of 76.3%.- DRAFT -

C-14 CLINICAL STUDY - DECREASE HFMonths from Randomization VisitPercent Free from HF AEAll patients implanted and active at the randomization follow-up; N=338.405060708090100NAcceptance Boundary338 313 300 290 279 267 2580123456Event-Free RateLower One-Sided 95% Confidence BoundFigure C-7. Time to Heart Failure Related Adverse EventDevice SafetyPrimary EndpointSystem-Related Complication Free Rate––The safety of the investigationalsystem was assessed by the system related complication free rate observedin the period between implant and six months post-implant in all patientsattempted or implanted. System related complication free rate is deﬁnedas the proportion of patients without a system related complication withinsix months post-implant. All patients who underwent an implant procedurewere included in the analysis. Table C-11 on page C-14 shows the systemrelated complication free rates by event type. A Kaplan-Meier analysis is alsopresented in (Figure C-8 on page C-15) to show time to events.Table C-11. System-related complication free rate at six monthsComplicationaNumberofEventsNumberof PatientsComplicationFree RateLower One-Sided95% ConﬁdenceBoundLV Lead 35 31 91.3 88.5RA Lead 11 9 97.5 95.7RV Lead 3 3 99.2 97.8PG 14 14 96.1 94.0Procedure 17 16 95.5 93.3Tota l b80 60 83.2 79.7a. All patients implanted or attempted; N=358.b. Includes patients in the Safety Arm.- DRAFT -

CLINICAL STUDY - DECREASE HF C-15The null hypothesis was to be rejected if the lower one-sided 95% conﬁdencebound for system related complication rate through six months post-implantwere greater than 70%. The system related complication free rate at six monthswas 83.2% with a lower 95% conﬁdence bound of 79.7%.Months from Implant 0 1 2 3 4 5 6 60 70 80 90 100 N 358 287 285 280 272 268 265 Acceptance Boundary Percent Free from System Comp All patients implanted or attempted; N=358. Event-Free Rate Lower One-Sided 95% Confidence Bound Figure C-8. Time to System Related ComplicationThe Holter SubstudyAncillary EndpointsSixty-nine patients at nine centers were included in the Holter Substudy. Datawere collected at the three-month visit at centers participating in the HolterSubstudy. Holter recordings were analyzed by a Holter core laboratory.Continuous Appropriate Pacing During Activities of Daily Living––Thesafety of CRT therapy provided by the investigational system was assessed bythe percent of time a patient is appropriately paced over a 24-hour period, asrecorded with a Holter monitor at the three-month visit. The appropriatenessof CRT delivery is deﬁned by whether the device delivers CRT in accordancewith the physician’s programming. The objective of this endpoint was todemonstrate that patients receive continuous appropriate pacing from thedevice during activities of daily living.It is expected that patients will receive pacing approximately 95% of the time onaverage. The null hypothesis was to be rejected if the lower one-sided 95%conﬁdence bound of the mean time paced were greater than 90%. Due to thenon-normality of the data, a non-parametric test of the median was performed,which compared the median to 90% instead of comparing the lower 95%conﬁdence bound of the mean to 90%.- DRAFT -

C-16 CLINICAL STUDY - DECREASE HFThe mean percentage of appropriately paced beats during activities of dailyliving was 99.5 ± 1.3 with a median of 100.0% (p <0.01) (Table C-12 on pageC-16).Table C-12. Continuous appropriate pacing during activities of daily livingStatisticaResultN69Mean ± SD 99.5 ± 1.3Median 100.0Range 93.1 - 100.0P-valueb<0.01a. All patients in the Holter Substudy; N=69.b. P-value calculated from a signed-rank test.Continuous Appropriate Pacing During Exercise––The safety of CRTtherapy provided by the investigational system was assessed by the percentof time a patient receives appropriate pacing during the patient’s three-monthCPX test, as recorded with a Holter monitor. The appropriateness of CRTdelivery is deﬁned by whether or not the device delivers CRT in accordancewith the physician’s programming. The objective of this endpoint was todemonstrate that patients receive continuous appropriate pacing from thedevice during exercise.It is expected that patients will receive pacing approximately 95% of the time onaverage. The null hypothesis was to be rejected if the lower one-sided 95%conﬁdence bound of the mean time paced were greater than 90%. Due to thenon-normality of the data, a non-parametric test of the median was performedcomparing the median to 90% instead of comparing the lower 95% conﬁdencebound to 90%.The mean percentage of appropriately paced beats during exercise was 99.4 ±1.9 with a median of 100.0% (p < 0.01) (Table C-13 on page C-16).Table C-13. Continuous appropriate pacing during exerciseStatisticaResultN67Mean ± SD 99.4 ± 1.9Median 100.0- DRAFT -

CLINICAL STUDY - DECREASE HF C-17Table C-13. Continuous appropriate pacing during exercise (continued)StatisticaResultRange 90.3 - 100.0P-valueb<0.01a. All patients in the Holter Substudy; N=67.b. P-value calculated from a signed-rank test.- DRAFT -

C-18 CLINICAL STUDY - DECREASE HF- DRAFT -

D-1CLINICAL STUDY - CONTAK CDAPPENDIX DCLINICAL STUDY POPULATIONSGuidant CRT-Ds, when compared to OPT alone, have been demonstrated withreasonable assurance, to be safe and effective in signiﬁcantly reducing: therisk of a composite of all-cause mortality or ﬁrst hospitalization by 20%, therisk of all-cause mortality by 36%, and heart failure symptoms in patients whohave moderate to severe heart failure (NYHA III/IV) including left ventriculardysfunction (EF ≤35%) and QRS duration ≥120 ms and remain symptomaticdespite stable, optimal heart failure drug therapy, based on the Guidantsponsored COMPANION clinical study. (Guidant devices were the only devicesstudied in the COMPANION clinical trial.)SUMMARYGuidant conducted the CONTAK CD Study to demonstrate the safety andeffectiveness of the CONTAK CD system and to demonstrate a reasonableassurance of the safety and effectiveness of biventricular stimulation, or cardiacresynchronization therapy (CRT), using the Guidant Model 1822 VENTAK CHFAICD and Model 1823 CONTAK CD CRT-D along with the EASYTRAK (Models4510/4511/4512/4513) coronary venous, steroid-eluting, single-electrodepace/sense lead.The CONTAK CD Study failed to prospectively demonstrate effectiveness of theCRT portion of the device. The CONTAK CD Study met the Lead and SystemEffectiveness endpoints as well as the Lead and System Safety endpoints.Subgroup analysis revealed a population of patients that had Class III/IV heartfailure at the time of randomization that appeared to have improvements oncertain functional endpoints, including the Peak VO2and the Six-Minute Hallwalk. A second study was performed (Focused Conﬁrmatory Study) using thissubgroup of patients to conﬁrm the effectiveness of CRT.OBSERVED ADVERSE EVENTSThe VENTAK CHF/CONTAK CD/EASYTRAK Biventricular Pacing Study(hereafter referred to as the CONTAK CD Study) was a prospective,randomized, controlled, multicenter, double-blind study conducted at 47 sites inthe United States and enrolled a total of 581 patients. Of these, 57 patientsinitially underwent a thoracotomy procedure to receive the Guidant Model 1822VENTAK CHF AICD; 7 patients underwent a repeat procedure to receive an- DRAFT -

D-2 CLINICAL STUDY - CONTAK CDEASYTRAK lead. An additional 510 patients initially underwent an implantprocedure to receive the Model 1823 CONTAK CD CRT-D along with theEASYTRAK (Models 4510/4511/4512/4513) coronary venous, single-electrodepace/sense lead for a total of 517 patients who underwent an EASYTRAKlead implant procedure. In 69 patients the EASYTRAK lead implant attemptwas unsuccessful.Table D-1 on page D-2 provides information on all adverse events reportedfrom implant through the randomization period in patients attempted orimplanted with the EASYTRAK lead. During this period, a total of 765 eventswere reported in 310 patients. Of these, 155 were classiﬁed as complications,and 610 were classiﬁed as observations.Table D-1. Adverse events through randomization period# Of Events(# of pts)%Complications(Patients)Complicationsper 100Device Months(Events)%Observations(Patients)Observationsper 100Device Months(Events)Total AdverseEvents765 (310) 23.4 (121) 6.0 (155) 51.8 (268) 23.5 (610)PG-Related EventsMigration of device 1 (1) 0.0 (0) 0.0 (0) 0.2 (1) 0.0 (1)Pacemaker-mediatedtachycardia (PMT)3 (3) 0.0 (0) 0.0 (0) 0.6 (3) 0.1 (3)Telemetry difﬁculty 1 (1) 0.2 (1) 0.0 (1) 0.0 (0) 0.0 (0)LV Lead-Related EventsLoss of capture 43 (41) 5.6 (29) 1.1 (29) 2.5 (13) 0.5 (14)Inappropriate shockdue to oversensing1 (1) 0.0 (0) 0.0 (0) 0.2 (1) 0.0 (1)Insulation breachobserved1 (1) 0.2 (1) 0.0 (1) 0.0 (0) 0.0 (0)Multiple counting 31 (22) 1.0 (5) 0.2 (5) 3.9 (20) 1.0 (26)Phrenicnerve/diaphragmstimulation15 (15) 0.4 (2) 0.1 (2) 2.5 (13) 0.5 (13)RA Lead-Related EventsLoss of capture 6 (6) 1.0 (5) 0.2 (5) 0.2 (1) 0.0 (1)Oversensing 3 (3) 0.0 (0) 0.0 (0) 0.6 (3) 0.1 (3)Undersensing 1 (1) 0.2 (1) 0.0 (1) 0.0 (0) 0.0 (0)RV Lead-Related Events- DRAFT -

CLINICAL STUDY - CONTAK CD D-3Table D-1. Adverse events through randomization period (continued)# Of Events(# of pts)%Complications(Patients)Complicationsper 100Device Months(Events)%Observations(Patients)Observationsper 100Device Months(Events)Loss of capture 10 (9) 0.6 (3) 0.1 (3) 1.2 (6) 0.3 (7)Elevated DFTs 6 (6) 0.4 (2) 0.1 (2) 0.8 (4) 0.2 (4)Inappropriate shockabove rate cutoff49 (38) 0.4 (2) 0.1 (2) 7.2 (37) 1.8 (47)Inappropriate shockdue to oversensing5 (4) 0.0 (0) 0.0 (0) 0.8 (4) 0.2 (5)Nonconversion ofVF1 (1) 0.2 (1) 0.0 (1) 0.0 (0) 0.0 (0)Oversensing 2 (2) 0.0 (0) 0.0 (0) 0.4 (2) 0.1 (2)Phantom shock 2 (2) 0.0 (0) 0.0 (0) 0.4 (2) 0.1 (2)Phrenicnerve/diaphragmstimulation5 (5) 0.4 (2) 0.1 (2) 0.6 (3) 0.1 (3)SubtotalDevice-RelatedEvents186 (135) 9.5 (49) 2.1 (54) 19.0 (98) 5.1 (132)Procedure-Related EventsAV block 7 (7) 0.0 (0) 0.0 (0) 1.4 (7) 0.3 (7)Coronary sinusdissection5 (5) 0.0 (0) 0.0 (0) 1.0 (5) 0.2 (5)Coronary venousperforation5 (5) 0.2 (1) 0.0 (1) 0.8 (4) 0.2 (4)Hematoma 11 (10) 0.8 (4) 0.2 (4) 1.2 (6) 0.3 (7)Hypotension 7 (7) 0.0 (0) 0.0 (0) 1.4 (7) 0.3 (7)Infection,post-operativewound7 (7) 0.6 (3) 0.1 (3) 0.8 (4) 0.2 (4)Pneumothorax 7 (7) 0.8 (4) 0.2 (4) 0.6 (3) 0.1 (3)Post surgicalwound discomfort10 (9) 0.2 (1) 0.0 (1) 1.5 (8) 0.3 (9)Renal failure 5 (5) 0.2 (1) 0.0 (1) 0.8 (4) 0.2 (4)Other 18 (18) 1.2 (6) 0.2 (6) 2.3 (12) 0.5 (12)SubtotalProcedure-RelatedEvents79 (71) 3.9 (20) 0.7 (17) 10.0 (51) 2.2 (56)- DRAFT -

D-4 CLINICAL STUDY - CONTAK CDTable D-1. Adverse events through randomization period (continued)# Of Events(# of pts)%Complications(Patients)Complicationsper 100Device Months(Events)%Observations(Patients)Observationsper 100Device Months(Events)Cardiovascular-Related EventsAV Block 3 (3) 0.0 (0) 0.0 (0) 0.6 (3) 0.1 (3)Arrhythmia - SVT 49 (42) 0.2 (1) 0.0 (1) 7.9 (41) 1.8 (48)Arrhythmia - VT 20 (17) 1.0 (5) 0.2 (5) 2.7 (14) 0.6 (15)Arrhythmia - brady 16 (14) 0.2 (1) 0.0 (1) 2.5 (13) 0.6 (15)Cardiac arrest 2 (2) 0.4 (2) 0.1 (2) 0.0 (0) 0.0 (0)Chest pain 30 (20) 1.0 (5) 0.2 (5) 3.1 (16) 1.0 (25)Coagulopathy 3 (3) 0.2 (1) 0.0 (1) 0.4 (2) 0.1 (2)Congestive heartfailure140 (91) 3.5 (18) 0.7 (18) 16.1 (83) 4.7 (122)Distalthromboemboli3 (2) 0.0 (0) 0.0 (0) 0.4 (2) 0.1 (3)Dizziness 17 (17) 0.0 (0) 0.0 (0) 3.3 (17) 0.7 (17)Dyspnea(shortness ofbreath)16 (13) 0.0 (0) 0.0 (0) 2.5 (13) 0.6 (16)Fatigue 10 (10) 0.0 (0) 0.0 (0) 1.9 (10) 0.4 (10)Hypertension 1 (1) 0.0 (0) 0.0 (0) 0.2 (1) 0.0 (1)Hypotension 11 (9) 0.2 (1) 0.0 (1) 1.7 (9) 0.4 (10)Myocardialinfarction2 (2) 0.0 (0) 0.0 (0) 0.4 (2) 0.1 (2)Pacemakersyndrome1 (1) 0.0 (0) 0.0 (0) 0.2 (1) 0.0 (1)Palpitations 2 (2) 0.0 (0) 0.0 (0) 0.4 (2) 0.1 (2)Pulmonary edema 6 (6) 0.4 (2) 0.1 (2) 0.8 (4) 0.2 (4)Shock 4 (4) 0.2 (1) 0.0 (1) 0.6 (3) 0.1 (3)Stroke syndrome orCVA4 (4) 0.0 (0) 0.0 (0) 0.8 (4) 0.2 (4)Syncope 9 (9) 0.0 (0) 0.0 (0) 1.7 (9) 0.3 (9)Thrombosis 3 (3) 0.0 (0) 0.0 (0) 0.6 (3) 0.1 (3)Vascular related 6 (6) 1.0 (5) 0.2 (5) 0.2 (1) 0.0 (1)- DRAFT -

CLINICAL STUDY - CONTAK CD D-5Table D-1. Adverse events through randomization period (continued)# Of Events(# of pts)%Complications(Patients)Complicationsper 100Device Months(Events)%Observations(Patients)Observationsper 100Device Months(Events)SubtotalCardiovascular-Related Events358 (200) 7.7 (40) 1.6 (42) 35.6 (184) 12.2 (316)Tota lNoncardiovascular-Related Events142 (92) 6.2 (32) 1.5 (39) 13.5 (70) 4.0 (103)DeathsA total of 109 deaths occurred during the study. These deaths occurred duringthe study periods as shown in Table D-2 on page D-5 along with the cause ofdeath as adjudicated by an independent events committee.Table D-2. Deaths that occurred during CONTAK CD studyCause of DeathStudy Perioda#ofptdeaths Cardiac:PumpFailureCardiac:ArrhythmicCardiac:OtherNon-cardiacUnknownAfter unsuccessful implantprocedure21 1 0 0 0Peri-operative (≤30 days) 10 52021Randomized therapy phase: NoCRTb16 9 0 1 3 3Randomized therapy phase: CRTb11 4 1 2 2 2Post-randomized therapy phasec70 26 511620Total 109 47 9 4 23 26a. All patients enrolled, N = 581.b. Day 31 to 120 for Phase I patients, day 31 to 210 for Phase II patients.c. Day 121 and beyond for Phase I patients, day 211 and beyond for Phase II patients.STUDY DESIGNThe CONTAK CD Study was a prospective, randomized, controlled,multi-center, double-blind study conducted at 47 sites in the United Statesand enrolled a total of 581 patients. All patients enrolled were intended to beimplanted with a device capable of delivering both CRT and treating ventriculartachyarrhythmias. Patients were randomized to CRT Off (VVI lower rate 40)- DRAFT -

$D-6 CLINICAL STUDY - CONTAK CDor CRT On (VDD). The study began as a crossover design (called "Phase I")and enrolled 248 patients with a primary endpoint of functional status withthree months of follow-up. The study was later modiﬁed to a parallel design(called "Phase II") and enrolled 333 patients with a longer, six-month follow-up.Thedatafromtheﬁrst three months of the crossover phase were pooledwith data obtained from the six-month parallel phase. The visit scheduleand testing requirements remained the same. Additionally, while the studyoriginally used the VENTAK CHF ICD in conjunction with epicardial leadsplaced via thoracotomy, the CONTAK CD CRT-D and EASYTRAK lead (placedtransvenously) were added to the protocol later in the study.INCLUSION/EXCLUSION CRITERIAPatients enrolled in the study were required to meet the following inclusioncriteria:• Meet the general indication for ICD implant• Symptomatic heart failure despite optimal drug therapy (ACE inhibitors withdiuretic and/or digoxin, as determined to be indicated and tolerated bythe patient’s physician-investigator)• Left ventricular ejection fraction ≤35%• QRS duration ≥120 ms•Age≥18 years• Normal sinus node functionPatients were excluded from the investigation if they met any of the followingcriteria:• Meet the general indications for permanent antibradycardia pacing,including pacemaker dependence• Have chronic, medically refractory atrial tachyarrhythmias• Require concomitant cardiac surgery• Are unable to undergo device implant, including general anesthesia ifrequired• Are unable to comply with the protocol and follow-up requirements,including exercise testing• Have a life expectancy of less than six months due to other medicalconditions• Have amyloid disease (amyloidosis)• Have hypertrophic obstructive cardiomyopathy- DRAFT -$

CLINICAL STUDY - CONTAK CD D-7• Require in-hospital continuous intravenous inotropes• Have pre-existing cardioversion/deﬁbrillation leads other than thosespeciﬁed in this investigational plan (unless the investigator intends toreplace them with permitted cardioversion/deﬁbrillation leads)• Women who are pregnant or not using medically accepted birth control• Have a mechanical tricuspid prosthesis• Involved in other cardiovascular clinical investigations of active therapyor treatmentFOLLOW-UP SCHEDULEThe follow-up schedule for the clinical study consisted of the followingcomponents:• Pre-implant visit––initial assessment of patient eligibility; taking of patienthistory.• Implant––implant of investigational devices and acute device testing.Randomization status (CRT or No CRT) was assigned for implementationafter a 30-day recovery period.• Recovery period––minimum 30-day period over which the patient recoveredfrom the implant procedure and had his/her heart failure medicationsadjusted, but with no CRT, regardless of the randomization assignment.• Post-recovery visit––ﬁrst visit after the Recovery Period in which patientsunderwent Special Testing to establish their baseline condition, after whichthe randomization assignment was implemented (CRT or No CRT).• Three- and six-month visit––evaluation of randomized therapy with SpecialTesting and device function at three- and six-months after the post-recoveryvisit.• Quarterly visits––After the six-month visit, patients were seen for routineevaluation of device function and patient condition.NOTE: Special Testing included a Symptom-Limited Treadmill Testwith measurement of oxygen uptake (Peak VO2), a Six-Minute Walk,Echocardiography, Holter monitoring, blood chemistry testing, and a Qualityof Life (QOL) questionnaire- DRAFT -

D-8 CLINICAL STUDY - CONTAK CDDEMOGRAPHIC DATAThe CONTAK CD Study included patients with symptomatic heart failuredespite optimal drug therapy as deﬁned in the inclusion criteria. The populationincluded patients who were NYHA Class II, III, or IV at the time of implant.Based upon the clinical results from the covariate analyses in this study and theinternal consistency of these clinical ﬁndings with those from other completedCRT studies, the patient subgroup with NYHA Class III/IV heart failure in thisstudy was examined further.• All Patients––all patients (NYHA Class II/III/IV at the time of implant)implanted with an investigational system (N = 501). Ten patients diedand one withdrew before the post-recovery visit. Therefore, therapyeffectiveness analyses used N = 490.• NYHA Class III/IV (Advanced Heart Failure)––this subgroup was deﬁnedas those patients with moderate to severe heart failure at the time of thePost-Recovery Visit (N = 227). A percentage of patients either had animprovement or worsening of their NYHA Class during the post-implantrecovery period. The patients in the Advanced Heart Failure subgroupwere only those who remained in NYHA Class III/IV at the end of thepost-recovery period. This subgroup was determined from interactionanalysis of preselected covariates with the functional status endpoints.ENDPOINTSThe CONTAK CD Study had three investigational elements consisting of thefollowing components:• CRT effectiveness– Primary––composite endpoint consisting of all-cause mortality,hospitalization for heart failure, and ventricular tachyarrhythmiarequiring device intervention– Secondary–Peak VO2derived from a symptom-limited exercise testand Quality of Life as measured by the Minnesota Living with HeartFailure Questionnaire®– Additional––Six-Minute Walk, NYHA Class, EchocardiographicAnalysis, Change in Norepinephrine, and Change in Heart Rate- DRAFT -

CLINICAL STUDY - CONTAK CD D-9• Lead and System Effectiveness– Lead––left ventricular pacing thresholds, biventricular sensing,biventricular lead impedance, and lead placement success rate– System––VF detection time and biventricular ATP effectiveness• Lead and System Safety– Lead––incidence of lead-related adverse events– System––incidence of severe, device-related adverse events andoperative mortalitySTUDY RESULTSPatient Accountability (Figure D-1 on page D-10)- DRAFT -

D-10 CLINICAL STUDY - CONTAK CDControl (No CRT)n = 245Assessable for effectivenessn = 224Assessable for effectivenessn = 232CRTn = 245Randomizedn = 49016230DeathUncorrected lead dislodgementHeart transplantWithdrew11011Six monthsn = 125Three monthsn = 99Six monthsn = 126Three monthsn = 106101DeathWithdrewImplantedn = 5011466Withdrew prior to surgery (“Intent”)Unable to implant lead (“Attempt”)Enrolledn = 581Figure D-1. Enrollment and follow-up of randomized patientsBaseline Characteristics––Includes all patients implanted; N = 501(Table D-3on page D-11, Table D-4 on page D-12).- DRAFT -

CLINICAL STUDY - CONTAK CD D-11Table D-3. Pre-implant assessmentAll Patients NYHA Class III/IVCharacteristicCRT(N = 248)No CRT(N = 253)P-valueaCRT(N = 117)No CRT(N = 110)P-valueaAge at Implant(years)N 248 253 117 110Mean ±SD66.0 ± 10.5 66.3 ± 10.5 0.73 66.1 ±10.565.8 ± 10.5 0.80Range 26.1 - 82.6 29.5 - 86.3 26.1 - 82.5 38.3 - 85.3Gender [N (%)] Male 210 (84.7) 211 (83.4) 0.70 90 (76.9) 86 (78.2) 0.82Female 38 (15.3) 42 (16.6) 27 (23.1) 24 (21.8)NYHA Class [N(%)]II 80 (32.3) 83 (32.8) 0.66 20 (17.1) 11 (10.0) 0.08III 148 (59.7) 144 (56.9) 85 (72.6) 78 (70.9)IV 20 (8.1) 26 (10.3) 12 (10.3) 21 (19.1)ConcomitantMedications [N(%)]ACE orARB212 (85.5) 224 (88.5) 0.31 95 (81.2) 98 (89.1) 0.10BetaBlocker119 (48.0) 117 (46.2) 0.70 53 (45.3) 44 (40.0) 0.42Digoxin 172 (69.4) 171 (67.6) 0.67 84 (71.8) 75 (68.2) 0.55Diuretic 217 (87.5) 210 (83.0) 0.16 108 (92.3) 95 (86.4) 0.15Qualifying LVEF(%)N 248 253 117 110Mean ±SD21.4 ± 6.6 21.5 ± 6.7 0.74 20.6 ± 6.4 21.1 ± 6.2 0.61Range 5.0 - 35.0 10.0 - 35.0 8.0 - 35.0 10.0 - 35.0PR Intervalb(ms) N 224 222 107 91Mean ±SD205± 42 202 ± 49 0.44 204 ± 41 200 ± 54 0.60Range 88 - 336 104 - 400 136 - 336 110 - 400Qualifying QRSDurationb(ms)N 226 224 109 93Mean ±SD160 ± 27 156 ± 26 0.06 164 ± 27 152 ± 24 < 0.01Range 120 - 240 120 - 264 120 - 240 120 - 222Resting Heart Rate( bpm)N 248 253 117 110- DRAFT -

D-12 CLINICAL STUDY - CONTAK CDTable D-3. Pre-implant assessment (continued)All Patients NYHA Class III/IVCharacteristicCRT(N = 248)No CRT(N = 253)P-valueaCRT(N = 117)No CRT(N = 110)P-valueaMean ±SD73 ± 12 75 ± 14 0.37 75 ± 13 74 ± 15 0.61Range 43 - 108 48 - 120 43 - 108 50 - 120Systolic BloodPressure (mmHg)N 247 253 116 110Mean ±SD118 ± 21 118 ± 21 0.95 116 ± 20 117 ± 23 0.72Range 79 - 197 70 - 190 79 - 191 74 - 190Diastolic BloodPressure (mmHg)N 247 253 116 110Mean ±SD67 ± 12 69 ± 12 0.27 68 ± 12 67 ± 14 0.85Range 31 - 100 40 - 109 31 - 100 40 - 109a. P-values for comparing means were calculated with Student’s t-test; p-values for comparing proportions were calculated withPearson’s chi-squared test.b. PR interval and QRS duration were not obtained for thoracotomy patients.Table D-4. Pre-implant historyAll Patients NYHA Class III/IVCharacteristicCRT(N = 248)No CRT(N = 253)P-valueaCRT(N = 117)No CRT(N = 110)P-valueaPrimaryTachyarrhythmia [N(%)]MonomorphicVT (MVT)148(59.7)136(53.8)0.44 72 (61.5) 48 (43.6) 0.03PolymorphicVT (PVT)16 (6.5) 20 (7.9) 7 (6.0) 7 (6.4)NonsustainedVT58 (23.4) 63 (24.9) 30 (25.6) 35 (31.8)VentricularFibrillation(VF)26 (10.5) 32 (12.6) 8 (6.8) 18 (16.4)Other 0 (0.0) 2 (0.8) 0 (0.0) 2 (1.8)Other Arrhythmias [N(%)]ParoxysmalAtrialFibrillation43 (17.3) 62 (24.5) 0.05 21 (17.9) 29 (26.4) 0.13Atrial Flutter 10 (4.0) 13 (5.1) 0.55 3 (2.6) 7 (6.4) 0.16- DRAFT -

CLINICAL STUDY - CONTAK CD D-13Table D-4. Pre-implant history (continued)All Patients NYHA Class III/IVCharacteristicCRT(N = 248)No CRT(N = 253)P-valueaCRT(N = 117)No CRT(N = 110)P-valueaArrhythmia/ConductionDisorder [N (%)]LBBB 133(53.6)138(54.5)0.83 59 (50.4) 59 (53.6) 0.55RBBB 35 (14.1) 31 (12.3) 21 (17.9) 14 (12.7)Non-Speciﬁc 80 (32.3) 84 (33.2) 37 (31.6) 37 (33.6)Etiology [N (%)] Ischemic 167(67.3)178(70.4)0.47 76 (65.0) 78 (70.9) 0.34Non-Ischemic 81 (32.7) 75 (29.6) 41 (35.0) 32 (29.1)a. P-values were calculated with Pearson’s chi-squared test.CRT EffectivenessHeart Failure Progression (Composite Index)––the Composite Index(primary endpoint) was a combination of three events: all-cause mortality,hospitalization for heart failure, and VT/VF event requiring therapy (Table D-5on page D-13). A committee consisting of three heart failure specialists andan electrophysiologist reviewed and adjudicated all patient deaths and allhospitalizations, deﬁned as an admission greater than 23 hours. Outpatientcare, emergency room care, and clinic visits less than 23 hours were collectedbut not considered to be hospitalizations for the purposes of analysis.Table D-5. Heart Failure Progression (Composite Index)CRT No CRTGroupaHeart Failure Mortalityor Morbidity Event N%N%Relative Reductionwith CRTAll Patients(N = 490)Death from any cause 114.5166.5 15% p = 0.35HF hospitalization 32 13.1 39 15.9VT/VF 36 14.7 39 15.9NYHA Class III/IV(N = 227)Death from any cause 11 9.4 11 10.0 22% p = 0.23HF hospitalization 23 19.7 27 24.5VT/VF 21 17.9 22 20.0a. All patients implanted and active 31 days post-implant.Twenty-seven patients died during the therapy phase. Mortality stratiﬁed bytreatment group and cause, as adjudicated by the Events Committee, is shown- DRAFT -

D-14 CLINICAL STUDY - CONTAK CDinTable D-6 on page D-14. The Kaplan-Meier curve, showing total survival bytreatment group, is shown in Figure D-2 on page D-14.TableD-6. Mortalitystratiﬁed by treatment group and causeDeathsaPatients with CRT(N = 245)Patients with No CRT(N = 245)Cardiac, pump failure 4 (1.6%) 9 (3.7%)Cardiac, arrhythmic 1 (0.4%) 0 (0.0%)Cardiac, other 2 (0.8%) 1 (0.4%)Noncardiac 2 (0.8%) 3 (1.2%)Unknown 2 (0.8%) 3 (1.2%)Tota l 11 (4.5%) 16 (6.5%)a. All patients implanted and active at 31 days post-implant; N = 490.6301009590850Total Survival (%)Time Post-randomization (Months)No CRTCRT245245242241239233127130125129123126122125p = 0.32CRTNo CRTPatients at RiskFigure D-2. Kaplan-Meier curveTable D-7 on page D-15 presents the reasons for hospitalization within thetreatment period as determined by the Events Committee. This table representsthe number of patients with each category of hospitalization. Patients may havemultiple hospitalizations that fall into different categories.- DRAFT -

CLINICAL STUDY - CONTAK CD D-15Table D-7. Patients hospitalized during treatment periodAll Patients NYHA Class III/IVReason for HospitalizationaCRT(N = 245)No CRT(N = 245)Total(N = 490)CRT(N = 117)No CRT(N = 110)Total(N = 227)Heart failure 32 39 71 23 27 50Cardiac, other 20 25 45 14 14 28Noncardiac 26 19 45 14 14 28Total Hospitalizations 66 70 136 40 46 86a. All patients implanted and active at 31 days post-implant; N = 490.Peak VO2––the Peak VO2was determined from a standardized protocol forexercise testing as a means of measuring a patient’s capacity for performingphysical activity. Figure D-3 on page D-15 and Table D-8 on page D-15 providethe change in Peak VO2.1513121411Time (months)Peak VO2 (ml/kg/min)06All Patients∆ = 0.8 ± 0.4p = 0.030Mean ± SECRT (n = 216)No CRT (n = 201)Time (months)06NYHA Class III/IV∆ = 1.8 ± 0.6p = 0.003CRT (n = 96)No CRT (n = 80)Figure D-3. Change in Peak VO2Table D-8. Change in Peak VO2All Patients NYHA Class III/IVPeak VO2(ml/kg/min) CRT(N = 216)No CRT(N = 201)P-valueaCRT(N = 96)No CRT(N = 80)P-valueaPost-recovery Visit 13.5 ± 0.2 13.5 ± 0.2 –– 12.0 ± 0.3 12.0 ± 0.3 ––- DRAFT -

D-16 CLINICAL STUDY - CONTAK CDTable D-8. Change in Peak VO2 (continued)All Patients NYHA Class III/IVPeak VO2(ml/kg/min) CRT(N = 216)No CRT(N = 201)P-valueaCRT(N = 96)No CRT(N = 80)P-valuea3 Months 14.3 ± 0.2 13.9 ± 0.2 0.206 12.8 ± 0.4 12.1 ± 0.4 0.0846 Months 14.4 ± 0.3 13.6 ± 0.3 0.030 13.8 ± 0.5 12.0 ± 0.5 0.003a. P-values reﬂect the between-group differences with respect to baseline.Six-Minute Walk––the Six-Minute Walk test is a measure of a patient’s abilityto sustain exercise during an activity similar to that which a patient may typicallyperform on a daily basis. For this test, patients are instructed to walk as faras possible in 6 minutes in a level corridor. Figure D-4 on page D-16 andTable D-9 on page D-16 provide the change in Six-Minute Walk.375325275350250Time (months)Six Minute Walk Distance (m)06All Patients∆ = 21 ± 10p = 0.043Mean ± SECRT (n = 224)No CRT (n = 220)Time (months)06NYHA Class III/IV∆ = 39 ± 18p = 0.029CRT (n = 99)No CRT (n = 90)300Figure D-4. Change in Six-Minute WalkTable D-9. Change in Six-Minute WalkAll Patients NYHA Class III/IVSixMinuteWalkDistance(meters)CRT(N = 224)No CRT(N = 200)P-valueaCRT(N = 99)No CRT(N = 90)P-valueaPost-recovery Visit 317 ± 5 317 ± 5 –– 268 ± 9 268 ± 9 ––- DRAFT -

CLINICAL STUDY - CONTAK CD D-17Table D-9. Change in Six-Minute Walk (continued)All Patients NYHA Class III/IVSixMinuteWalkDistance(meters)CRT(N = 224)No CRT(N = 200)P-valueaCRT(N = 99)No CRT(N = 90)P-valuea3 Months 348 ± 7 331 ± 8 0.058 312 ± 12 280 ± 12 0.0286 Months 353 ± 8 332 ± 8 0.043 327 ± 14 288 ± 15 0.029a. P-values reﬂect the between-group differences with respect to baseline.Quality of Life (QOL)––QOL was assessed using the 21-question MinnesotaLiving with Heart Failure Questionnaire®. Each question is answered by thepatient, ranking each item on a scale ranging from 0 to 5. A lower total scoreindicates an improved quality of life. Figure D-5 on page D-17 and Table D-10on page D-17 provide the change in QOL.3040503555Time (months)QOL Score (points)06All Patients∆ = -2.1 ± 2.4p = 0.40Mean ± SECRT (n = 234)No CRT (n = 225)Time (months)06NYHA Class III/IV∆ = -10.1 ± 4.2p = 0.017CRT (n = 107)No CRT (n = 96)45Figure D-5. Change in Quality of LifeTable D-10. Change in Quality of LifeAll Patients NYHA Class III/IVQOL (points)CRT(N = 234)No CRT(N = 225)P-valueaCRT(N = 107)No CRT(N = 96)P-valueaPost-recovery Visit 41.8 ± 1.1 41.8 ± 1.1 –– 52.7 ± 1.5 52.7 ± 1.5 ––- DRAFT -

D-18 CLINICAL STUDY - CONTAK CDTable D-10. Change in Quality of Life (continued)All Patients NYHA Class III/IVQOL (points)CRT(N = 234)No CRT(N = 225)P-valueaCRT(N = 107)No CRT(N = 96)P-valuea3 Months 36.6 ± 1.5 37.3 ± 1.6 0.711 41.9 ± 2.4 47.5 ± 2.6 0.0786 Months 34.8 ± 1.8 36.9 ± 1.8 0.395 37.2 ±- 3.1 47.3 ± 3.2 0.017a. P-values reﬂect the between-group differences with respect to baseline.NYHA Class––the determination of New York Heart Association (NYHA) Classis based on mutual assessment by the patient and the patient’s physician ofthe patient’s heart failure symptoms both at rest and while performing ordinaryphysical activity. NYHA Class was determined at each follow-up visit by aphysician who was blinded to the patient’s randomized therapy. Figure D-6 onpageD-18andTableD-11onpageD-18providethechangeinNYHAClassresults.6040200Change in NYHA Class (%)CRT(n = 109)No CRT(n = 116)CRT(n = 45)No CRT(n = 48)Improve ≥ 2 ClassesWorsen ≥ 1 ClassImprove = 1 ClassNo changeAll Patients(NYHA Class II / III / IV)p = 0.10NYHA Class III / IVp = 0.006Figure D-6. Change in NYHA ClassTable D-11. Change in NYHA ClassAll Patients NYHA Class III/IVCRT(N = 109)No CRT(N = 116)CRT(N = 45)No CRT(N = 48)Change in NYHA ClassN%N%P-valueaN%N%P-valueaImprove 2 or More Classes 12 11.0 2 1.7 0.10 12 26.7 2 4.2 0.006Improve1Class 27 24.8 35 30.2 21 46.7 24 50.0No Change 56 51.4 59 50.9 10 22.2 18 37.5- DRAFT -

CLINICAL STUDY - CONTAK CD D-19Table D-11. Change in NYHA Class (continued)All Patients NYHA Class III/IVCRT(N = 109)No CRT(N = 116)CRT(N = 45)No CRT(N = 48)Change in NYHA ClassN%N%P-valueaN%N%P-valueaWorsen 1 Class 13 11.9 19 16.4 2 4.4 4 8.3Worsen2orMoreClasses 10.910.9 00.000.0a. P-value was calculated from Mantel-Haenszel test and reﬂects the between-group differences with respect to baseline.Echocardiography––several echocardiography (echo) variables wereidentiﬁed to assist in measuring the possible hemodynamic impact of CRT(Table D-12 on page D-19). The limitation of this data is that patientsare measured while at rest, and therefore, the data may not reﬂect anyhemodynamic beneﬁt that may be observed when patients are exercising andperforming their daily activities.Table D-12. Echocardiography resultsCRT No CRT Between GroupsParameter TimepointNMean ± SE NMean ± SE Mean ± SE P-valueAll PatientsLVIDd (mm) Post-recoveryVisit228 70.4 ± 0.5 219 70.4 ± 0.5 0 ––Change at 6Months228 -3.4 ± 0.6 219 -0.3 ± 0.6 -3.1 ± 0.9 < 0.001LVIDs (mm) Post-recoveryVisit228 58.3 ± 0.5 219 58.3 ± 0.5 0 ––Change at 6Months228 -4.0 ± 0.7 219 -0.7 ± 0.7 -3.3 ± 0.9 < 0.001LVEF (%) Post-recoveryVisit222 27.8 ± 0.3 216 27.8 ± 0.3 0 ––Change at 6Month222 5.1 ± 0.7 216 2.8 ± 0.7 2.4 ± 1.0 0.020NYHA Class III/IVLVIDd (mm) Post-recoveryVisit104 71.2 ± 0.7 92 71.2 ± 0.7 0 ––Change at 6Months104 -4.9 ± 1.0 92 -0.2 ± 1.1 -4.7 ± 1.5 0.001LVIDs (mm) Post-recoveryVisit104 59.2 ± 0.7 92 59.2 ± 0.7 0 ––- DRAFT -

D-20 CLINICAL STUDY - CONTAK CDTable D-12. Echocardiography results (continued)CRT No CRT Between GroupsParameter TimepointNMean ± SE NMean ± SE Mean ± SE P-valueChange at 6Months104 -5.4 ± 1.1 92 -0.6 ± 1.1 -4.8 ± 1.5 0.002LVEF (%) Post-recoveryVisit99 26.9 ± 0.5 91 26.9 ± 0.5 0 ––Change at 6Months99 6.0 ± 1.1 91 2.3 ± 1.2 3.7 ± 1.7 0.029Measures of Sympathetic Tone––Mean Norepinephrine levels (Table D-13 onpage D-20) and Mean Heart Rate (Table D-14 on page D-20) were examined asmarkers of how CRT may inﬂuence the excessive sympathetic drive associatedwith chronic heart failure.Table D-13. Mean Norepinephrine resultsAll Patients NYHA Class III/IVNorepinephrine (pg/mL)CRT(N =228)No CRT(N =217)P-value CRT(N =104)No CRT(N = 90)P-valuePost-recovery Visit 663 ± 19 663 ± 19 –– 720 ± 31 720 ± 31 ––3 Months 651 ± 31 681 ± 32 0.479 685 ± 55 743 ± 60 0.4636 Months 658 ± 40 738 ± 41 0.143 681 ± 75 827 ± 79 0.163Table D-14. Mean heart rate resultsAll Patients NYHA Class III/IVHeart Rate (bpm)CRT(N = 240)No CRT(N = 233)P-value CRT(N = 113)No CRT(N = 101)P-valuePost-recovery Visit 72.3 ± 0.6 72.3 ± 0.6 –– 74.5 ± 1.0 74.5 ± 1.0 ––3 Months 70.8 ± 0.8 72.1 ± 0.8 0.20 74.1 ± 1.2 73.9 ± 1.3 0.946 Months 69.4 ± 1.0 70.2 ± 1.0 0.58 70.6 ± 1.6 72.5 ± 1.6 0.40EASYTRAK Lead and System EffectivenessIt was hypothesized that the upper tolerance limit of the chronic left ventricularpacing threshold of the EASYTRAK lead be less than 5.5 V to ensure thatan adequate safety margin exists. Chronic left ventricular pacing thresholds- DRAFT -

CLINICAL STUDY - CONTAK CD D-21shown in Figure D-7 on page D-21 and Table D-15 on page D-21 are wellwithin this limit.6420LV Pacing Threshold (V at 0.5 ms)02418126Acceptance Boundary90% Tolerance IntervalImplant Duration (months)Figure D-7. EASYTRAK lead threshold measurementsTable D-15. EASYTRAK lead threshold measurementsStatisticaImplant 3 Months 6 Months 12 Months 18 Months 24 MonthsN 435 347 330 233 103 25Mean ± SD 1.8 ± 1.2 1.7 ± 1.3 1.9 ± 1.5 1.8 ± 1.2 1.8 ± 1.1 2.0 ± 1.2Range 0.2 - 7.5 0.2 - 7.5 0.2 - 7.5 0.4 - 7.5 0.6 - 7.5 0.6 - 5.0Upper Tolerance Limit 3.8 3.8 4.3 3.8 3.7 3.9a. EASYTRAK lead models: 4511, 4512, and 4513Mean chronic biventricular R-wave amplitudes are measured as a combinationof the R-waves from both the right ventricle (commercially available ENDOTAKlead) and left ventricle (EASYTRAK lead). It was hypothesized that the meanbiventricularR-waveamplitudebegreaterthan5mVtoensurepropersensing.InFigureD-8onpageD-22andTableD-16onpageD-22,theperformanceofthe EASYTRAK lead system was signiﬁcantly above this value (p < 0.01).- DRAFT -

D-22 CLINICAL STUDY - CONTAK CD1050BiV R-wave Amplitude (mV)02418126Acceptance BoundaryImplant Duration (months)15MeanFigure D-8. EASYTRAK biventricular-sensed R-wave amplitudeTable D-16. EASYTRAK biventricular-sensed R-wave amplitudeStatisticaImplant 3 Months 6 Months 12 Months 18 Months 24 MonthsN 433 346 326 220 99 23Mean ± SD 10.0 ± 5.2 9.9 ± 4.4 9.9 ± 4.5 9.8 ± 4.4 8.9 ± 3.5 8.5 ± 3.3Upper Tolerance Limit 1.9 - 25.0 1.4 - 25.0 1.7 - 25.0 1.2 - 25.0 2.6 - 20.4 2.2 - 13.6The impedance measured by the CONTAK CD device is the parallelcombination of the left ventricular (EASYTRAK) and right ventricular(ENDOTAK) leads simultaneously. Therefore, the biventricular lead impedancewill be substantially less than that of either lead alone. It was hypothesizedthat the lower limit of the 95% conﬁdence interval of the mean chronicbiventricular lead impedance would be greater than 200 Ωto ensure properpulse generator function. The lower limit of the 95% conﬁdence interval of thechronic biventricular lead impedance exceeds this value (Figure D-9 on pageD-23, Table D-17 on page D-23).- DRAFT -

CLINICAL STUDY - CONTAK CD D-233002000BiV Lead Impedance (Ω)02418126Acceptance BoundaryImplant Duration (months)400Lower Bound of 95% CIFigure D-9. EASYTRAK biventricular pacing impedanceTable D-17. EASYTRAK biventricular pacing impedanceStatisticaImplant 3 Months 6 Months 12 Months 18 Months 24 MonthsN 436 355 336 237 107 26Mean ± SD 340 ± 46 352 ± 47 349 ± 50 351 ± 51 347 ± 46 356 ± 67Range 243 - 550 248 - 519 186 - 534 237 - 513 254 - 507 267 - 52095% CI (336, 344) (347, 357) (344, 355) (345, 358) (338, 356) (329, 383)EASYTRAK Lead Placement Success Rate––the EASYTRAK lead wasimplanted in 448/517 (87%) of patients who underwent the implant procedure.Table D-18 on page D-23 shows the reasons for inability to place theEASYTRAK lead. Table D-19 on page D-24 provides the EASYTRAK leadimplant success rate.Table D-18. Reasons for unsuccessful EASYTRAK lead implantReason #ofptsa%Inability to locate or cannulate the coronary sinus 29 42Dislodgment of EASYTRAK lead while removing guide catheter 13 18.8Inability to advance the lead to a stable position 11 15.9Inability to obtain adequate pacing thresholds 6 8.7Procedure stopped due to coronary sinus dissection or perforation 57.2Procedure stopped due to transient AV block 1 1.4Procedure stopped due to venous perforation during subclavian stick 11.4Reason not stated 11.4- DRAFT -

D-24 CLINICAL STUDY - CONTAK CDTable D-18. Reasons for unsuccessful EASYTRAK lead implant (continued)Reason #ofptsa%Extracardiac stimulation 1 1.4Inability to place an atrial pace/sense lead 11.4Tota l 69 100a. Patients with unsuccessful attempt to implant EASYTRAK lead; N = 69.Table D-19. EASYTRAK lead placement success rateMeasurement All ProceduresaNumber of patients implanted or attempted 517Number of placements of the EASYTRAK Leadb448Rate 87%95% CI (84%, 90%)a. All patients implanted or attempted with EASYTRAK lead; N = 517.b. Deﬁned as an EASYTRAK implant procedure that is concluded with the implant of the investigational cardiac resynchronizationsystem.Although some situations such as patient anatomy and poor thresholds cannotbe avoided, increased investigator experience with the EASYTRAK lead andaccessories was associated with improved success, decreased total proceduretime (measured skin-to-skin), and decreased ﬂuoroscopy exposure time(FigureD-10onpageD-24).95Implant Success Rate (%)90858008-15>154-71-3250Mean Procedure Time (min)20015010008-15>154-71-360Mean Fluoroscopy Exposure Time (min)50403008-15>154-71-3Mean ± SEMean ± SEImplants (by order of enrollment)Figure D-10. EASYTRAK success rate, procedure time, and ﬂuoroscopy exposure time- DRAFT -

CLINICAL STUDY - CONTAK CD D-25Biventricular ATP Conversion Effectiveness Performance––the conversionrate of induced monomorphic ventricular tachycardia (MVT) was 64% andthat of spontaneous MVT was 88%.Ventricular Tachyarrhythmia Detection Time––the VENTAK CHF andCONTAK CD devices sense events from both ventricles simultaneously.Ventricular tachyarrhythmia detection time was analyzed to determine if theadditional lead had an adverse effect on sensing VT/VF. Guidant’s ICDstypically have a detection time of two seconds. The VF detection time of 2.1± 0.6 seconds was statistically signiﬁcantly lower than 6 seconds (p < 0.01),demonstrating that there was no statistically signiﬁcant prolongation of inducedVF detection times with the additional left ventricular lead1.Therewerealsono adverse events reported in which a VENTAK CHF or CONTAK CD failed todetect a spontaneous ventricular tachyarrhythmia.EASYTRAK Lead and System SafetyEASYTRAK Lead Safety––safety was established using the rate of adverseevents that are either related to the EASYTRAK lead or to the implantprocedure necessary to place the EASYTRAK lead.An EASYTRAK lead implant procedure was performed in 517 patients with 448patients (86.7%) being successfully implanted with the EASYTRAK lead. Theupper boundary of the 95% conﬁdence interval was hypothesized to be lessthan 23% at six months (Table D-20 on page D-25).Table D-20. Lead-related adverse events at six monthsPatient Population N Event Rate (%) 95% CIAll Patients 517 12.2 (9.4, 15.0)NYHA Class III/IV 201 17.4 (12.7, 22.7)Fifty-three lead-related adverse events were reported during the clinicalinvestigation of the EASYTRAK lead among the 448 patients who wereimplanted with an EASYTRAK lead. Twenty-seven procedure-related adverseevents were reported among the 517 patients who underwent the implantprocedure for an EASYTRAK lead.2The overall lead-related adverse event1. Detection time at implant with legally marketed Guidant ICD devices is typically two seconds,and investigators have stated that an additional delay of 3 to 5 seconds would be a clinicallysigniﬁcant event. The expected detection time is 2 seconds (95% CI: [0, 6 sec]).2. For purposes of deﬁning event rates, a denominator of 448 will be used for those adverseevents that pertain to chronically implanted EASYTRAK leads, and a denominator of 517 willbe used for those adverse events that pertain to the implant procedure of the EASYTRAK lead.- DRAFT -

$D-26 CLINICAL STUDY - CONTAK CDratewas14.5%(95%CI[11.5–17.5%]). TableD-21onpageD-26reportslead-related adverse events observed during the CONTAK CD Study.Table D-21. EASYTRAK lead-related adverse eventsAdverse EventsaTotal % of pts (95% Cl)Lead-Related, N = 448Loss of capture/lead dislodgment 31b6.9 (4.6–9.3)Ventricular oversensing 11 2.5 (1.0–3.9)Extracardiac stimulation 9 2.0 (0.7–3.3)Insulation breach 2 0.4 (0.0–1.1)Procedure-Related, N = 517Transient AV block 6 1.2 (0.2–2.1)Coronary venous dissection 51.0 (0.1–1.8)Coronary venous perforation 51.0 (0.1–1.8)Transient renal failure 51.0 (0.1–1.8)Pericardial effusion 20.4 (0.0–0.9)Finishing wire left in lead 10.2 (0.0–0.6)Right ventricular lead dislodgment 1 0.2 (0.0–0.6)Guide wire fracture 10.2 (0.0–0.6)Hypotension due to blood loss 1 0.2 (0.0–0.6)Total (unique patients) 75 14.5 (11.5–17.5)a. All patients implanted, N = 448; All patients attempted, N = 517.b. Twenty-six events were successfully corrected in a repeat procedure.The most common of the 53 lead-related adverse events (>1% incidence)included the following:• Loss of left ventricular capture (31 patients, 6.9%)• Ventricular oversensing (11 patients, 2.5%)• Extracardiac stimulation (9 patients, 2.0%)These events were typically resolved with surgical intervention.The most common of the 27 procedure-related adverse events (> 1%incidence) included the following:• Coronary venous trauma (10 patients, 2.0%)• Transient atrioventricular block (6 patients, 1.2%)- DRAFT -$

CLINICAL STUDY - CONTAK CD D-27• Transient renal failure (5 patients, 1.0%)These events were typically resolved without intervention and no permanentlong-term sequelae were reported.Severe, Device-Related Adverse Events and Operative Mortality––theincidence of severe, device-related events was reported in 7 of 567 patients(1.2%); this was signiﬁcantly less than the hypothesized rate of 20% (p <0.01) (Table D-22 on page D-27). Table D-23 on page D-27 reports system,device-related, severe adverse events observed during the CONTAK CD Study.Table D-22. Adverse events and operative mortalityMeasurementaN%95%CISevere, Device-Related Adverse Events (Type I)b71.2 (0.3, 2.1)All-Cause Operative Mortality (< = 30 Days Post Implant) 12 2.1 (0.9, 3.3)a. All patients attempted or implanted, N = 567b. Percent is of patients with at least one event.Table D-23. System, device-related, severe adverse eventsAdverse Eventa# of pts % of pts (95% CI)Te l em e t ry d if ﬁculty; device explanted 20.4 (0.0–0.9)Ventricular tachycardia during CPX testing 10.2 (0.0–0.5)Coronary sinus perforation 10.2 (0.0–0.5)Inappropriate shock due to oversensing 1 0.2 (0.0–0.5)Lead dislodgment 1 0.2 (0.0–0.5)Anaphylaxis in association with use of a pulmonary arterycatheter10.2 (0.0–0.5)a. All patients attempted or implanted, N = 567Operative mortality, deﬁnedasdeathfromanycausewithin30daysofimplant,was reported in 12 of 567 patients (2.1%) undergoing the implant procedure.The outcome is signiﬁcantly less than the hypothesized rate of 9% (p < 0.01).Table D-24 on page D-28 reports the cause of death for operative mortality.- DRAFT -

D-28 CLINICAL STUDY - CONTAK CDTable D-24. Cause of death for operative mortalityCause of Death ImplantsN=501AttemptsN=66TotalaN=567Cardiac: pump failure 516Cardiac: arrhythmic 213Noncardiac 2 0 2Unknown 1 0 1Tota l 10 2 12a. All patients attempted or implanted, N = 567.System Safety Proﬁle––analysis of system safety was performed on thecomplication-free rate of device-related adverse events, regardless of whetheror not they were related to the investigational device (Figure D-11 on pageD-29). Table D-25 on page D-28 outlines the device related complications. Thisstudy used an acceptance criterion such that the lower boundary of the 95%conﬁdence interval could not be less than 70%.Table D-25. Device-related complicationsComplicationa#ofpts %ofptsAll patients implanted (N = 448)Loss of LV capture 31 6.9Loss of right atrial capture 71.6Ventricular oversensing 6 1.3Extracardiac stimulation 51.1All patients attempted or implanted (N = 517)Infections 71.4a. This table represents patients attempted or implanted with the EASYTRAK lead; most common (> 1%) device-relatedcomplications reported.- DRAFT -

CLINICAL STUDY - CONTAK CD D-2990 80 70 85 0 Complication-free Rate (%) 75 Acceptance Boundary All Patients (NYHA Class II / III / IV) NYHA Class III / IV Figure D-11. System safetyVeriﬁcation of CRT DeliveryThe delivery of biventricular pacing throughout the CONTAK CD Study wasconﬁrmed by comparing the programmed device output to the biventricularpacing threshold and demonstrating that capture was maintained in dailyactivities and during exercise.The investigational plan recommended programming the device output to atleast twice the biventricular pacing voltage threshold. Electrocardiograms(ECGs) from Holter Monitors during daily activities were received and analyzedto verify that total capture was maintained at the 3-month and 6-month visitsand to ensure that the safety margin was adequate. Cardiopulmonary exercisetests (CPX) were performed on patients who were randomized to receive CRTtherapy at 3- and 6- month visits.• In 623 evaluations of safety margin at baseline, three-, and six-months, thedevice output was programmed to deliver a voltage approximately threetimes that necessary to stimulate both ventricles.• A total of 1139 Holter monitors were placed throughout the study atbaseline, three-, and six-months. The tests indicated only 4 instances(0.4%) of inappropriate pacing or sensing that were all corrected withdevice programming.• A total of 316 CPX tests at the three- and six-month follow-up visits wereperformed in patients with CRT who also had interpretable ECG results. Ofthese, 277 (88%) had continuous CRT delivery throughout exercise. Theremaining 39 patients (12%) had continuous CRT delivery until the sinusrate exceeded the maximum tracking rate (MTR).- DRAFT -

D-30 CLINICAL STUDY - CONTAK CDFOCUSED CONFIRMATORY STUDYStudy DesignThe Focused Conﬁrmatory Study (FCS) was a prospective, multicenter studyconducted in the United States in 127 patients who participated in an exerciseperformance study. The purpose of the FCS was to conﬁrm effectivenessresults related to functional capacity measures, speciﬁcally the Peak VO2and6-Minute Hall Walk, previously reported in the NYHA Class III/IV subgroup ofthe CONTAK CD Study.CRT was provided in the same manner for the FCS as for the CONTAK CDStudy. The EASYTRAK lead, along with market approved right atrial and rightventricular leads were used to provide biventricular stimulation.Demographic DataThe patients in the FCS had the same heart failure indications as the patientsin the NYHA Class III/IV subgroup of the CONTAK CD Study; i.e., patientinclusion criteria included NYHA Class III or IV while on drug therapy, QRSduration ≥120 ms, and Left Ventricular Ejection Fraction (LVEF) ≤35%.A baseline physical assessment and functional measures were performed priorto CRT system implant. Patients were eligible for participation in the study ifthey were capable of walking between 150 and 425 meters. In addition toa Six-Minute Walk test, other special tests were performed prior to implantconsisting of a symptom-limited treadmill test and completion of the MinnesotaLiving with Heart Failure Questionnaire®to assess Quality of Life. CRT therapywas enabled immediately upon device implant. Patients were followed at oneweek, one month, three months, six months and every three months thereafterfor a routine physical assessment and device evaluation. Special testing asdeﬁned above was repeated at three months and six months post-implant.Prior to study entry, patients were stable on optimal heart failure medications(ACE inhibitors or substitute > 1 month and beta blockers > 3 months). Patientswere excluded if they were indicated for either a pacemaker or ICD or if theywere hospitalized for heart failure in the month prior to enrollment.The patient characteristics at study entry are summarized in Table D-26 onpage D-31.- DRAFT -

$CLINICAL STUDY - CONTAK CD D-31Table D-26. Pre-implant characteristics of study patientsCharacteristics All Patients Receiving CRTAge (years) 61 ± 12Male Gender (%) 69NYHA Class III (%) 94Ischemic Etiology (%) 49Resting heart rate (bpm) 73 ± 12QRS width (ms) 159 ± 27LBBB/NSIVCD (%) 91Heart failure medications (%)ACE inhibitor or ARB 91Beta blockers 77Digoxin 76Diuretics 98Inclusion CriteriaInclusion criteria included:• Moderate or severe heart failure, deﬁned as symptomatic heart failurefor at least six months with NYHA Class III or IV symptoms at the timeof enrollment, AND at least one of the following events in the previous12 months:– Hospitalization for heart failure management– Outpatient visit in which intravenous (IV) inotropes or vasoactiveinfusion were administered continuously for at least 4 hours– Emergency room visit of at least twelve hours duration in which IVheart failure medications were administered (including diuretics)•QRS≥120 ms and PR interval > 150 ms from any two leads of a 12-leadECG• Left ventricular ejection fraction ≤35%• Left ventricular end diastolic dimension ≥60 mm (required only if LVEFmeasured by echo)- DRAFT -$

D-32 CLINICAL STUDY - CONTAK CD•Age≥18 years• Optimal pharmacologic therapy for heart failure• Abletowalkbetween150and425minaSix-MinuteWalktestMajor Differences Between CONTAK CD and Focused ConﬁrmatoryStudy PatientsThe CONTAK RENEWAL 3, CONTAK RENEWAL, and CONTAK CD devicesprovide the same cardiac resynchronization therapy (biventricular pacing) andhave the same Indications for Use. Therefore, the CONTAK CD clinical trialdata used to support CONTAK CD is also applicable to CONTAK RENEWALand CONTAK RENEWAL 3. The primary difference between CONTAK CDdevices and CONTAK RENEWAL and CONTAK RENEWAL 3 devices is thatCONTAK CD utilizes an electrically common RV and LV sensing/pacing circuitwhereas CONTAK RENEWAL and CONTAK RENEWAL 3 incorporate anindependent RV and LV sensing/pacing circuit. Additional clinical analysiswas also conducted with CONTAK RENEWAL to provide conﬁrmation that theindependent sensing and pacing capability did not adversely affect the abilityof the device to detect ventricular tachyarrhythmias or provide continuousbiventricular pacing therapy.Some of the major differences between the study populations included:• Patients were excluded from the FCS if they were indicated for eithera pacemaker or implantable cardioverter deﬁbrillator (ICD). Patients inthe CONTAK CD Study were excluded if they met the indications for apacemaker; however, they were required to meet the general indicationsfor an ICD.• Patients were excluded from the FCS if they were hospitalized for heartfailure in the month prior to enrollment; whereas, there was no exclusionfor hospitalization for heart failure in the month prior to enrollment for theCONTAK CD patients.• Patients in the FCS must have been on stable, optimal heart failuremedications, including beta blocker therapy for three months, prior to studyentry. Patients in the CONTAK CD Study could be optimized on drugtherapy between the time from device implant until the treatment phase(either CRT or No CRT) began.- DRAFT -

CLINICAL STUDY - CONTAK CD D-33• Patients in the FCS had baseline measurements performed prior to implant.Patients in the CONTAK CD Study had baseline measurements performedpost-implant, but before programming of the randomized therapy.• Seventy-seven percent of patients in the FCS (98 of N = 127) were on betablockers compared to 42% in the CONTAK CD Study (95 of N = 227).• Forty-nine percent of patients in the FCS (62 of N = 127) had ischemicetiology compared to 68% in the CONTAK CD Study (154 of N = 227).EndpointsThe primary endpoints of the study were Peak VO2and Six-Minute Walkdistance. The study was designed to show a mean change of at least1ml/kg/min and a 95% lower conﬁdence bound (LCB) at least 0.5 ml/kg/min.The study was also designed to detect a statistically signiﬁcant improvementin the Six-Minute Walk distance at a one-sided signiﬁcance level of 0.10.Additionally, two ancillary analyses of Quality of Life Score and NYHA Classhad to demonstrate a change that was directionally favorable towards CRTusing descriptive statistics.Study ResultsStudy results for the Focused Conﬁrmatory Study include the following:•PeakVO2––a statistically signiﬁcant improvement from baseline of 0.94 ±0.30 ml/kg/min with a 95% LCB of 0.45 was observed in Peak VO2after sixmonths of CRT• Six-Minute Walk––statistically signiﬁcant improvements versus baselinewere observed in Six-Minute Walk distance after six months of CRT with anobserved mean improvement of 50.9 ± 10.4 m with a 95% LCB of 37.6 m• Quality of Life––consistent with the other analyses, a statistically signiﬁcantimprovement of 23.9 ± 2.6 points was observed in the Quality of Life scoreafter six months of CRT with a 95% LCB of 19.7 points• New York Heart Association Class––after six months of CRT, a statisticallysigniﬁcant improvement in NYHA Class was observed with 60.4% ofpatients improving one or more NYHA Class- DRAFT -

D-34 CLINICAL STUDY - CONTAK CD- DRAFT -

E-1CLINICAL STUDY - CONTAK RENEWALAPPENDIX ECLINICAL STUDY POPULATIONSGuidant CRT-Ds, when compared to OPT alone, have been demonstrated withreasonable assurance, to be safe and effective in signiﬁcantly reducing: therisk of a composite of all-cause mortality or ﬁrst hospitalization by 20%, therisk of all-cause mortality by 36%, and heart failure symptoms in patients whohave moderate to severe heart failure (NYHA III/IV) including left ventriculardysfunction (EF ≤35%) and QRS duration ≥120 ms and remain symptomaticdespite stable, optimal heart failure drug therapy, based on the Guidantsponsored COMPANION clinical study. (Guidant devices were the only devicesstudied in the COMPANION clinical trial.)SUMMARYGuidant conducted the CONTAK RENEWAL Study, which demonstrated thedevice’s ability to appropriately detect ventricular tachyarrhythmias with anindependent sensing conﬁguration. Finally, the CONTAK RENEWAL HolterStudy was conducted to provide conﬁrmation of the device’s ability to providecontinuous biventricular pacing on both a daily basis and during exercise.STUDY DESIGNThe CONTAK RENEWAL Study was a prospective, multi-center,non-randomized evaluation conducted in Europe and enrolled a total of 45patients. The purpose of the study was to verify that the CONTAK RENEWALdevice performs according to speciﬁcation.INCLUSION/EXCLUSION CRITERIAPatients who were enrolled in the study were required to meet the followinginclusion criteria:• Symptomatic heart failure• Left ventricular dysfunction•WideQRS• At risk for sudden cardiac death• 18 years or of legal age in order to give informed consent according tonational laws- DRAFT -

E-2 CLINICAL STUDY - CONTAK RENEWAL• Able to understand the nature of the procedure• Available for follow-up on a regular basis at an approved investigationalcenterPatients were excluded from the investigation if they met any of the followingcriteria:• Life expectancy of less than six months due to other medical conditions• For women: Pregnancy or absence of medically accepted birth control• Inability or refusal to sign the Patient Informed Consent• Inability or refusal to comply with the follow up schedule or protocolrequirements• Mechanical tricuspid prosthesis• Currently enrolled in another investigational study, including druginvestigations• Hypertrophic Obstructive Cardiomyopathy• Are unable to undergo device implant, including general anesthesia ifrequired• Have pre-existing leads other than those speciﬁed in the investigationalplan (unless the investigator intended to replace them with the permittedleads)DEMOGRAPHIC DATAThe patient characteristics at study entry are summarized in Table E-1 on pageE-2.Table E-1. Pre-implant characteristics of study patientCharacteristicsaPatient DataN patients implanted 44Gender Male (91%), Female (9%)Age (years) 65 ± 9NYHA II (14%), III (77%), IV (9%)LVEF (%) 22 ± 6BBB LBBB/NSIVCD (86%),RBBB (14%)Etiology Ischemic (56%), Non-ischemic (44%)QRS Width 172 ± 24 ms- DRAFT -

CLINICAL STUDY - CONTAK RENEWAL E-3Table E-1. Pre-implant characteristics of study patient (continued)CharacteristicsaPatient DataPR Interval 211 ± 49 msResting HR 70 ± 12 bpma. Continuous measures are reported as means ± standard deviations.VENTRICULAR TACHYARRHYTHMIA DETECTION TIMEThe CONTAK RENEWAL device has independent Left Ventricular and RightVentricular Sensing. Ventricular tachyarrhythmia detection time was analyzedto determine if the sensing conﬁguration had any effect on sensing VT/VF.Based on previous clinical studies of the VENTAK AV family, upon which theICD function of CONTAK CD and CONTAK RENEWAL are built, Guidant’sICDs typically have a VF detection time of approximately two seconds. The VFdetection time of 2.4 ± 0.5 seconds in the RENEWAL study was statisticallylower than 6 seconds (p < 0.01), demonstrating that there was no statisticallysigniﬁcant prolongation of induced VF detection times with the independentsensing conﬁguration.1There were no adverse events reported in whicha CONTAK RENEWAL device failed to detect a spontaneous ventriculartachyarrhythmia.HOLTER STUDY - CONTAK RENEWALStudy DesignThe CONTAK RENEWAL Holter Study was a prospective, multi-center,non-randomized evaluation conducted in Europe, in which 46 patientscompleted testing. The purpose of the study was to demonstrate continuousappropriate biventricular (BiV) pacing over a 24 hour period and duringexercise using Holter monitor recordings. All patients had been implanted witha CONTAK RENEWAL for a minimum of one month at the time of the studyinitiation.1. Detection time at implant with legally marketed Guidant ICD devices is typically two seconds,and investigators have stated than an additional delay of 3 to 5 seconds would be a clinicallysigniﬁcant event. The expected detection time is 2 seconds (95% CI: [0, 6 sec]).- DRAFT -

E-4 CLINICAL STUDY - CONTAK RENEWALInclusion/Exclusion CriteriaPatients who were enrolled in the study were required to meet the followinginclusion criteria:• Availability for 24 hours follow-up at an approved study center• Willingness and ability to participate in all testing associated with this study• Age 18 or above, or of legal age to give informed consent as speciﬁedby national law• Implanted with the CONTAK RENEWAL system for at least 1 month• Stable when programmed according to labeled recommendations forcontinuous BV pacing• Sinus rhythm at follow-up• Active atrial lead implantedPatients were excluded from the investigation if they met any of the followingcriteria:• Life expectancy of less than six months due to other medical conditions• Concurrent participation in any other clinical study, including drug study•Inatrialﬁbrillation at follow-up• Inability or refusal to sign the Patient Informed Consent• Inability or refusal to comply with the follow-up schedule• Known pregnancyDemographic DataThe patient characteristics at study entry are summarized in Table E-2 on pageE-4.Table E-2. Pre-implant characteristics of study patientsCharacteristics Patient DataN patients 46Gender Male: 40 (87%), Female: 6 (13%)Age (years) 60.9 ± 9.0NYHA at implant [N (%)] I0(0%)II 5(10.9%)III 34 (73.9%)- DRAFT -

CLINICAL STUDY - CONTAK RENEWAL E-5Table E-2. Pre-implant characteristics of study patients (continued)Characteristics Patient DataIV 7(15.2)%NYHA current [N (%)] I9(19.6%)II 25 (54.3%)III 11 (23.9%)IV 1 (2.2%)Duration implanted (months) Mean ± SD 8.3 ± 4.1Range 1.5 – 15.0Median 9.0Programming ParametersRefer to the Pacing Therapies chapter for information about programming tomaintain CRT. Programming recommendations in this study were consistentwith the recommendations in that chapter.EndpointsThe study had the following primary endpoints:• Continuous appropriate BiV pacing during activities of daily living• Continuous appropriate BiV pacing during exerciseThe mean percentage of sinus beats appropriately BiV paced was measured bya Holter monitor over a 24 hour period and during exercise. Exercise intensitywas measured using the Borg rating of perceived exertion (RPE) 6-20 scale.PatientswereaskedtoexercisetoaBorglevelof15(difﬁcult). The exerciseprotocol used was left to the discretion of the physician based on the patients’functional status. The type of exercise performed, duration and intensity ofexercise testing is listed in Table E-3 on page E-5 and Table E-4 on page E-6.Table E-3. Type of exercise testing performedExercise Performed Number of PatientsBicycle Ergometry 24 (52.2%)Hall Walk 8(17.4%)- DRAFT -

E-6 CLINICAL STUDY - CONTAK RENEWALTable E-3. Type of exercise testing performed (continued)Exercise Performed Number of PatientsStair Climbing 14 (30.4%)Tota l 46Table E-4. Duration and intensity of exercise testingResults (N = 46)Borg RPE Rating Obtained Mean ± SD 15 ± 1Median 15Range 7 – 18Duration of Exercise (minutes) Mean ± SD 6.6 ± 3.3Median 6.0Range 1 – 17Maximum HR Obtained (bpm) Mean ± SD 103 ± 20Median 105Range 60 – 156Study ResultsPacing during activities of daily livingThe mean percentage of appropriately continuously paced beats during dailyliving was calculated as 99.6 ± 1.3% with a median of 100% and is summarizedin Table E-5 on page E-6. Continuous appropriate BiV pacing is deﬁned aspacing provided between the lower rate limit and the MTR, excluding PVCs.Table E-5. Activities of daily living: continuous appropriate BiV pacingStatistic P-valueaMean ± SD 99.6 ± 1.3 –Range 91.4 – 100 –Medianb100 <0.01a. The p-value is based on the sign-rank test.b. Due to the non-normality of the data a non-parametric test of the median was performed comparing the median to 90%.- DRAFT -

CLINICAL STUDY - CONTAK RENEWAL E-7Pacing During ExerciseThe mean percentage of appropriately continuously paced beats duringexercise was calculated as 98.3 ± 5.6% with a median of 100% and issummarized in Table E-6 on page E-7. Continuous appropriate BiV pacingis deﬁned as pacing provided between the lower rate limit and the MTR,excluding PVCs.Table E-6. Exercise: continuous appropriate BiV pacingStatistic P-valueMean ± SD 98.3 ± 5.6 –Range 68.1 – 100 –Median 100 <0.01Device CountersFinally, during the study CONTAK RENEWAL device counters were foundto correlate highly to the data collected on the independent Holter monitors(Table E-7 on page E-7).Table E-7. Correlation between holter and deviceMean ± SD Correlation (P-value)Holter 97,536 ± 13,307 0.97 (<0.01)Device 100,143 ± 13,373 –- DRAFT -

E-8 CLINICAL STUDY - CONTAK RENEWAL- DRAFT -

F-1CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATIONALGORITHM VALIDATION STUDYAPPENDIX FCLINICAL STUDY DESIGNNOTE: The SmartDelay optimization feature was previously known as ExpertEase for Heart Failure (EEHF+).This clinical investigation was a 50 patient, multi-center, acute hemodynamicstudy at 5 centers in the United States to validate the performance of ExpertEase for Heart Failure AV delay optimization algorithm (EEHF+1).The main purposes of this clinical investigation were: (A) to test prospectivelythe effectiveness of EEHF+ in optimizing LV dP/dtmax (maximum rate of LVpressure change) for biventricular (BV) CRT in atrial sensing and pacingmodes; and (B) to evaluate and compare LV dP/dtmax and stroke volume asmeasured by AoVTI2at AV delays determined by a CRT optimization method(EEHF+, Echo) and also by a series of population ﬁxed values, for BV CRTin atrial sensing and pacing modes.The study consisted of three phases in the following order: an acute test(during implant), a device implant, and an echocardiography study. In theacute test the LV pressure data was collected invasively at various stimulationmode/site/AV delay combinations. After the implant of a CRT/CRT-D deviceusing standard procedures, a standard non-invasive Doppler echo procedurewas performed, in which Doppler ﬂow-velocity proﬁles from aortic, mitral,and pulmonary valves were collected during BV CRT at various stimulationmode/AV delay combinations.Inclusion/Exclusion CriteriaPatients enrolled in this study were required to meet the criteria for aCRT/CRT-D device implant at the time of implant (Sept 2003 - Oct 2004).Patients were excluded from the study if they met any of the following criteria:1. From this point on, Expert Ease for Heart Failure AV delay optimization algorithm is referredas EEHF+.2. Otto CM, Pearlman AS, Comess K, Reamer R, Janko C, Huntsman L. Determination ofthe stenotic aortic valve area in adults using Doppler echocardiography. J Am Coll Cardiol1986;7:509-17.- DRAFT -

F-2 CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDY• Patients in AF that can not be cardioverted for the study• Sustained, uncontrolled ventricular tachycardia• Frequent ectopic activity that makes stable hemodynamic measurementsinfeasible• Sinus rhythm < 30 bpm or > 100 bpm• Complete AV node block• Acute severe heart failure exacerbation• Severeaorticvalvularstenosis(valvearea<1.0squarecm)• Hypertrophic obstructive cardiomyopathy• CABG within 2 weeks• Congenital heart disease• Pregnancy• Patient involved in other clinical investigations of active therapy or treatment• Patient at unacceptably high risk for catheterization (a patient who wouldnot medically be indicated for an EP study or diagnostic catheterization)STUDY RESULTSPatient AccountabilityFifty patients were enrolled in the study. Forty-one patients had valid acutehemodynamic tests completed and thirty-eight patients had valid echo testscompleted. Among the 9 patients with invalid acute hemodynamic tests, 7were attempts, and 2 completed the acute test but with invalid results (one ofthem had an unstable atrial rate and the other had 2:1 AV conduction). A validecho was deﬁned as a subject who had a valid acute hemodynamic test andalso completed the echo test.Patient CharacteristicsThe subject demographics are shown below (Table F-1 on page F-3).- DRAFT -

CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDY F-3Table F-1. Subject DemographicsCharacteristic Measurement ResultAge at Implant Number of subjects 50Mean ± SD 68.1 ± 10.5Range [47.0, 85.0]Gender [N (%)] Female 12 (24.0)Male 38 (76.0)NYHA Class [N (%)] II 1(2.0)III 49 (98.0)LVEF Number of subjects 50Mean ± SD 26.6 ± 6.6Range [5.0, 35.0]Conduction Disorder LBBB 38 (86.4)RBBB 12 (27.3)LV DP/dtmax ResultsCorrelation between the LV dP/dtmax at the EEHF+ recommended AVdelay and maximum achievable LV dP/dtmax• For the regression analysis, the 95% conﬁdence intervals of the regressionslope were [0.98, 1.07] and [0.94, 1.10] for atrial sensing and pacing. Thecorresponding intercept values were [-2.07, -0.86] and [-3.73, -0.76] foratrial sensing and pacing (Figure F-1 on page F-4). The ability of EEHF+ tosuggest an AV delay that maximizes %LV dP/dtmax for both atrial sensingand atrial pacing is demonstrated in the regression plots. For patientswith a near-zero maximum improvement in %LV dP/ dtmax from baseline,the %LV dP/dtmax at the AV delay estimated by EEHF+ was close to themaximum achievable %LV dP/dtmax as indicated by the small intercept(-1.47 for atrial sensing and -2.25 for atrial pacing).- DRAFT -

F-4 CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDYCorrelation between maximum achievable %LV dP/dtmax and the %LV dP/dtmax achieved with theEEHF+ delay for both atrial sensing (left, n=38) and atrial pacing (right, n=36).Figure F-1. Correlation between achievable and achievedComparison of EEHF+ recommended AV delay to ﬁxed AV delays of 100ms, 120 ms, 140 ms or 160 ms in achieving LV dP/dtmax• Differences between the %LV dP/dtmax achieved with EEHF+ and the%LV dP/dtmax achieved with ﬁxed AV delays were plotted for each ﬁxedAV delay (Figure F-2 on page F-5). A negative value indicated that the%LV dP/dtmax achieved with EEHF+ was higher. See the tables below forfurther details about the number of subjects, mean, standard deviation,p-value, and conﬁdence interval for each comparison (Table F-2 on pageF-5, Table F-3 on page F-6).- DRAFT -

CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDY F-5Differences between %LV dP/dtmax achieved with EEHF+ and with ﬁxed AV delays of 100 ms, 120ms, 140 ms or 160 ms for atrial sensing (left) and atrial pacing (right). A negative value indicatesthat the EEHF+ algorithm is better. The box represents the mean and error bars represent 95%CI of mean.Figure F-2. Differences, achieved with EEHF+ and ﬁxed AV delaysTable F-2. Differences between maximal achievable %LV dP/dt max and that achieved using EEHF+ and aﬁxed AV delay of 100 ms, 120 ms, 140 ms and 160 ms, during atrial sensingn, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ 100 ms Paired difference P-value38, -1.2 ± 1.3, (-1.6, -0.8) 38, -2.7 ± 2.9, (-3.6, -1.8) 38, -1.4 ± 2.8, (-2.3, -0.6) 0.0025EEHF+ 120 ms Paired difference P-value38, -1.2 ± 1.3, (-1.6, -0.8) 38, -2.2 ± 2.2, (-2.9, -1.5) 38, -1.0 ± 2.3, (-1.7, -0.2) 0.0130EEHF+ 140 ms Paired difference P-value36, -1.3 ± 1.3, (-1.7, -0.8) 36, -2.1 ± 2.0, (-2.8, -1.5) 36, -0.9 ± 2.3, (-1.6, -0.1) 0.0279- DRAFT -

F-6 CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDYTable F-2. Differences between maximal achievable %LV dP/dt max and that achieved using EEHF+ and aﬁxed AV delay of 100 ms, 120 ms, 140 ms and 160 ms, during atrial sensing (continued)n, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ 160 ms Paired difference P-value33, -1.1 ± 1.3, (-1.6, -0.7) 33, -3.6 ± 3.8, (-4.9, -2.3) 33, -2.5 ± 4.0, (-3.8, -1.1) 0.0013Table F-3. Differences between maximal achievable %LV dP/dt max and that achieved using EEHF+ and aﬁxed AV delay of 100 ms, 120 ms, 140 ms and 160 ms, during atrial pacingn, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ 100 ms Paired difference P-value36, -2.0 ± 2.6, (-2.8, -1.1) 36, -8.7 ± 5.1, (-10.3, -7.0) 36, -6.7 ± 4.8, (-8.3, -5.2) < 0.0001EEHF+ 120 ms Paired difference P-value36, -2.0 ± 2.6, (-2.8, -1.1) 36, -6.5 ± 4.7, (-8.0, -4.9) 36, -4.5 ± 4.2, (-5.9, -3.2) < 0.0001EEHF+ 140 ms Paired difference P-value36, -2.0 ± 2.6, (-2.8, -1.1) 36, -4.7 ± 4.2, (-6.1, -3.3) 36, -2.7 ± 3.5, (-3.9, -1.6) < 0.0001EEHF+ 160 ms Paired difference P-value36, -2.0 ± 2.6, (-2.8, -1.1) 36, -3.4 ± 3.6, (-4.5, -2.2) 36, -1.4 ± 2.8, (-2.3, -0.5) 0.0049Comparison of EEHF+ recommended AV delay to the echo-based Ritterand AoVTI methods in achieving LV dP/dtmax• Differences between the %LV dP/dtmax achieved with EEHF+ and the %LVdP/dtmax achieved with the Ritter and AoVTI echo methods were plotted forthe two echo methods (Figure F-3 on page F-7). A negative value indicatedthat the LV dP/dtmax achieved with EEHF+ was higher. See the tablesbelow for further details about the number of subjects, mean, standarddeviation, p-value, and conﬁdence interval for each comparison (Table F-4on page F-7, Table F-5 on page F-7).- DRAFT -

CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDY F-7Differences between %LV dP/dtmax achieved with EEHF+ and with two echo-based methods: theRitter method and the AoVTI method for atrial sensing (left) and atrial pacing (right). A negativevalue indicates that EEHF+ was better. The box represents the mean and error bars represent95% CI of mean.Figure F-3. Differences, achieved with EEHF+ and echo-based methodsTable F-4. Differences between maximal achievable %LV dP/dt max and that achieved from the EEHF+, theRitter method, and the AoVTI method during atrial sensingn, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ Ritter method Paired difference P-value35, -1.3 ± 1.3, (-1.7, -0.8) 35, -2.5 ± 2.7, (-3.3, -1.6) 35, -1.2 ± 3.0, (-2.1, -0.2) 0.0259EEHF+ AoVTI method Paired difference P-value33, -1.3 ± 1.3, (-1.8, -0.8) 33, -1.7 ± 1.9, (-2.3, -1.0) 33, -0.4 ± 1.6, (-0.9, 0.2) 0.2036Table F-5. Differences between maximal achievable %LV dP/dt max and that achieved from the EEHF+, theRitter method, and the AoVTI method during atrial pacingn, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ Ritter method Paired difference P-value33, -2.0 ± 2.7, (-2.9, -1.1) 33, -7.4 ± 5.5, (-9.3, -5.5) 33, -5.4 ± 5.1, (-7.2, -3.7) < 0.0001- DRAFT -

F-8 CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDYTable F-5. Differences between maximal achievable %LV dP/dt max and that achieved from the EEHF+, theRitter method, and the AoVTI method during atrial pacing (continued)n, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ AoVTI method Paired difference P-value34, -2.0 ± 2.6, (-2.8, -1.1) 34, -2.8 ± 2.9, (-3.8, -1.8) 34, -0.9 ± 2.6, (-1.8, 0.0) 0.0617AoVTI ResultsThere was a large variance in the difference in %AoVTImax achieved by all themethods evaluated in this study, which is consistent with the inherent variabilityof the AoVTI measurements3.Comparison of EEHF+ recommended AV delay to ﬁxed AV delays of 100ms, 120 ms, 140 ms, and 160 ms in achieving AoVTImax• As shown in the tables below, there was a large variance in the differencein %AoVTImax achieved by EEHF+ and the ﬁxed AV delays for both atrialsensing and pacing; the tables also provide further details about thenumber of subjects, mean, standard deviation, p-value, and conﬁdenceinterval(TableF-6onpageF-8,TableF-7onpageF-9).Table F-6. Differences between maximal achievable %AoVTI and that achieved using the EEHF+ and a ﬁxedAV delay of 100 ms, 120 ms, 140 ms, and 160 ms during atrial sensing.n, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ 100 ms Paired difference P-value36, -8.5 ± 8.3, (-11.2, -5.8) 36, -6.6 ± 5.7, (-8.4, -4.7) 36, 1.9 ± 9.5, (-1.2, 5.0) 0.2323EEHF+ 120 ms Paired difference P-value35, -7.7 ± 7.0, (-10.0, -5.4) 35, -5.1 ± 6.9, (-7.3, -2.8) 35, 2.7 ± 8.9, (-0.3, 5.6) 0.0837EEHF+ 140 ms Paired difference P-value36, -8.5 ± 8.3, (-11.2, -5.8) 36, -6.1 ± 4.4, (-7.5, -4.6) 36, 2.4 ± 9.3, (-0.6, 5.5) 0.12963. Otto CM, Pearlman AS, Comess K, Reamer R, Janko C, Huntsman L. Determination ofthe stenotic aortic valve area in adults using Doppler echocardiography. J Am Coll Cardiol1986;7:509-17.- DRAFT -

CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDY F-9Table F-6. Differences between maximal achievable %AoVTI and that achieved using the EEHF+ and a ﬁxedAV delay of 100 ms, 120 ms, 140 ms, and 160 ms during atrial sensing. (continued)n, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ 160 ms Paired difference P-value36, -8.5 ± 8.3, (-11.2, -5.8) 36, -7.3 ± 6.4, (-9.4, -5.2) 36, 1.2 ± 9.8, (-2.0, 4.4) 0.4769Table F-7. Differences between maximal achievable %AoVTI and that achieved using the EEHF+ and a ﬁxedAV delay of 100 ms, 120 ms, 140 ms, and 160 ms during atrial pacingn, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ 100 ms Paired difference P-value35, -6.6 ± 5.2, (-8.3, -4.8) 35, -11.9 ± 6.4, (-14.0,-9.8)35, -5.4 ± 9.3, (-8.4, -2.3) 0.0017EEHF+ 120 ms Paired difference P-value34, -6.4 ± 5.3, (-8.2, -4.7) 34, -9.3 ± 7.1, (-11.7, -6.9) 34, -2.8 ± 9.9, (-6.2, 0.5) 0.1046EEHF+ 140 ms Paired difference P-value35, -6.6 ± 5.2, (-8.3, -4.8) 35, -7.9 ± 5.3, (-9.7, -6.2) 35, -1.4 ± 7.7, (-3.9, 1.2) 0.2977EEHF+ 160 ms Paired difference P-value35, -6.6 ± 5.2, (-8.3, -4.8) 35, -7.1 ± 5.4, (-8.9, -5.3) 35, -0.5 ± 7.7, (-3.1, 2.0) 0.6896Comparison of EEHF+ recommended AV delay to echo-based Rittermethod in achieving AoVTImax• As shown in the tables below, there was a large variance in the differencein %AoVTI obtained with EEHF+ and that obtained with Ritter method inatrial sensing mode and atrial pacing mode; the tables also provide furtherdetails about the number of subjects, mean, standard deviation, p-value,and conﬁdence interval (Table F-8 on page F-10, Table F-9 on page F-10).- DRAFT -

F-10 CLINICAL STUDY - SUMMARY OF CRT OPTIMIZATION ALGORITHM VALIDATION STUDYTable F-8. Differences between maximum achievable %AoVTI and that achieved with EEHF+ and Ritterduring atrial sensingn, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ Ritter method Paired difference P-value36, -8.5 ± 8.3, (-11.2, -5.8) 36, -6.5 ± 5.6, (-8.3, -4.6) 36, 2.0 ± 9.1, (-0.9, -5.0) 0.1895Table F-9. Differences between maximum achievable %AoVTI and that achieved with EEHF+ and Ritterduring atrial pacingn, mean ± std, 95% CI n, mean ± std, 95% CI n, mean ± std, 95% CI Paired t-testEEHF+ Ritter method Paired difference P-value34, -6.8 ± 5.2, (-8.5, -5.0) 34, -10.4 ± 6.6, (-12.6,-8.1)34, -3.6 ± 9.0, (-6.7, -0.6) 0.0255CONCLUSIONSThe results of the CRTAVO study are summarized as follows:• The algorithm recommended AV delays that maximized global contractilefunction as measured by LV dP/dt.• The EEHF+ algorithm recommended an AV delay that increased acutehemodynamic responses in terms of %LV dP/dtmax, as compared to ﬁxedAV delays of 100 ms, 120 ms, 140 ms or 160 ms.• The EEHF+ algorithm recommended an AV delay that increased acutehemodynamic responses in terms of %LV dP/dtmax, as compared to AVdelay recommended by Ritter method.- DRAFT -

$G-1CLINICAL STUDY - VITALITYAPPENDIX GCLINICAL STUDY POPULATIONSGuidant CRT-Ds, when compared to OPT alone, have been demonstrated withreasonable assurance, to be safe and effective in signiﬁcantly reducing: therisk of a composite of all-cause mortality or ﬁrst hospitalization by 20%, therisk of all-cause mortality by 36%, and heart failure symptoms in patients whohave moderate to severe heart failure (NYHA III/IV) including left ventriculardysfunction (EF ≤35%) and QRS duration ≥120 ms and remain symptomaticdespite stable, optimal heart failure drug therapy, based on the Guidantsponsored COMPANION clinical study. (Guidant devices were the only devicesstudied in the COMPANION clinical trial.)CHRONIC IMPLANT STUDY - VITALITYThe purpose of this study was to evaluate the safety and effectiveness ofGuidant VITALITY family devices with Automatic Intrinsic Rhythm ID. Thisclinical study was a single-arm, prospective, multi-center study. There werea total of 100 patients enrolled at 21 US investigational centers betweenDecember 3, 2002 and January 10, 2003.Patient PopulationOne hundred patients were enrolled in this study and 96 patients receivedinvestigational devices. The mean age of the patients implanted with theVITALITY device was 67.3 ± 10.8 years old. The mean left ventricular ejectionfraction was 30.4% (range 11.0% - 71.0%). Seventy-eight (78) patients (81.3%)were male. The primary cardiovascular disease (42.1%) was coronary arterydisease (CAD) and the primary tachyarrhythmia (38.5%) was monomorphicventricular tachycardia (MVT).MethodsA prospective, multi-center, nonrandomized clinical study evaluated the safetyand effectiveness of the VITALITY device in humans. Ninety-six patientsselected from the investigator’s general patient population who met theindications for use of the VITALITY device were followed through pre-discharge,2-week and 1-month follow-ups and continued every 3 months thereafter untilstudy closure.- DRAFT -$

G-2 CLINICAL STUDY - VITALITYResultsA total of 100 patients were enrolled in this study. Of these, 96 patients weresuccessfully implanted, with 4 intents. Ninety-three (93) patients ﬁnished their1-month follow-up per the study protocol. All primary and secondary endpointsof this study were met. The results from this study provide evidence of thesafety and effectiveness of the VITALITY with Automatic Intrinsic Rhythm IDalgorithm (Table G-1 on page G-2).Table G-1. VITALITY Chronic Study ResultsSafety EndpointsVT/VF Detection Time 3.43 secondsPrimary EndpointsSensitivityInduced VT/VF 100%Spontaneous VT/VF 100%Speciﬁcity––InducedRhythm Physician/Annotation Device Decision–SVT SpeciﬁcityAtrial Fibrillation 71 68 95.8%Atrial Flutter 94 88 93.6%Sinus Tachycardia 7571.4%Total Induced 172 161 93.6%Speciﬁcity–SpontaneousRhythm Physician/Annotation Device Decision–SVT SpeciﬁcityAtrial Fibrillation 65 65 100%Atrial Flutter 31 28 90.3%Sinus Tachycardia 37 32 86.5%Other 77100%Total Spontaneous 140 132 94.3%aCombinedSpeciﬁcityb312 293 93.9%Secondary EndpointsAcute Automatic Rhythm ID Accuracy (2 weeks) 100%- DRAFT -

$CLINICAL STUDY - VITALITY G-3Table G-1. VITALITY Chronic Study Results (continued)Automatic Rhythm ID Accuracy (1 month) 97.7%Manual Rhythm ID Accuracy (1 month) 100%a. GEE adjusted speciﬁcity = 93.7%b. Combined speciﬁcity includes both Induced and Spontaneous data.ACUTE STUDY - VITALITYThe VITALITY ICD was compared to a commercially available ICD (VENTAKPRIZM‚ or VENTAK PRIZM 2 ICD) in an acute (nonimplant) paired study of 50patients enrolled at nine investigating centers between March 8, 2001 and July24, 2001. A total of 47 patients were tested with the study device, followed bya control device at the time of a Guidant commercially approved (VENTAKPRIZM, model 1851 or VENTAK PRIZM 2, model 1861) implantation.The purpose of the acute study was to demonstrate that the addition of an SVTdetection enhancement and brady features did not adversely impact normalICD sensing and detection functionality. A total of 50 patients were testedin nine U.S. centers.Patients studiedThe patients (38 M/9 F) had a mean age of 66 years (range 37 to 90) anda left ventricular ejection fraction of 32% (range 10% to 62%). Most (40%)presented with monomorphic ventricular tachycardia (MVT) and nonsustainedVT as their primary arrhythmia. Of the patients studied, 87 percent presentedwith coronary artery disease or ischemic cardiomyopathy.Methods and statisticsThe acute study was done in the operating room or electrophysiologylaboratory without implantation of the study device. The primary endpoint wasto determine that VT/VF detection time for induced episodes is within twoseconds of the VENTAK PRIZM or VENTAK PRIZM 2 detection time.ResultsA total of 50 patients were enrolled in the acute study. Of those, 47patients were successfully tested with the system per study protocol; therewere two attempted procedures and one intent. There was one clinicalcomplication and two observations reported in the acute study, all of whichwere non-investigational device related. No patient deaths were reported.- DRAFT -$

G-4 CLINICAL STUDY - VITALITYThe VT/VF detection time of the VITALITY ICD was found to be within twoseconds of the VENTAK PRIZM 2 detection time, leading to the conclusion thatactivating the additional VITALITY features does not have a negative effect onthe existing ICD sensing and detection functionality (Table G-2 on page G-4).Table G-2. Acute study resultsStudy Endpoint VITALITY(Mean ± std)NVENTAKPRIZM 2 DR(Mean ± std)NVT/VF Detection Time (seconds) 3.60 ± 0.60N=473.52 ± .057N=47p-value: <0.001- DRAFT -

H-1CLINICAL STUDY - SUMMARY OF GDT1000 SENSING ACUTESTUDYAPPENDIX HCLINICAL STUDY POPULATIONSGDT1000 study included patients indicated for a CRT-D device. Excluded fromthe study were patients meeting any of the following criteria:• Having no intrinsic P and/or R waves at implant• Having a pre-existing unipolar pacemaker that was not to beexplanted/abandoned• Enrolled in a concurrent study that would confound the study results• Having ventricular tachyarrhythmias associated with a reversible cause(e.g., digitalis toxicity, hypoxia, sepsis, transient electrolyte imbalance,acute myocardial infarction, electrocution, or drowning)• Women who were pregnant or planned to become pregnant• Having a prosthetic mechanical tricuspid heart valveSTUDY METHODSThis clinical investigation was a 50 patient, multi-center, acute study conductedat seven (7) centers in the United States. The main purpose of this clinicalinvestigation was to characterize the performance of the new Automatic GainControl (AGC) sensing platform, with the Dynamic Noise Adjustment (DNA)feature, that is used in both COGNIS and TELIGEN devices. The AGCsensing platform was studied using a Guidant Acute Sensing Device (GASD)system, a non-implantable device containing the COGNIS/TELIGEN systemboard, hardware, and ﬁrmware required for sensing intracardiac signals. Thestudy enrolled a total of 50 patients and was conducted in two phases. In theﬁrst phase, 28 of 30 patients completed protocol testing. The algorithm wasmodiﬁed after the ﬁrst phase, and it was re-evaluated in the second phase, inwhich 17 of 20 patients completed protocol testing. Of the ﬁve patients whodid not complete testing in the two phases, three were attempts, and two wereintents.- DRAFT -

H-2 CLINICAL STUDY - SUMMARY OF GDT1000 SENSING ACUTE STUDYProtocol TestingFour scenarios were tested, including different combinations of sensed atrialsignals (AS), paced atrial signals (AP), sensed ventricular signals (VS), andpaced ventricular signals (VP), i.e., AS/VS, AS/VP, AP/VS, and AP/VP. Sensingalgorithm performance was analyzed from patients’ real-time electrograms(EGM) and electronic signals. Primary analysis was performed by visuallyreviewing the EGM and markers of the printed strips for proper sensing, as wellas for instances of undersensing and oversensing.Additional analysis included tabulating the sensed and paced events storedin the patient data ﬁles from the patient CD-ROM. During this tabulation,unexpected events were noted. An example of an unexpected event is asensed event during an AP/VP testing scenario. The sensed event could be areal event, such as a PVC, or an oversensed event. These unexpected eventswere evaluated by viewing the electronic signals stored in the patient data ﬁlesand correlating these signals to the printed strips.Statistical AnalysisThe sensitivity, speciﬁcity, positive predictive value, rate of oversensing, andrate of undersensing of the sensing algorithm were analyzed for each chamber.A true positive (TP) is the number of intrinsic/paced signals appropriatelysensed, a false positive (FP) is the number of intrinsic/paced signals fromthe opposite chamber oversensed, a false negative (FN) is the number ofintrinsic/paced signals undersensed, and a true negative (TN) is the numberof intrinsic/paced signals from the opposite chamber appropriately notsensed. The sensing sensitivity was calculated as TP/(TP+FN), speciﬁcityas TN/(TN+FP), positive predictive value (PPV) as TP/(TP+FP), rate ofoversensing as FP/(TP+FP), and rate of undersensing as FN/(TP+FN).The sensing performance results from the ﬁrst phase of the study are providedand compared to the results from the second phase in order to demonstratethe improvement in the operation of the updated sensing algorithm followingthe between-phase changes. Results from the second phase of the study arethe most clinically relevant, as they reﬂect the performance of the ﬁnal sensingalgorithm implemented in the COGNIS/TELIGEN devices.The GDT1000 protocol did not pre-specify acceptable sensitivities, speciﬁcities,PPV, rates of oversensing, or rates of undersensing for the RA, RV, and LVchannels.- DRAFT -

CLINICAL STUDY - SUMMARY OF GDT1000 SENSING ACUTE STUDY H-3STUDY RESULTSPatient CharacteristicsThe table below shows the characteristics of the patients implanted orattempted (Table H-1 on page H-3).Table H-1. All patients implanted or attempted, Phase 1 and Phase 2Characteristic Measurement Phase 1 Result (N=29) Phase 2 Result (N=19)Age at implant Mean ± SD 65.8 ± 12.2 68.1 ± 9.6Range [44.6, 85.5] [51.3, 81.8]Gender [N (%)] Female 14 (48.0) 14 (74.0)Male 15 (52.0) 5 (26.0)NYHA Class [N (%)] III 27 (93) 19 (100)IV 2(7) 0(0)LVEF (%) Mean ± SD 22.4 ± 7.7 23.5 ± 6.4Range [10.0, 35.0] [15.0, 35.0]QRS Duration Mean ± SD 161± 29 149± 30Range [124, 248] [106, 220]Cardiac Disease [N (%)] NonischemicCardiomyopathy14 (48) 7 (37)Ischemic Cardiomyopathy,CAD10 (34) 9 (47)Hypertension 3(10) 0(0)Coronary Artery Disease(CAD)1(3) 0(0)Ischemic Cardiomyopathy,no CAD1 (3) 3 (16)Valvular Heart Disease 0(0) 1(5)Other 0 (0) 1 (5)Lead PositionIn this study, the position of each lead was per physician’s discretion. A majorityof the atrial leads in the ﬁrst/second phase of the study were placed in theright atrial appendage (19/12) with the remaining placed in the lateral wall(5/2), septal wall (2/3), and unspeciﬁed location (1/0). A majority of the rightventricular leads were implanted in the right ventricular apex, with the remaining- DRAFT -

H-4 CLINICAL STUDY - SUMMARY OF GDT1000 SENSING ACUTE STUDYplaced in the septal wall (0/1) and unspeciﬁed location (1/0). A majority ofthe left ventricular leads were implanted in the lateral, postero-lateral, orposterior wall (21/15), with the remaining placed in an antero-lateral, anterior,or postero-septal location (4/3).Lead ConﬁgurationsIn this study, both RA and RV leads used a bipolar conﬁguration, which was notprogrammable. The LV lead conﬁguration programming was per physician’sdiscretion. In the ﬁrst phase, 20 patients had LV sensing programmed to theLVtip>>LVring conﬁguration, four to LVtip>>RVcoil, and one to LVtip>>Can.In the second phase, 13 patients had LV sensing programmed to theLVtip>>LVring conﬁguration, and four to LVtip>>RVcoil.Lead PerformanceThe lead performance, including pacing threshold, pacing impedance andsensing amplitude, were measured at implant by a commercially availablePacing System Analyzer (PSA). The results are provided in the table below(Table H-2 on page H-4).Table H-2. Lead performanceMeasurement Lead Location Number ofLeads: Phase 1Mean ± SD:Phase 1Number ofLeads: Phase 2Mean ± SD:Phase 2PacingImpedance (Ω)Left Ventricle 25 1034 ± 394 18 779 ± 227Right Atrium 28 520 ± 161 17 519 ± 112Right Ventricle 29 816 ± 263 18 649 ± 206PacingThreshold (V)Left Ventricle 25 1.9 ± 1.4 18 1.3 ± 1.0Right Atrium 28 1.1 ± 0.7 16 1.2 ± 0.6Right Ventricle 29 1.0 ± 0.4 18 0.8 ± 0.3SensingAmplitude (mV)Left Ventricle 25 14.1 ± 7.6 18 13.2 ± 7.3Right Atrium 28 2.9 ± 1.5 16 3.7 ± 3.3Right Ventricle 29 12.3 ± 6.2 18 13.4 ± 7.0Sensing PerformanceIn the ﬁrst phase of the study, a total of 55,207 signals were recorded, including54,151 appropriate sensed intrinsic and paced beats and 1,056 inappropriate- DRAFT -

CLINICAL STUDY - SUMMARY OF GDT1000 SENSING ACUTE STUDY H-5sensed events (223 undersense and 833 oversense events). The sensingalgorithm used in the ﬁrst phase achieved the sensitivities, speciﬁcities, positivepredictive values (PPV), rates of undersensing (1-sensitivity), and rates ofoversensing (1-PPV) are summarized in the table below (Table H-3 on pageH-5).Table H-3. Summary of Sensing Performance - First PhaseSensitivity(Rate ofUndersensing)Speciﬁcity PositivePredictiveValue(Rate ofOversensing)AppropriateSensedBeatsInappropriateSensedBeats:UndersenseInappropriateSensedBeats:OversenseRight AtrialChannel100% (0%) 96.81% 97.03%(2.97%)19,478 0 615RightVentricularChannel100% (0%) 98.94% 98.86%(1.14%)18,439 0 216LeftVentricularChannel98.63%(1.37%)99.99% 99.99%(0.01%)16,054 223 2Totals 54,151 223 833In the second phase of the study, a total of 35,998 signals were recordedincluding 35,831 appropriate sensed intrinsic and paced beats and 171inappropriate sensed events (2 undersense and 169 oversense events).The upgraded sensing algorithm used in the second phase achievedthe sensitivities, speciﬁcities, positive predictive values (PPV), rates ofundersensing (1-sensitivity), and rates of oversensing (1-PPV) summarized inthe table below; the table also summarizes the results of the analysis from thesecond phase (Table H-4 on page H-5).Table H-4. Summary of Sensing Performance - Second PhaseSensitivity(Rate ofUndersensing)Speciﬁcity PositivePredictiveValue(Rate ofOversensing)AppropriateSensedBeatsInappropriateSensedBeats:UndersenseInappropriateSensedBeats:OversenseRight AtrialChannel100% (0%) 98.64% 98.54%(1.46%)11,372 0 168RightVentricularChannel100% (0%) 100% 100% (0%) 12,230 0 0- DRAFT -

H-6 CLINICAL STUDY - SUMMARY OF GDT1000 SENSING ACUTE STUDYTable H-4. Summary of Sensing Performance - Second Phase (continued)Sensitivity(Rate ofUndersensing)Speciﬁcity PositivePredictiveValue(Rate ofOversensing)AppropriateSensedBeatsInappropriateSensedBeats:UndersenseInappropriateSensedBeats:OversenseLeftVentricularChannel99.98%(0.016%)99.99% 99.99%(0.008%)12,227 2 1Totals 35,831 2 169Comparing the performance between the two phases, there were 1,056inappropriate sensing events out of 55,027 signals (1.919%) in the ﬁrst phaseof the study, and a total of 171 inappropriate sensing events out of 35,998signals (0.475%) in the second phase of the study, reﬂecting a 75.2% reductionin inappropriate sensing events from phase one to two.Oversense EventsDuring the analysis of the ﬁrst phase data, some unexpected oversense eventswere identiﬁed. There were a total of 831 oversense events in the RA (615) andRV (216) channels in phase one. The majority of the RA and RV oversenseevents were attributed to an artiﬁcial event introduced while pacing. This typeof oversense was observed in 6 patients in the RA channel and 7 patients inthe RV channel. The results for the second phase of the study demonstratedthat oversensing artiﬁcial events observed in the ﬁrst phase of the study weresuccessfully eliminated by using the upgraded GASD system. There were noartiﬁcial events introduced in the second phase of the study.In the second phase, a total of 168 oversense events in the RA channel wereobserved in one patient. This patient had an intrinsic P-R interval greaterthan 300 ms. In order to complete the AP/VS test scenario, the device wasprogrammedwithaLRL=80bpmandAVDelay=300ms,whichisthemaximum allowable AV Delay in a CRT-D device. At the end of the AV Delay,no intrinsic activity occurred and the device paced both ventricles. These pacedbeats were oversensed by the atrial channel. If the atrial blanking period wereprogrammed to a larger value, atrial oversensing would have been eliminated.Therefore, excluding this patient’s AP/VS test scenario from the analysis, therewere no undersensing or oversensing events in the RA channel.In one patient, the single oversense event in the LV channel was caused bynoise.- DRAFT -

CLINICAL STUDY - SUMMARY OF GDT1000 SENSING ACUTE STUDY H-7Undersense EventsLV undersense events (223) in the ﬁrst phase of the study primarily occurredin one patient whose LV intrinsic amplitude was less than 1.0 mV, which ismuch smaller than the clinically acceptable threshold. This small LV intrinsicamplitude resulted in undersensing some LV events.Two LV undersense events occurred in the second phase of the study. Apotential cause for the LV undersense events was premature ventricularcontractions while atrial pacing.CONCLUSIONSThis acute study demonstrated excellent sensitivity (100% in the RA, 100% inthe RV, and 99.98% in the LV channels), speciﬁcity (98.64% in the RA, 100% inthe RV, and 99.99% in the LV channels), and positive predictive values (98.54%in the RA, 100% in the RV, and 99.99% in the LV channels). In conclusion, thenew sensing platform evaluated in the GDT1000 study will be implemented inCOGNIS/TELIGEN devices.By excluding one patient’s oversensed events that could be eliminated byprogramming a longer atrial blanking period, the modiﬁed speciﬁcity andpositive predictive values in the RA channel are 100% and 100% (oversensingeliminated). While the GDT1000 protocol did not pre-specify acceptablesensitivities, speciﬁcities, or PPV values, a Sensing Tape Testing DAT reportfor COGNIS/TELIGEN on ﬁle at Boston ScientiﬁcCRM1reported a 99.96%sensitivity (0.04% Undersensing) and a 99.97% positive predictive value(0.03% Oversensing) for the RV channel in normal sinus rhythm. Using the RVvalues as a benchmark (RA and LV values were not calculated), the results ofthis study compare favorably.1. Sensing Tape Testing Design Analysis Test report 100019-687 Revision A describes testingperformed in which the COGNIS/TELIGEN sensing platform is modeled and compared to aprevious Guidant device, CONTAK RENEWAL TR. The analysis was performed using 219patient rhythms from the Gold Development Database, including normal sinus rhythm andatrial and ventricular arrhythmias.- DRAFT -

H-8 CLINICAL STUDY - SUMMARY OF GDT1000 SENSING ACUTE STUDY- DRAFT -

$INDEXSymbols50 Hz/manual burst pacing 8-9AA-blankafter RV-sense 5-57after V-pace 5-57A-tachy response (ATR)mode switch 5-27ABM (Autonomic Balance Monitor) 7-4Accelerate, in zone 4-3Accelerometer 5-21activity threshold 5-22reaction time 5-23recovery time 5-26response factor 5-24Activity threshold 5-22Adaptive-rate pacing 5-20Adverse event 1-19AFib rate threshold 3-29, 3-36, 3-38Amplitude 5-16ATP (antitachycardia pacing) 4-18intrinsic test 6-7Application screen 2-5Arrhythmia logbook 7-5episode detail 7-7events summary 7-7interval 7-11stored EGM 7-8ATP (antitachycardia pacing) 4-10amplitude 4-18burst cycle length (BCL) 4-14burst scheme 4-15commanded, EP test 8-11coupling interval 4-12minimum interval 4-14number of bursts 4-11pulse count 4-11pulse width 4-18ramp scheme 4-15ramp/scanscheme4-17redetection after ATP 3-18scan scheme 4-16time-out 4-18ATR (atrial tachy response)atrial ﬂutter response 5-34biventricular trigger 5-33duration 5-29end of ATR episode 5-32entry count 5-29exit count 5-30LRL, fallback 5-31maximum pacing rate 5-33mode switch 5-27mode, fallback 5-30PMT termination 5-35rate threshold 5-29time, fallback 5-31ventricular rate regulation 5-32VTR (ventricular tachy response) 5-32Atrialrefractory period, post ventricular atrial(PVARP) 5-52use of atrial information 3-5Atrial arrhythmia rate threshold 5-29Atrial ﬂutter response 5-34Atrial tachyATRmodeswitch5-27atrial ﬂutter response 5-34PMT termination 5-35ventricular rate regulation 5-32Attention conditions, yellow 2-10Automatic Intrinsic Rhythm ID 3-8AV delay 5-46paced 5-46sensed 5-48- DRAFT -$

BBackup VVI pacing during atrial stimulation,EP test 8-2BatteryBeginning of life (BOL) 6-2Explant status 6-2icon 2-9indicator 6-2status 6-2Beepduring capacitor charge 6-5feature setup 10-4Biventricular trigger 5-33maximumpacingrate5-34BlankingA-blank after RV-sense 5-57A-blank after V-pace 5-57Blanking; Noiserejection, blanking 5-56BurstATP (antitachycardia pacing) 4-11cycle length (BCL) 4-14minimum interval 4-14number of bursts 4-11pacing, 50 Hz/manual burst 8-9parameter 4-11pulse count 4-11scheme 4-15Buttons, software 2-8CCapacitordeformation 4-22, 6-5re-formation 6-5Characteristics as shipped 1-23Charge time 4-22measurement 6-5Checkicon 2-9CommandedATP, EP test 8-11shock, EP test 8-10therapy, EP test 8-10Committed shock 3-11, 4-23Communication, telemetryradio frequency (RF) 2-12Connection, lead to pulse generator 1-21, 9-6Content, package 1-22Contraindications 1-6Counterbrady/CRT 7-16therapy history 7-15ventricular 7-15Coupling interval 4-12decrement 4-12CRT (cardiac resynchronization therapy)delivery zone 3-5DDatadisk 2-11patient 2-11Decelerate,inzone4-3Decrementcoupling interval 4-12ramp scheme 4-15scan scheme 4-16Deﬁbrillationbackup deﬁbrillator, safety mode 2-19DemonstrationProgrammer/recorder/monitor (PRM)mode 2-6Description, device 1-4Detail icon 2-8DetectionAFib rate threshold 3-29duration 3-15enhancement 3-7, 3-23episode 3-20- DRAFT -

onset 3-34rate sensing 3-3rate threshold 3-4reconﬁrmation/committed shock 4-23redetection 3-11stability 3-32sustained rate duration (SRD) 3-35tachyarrhythmia 3-1tachyarrhythmia, safety mode 2-19Vrate>Arate3-28vector timing and correlation 3-27ventricular, initial 3-6window 3-13Devicecharacteristics as shipped 1-23description 1-4mode 3-2programming recommendation 5-2speciﬁcation 1-21storage 1-8Diagnosticbattery status 6-2heart rate variability (HRV) 7-12histogram 7-11lead test 6-6patient triggered monitor 7-17Diskdata 2-11read 2-11save 2-11Disposal of pulse generator 10-8DIVERT THERAPY 2-16Duration 3-15ATR (atrial tachy response) 5-29post-shock 3-18redetection 3-18EECG (electrocardiogram)display 2-7EGM (electrogram)display 2-7Electrocauterymode 3-3Electrode, lead conﬁguration 5-42Electromagnetic interference (EMI) 1-12EMI (electromagnetic interference) 1-12EndofATRepisode5-32Energyshock 4-21Enhancementdetection 3-7, 3-23Entry count 5-29EP test (electrophysiologic test) 8-2ATP, commanded 8-11backup VVI pacing during atrialstimulation 8-2burst pacing, 50 Hz/manual 8-9commanded therapy 8-10ﬁbrillation 8-5induction 8-4mode, temporary 8-2programmed electrical stimulation(PES) 8-7shock on T 8-6shock, commanded 8-10VFib 8-5Episode 3-20end of ATR 5-32nontreated 3-20, 7-15treated 3-20, 7-15ventricular 3-20Eventadverse, potential 1-19counter 7-15icon 2-9therapy history 7-2Eventssummary 7-7Exit count 5-30Explantation 10-8- DRAFT -

FFallback, atrial mode switchLRL 5-31mode 5-30time 5-31Federal Communications Commission(FCC) 1-25FibrillationVFib induction 8-5Follow-upexamination, routine 10-2predischarge 10-2test 10-2HHeart failure 5-2Heart rate variability (HRV) 7-12Histogram 7-11Horizontal slidericon 2-9IIconbattery 2-9check 2-9details 2-8event 2-9horizontal slider 2-9increment and decrement 2-9lead 2-8patient 2-8patient information 2-11Programmer/recorder/monitor (PRM)mode indicator 2-6run 2-9scrolling 2-10sorting 2-9vertical slider 2-9Identiﬁer, x-ray 1-24Impedance test, lead 6-8Implantpost, follow-up 10-2post, information 10-3predischarge follow-up 10-2procedure 9-2Increment and decrementicon 2-9Indications and usage 1-6Indications Based Programming (IBP) 2-2Induction, EP test 8-4Industry Canada (IC) 1-25Informationimplant 2-11, 9-2lead 2-11patient 2-11patient counseling 1-28post implant 10-2related 1-5warranty 1-27Interrogate 2-13Intervalarrhythmia logbook 7-11coupling, ATP 4-12minimum, burst cycle length 4-14Intrinsic amplitude test 6-7Items included in package 1-22LLast delivered shock 6-6Latitude 2-2Leadconﬁguration 5-42connection to pulse generator 1-21connection to pulse generator (PG) 9-6icon 2-8impedance 6-8- DRAFT -

$intrinsic amplitude 6-7pace threshold 6-8test 6-6Left ventricular protection period (LVPP) 5-55Left ventricular refractory period (LVRP) 5-55Logbook 7-5Longevitypulse generator 1-26Lower rate limit (LRL) 5-10LV offset 5-14LV-blank after A-pace 5-57MMagnet 1-8electromagnetic interference (EMI) 1-12feature setup 10-5inhibit tachy therapy 10-5static magnetic ﬁelds 1-17Magnetic ﬁelds 1-17Magnetic Resonance Imaging (MRI) 1-7Maintaining cardiac resynchronizationtherapymaintaining CRT 5-4Manual programming 2-5Manual/50 Hz burst pacing 8-9Maximumpacing rate 5-33, 5-34sensor rate (MSR) 5-13tracking rate (MTR) 5-11Maximum pacing raterate smoothing 5-42Measurementlead, baseline 9-5Mechanical speciﬁcation 1-21Memory, pulse generator 2-11Minimuminterval 4-14Modedevice 3-2electrocautery 3-3fallback ATR (atrial tachy response) 5-30pacing 5-7Programmer/recorder/monitor (PRM) 2-6temporary, EP test 8-2ventricular tachy 3-2MTR (maximum tracking rate) 5-4NNoiseresponse 5-60Nominal parameter setting A-1Number of bursts 4-11pulse count 4-11OOnset 3-10, 3-34, 3-38, 3-39PPaceSTAT PACE 2-18Pace threshold test 6-8pacingCRT (cardiac resynchronizationtherapy) 5-4Indications Based Programming(IBP) 2-2Pacingadaptive-rate 5-20amplitude 5-16ATRmodeswitch5-27AV delay 5-46backup pacemaker in safety mode 2-19backup VVI during atrial stimulation 8-2burst, 50 Hz/manual 8-9- DRAFT -$

$chamber, ventricular 5-14lower rate limit (LRL) 5-10LV offset 5-14maximum sensor rate (MSR) 5-13maximum tracking rate (MTR) 5-11mode 5-7noise response 5-60Parameter, BASIC 5-7post therapy 5-17, 5-18post-therapy 5-17programming recommendation 5-2pulse width 5-15refractory 5-52runaway protection 5-11sensitivity 5-16sensor 5-19SmartDelay optimization 5-50temporary 5-18therapy 5-6Packagecontent 1-22symbol on 1-22Parameter, characteristics 1-23Patientcounseling information 1-28handbook 1-29information icon 2-8Patient triggered monitor 7-17PES (programmed electrical stimulation) 8-7PMT (pacemaker-mediated tachycardia)termination 5-35Polarityshock 4-22Post implant information 10-2, 10-3beeper feature 10-4magnet feature 10-5sensitivity adjustment 10-3Post-shockdetection parameter 3-12duration 3-18pacing 5-17, 5-18Post-therapy pacing 5-17Precautions 1-8Predischarge follow-up 10-2Premature ventricular contraction (PVC) 5-54Prescriptiontherapy 4-2Printreport 2-12Printerexternal 2-12Program 2-2Programmable option, parameter A-1Programmer/recorder/monitor (PRM) 2-2controls 2-5modes 2-6software terminology 2-5use of color 2-10programming recommendation 5-2Programming recommendation 2-2, 2-5Protectionperiod, left ventricular (LVPP) 5-55runaway 5-11Pulse amplitude 5-16Pulse count 4-11Pulse generator (PG)disposal 10-8explantation 10-8implant 9-2longevity 1-26memory 2-11setscrew locations 9-6suture hole 9-6Pulse width 5-15ATP (antitachycardia pacing) 4-18PVARP (post ventriuclar atrial refractoryperiod) 5-52after PVC (premature ventricularcontraction) 5-54PVC (premature ventricular contraction) 5-54QQUICK CONVERT ATP 4-20- DRAFT -$

$QuitEnding a telemetry session 2-14RRadio frequency (RF)ending telemetry 2-14interference 2-15operating temperature, telemetry 2-14starting telemetry 2-13telemetry 2-12Ramp scheme 4-15Ramp/Scan scheme 4-17Rateadaptive 5-20AFib threshold 3-29calculation 3-4lower limit (LRL) 5-10maximum sensor 5-13maximum tracking 5-11sensing 3-3sustained rate duration (SRD) 3-35threshold, ventricular 3-4Vrate>Arate3-28ventricular 3-4zone 3-4Rate enhancement, pacing 5-35rate hysteresis 5-36rate smoothing 5-38tracking preference 5-36Rate smoothing 5-38down 5-41Maximum pacing rate 5-42up 5-41Rate threshold, ATR 5-29Reaction time 5-23Read disk 2-11Reconﬁrmation 3-11, 4-23Recovery time 5-26Red warning conditions 2-10Redetection 3-11after ATP delivery 3-18, 4-8after shock delivery 3-18, 4-9duration 3-18ventricular 4-8Reform, capacitor 6-5Refractoryatrial, post ventricular (PVARP) 5-52blanking and noise rejection 5-56left ventricular (LVRP) 5-55left ventricular protection period 5-55PVARP after PVC 5-54right ventricular (RVRP) 5-54Refractory; pacingrefractory 5-52Related information 1-5Reliability 1-27Report, printed 2-7, 2-11ECG/EGM 2-7Response factor 5-24RhythmID, automatic intrinsic 3-8Right ventricular refractory (RVRP) 5-54Runicon 2-9Runaway protection 5-11RV-blank after A-pace 5-56SSafety core 2-18Safety mode 2-18Safety Tachy Mode 2-20Save disk 2-11Scan scheme 4-16Screen, programmer application 2-5Scrollingicon 2-10SecurityZIP telemetry 2-14Sensing, rate 3-3Sensitivity 5-16- DRAFT -$

Boston Scientific CRMN11906 Implantable Defibrillator User Manual Cognis Part 4 Manual

Boston Scientific Corporation Implantable Defibrillator Cognis Part 4 Manual

Contents

Cognis Part 4 Manual

Navigation menu

Namespaces

Views

Navigation