HTA. Supporting Informed Decisions. Canadian Agency for Drugs and Technologies in Health

Transcription

1 Canadian Agency for Drugs and Technologies in Health Agence canadienne des médicaments et des technologies de la santé technolo g y o v e r v i e w HTA Issue 26 March 2007 Overview of Implantable Cardiac Defibrillators for Primary Prevention of Sudden Cardiac Death in High Risk Patients: Clinical Efficacy, Cost-Effectiveness, and Psychosocial Issues Supporting Informed Decisions

2 Until April 2006, the Canadian Agency for Drugs and Technologies in Health (CADTH) was known as the Canadian Coordinating Office for Health Technology Assessment (CCOHTA). Publications can be requested from: CADTH Carling Avenue Ottawa ON Canada K1S 5S8 Tel. (613) Fax. (613) or download from CADTH s web site: Cite as: Ho C, Li H, Noorani H, Cimon K, Campbell K, Tang A, Birnie D. Overview of implantable cardiac defibrillators for primary prevention of sudden cardiac death in high risk patients: clinical efficacy, cost-effectiveness, and psychosocial issues [Technology overview no 26]. Ottawa: Canadian Agency for Drugs and Technologies in Health; Production of this report is made possible by financial contributions from Health Canada and the governments of Alberta, British Columbia, Manitoba, New Brunswick, Newfoundland and Labrador, Northwest Territories, Nova Scotia, Nunavut, Ontario, Prince Edward Island, Saskatchewan, and Yukon. The Canadian Agency for Drugs and Technologies in Health takes sole responsibility for the final form and content of this report. The views expressed herein do not necessarily represent the views of Health Canada or any provincial or territorial government. Reproduction of this document for non-commercial purposes is permitted provided appropriate credit is given to CADTH. CADTH is funded by Canadian federal, provincial, and territorial governments. Legal Deposit 2007 National Library of Canada ISSN: (print) ISSN: (online) H0471 March 2007 PUBLICATIONS MAIL AGREEMENT NO RETURN UNDELIVERABLE CANADIAN ADDRESSES TO CANADIAN AGENCY FOR DRUGS AND TECHNOLOGIES IN HEALTH CARLING AVENUE OTTAWA ON K1S 5S8

3 Canadian Agency for Drugs and Technologies in Health Overview of Implantable Cardiac Defibrillators for Primary Prevention of Sudden Cardiac Death in High Risk Patients: Clinical Efficacy, Cost-Effectiveness, and Psychosocial Issues March 2007 We thank Cathy Allison for her assistance in creating this overview from a longer report authored by Ho C, et al. This overview is based on a Technology Report by: Ho C, Li H, Noorani H, Cimon K, Campbell K, Tang A, Birnie D. Implantable cardiac defibrillators for primary prevention of sudden cardiac death in high risk patients: a meta-analysis of clinical efficacy and a review of cost-effectiveness and psychosocial issues [Technology report issue 81]. Ottawa: Canadian Agency for Drugs and Technologies in Health; CADTH takes sole responsibility for the final form and content.

4 REPORT IN BRIEF March 2007 Implantable Cardiac Defibrillators for Primary Prevention of Sudden Cardiac Death in High Risk Patients: A Meta-Analysis of Clinical Efficacy and a Review of Cost-Effectiveness and Psychosocial Issues Technology An implantable cardiac defibrillator (ICD) is a small electronic device implanted under the skin in the chest. It can be used to detect abnormal and potentially life-threatening heart rhythms, then to deliver a shock to restore normal rhythm. Condition Sudden cardiac death (SCD) due to an abnormal heart rhythm is a leading cause of cardiacrelated death. Only 5% of patients survive a cardiac arrest. Issues The use of ICDs in patients who have survived an episode of SCD (secondary prevention) is established. ICDs may also be used to prevent a first episode of SCD in patients at risk (primary prevention) but whether they should be used for these patients remains questionable. The costs associated with ICD use are high compared to those for antiarrhythmic medications, and psychosocial issues such as anxiety and depression may be associated with the use of ICDs. Questions about the allocation of health resources are raised when considering ICDs for the primary prevention of SCD, with the demand already outstripping capacity in many jurisdictions. To address these issues, the clinical effectiveness and cost-effectiveness of ICDs for the primary prevention of SCD in different patient groups must be determined. Methods and Results We used systematic, defined literature searches to identify published and unpublished studies. Ten randomized controlled trials (RCTs) reporting clinical outcomes for ICDs in primary prevention were systematically reviewed. Psychosocial and ethical issues, and the cost-effectiveness of ICD treatment, were examined using the literature review, and a budget impact analysis was performed. Implications for Decision Making ICDs, with optimal pharmacologic therapy, significantly reduce SCD and all-cause death in patients at high risk. ICDs are effective in reducing SCD in patients with ischemic and non-ischemic heart disease. ICD therapy is expensive but some reviewed studies showed that ICDs are cost-effective if the willingness to pay is $50,000 per qualityadjusted life year (QALY). More study with a broader target population is warranted. Poor psychosocial outcomes in ICD patients may occur as a result of their underlying cardiac condition, rather than as a direct response to ICDs. ICD therapy for primary prevention of SCD would result in a substantial budget impact. There are also infrastructure issues that would need to be addressed in order to meet demand. This summary is based on a health technology assessment available from CADTH s web site ( Ho C, Li H, Noorani H, Cimon K, Campbell K, Tang A, Birnie D. Implantable cardiac defibrillators for primary prevention of sudden cardiac death in high risk patients: a meta-analysis of clinical efficacy and a review of cost-effectiveness and psychosocial issues. Canadian Agency for Drugs and Technologies in Health (CADTH) Carling Avenue, Ottawa ON Canada K1S 5S8 Tel: Fax: CADTH is an independent, not-for-profit organization that supports informed health care decision making by providing unbiased, reliable information about health technologies.

5 1 Introduction Sudden cardiac arrest due to abnormal heart rhythm (arrhythmia) is a leading cause of cardiacrelated death. Most sudden cardiac deaths (SCDs) are caused by fatal arrhythmias, particularly ventricular fibrillation and ventricular tachycardia, which are common complications of coronary artery disease. Only 5% of patients survive a cardiac arrest. Survivors have a high risk of dying within a few years, and more than half of these deaths are due to recurrent sudden cardiac arrest. 1 Several approaches have been used to reduce arrhythmia mortality. 2 The key interventions for patients at a low risk of SCD include aggressive management of coronary heart disease and the placement of automatic external cardiac defibrillators in the community. For patients at a high risk of SCD, there has been increasing interest in using implantable cardiac defibrillators (ICDs) to treat cardiac arrhythmias. High-risk patients include those with a prior heart attack (myocardial infarction) and a left ventricular ejection fraction (LVEF) that is 35%; patients with congestive heart failure (CHF); or patients with a dilated cardiomyopathy. More than three million patients in North America meet the high-risk criteria, with approximately 400,000 new patients diagnosed each year. 3 An ICD is a small electronic device that is implanted under the skin in the chest during a surgical procedure that lasts between one to two hours. ICDs consist of a pulse generator (containing electronic circuitry and a battery) with one or more leads that are inserted into the heart through a vein. The device acts as a pacemaker and a defibrillator: it first detects abnormal and potentially life-threatening heart rhythms, and then delivers a high-voltage shock to the heart to restore normal rhythm. An ICD costs approximately C$20,000 4 and has a functional life of five to six years, 5,6 at which time additional surgery is required to replace the battery or the device. The use of ICDs is established for the secondary prevention of SCD in patients who have survived a life-threatening arrhythmia. 7 Because only 5% of patients survive these events, there is increasing interest in identifying people at high risk who are using an ICD for the primary prevention of SCD before an event occurs. The costs of ICD therapy are high compared with those of antiarrhythmic medications, and an increasing demand for the devices is exceeding available funding in many Canadian jurisdictions. Approximately 85,000 untreated Canadians could undergo ICD implantation for the primary prevention of arrhythmia death, with the number of new cases expected to be 3,700 annually. 8 If everyone who might benefit from an ICD were to receive one, the financial toll on the Canadian health care system would be substantial. 7 This raises moral questions about who should receive the device and places physicians in the difficult position of making this decision. ICD recipients may also face psychosocial issues after implantation. Lifestyle changes, such as the loss of driving privileges, could result in reduced personal freedom and lower quality of life (QoL). 1

6 2 Objective The objective of this report is to review the clinical, economic, ethical, and psychosocial issues of ICD therapy compared with conventional treatment for the primary prevention of SCD in patients who are at high risk of having life-threatening cardiac arrhythmia. 3 Clinical Review Methods The clinical review is based on a systematic review and meta-analysis of randomized controlled trials (RCTs) identified through a systematic literature search. A protocol was written a priori and followed throughout the review process. Clinical studies were included if they met all the following criteria: study design was RCT population group was patients at high risk for SCD, without a prior history of cardiac arrhythmia intervention was ICD comparator was conventional medical therapy outcomes were rates of death from any cause (all-cause death); rates of all-cause death in patients with ischemic disease (reduced blood supply to the heart muscle); rates of allcause death in patients with non-ischemic disease; rates of SCD; rates of SCD in patients with ischemic heart disease; rates of SCD in patients with non-ischemic disease. Two reviewers independently selected reports for inclusion and subsequently extracted clinical outcome data from relevant reports using a form that was designed a priori to capture specific information. Disagreements were resolved by discussion. A quality appraisal assessment form, modified from the work of Hailey et al. 9 was used to assess the design and performance of included studies. This score was compared to the Jadad quality assessment scale. 10 The clinical outcomes investigated were the numbers of all-cause death and SCD in patients with ischemic or non-ischemic heart disease. The safety endpoints included the number of perioperative incidents with ICD implantation and complications caused by bleeding, infection, or device dislocation. Results A total of 2,273 citations were identified in our literature search. After eliminating reports that did not meet our selection criteria, we had 11 relevant reports that described 10 unique RCTs, which enrolled more than 7,000 patients. All trials were funded by industry. The RCTs were of variable quality, with an average Jadad quality score of 2 out of 5. Because patients and physicians cannot be blinded regarding an ICD implantation, a low Jadad score is expected. When the trials were assessed using the modified Hailey scale, the average quality score was 12.5 out of 15, indicating a high degree of confidence in trial design and findings. Most studies included patients 18 years old with New York Heart Association (NYHA) class II or III heart failure and a left ventricular ejection fraction of 35%. 2

7 Our clinical review revealed three principal findings. The rate of death, overall or cardiac in origin, was high in patients who had no previous history of arrhythmia, but who were considered to be at high risk due to underlying cardiac disease. Among the patients who were treated with conventional medical therapy, the rates were 25% for all-cause death and 11% for SCD over a median follow-up period of 45 months. Among the patients who received an ICD implant in addition to conventional therapy, the rates decreased to 19% for all-cause death and 4% for SCD. Treatment with an ICD implant in addition to conventional therapy, resulted in statistically significantly reductions in the RRs of all-cause death by 28% and of SCD by 67%. The absolute risk reduction (ARR) was 6% for all-cause death and 7% for SCD. A subgroup analysis showed that patients with ischemic heart disease had higher rates of death than patients with non-ischemic heart disease. a) SCD There was a 67% reduction in the risk of SCD (RR=0.33, ARR 9%) among patients with ischemic disease who received ICDs compared with conventional medical therapy, and a 74% reduction (RR=0.26, ARR 4%) in the risk of SCD in patients without ischemia. These reductions were statistically significant. b) All-cause death There was a 32% decrease in the risk of all-cause death (RR=0.68, ARR 7%) among patients with ischemic disease who received ICDs compared with conventional medical therapy, while for patients with non-ischemic disease, ICD therapy resulted in a 24% reduction in the risk of allcause death (RR=0.76, ARR 5%). These risk reductions failed to reach statistical significance, likely because of an inadequate sample size. All-cause death is the more reliable clinical outcome, because bias can be introduced if deaths are systematically misclassified, particularly when studies are not double-blinded. A numbers-needed-to-treat analysis showed that ICDs would need to be implanted into 14 patients to prevent one SCD in the overall population, and in 12 and 28 patients with ischemic and non-ischemic heart disease respectively. Adverse events were poorly reported in the trials. The most common complications reported with ICD use were infections associated with the surgical procedure and inappropriate shocks due to device malfunction. 4 Economic Review Methods Two reviewers independently selected relevant economic studies that had been identified through an electronic literature search. Full and partial economic studies were eligible for inclusion. The two reviewers then extracted data using a form that was designed to capture 3

8 specific economic information. A 35-item checklist developed by the British Medical Journal (BMJ) 22 was used to assess the quality of included economic studies. Given the heterogeneity among the six included economic studies, no attempt was made to pool the results. A qualitative systematic review was performed instead, with a tabulation of study characteristics and results. The robustness of the main economic findings in the base case was tested with sensitivity analyses. Results Of the 408 economic studies identified in our literature search, six full economic studies and one partial economic evaluation 29 met our inclusion criteria. Four of the full economic studies 24-26,28 carried out cost-effectiveness analyses and cost-utility analyses. One economic study 27 was based on a clinical effectiveness trial, and the remaining six were model-based evaluations ,28,29 Most of the included studies performed well against the BMJ checklist, apart from a lack of detail on statistical tests and justification for the choice of sensitivity analysis variables. Economic outcomes were reported in incremental cost per life-year gained and incremental cost per quality-adjusted life-year (QALY) gained. The base-case results suggest that ICD therapy generally costs more than control therapy, primarily because of the expensive implantation, but it is more effective in treating patients without prior clinical arrhythmia. If effectiveness was measured by life-year gained, most of the incremental cost-effectiveness ratio (ICER) estimates for patients were below or slightly above the commonly used willingnessto-pay threshold of US$50,000 per life-year gained. Most of the base-case economic studies included patients with ischemic heart disease and a low ejection fraction (EF). Effectiveness was also measured by QALY gained. Most economic studies had ischemic samples with low LVEG. Their evaluated ICER of ICD compared with control therapy ranged from US$24,50 per life-year (US$34,000 per QALY) 28 to US$63,300 per life-year (US$71,800 per QALY) 24 for the group with LVEF 0.3. For the ischemic patients with higher LVEF, one identified economic study 24 reported that the ICER of ICD therapy over amiodarone was US$173,400 per life-year (US$195,700 per QALY) for a group with LVEF 0.31 to 0.40 and US$501,500 per life-year (US$557,900 per QALY) for a group with LVEF >0.4. One-way, threshold, scenario, and probabilistic sensitivity analyses were used to examine the robustness of individual base-case results. The sensitivity analyses indicated that the costeffectiveness of ICD treatment, as measured by QALY, was driven mainly by ICD efficacy, ICD implantation cost, and the patient s life utility. Utility is defined as a measure of the strength of patients preferences for certain health states on a scale of 0 to 1, with 1 being perfect health. 4

9 5 Ethical and Psychosocial Review Methods Studies that addressed the ethical and psychological issues related to ICD therapy were included. Only reports published from 1995 onward were included, because of advances in ICD technology. Two reviewers independently selected studies. Because the ethical and psychosocial review is mainly qualitative, no data extraction form was used. Two reviewers independently verified extracted data for accuracy. The studies included were not assessed for quality. Results A total of 95 citations were identified in our literature search for ethical and psychosocial studies. No studies assessed the ethical implications of using ICDs for primary prevention, so we included eight relevant studies that addressed the ethical issues of the use of ICD therapy for secondary prevention, including four review papers; 7,30-32 one commentary on the adoption of new technologies; 33 one editorial; 34 a textbook chapter; 35 and a Canadian case report. 36 The need for ICD devices exceeds funding in many Canadian jurisdictions. Physicians have been put into the difficult moral dilemma of rationing ICDs, which is in opposition to their traditional role as patient s advocate. A more fair decision-making process is required, including guidelines for resource allocation decisions. 31 Of the 11 studies addressing the psychosocial issues, 21,37-46 two evaluated the impact on QoL of ICD therapy for primary prevention. 21,41 One study 41 concluded that ICD therapy was associated with poorer QoL, and another 21 found that QoL scores were similar between patients treated with ICD or standard medical therapy. A meta-analysis of 20 studies 43 that assessed ICDs for primary and secondary prevention found no significant differences in psychosocial outcomes between patients who received ICDs compared with those who received standard medical therapy. The meta-analysis had limitations, and the findings suggest that poor psychosocial outcomes may result from the underlying cardiac condition rather than as a direct response to ICD implantation and its therapy. The available evidence about psychosocial outcomes is weak, and more study is required on the effect on QoL for patients who receive ICDs for primary prevention. The most common psychological problems reported after ICD implants for secondary prevention are anxiety, depression, fear of shock, and fear of death. 42,43,46 6 Limitations There were limitations to the clinical review. The quality of the included RCTs was low, based on the Jadad quality assessment scale. 10 This scale unduly penalizes studies of devices and surgical techniques because the doubleblinding of patients and physicians is not possible. When the modified Hailey appraisal form 9 was applied, studies were shown to have a higher degree of quality. 5

10 Our meta-analysis of the rates of death among patients with or without ischemic disease, excluded 1,676 patients from one large study 12 because the actual number of deaths was not reported for each subgroup. Observational studies were excluded from our review. These could give a clearer picture of real-world effectiveness and safety, including issues related to ICD advisories and recalls due to defective or malfunctioning devices. It was beyond this report s scope to compare standard ICDs with limited-use ICDs. Nor did we examine the role that biventricular pacing may play, alone or combined with an ICD. There were limitations to the economic review. Although the quality of most of the included studies was good (as assessed by the BMJ checklist 22 ), most samples represented subsets rather than the entire targeted population. While the clinical trials associated with the economic studies had follow-up periods from 20 to 45 months; the economic studies usually used a lifetime horizon. It is difficult to say if the short-term results would apply to a longer-term situation. No primary Canadian economic studies compared ICDs with standard medical therapy for the primary prevention of SCD. A limitation of our budget impact analysis is that its precision is based on the quality of mortality data reported in the literature. For the psychosocial and ethical review, few studies examined the impact that prophylactic ICD implantation has on a patient s QoL. Further examination is required, because psychosocial outcomes are poor, with self-reports of lower QoL. 7 Health System Implications The rate of death in patients at a high risk of ventricular arrhythmia, who are treated with conventional medical treatment, remains significant. ICD therapy has a significant benefit in the primary prevention of life-threatening ventricular arrhythmia in high-risk patients. The budget impact for ICD treatment was determined by multiplying the per-person cost (including hospitalization for the device implant and follow-up care) by the number of ICD candidates. This was compared with a treatment scenario without ICD use, where the budget included the number of additional cardiac deaths that would have been prevented by ICD prophylactic use and the costs associated with SCD. A budget impact analysis usually assumes a horizon of three to five years. Because the life expectancy of an ICD recipient is around 10 years, and the lifespan of the ICD device or battery is approximately five years, 5,6 we used a seven-year horizon. We used an annual inflation rate of 3% and a currency based on the 2005 Canadian dollar. We estimated that during the seven-year time horizon, the total cost was C$48,119, largely due to the cost of the ICD implant in the first year and the cost of replacing the device or battery in 6

11 the sixth year. The estimated cost of each SCD that would have been prevented by ICD treatment ranged from C$300 to C$6,500, depending on what health care services were used. It is estimated that 85,000 untreated Canadians are candidates for ICD implantation in the primary prevention of a life-threatening arrhythmia, with the number of new cases expected to be 3,700 annually. 8 Ideally, the budget impact analysis would be based on this number, but the infrastructure that is needed to implant tens of thousands of ICDs in a given year is limited. 4 We based our budget impact analysis on a more realistic forecast that ICDs would be provided to 3,500 in ; 4,500 in ; 5,500 in ; and 6,500 in For , , and , it was assumed that the number of new ICDs would remain at 6,500 annually. 4 From the perspective of the health care system, if the cost associated with preventing SCD was C$300 per case, the estimated budget impact of ICDs (relative to standard medical therapy) would be C$88.6 million, C$332.4 million, C$634.4 million, C$834.4 million, and C$1 billion respectively, during one-, three-, five-, six-, and seven-year time horizons. If the cost associated with SCD prevention was assumed to be C$6,500 per case, the corresponding budget impact of ICD implantation (relative to standard medical therapy) would be C$330, 000; C$2.26 million; C$6.37 million; C$9.67 million; and C$10 million less than that seen when the cost of preventing SCD was assumed to be $300. The use of ICD therapy for the primary prevention of SCD would result in a substantial budget impact due to the cost of the device and the need for its replacement in five to six years. Advisories and recalls of defective or malfunctioning ICD devices have been a concern. A retrospective Canadian study surveyed 17 Canadian centres to examine the issue of ICD advisories or recalls. Between October 2004 and October 2005, of the 2,915 patients who were affected by device advisories, 533 underwent device replacement. 47 The complication rates related to ICD replacement were found to be similar to those encountered with initial implantation in the large RCTs, ranging from 2.5% to 15.2%. An analysis of US Food and Drug Administration reports examined ICD generator malfunctions that were reported from 1990 to The three-year replacement rate of malfunctioning ICDs, from 2000 to 2002, was 26.8 per 1,000 implants (2.7%), which is three times the replacement rate in the mid-1990s. Careful monitoring of device performance, diligent reporting of adverse events, and the return of defective devices to the manufacturer for analysis are essential. In 2005, the Canadian Heart Rhythm Society and the Canadian Cardiovascular Society developed a joint position paper 49 to make recommendations about ICD use for specific patients, based on the presence or absence of ischemic heart disease, LVEF, and other characteristics, such as the time elapsed after a heart attack. Our report reviewed the available clinical efficacy and cost-effectiveness studies, but these were based on a subset of the patient population that was included in the Canadian position paper. The ICER of ICD versus control therapy would correspondingly change if the population in the position paper was used in the economic evaluation. If Canadian ICD policy is to be based on recommendations made in this joint 7

12 position paper, a primary study with a broader target population is warranted to more accurately determine the cost-effectiveness of ICD use for primary prevention in Canada. 8 Conclusions There is evidence that the use of ICD therapy with optimized drug therapy reduces all-cause death and SCD when used for the primary prevention of life-threatening ventricular arrhythmias in patients who are at high risk due to heart disease. In these patients, the risk of all-cause death is reduced by one-third and the risk of SCD by two-thirds compared with conventional medical treatment. ICD treatment is effective in reducing the risk of SCD in ischemic and non-ischemic patients. The clinical review has several strengths. It is a systematic review and meta-analysis carried out independently from vested interest. Its research questions and methods were predefined in a protocol. Only evidence from the highest level of evidence, RCTs, was considered. All included trials evaluated the use of ICD compared with conventional treatment, not placebo. This is important because patients with ICDs are also treated with conventional medical therapy, meaning that the differences in outcomes observed in our review are due to the ICD. The seven identified economic studies result in various ICER estimates. Whether ICD treatment is cost-effective compared with conventional therapy depends on the threshold of willingness-topay for one life-year gained or one QALY gained. If the threshold was US$50,000 per QALY gained, some reviewed studies showed that ICD treatment was cost-effective compared with conventional therapy, while others did not. In 2005, the Canadian Cardiovascular Society and the Canadian Heart Rhythm Society issued a position paper 49 with recommendations for ICD use, based on specific characteristics. The study populations in our clinical and economic reviews are subsets of the population described in the position paper. If Canadian ICD policy is to be based on recommendations made in the joint position paper, a primary study with a broader target population is warranted to more accurately determine the cost-effectiveness of ICD use for primary prevention. For the primary prevention of SCD in high-risk patients, the use of the expensive ICD device, which requires device or battery replacement after six years, would result in a substantial budget impact for the Canadian health care system, compared with the increased cardiac event risk and associated costs if patients received conventional medical therapy only. The impact of ICDs on QoL among primary prevention patients requires more study; it is unknown whether the psychosocial issues in ICD patients result from the underlying cardiac condition, or from a direct response to implantation of the device and therapy. 8

13 9 References 1. Stiell IG, et al. JAMA 1999;281(13): Myerburg RJ. J Cardiovasc Electrophysiol 2001;12(3): Stecker EC, et al. N Engl J Med 2002;347(5): Simpson CS, et al. Can J Cardiol 2005;21 Suppl A:19-24A 5. Manolis AS, et al. Europace 2005;7:95 6. Hauser RG. J Am Coll Cardiol 2005;45(12): Noorani HZ, et al. Can J Cardiol 2000;16(10): Gillis AM. Can J Cardiol 2005;21 Suppl A: Hailey D, et al. J Telemed Telecare 2004;10(6): Jadad AR, et al. Control Clin Trials 1996;17(1): Bänsch D, et al. Circulation 2002;105(12): Available: << 12. Bardy GH, et al. N Engl J Med 2005;352(3): Bigger JT. N Engl J Med 1997;337(22): Buxton AE, et al. N Engl J Med 1999;341(25): Greenberg H, et al. J Am Coll Cardiol 2004;43(8): Hohnloser SH, et al. N Engl J Med 2004;351(24): Kadish A, et al. N Engl J Med 2004;350(21): Moss AJ, et al. N Engl J Med 2002;346(12): Moss AJ, et al. N Engl J Med 1996;335(26): Raviele A, et al. Europace 2005;7(4): Strickberger S Adam, et al. J Am Coll Cardiol 2003;41(10): Drummond MF, et al. BMJ 1996;313(7052): Available: << 23. Chen L, et al. Cardiovasc Drugs Ther 2004;18(2): Sanders GD, et al. Ann Intern Med 2001;135(10): Sanders GD, et al. Technol Eval Cent Assess Program 2004;19(3): Al Khatib SM, et al. Ann Intern Med 2005;142(8): Mushlin A, I, et al. Circulation 1998;97(21): Sanders GD, et al. N Engl J Med 2005;353(14): Anderson MH, et al. Br Heart J 1993;69(1): Silver MT. Am Heart Hosp J 2005;3(3): Simpson CS, et al. Can J Cardiol 2005;21(7): Berger JT. Ann Intern Med 2005;142(8): Gafni A, et al. CMAJ 1993;148(6): Brennan FJ. Pacing Clin Electrophysiol 2004;27(7):

14 35. Farrell CD. Appendix III: Risk measurement and assessment: breast and ovarian cancer. In: Genetic susceptibility to breast and ovarian cancer: assessment, counseling and testing guidelines. Bethesda (MD): American College of Medical Genetics; Cervini P, et al. Can J Cardiol 2003;19(13): Kohn CS, et al. Pacing Clin Electrophysiol 2000;23(4 Pt 1): Chevalier P, et al. Eur Heart J 2004;25(Suppl S): Irvine J, et al. Am Heart J 2002;144(2): Schron EB, et al. Circulation 2002;105(5): Namerow PB, et al. Pacing Clin Electrophysiol 1999;22(9): Sola CL, et al. Mayo Clin Proc 2005;80(2): Burke JL, et al. Br J Health Psychol 2003;8(Pt 2): Jung W, et al. Eur Heart J 1997;18(8): Shea JB. AACN Clinical Issues 2004;15(3): McCready MJ, et al. Card Electrophysiol Rev 2003;7(1): Gould PA, et al. JAMA 2006;295(16): Maisel WH, et al. JAMA 2006;295: Tang AS, et al. Can J Cardiol 2005;21 Suppl A: