Posted as supplied by the authors

Detail description of methods

Data Sources

The study population consisted of respondents to the Canadian Heart Health Survey (CHHS), a cross-sectional, population-based survey that contained biophysical measures including blood pressure, weight, height, and blood lipid levels.¹

The CHHS, conducted between 1990 and 1992, is well suited to estimate the potential effectiveness of statins since this was before they became a standard mode of therapy thus

providing an estimate of the risk of coronary heart disease (CHD) outcomes without treatment. Additionally, there is no strong evidence that the CHD risk of Canadians has appreciably changed in the intervening period, given that the incidence of disease is stable² and, except for smoking, risk factors are either unchanged or have shown small prevalence increases.³

The target population of this survey included all Canadian males and females between the ages of 20 and 74 years. The survey was a stratified, two-stage, probability sample and consisted of a home interview, followed by a clinic visit where physical measures, including fasting blood samples, were taken. All blood lipid analyses were done at the Lipid Research Laboratory of the University of Toronto and discussed in detail elsewhere.⁴ The response rate for the home interview was 78% (N = 23,129). We used a subsample of the respondents who where asked questions about their family history (N=9,286). Of these, 6760 respondents (73% of those interviewed) provided a fasting blood sample during the clinic interview and had completed responses to the questions pertaining to CVD risk.

Baseline CHD mortality risk estimates are necessary for estimations population CHD death and strategy impact. Framingham risk models, however, are not available for people with CHD or cardiovascular disease (CVD). For people with CVD, the baseline risks of CHD death was estimated using a population-based prevalence cohort of CVD patients from the Canadian Institutes for Health Information (CIHI) hospital discharge abstract database. This data has been used previously to examine CVD incidence and survival.^5,6 We identified all persons aged 18 to 74 years in Ontario who were discharged from 1988 to 1992 with either ischemic heart disease (International Classification of Diseases, Ninth Revision [ICD-9] codes 410-4), cerebrovascular (433-6) or peripheral arterial disease (440, 444) for any reason (most responsible diagnosis, diagnosis contributing to hospital stay or pre-existing condition). Their age-specific 10-year CHD mortality rates were calculated from 1992 (Table B) and applied to all Canadians with CVD.

Estimating the rates of CHD-related death on the basis of patients who were discharged within the previous 4 years of the observation period (1992-2002) will overrepresent the number of patients who had a recent CVD event. This has a potential to overestimate the mortality rates of CHHS respondents with prevalent CVD. The extent of this potential bias was examined by comparing the 5-year mortality rate (1997-2002) between rates calculated using 4 years (1992-1997) and 9 years (1988-1997) of wash-in data. The rates for the longer wash-in period changed little from 15.5% to 16.2%.

In this study, CHD death was estimated using a proportion of all-cause mortality for CVD patients, since cause of death is not known for hospitalized patients in Ontario. The proportion of all deaths among people with CVD that were attributed to CHD has been examined in clinical trials for statins.^7-10 The estimate we used (47% proportional mortality of CHD) most closely reflected a population-based CVD prevalence cohort¹⁰. The proportion of deaths attributed to CAD from 4 studies examined ranged from 47% to 74%.^7-10

Analysis

Figure 1 illustrates the methods used to identify individuals for screening and treatment using the 3 CHD prevention strategies.

For the baseline risk strategy, we followed the NZ guidelines screening and treatment recommendations.¹¹ People were "screened" for treatment if they were men over age 45, or women over age 55, or 10 years earlier if has more than one risk factor (hypertension, smoking, obesity, or family history of CVD). Familial hyperlipidemia, familial hypercholesterolemia or renal diseases were not considered in the risk assessment because it could not be identified in the CHHS population. We assumed people were treated if they had CVD or a 5-year risk of developing CVD greater than 15% as calculated from the Framingham based risk chart provided by the guideline or has total cholesterol or TC:HDL ratio greater than 8. For the single raised risk factor strategy, we assumed that people had there lipid levels screened according to the NZ guidelines and then treated everyone with a total cholesterol level greater than 6.2 mmol/l.

Estimates of the number of individuals with diabetes and CVD generated in this study were based on the proportion of respondents who self-reported having these diseases in the CCHS. Self-reported DM and CVD likely under represents physician-diagnosed disease status^12,13 and therefore may under estimate the size and population health benefit of statin therapy in the high-risk population. The proportion of CCHS respondents who reported having diabetes (3.1%) was lower than estimates of the prevalence of physician diagnosed diabetes in the Ontario population (4.5%).¹⁴ However, the magnitude of the underestimation is likely to be small since the coefficients for diabetes status in the Framingham equations used to predict CHD risk are relatively small. Estimate of CVD prevalence in the CHHS is similar to estimates of the prevalence of CVD in the Canadian population reported in other studies and therefore potential error is likely small.

The baseline risk of CHD was calculated in the same way for all three strategies. For people without pre-existing CVD, CHD mortality risk was estimated using the Framingham risk equations created and validated and applied for this purpose.¹⁵ For individuals with CVD, observed 10 year mortality rate was used.

In the baseline risk strategy, people were assumed to be treated with statins if they had: CVD or diabetes; a baseline risk greater than 15%, or; a cholesterol level exceeded 8 mmol/L or total cholesterol:HDL ratio exceeded 8. The relative effectiveness of statins in avoiding a CHD event was assumed to be 27%, as reported by La Rosa et al.¹⁶ The number of CHD events avoided was calculated by summing, for those recommended treatment, the product of baseline risk and the relative effectiveness of statins. For the single risk factor strategy, we used the NZ screening recommendations, but treated those with TC above 6.2 mmol/L. For the population health strategy, we estimated how many CHD deaths would be reduced if TC in the entire population decreased by 2%. The 2% TC decrease corresponds to the decrease observed in the 21 country MONICA study over 10 years.^17,18 In the population health strategy, we estimated the benefit of a 2% reduction in TC by lowering each person�s TC by 2% and then we re-estimated the number of CHD deaths using the Framingham equation. For people with CVD, we estimated a 2.7% reduction in CHD deaths, which corresponded to the observed mortality reduction that could be expected with a 2% reduction in TC from treatment trials.¹⁶

The survey sampling weights were applied to generate estimates for all Canadians 20 to 74 years of age.

Study Limitations

Our study has several potential limitations. We estimated the impact of the three different strategies using a risk algorithm and the observed relative risk reduction of CHD death through statins and incremental decreases to TC. The impact of one strategy will be overestimated compared to the other strategies if: 1) the risk algorithm overestimates the baseline risk of CHD death for that strategy�s target population and/or; 2) we over estimated relative risk reduction of the strategy�s intervention (statins or cholesterol reduction). The first potential bias arising from the CHD death risk algorithm should be small because we used the same method for all strategies. There are concerns that the Framingham algorithm overestimates CHD risk^19,20, but these concerns are for both high and low-risk populations, so any error should similarly affect all strategies. For the second potential bias, we assumed the relative benefit of statins for reducing CHD mortality was 27%.¹⁶ Other meta-analyses and reviews of statins have a range of statin impact of CHD mortality between 0% for primary prevention (people at low-risk of CHD death)²¹ to as high as 40% in secondary prevention (people at high-risk of CHD death).^22,23 The overall impact of statins is mostly influenced by changes in the relative benefit for high-risk people. If we had used a higher statin relative benefit, the potential number of deaths prevented would be correspondingly higher for the single risk factor and baseline risk strategies. Similarly, the impact of the population health strategy is directly related to the relative reduction in CHD death that is expected with a decrease in TC. We assumed a 2.7% CHD death reduction for each 2% TC decrease. Reviews of the relationship between CHD death and TC have a surprisingly similar estimates, regardless of whether the relationship was estimated using information from drug trials or observational studies.^24-26

�

Reference List

1. MacLean DR, Petrasovits A, Nargundkar M, Connelly PW, MacLeod E, Edwards A et al. Canadian heart health surveys: a profile of cardiovascular risk. Survey methods and data analysis. Canadian Heart Health Surveys Research Group. CMAJ 1992;146:1969-74.

2. Tu JV, Austin PC, Filate WA, Johansen HL, Brien SE, Pilote L et al. Outcomes of acute myocardial infarction in Canada. Can J Cardiol 2003;19:893-901.

3. Tanuseputro P, Manuel DG, Leung M, Nguyen K, Johansen H. Risk factor for cardiovascular disease in Canada. Can J Cardiol 2003;19:-1249.

4. Connelly PW, MacLean DR, Horlick L, O'Connor B, Petrasovits A, Little JA. Plasma lipids and lipoproteins and the prevalence of risk for coronary heart disease in Canadian adults. Canadian Heart Health Surveys Research Group. CMAJ 1992;146:1977-87.

5. Jackevicius CA, Mamdani M, Tu JV. Adherence with statin therapy in elderly patients with and without acute coronary syndromes. JAMA 2002;288:462-7.

6. Tu JV, Pashos CL, Naylor CD, Chen E, Normand SL, Newhouse JP et al. Use of cardiac procedures and outcomes in elderly patients with myocardial infarction in the United States and Canada. N Engl J Med 1997;336:1500-5.

7. LIPID Study Group. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. N Engl J Med 1998;339:1349-57.

8. Lewis SJ, Sacks FM, Mitchell JS, East C, Glasser S, Kell S et al. Effect of pravastatin on cardiovascular events in women after myocardial infarction: the cholesterol and recurrent events (CARE) trial. J Am Coll Cardiol 1998;32:140-6.

9. Sacks FM, Pfeffer MA, Moye LA, Rouleau JL, Rutherford JD, Cole TG et al. The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events Trial investigators. N Engl J Med 1996;335:1001-9.

10. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: a randomised placebo-controlled trial. Lancet 2002;360:7-22.

11. New Zealand Guidelines Group. The assessment and management of cardiovascular risk. 2003. Wellington, New Zealand, http://www.nzgg.org.nz/Last Accessed Sept 12, 2005.

12. Hux JE, Ivis F, Flintoft V, Bica A. Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm. Diabetes Care 2002;25:512-6.

13. Manuel DG, Leung M, Nguyen K, Tanuseputro P, Johansen KS. Burden of Cardiovascular Disease in Canada. Can J Cardiol 2003;19:997-1004.

14. Manuel DG, Schultz SE. Using Linked Data to Calculate Summary Measures of Population Health: Health-adjusted Life Expectancy of People with Diabetes Mellitus. Popul Health Metr 2004;2:4.

15. Anderson KM, Odell PM, Wilson PW, Kannel WB. Cardiovascular disease risk profiles. Am Heart J 1991;121:293-8.

16. LaRosa JC, He J, Vupputuri S. Effect of statins on risk of coronary disease: a meta-analysis of randomized controlled trials. JAMA 1999;282:2340-6.

17. Kuulasmaa K, Tunstall-Pedoe H, Dobson A, Fortmann S, Sans S, Tolonen H et al. Estimation of contribution of changes in classic risk factors to trends in coronary-event rates across the WHO MONICA Project populations. Lancet 2000;355:675-87.

18. Tunstall-Pedoe H, Vanuzzo D, Hobbs M, Mahonen M, Cepaitis Z, Kuulasmaa K et al. Estimation of contribution of changes in coronary care to improving survival, event rates, and coronary heart disease mortality across the WHO MONICA Project populations. Lancet 2000;355:688-700.

19. Ferrario M, Chiodini P, Chambless LE, Cesana G, Vanuzzo D, Panico S et al. Prediction of coronary events in a low incidence population. Assessing accuracy of the CUORE Cohort Study prediction equation. Int J Epidemiol 2005;34:413-21.

20. Brindle P, Emberson J, Lampe F, Walker M, Whincup P, Fahey T et al. Predictive accuracy of the Framingham coronary risk score in British men: prospective cohort study. BMJ 2003;327:1267.

21. Do Statins Have a Role in Primary Prevention? Therapeutics Letter 48. 2003. Vancouver, University of British Columbia.

22. Cheung BM, Lauder IJ, Lau CP, Kumana CR. Meta-analysis of large randomized controlled trials to evaluate the impact of statins on cardiovascular outcomes. Br J Clin Pharmacol 2004;57:640-51.

23. Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344:1383-9.

24. Gould AL, Rossouw JE, Santanello NC, Heyse JF, Furberg CD. Cholesterol reduction yields clinical benefit: impact of statin trials. Circulation 1998;97:946-52.

25. Bucher HC, Griffith LE, Guyatt GH. Systematic review on the risk and benefit of different cholesterol-lowering interventions. Arterioscler Thromb Vasc Biol 1999;19:187-95.

26. Marchioli R, Marfisi RM, Carinci F, Tognoni G. Meta-analysis, clinical trials, and transferability of research results into practice - The case of cholesterol-lowering interventions in the secondary prevention of coronary heart disease. Archives of Internal Medicine 1996;156:1158-72.

�

Posted as supplied by the authors

Table A The number of CHD deaths avoided over ten years with three strategies**

Risk group	Estimated Population	(%)	Number of people"treated"	Treated in each risk group (%)	Number of CHD deaths avoided over 10 years**	Deaths avoided per 100,000 population**	NNT- CHD Death over10 years	Cumulative percentage treated
Single Risk Factor Approach	�	�	�	�	�	�	�	�
History of CVD	499,000	(5.4)	155,000	31.1	5,370	1,080	29	1.3
10-year risk of CHD* > 10%	184,000	(2.0)	77,200	42.1	2,720	1,480	28	1.9
10-year risk of CHD >5-10%	530,000	(5.7)	162,000	30.5	3,230	609	50	3.2
10-year risk of CHD >1-5%	1,840,000	(19.8)	550,000	29.8	3,750	204	147	7.6
10-year risk of CHD 1% or less	9,290,000	(100.0)	431,000	4.6	422	5	1,020	11.1
Total	12,300,000	�	1,370,000	11.1	15,500	125	89	�
Baseline Risk Approach	�	�	�	�	�	�	�	�
History of CVD	499,000	(5.4)	499,000	100.0	16,800	3,370	30	4.0
10-year risk of CHD* > 10%	184,000	(2.0)	184,000	100.0	6,380	3,470	29	5.5
10-year risk of CHD >5-10%	530,000	(5.7)	458,000	86.4	9,060	1,710	51	9.2
10-year risk of CHD >1-5%	1,840,000	(19.8)	416,000	22.6	3,470	188	120	12.6
10-year risk of CHD 1% or less	9,290,000	(100.0)	36,000	0.4	58	1	616	12.9
Total	12,300,000	�	1,590,000	12.9	35,800	290	45	�
Rose Population Health Approach	�	�	�	�	�	�	�	�
History of CVD	499,000	(5.4)	499,000	100.0	1,490	299	334	4.0
10-year risk of CHD* > 10%	184,000	(2.0)	184,000	100.0	493	269	372	5.5
10-year risk of CHD >5-10%	530,000	(5.7)	530,000	100.0	1,020	192	521	9.8
10-year risk of CHD >1-5%	1,840,000	(19.8)	1,840,000	100.0	1,590	86	1,160	24.7
10-year risk of CHD 1% or less	9,290,000	(100.0)	9,290,000	100.0	571	6	16,300	100.0
Total	12,300,000	�	12,300,000	100.0	5,160	42	2,390	�

*CHD death, as predicted by Framingham equations (Anderson, 1991)

**Assuming 100% community effectiveness for the single risk factor and reversible risk approaches, and a 2% total cholesterol reduction for the Geoffrey Rose approach

�

��

Posted as supplied by the authors

Table B Probabilities of coronary heart disease (CHD) death for patients discharged from hospital for cardiovascular diseases, Ontario, 1988-1992

�	10-year Death Rate from CHD (1992-2002) (%)
Age-Groups	Both	Female	Male
All	13.9	13.3	14.2
18-44	4.8	4.3	5.0
45-54	6.3	5.8	6.3
55-59	8.8	8.4	9.0
60-64	11.6	11.0	11.9
65-69	15.2	13.4	16.0
70-74	20.3	18.2	21.4