Q J Med 1999; 92: 379-385
© 1999 Association of Physicians
Prediction of coronary risk for primary prevention of coronary heart disease: a comparison of methods
From the Department of Clinical Pharmacology and Therapeutics, Royal Hallamshire Hospital, Sheffield, UK
Received 23 April 1999
Professor L.E. Ramsay, Clinical Pharmacology and Therapeutics, Floor L, Royal Hallamshire Hospital, Glossop Road, Sheffield S10 2JF. e-mail: d.colley{at}sheffield.ac.uk
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Most recent guidelines advise targeting of lipid lowering for primary prevention at those at high absolute coronary (CHD) risk. We compared the accuracy of five CHD risk assessment methods in identifying such patients: one based on total cholesterol
6.5 mmol/l plus two risk factors, and four based on the Framingham risk function (the European Task Force chart and Sheffield table, both using total cholesterol and the New Zealand chart and modified Sheffield table, both using total : HDL cholesterol ratio) for predicting CHD event risk 2% per year, calculated by an independent risk function, PROCAM, in 126 treated hypertensive men. Cholesterol threshold plus two risk factors had sensitivity 59% and specificity 63%, did not identify some very high-risk patients, and identified very low-risk patients. Framingham-based methods using total cholesterol alone had sensitivity 90–98% and specificity 37–43%, and identified high-risk patients well, but identified some patients at very low risk. Methods based on total : HDL cholesterol ratio had sensitivity 90–98% and specificity 60–63%, and did not identify incorrectly patients at very low CHD risk. Methods based on cholesterol threshold and counting of risk factors are too inaccurate for targeting drug therapy for primary prevention of CHD. Framingham-based methods should incorporate HDL-cholesterol as the total : HDL cholesterol ratio. |
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Recent outcome trials with HMG Co-A reductase inhibitors (statins) have radically changed therapeutic policies for cholesterol reduction. Secondary prevention in those with established coronary or atherosclerotic disease and serum cholesterol
5.0 mmol/l is accepted as the first priority.1,2 However, guidance on statin treatment for primary prevention of coronary heart disease (CHD) is not uniform.3 When recommendations of the British Hyperlipidemia Association,4 European Joint Task Force,5 US National Cholesterol Education Program6 (NCEP), or a Standing Medical Advisory Committee7 (SMAC) are applied to British adults, they identify for treatment as few as 3.7% or as many as 13.5% of the population.3 Moreover, the individuals identified by the different guidelines are not the same.3 These inconsistencies could reflect differences in interpreting the evidence, differences in policy, or differences in methods of implementing policy.8 Interpretation of the evidence on statin treatment appears straightforward. Statins reduce major CHD events by 33% in those with LDL-cholesterol >3.2 mmol/l, and the relative risk reduction is constant between trials and in subgroups within trials.9,10 A constant relative risk reduction means that benefit from treatment is related to the absolute CHD risk. Absolute CHD risk therefore determines benefit to the individual (NNT), cost-effectiveness, the proportion of the population treated, and the total cost of statin treatment.9–11 A general consensus that statin treatment should be targeted at absolute CHD risk12 is echoed in all recent guidelines.1,2,5,7 The guidelines differ principally in the policies set for treatment, and on methods of identifying people to be treated. Two of them4,6 advise treatment based on thresholds of total or LDL-cholesterol plus counting of additional risk factors such as smoking, diabetes or hypertension. They do not state explicitly the CHD risk threshold identified, but it equates to a CHD event risk of approximately 2% per year, at least in middle-aged men.13,14 The other two guidelines5,7 target specified thresholds of absolute CHD risk, and advise formal risk estimation using simple versions of the Framingham risk function based on multiple risk factors. The European Joint Task Force guidelines1 advise treatment at a CHD event risk of 20% over 10 years (which equates to a risk of approximately 2% per year), or a CHD event risk of 20% by the age of 60 years, a recommendation apparently based on consensus. Standing Medical Advisory Committee guidance7 is based on the Sheffield table10,15 and advises primary prevention at a CHD event risk of 3% per year, a recommendation based on explicit analysis of the cost-effectiveness, total cost and volume of treatment for the British population.9,10
The appropriate policy on absolute CHD risk threshold may properly differ between countries depending on the population prevalence of CHD risk and availability of healthcare resources.9 Regardless of the policy adopted, what method will identify best those with the level of absolute CHD risk that warrants treatment? The method selected needs to be simple for acceptance in ordinary practice, accurate, and valid for estimating absolute CHD risk. When considering accuracy the balance between sensitivity and specificity is important. Statin treatment for primary prevention might now be justifiable on harm–benefit criteria at a CHD event risk as low as 0.6% per year.16 The emphasis should therefore be to identify and treat all those above the specified CHD risk threshold because no harm will come from treating `incorrectly' those below that threshold, provided their CHD risk is not very low. The method of identifying people for treatment should therefore have high sensitivity, with some sacrifice of specificity if necessary. We have compared the accuracy of five methods for estimating absolute CHD risk, one using a cholesterol threshold plus simple counting of risk factors,4,6,13 and four based on the Framingham risk function.17 Two of the Framingham-based methods use total cholesterol only (European Joint Task Force chart1,5 and Sheffield table10,15), and two use the total : HDL cholesterol ratio (New Zealand chart18 and modified Sheffield table19). The validity of risk estimates based on the Framingham function was examined by comparing them to estimates of CHD risk by the PROCAM risk function, which is derived from a prospective study in a German population.20
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Risk assessment methods
The European Joint Task Force chart1 estimates CHD risk using age, sex, systolic blood pressure, smoking and total cholesterol. The original Sheffield table10 used the same variables plus diabetes and left ventricular hypertrophy on ECG. In an updated Sheffield table,19 serum total : HDL cholesterol ratio replaces total cholesterol. The New Zealand chart18 also uses total : HDL cholesterol ratio and these other risk factors, but estimates cardiovascular event risk rather than CHD event risk. The method for cholesterol threshold plus counting of risk factors was that used in a subgroup analysis of WOSCOPS,13 requiring a total cholesterol
6.5 mmol/l plus any two additional CHD risk factors, with age >55 years in men counted as one risk factor. This is similar to recommendations by the British Hyperlipidemia Association4 and American National Cholesterol Education Program,6 although the latter is based on LDL-cholesterol rather than total cholesterol. The European and threshold plus counting methods cannot target a CHD event risk of 3% per year, and for the purpose of this analysis the Sheffield tables were recast to target a CHD event risk of 2% per year.9 The CHD risk of each patient was estimated using all five risk assessment methods in accordance with published instructions on their use. The external standard was CHD risk calculated by the full PROCAM risk function20 using age, blood pressure, smoking, diabetes, total and HDL-cholesterol, and family history. This analysis was restricted to men, because CHD risk in women could not be calculated by PROCAM at the time. CHD event risk was also calculated for men and women using the full Framingham risk function.17
Patients
Complete data to enable calculation of CHD risk by the PROCAM and Framingham risk functions were collected prospectively for 216 consecutive treated hypertensive patients who were aged 35–70 years, had total cholesterol 5.5 mmol/l, and were free of vascular complications by clinical evaluation or ECG. The variables measured and recorded were age, sex, smoking habit and family history (both by structured questions), blood pressure (mean of two measurements), diabetes (fasting glucose >7.8 mmol/l, or treated), and ECG left ventricular hypertrophy by criteria used in Framingham. The characteristics of the study population are shown in Table 1
. The mean CHD event risk calculated by the full Framingham function was 2.5% (SD 1.2%) per year in men (n=126); 1.4% (0.9%) per year in women (n=90); and 2.0% (1.2%) per year for all subjects.
|
Analysis
The principle endpoint was the accuracy of the different risk assessment methods in men for predicting a CHD event risk
2.0% per year calculated by the full PROCAM risk function. The predictive value of HDL-cholesterol (in the total : HDL cholesterol ratio) was examined further by comparing the accuracy of Sheffield tables based on total cholesterol alone (using population mean values for HDL-cholesterol) or the total : HDL cholesterol ratio for estimating CHD event risk 3.0% per year calculated by the full Framingham risk function. Sensitivity, specificity and 95%CIs were calculated by standard methods.21 |
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Comparison of methods
The sensitivity and specificity for each method, and the mean CHD event risks in those treated or not treated, are shown in Table 2
. The method based on cholesterol threshold 6.5 mmol/l plus two CHD risk factors had a sensitivity of 59% and specificity 63% versus PROCAM, with mean annual CHD event risk of 2.7% for those treated and 1.9% for those not treated. All four methods based on Framingham had sensitivity between 90–98% for predicting a CHD event risk of 2% per year by PROCAM. Those based on total cholesterol alone5,10 had lower specificity (37% and 43%) than those using the total : HDL cholesterol ratio18,19 (60% and 63%). The two methods based on total : HDL cholesterol ratio18,19 identified men with a mean CHD event rate of 3.0% per year for treatment, and with rates of 0.7 and 1.0% per year in those not treated. The scatters for the cholesterol threshold plus risk factor counting method and the Sheffield tables based on total cholesterol or the total : HDL cholesterol ratio are shown in Figure 1. The cholesterol threshold plus risk factor counting method failed to treat patients with high and even extremely high CHD risk, and often identified for treatment patients at very low CHD risk, below 0.6% per year. This method failed to classify five men for treatment despite them actually having an annual CHD risk >5% by the PROCAM function. These men had serum cholesterol <6.5 mmol/l but had numerous other risk factors. All five men were >60 years old; three had systolic blood pressure 160 mmHg; two were smokers; four were diabetic; three had a positive family history of ischaemic heart disease; and four had serum HDL concentration 
1.0 mmol/l. The table based on total cholesterol alone failed to treat some patients above the 2% per year threshold, and also treated patients with CHD risk below 0.6% per year. The table incorporating HDL-cholesterol in the ratio treated almost all patients with CHD event risk 2% per year, and did not treat any patient with CHD risk below 0.6% per year.
|
|
Importance of total : HDL cholesterol ratio
Figure 2
shows the outcome of targeting a CHD event risk of 3% per year, calculated by the full Framingham function, using Sheffield tables based on total cholesterol alone or the total : HDL cholesterol ratio in men and women combined. The latter table incorporates all the variables in the Framingham risk function, with only blood pressure approximated, and as anticipated the sensitivity (100%) and specificity (94%) are very high. Those identified for treatment have a mean CHD event risk much higher than that of those untreated (4.1% vs. 1.8% per year). Use of total cholesterol alone also identifies high (3.6%) and low (1.5%) risk groups, has no important effect on specificity (98%), but reduces the sensitivity markedly to 45%. Virtually all those treated have a CHD event risk 3% per year, but it fails to treat many patients who are above the preset threshold. Most but not all of the high-risk men not treated have CHD event risks between 3–4% per year.
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
The Framingham risk function has been available for several years, but was used little in ordinary practice until recently. In Britain a national survey of local policies for cholesterol management in 1994 showed that only 2% included a formal risk assessment method.22 However, methods based on Framingham are now in widespread use. Reasons include acceptance that individual benefit, cost-effectiveness, volume and total cost of statin treatment are determined by absolute CHD risk;9,11 and wide dissemination of simple Framingham-based methods in national and international guidelines.1,5,7,14,18,19 Absolute CHD risk is the correct starting point when considering statin treatment, not thresholds of cholesterol or other lipid fractions.9,23 Moreover, intuitive or informal assessment of absolute CHD risk by doctors is highly inaccurate.24 Without formal risk assessment, prescribing policies are very inconsistent, and also prejudiced against certain high-risk groups such as the overweight, smokers, and older people.25 Some method of formal CHD risk assessment is clearly essential for rational prescribing of statins.
Some guidelines use thresholds of total or LDL-cholesterol along with simple counting of risk factors,4,6 and indeed this policy is still advocated in the Monthly Index of Medical Specialities (MIMS),26 a prescribing aid widely used by British doctors. In our study, this method was distinctly less accurate than Framingham-based methods that count and weight risk factors for CHD. It failed to treat patients with high and even extremely high CHD risk, while identifying for treatment many with a CHD event risk below 0.6% per year, the present limit of outcome trial evidence for statins.16 Low accuracy was reported in previous studies,23,27 and accuracy declines with increasing age beyond 35 years in men and 55 years in women.27 Use of LDL-cholesterol or total cholesterol thresholds plus risk factor counting identifies a high proportion of the adult population for statin treatment,3,28,29 for example more than 20% of those aged over 45 years in the US.29 Many of those identified will have very low CHD risk, and many at high CHD risk will not be treated. Methods of lipid thresholds plus counting of risk factors continue to be used because it is thought that doctors may not understand newer methods of formal CHD risk estimation.30 That is not our experience. Provided the method of formal CHD risk assessment is simple, it is easier to classify patients by only two numbers, cholesterol and CHD event risk, than to juggle several categorical variables. Methods that use lipid thresholds plus counting of risk factors are unacceptably inaccurate for clinical practice.
A recent comparison of three risk-assessment methods (NCEP guidelines,6 Sheffield table using total cholesterol,10 and European guidelines5) was carried out in a series of 570 patients without clinical evidence of atherosclerosis referred to a lipid clinic.28 This showed that the different methods varied markedly in their assessment of CHD risk. However one of the methods tested10 was not used correctly, and in addition, specificity was entirely ignored. The present study confirms the high sensitivity of the Joint European Task Force chart, but shows that it has very low specificity.
The comparison of Framingham-based methods against CHD risk calculated by PROCAM shows that the Framingham function has external validity, adding to a large body of evidence that it predicts CHD risk accurately in British and northern European populations despite secular changes in CHD incidence.31 The simple Framingham-based methods are all capable of detecting most high-risk individuals for treatment, but incorporation of HDL-cholesterol, as the total : HDL cholesterol ratio, greatly enhances the accuracy of risk prediction with either the PROCAM or Framingham risk function as the gold standard. The Framingham function appears robust to many simplifications or approximations, for example omitting left ventricular hypertrophy or categorizing blood pressure, but not to omission of HDL-cholesterol concentration. British doctors may be relatively unfamiliar with the total : HDL cholesterol ratio, and many laboratories do not measure or report this parameter routinely, but the accuracy of CHD risk prediction and statin prescribing would be greatly enhanced by its use.32 LDL-cholesterol is favoured in the US and by some specialists in Britain and Europe because it is the proximate cause of CHD. However, use of LDL-cholesterol rather than total cholesterol does not improve risk prediction by the Framingham function,14 and therefore offers no advantage.
Any of the Framingham-based methods studied will improve the prescribing accuracy of doctors who are unable or unwilling to use the full Framingham risk function. The New Zealand chart18 estimates cardiovascular risk rather than CHD risk, a fact not widely appreciated in practice or even by experts. Despite this, it estimates CHD risk accurately, because cardiovascular risk and CHD risk correlate highly in the Framingham function, but with a ratio of approximately 4 : 3. Use of the New Zealand chart at a threshold of 2.0% per year in this study actually targets a CHD event rate of about 1.5% per year. The Sheffield tables10,15,19 are designed to assess CHD risk after control of any hypertension, because antihypertensive treatment is currently more cost-effective than statin treatment. They will be less accurate when hypertension is untreated or uncontrolled, although the Sheffield table using the total : HDL cholesterol ratio had a sensitivity of 89% and specificity of 97% for predicting CHD risk in a large population of diabetics not selected for blood pressure level.33 The estimates of sensitivity and specificity for the Framingham-based methods in the selected high-risk patients we studied should be generalizable, but the positive and negative predictive values for all these methods need to be examined in much larger and unselected samples of the general population.
There is general consensus that statin treatment should be targeted at absolute CHD risk,12 and at least in Britain and Europe, a consensus that formal methods of estimating CHD risk should be used.1,4,7 However, the problem of setting a policy remains—what threshold of CHD event risk should be set for primary prevention with statins? In the four guidelines assessed by Unwin et al.,3 the policy was set by consensus alone in three, but this was not so for guidance based on the Sheffield table.7,10 This was developed after the implications of statin treatment at different levels of CHD risk had been set out explicitly for examination and debate,9,10 an important fact that has been overlooked entirely by some.3,34 Statin prescribing in Britain is not constrained by benefit to the individual (NNT) or even primarily by cost-effectiveness. The main constraint by far is the high proportion of the adult population likely to need statin treatment and, stemming from this, the very high total cost of treatment and the resource implications for primary care.9,10 Complete implementation of statin treatment for secondary prevention plus primary prevention at a CHD event rate 3% will entail treating about 8% of all adults,9 and some believe that even this is not feasible. This policy still seems appropriate for the UK, but treatment at lower CHD risk levels may be appropriate in the future, and at present for other populations with a lower prevalence of CHD risk factors or CHD risk. The SMAC guidance7 was based on a Sheffield table using total cholesterol alone and cholesterol threshold 5.5 mmol/l, which actually extrapolated beyond the evidence available then for primary prevention. High specificity at the expense of sensitivity (Figure 2
) was appropriate at that time, but is now too conservative in light of new evidence on primary prevention.16 The total cholesterol threshold for treatment should be lowered to 5.0 mmol/l,16 and a method with high sensitivity for CHD risk prediction is required, at the expense of specificity if necessary. Methods based on the total : HDL cholesterol ratio (Figure 2) should be preferred.
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Wood D, De Backer G, Faergeman O, Graham I, Mancia G, Pyörälä K. Prevention of coronary heart disease in clinical practice. Recommendations of the Second Joint Task Force of European and other societies on coronary prevention. Eur Heart J 1998; 19:1434–503.
2.
Grundy SM, Balady GJ, Criqui MH, Fletcher G, Greenland P, Hiratzka LF, Houston-Miller N, Kris-Etherton P, Krumholz HM, LaRosa J, Ockene IS, Pearson TA, Reed J, Washington R, Smith SC. Primary prevention of coronary heart disease: guidance from Framingham. A statement for healthcare professionals from the AHA Task Force on risk reduction. Circulation 1998; 97:1876–87.
3.
Unwin N, Thomson R, O'Byrne AM, Laker M, Armstrong H. Implications of applying widely accepted cholesterol screening and management guidelines to a British adult population: cross sectional study of cardiovascular disease and risk factors. Br Med J 1998; 317:1125–30.
4. Betteridge D, Dodson P, Durrington P, Hughes E, Laker M, Nicholls D, et al. Management of hyperlipidaemia: guidelines of the British Hyperlipidaemia Association. Postgrad Med J 1993; 69:359–69.[Abstract]
5.
Pyörälä K, De Backer G, Graham I, Poole-Wilson P, Wood D, on behalf of the task force. Prevention of coronary heart disease in clinical practice: recommendations of the task force of the European Society of Cardiology, European Atherosclerosis Society and European Society of Hypertension. Eur Heart J 1994; 15:1300–31.
6. National Cholesterol Education Programme. Second report of the expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult treatment panel 2). Circulation 1994; 89:1333–45.[Medline]
7. Standing Medical Advisory Committee. The use of statins. London, Standing Medical Advisory Committee 1997.
8.
Fahey T. Assessing heart disease risk in primary care. Cholesterol lowering should be just one part of a multiple risk factor intervention. Br Med J 1998; 317:1093–4.
9. Haq IU, Ramsay LE, Pickin DM, Yeo WW, Jackson PR, Payne JN. Lipid-lowering for the prevention of coronary heart disease: what policy now? Clin Sci 1996; 91:399–413.[Medline]
10. Ramsay LE, Haq IU, Jackson PR, Yeo WW, Pickin DM, Payne JN. Targeting lipid-lowering drug therapy for primary prevention of coronary disease: an updated Sheffield table. Lancet 1996; 348:387–8.[ISI][Medline]
11. NHS Centre for Reviews and Dissemination, University of York. Cholesterol and coronary heart disease: screening and treatment. Effective Health Care 1998; 4:1–16.
12.
Grundy SM. Statin trials and goals of cholesterol-lowering therapy. Circulation 1998; 97:1436–9.
13. West of Scotland Coronary Prevention Group. West of Scotland Coronary Prevention Study: identification of high-risk groups and comparison with other cardiovascular intervention trials. Lancet 1996; 348:1339–42.[ISI][Medline]
14.
Wilson PWF, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998; 97:1837–47.
15. Haq IU, Jackson PR, Yeo WW, Ramsay LE. Sheffield risk and treatment table for cholesterol lowering for primary prevention of coronary heart disease. Lancet 1995; 346:1467–71.[ISI][Medline]
16.
Downs GR, Clearfield M, Weiss S, Whitney E, Shapiro DR, Beere PA, Langendorfer A, Stein EA, Kruyer W, Gotto AM, for the AFCAPS/TexCAPS Research Group. Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels: results of AFCAPS/TEXCAPS. Air Force/Texas Coronary Atherosclerosis Study. JAMA 1998; 279:1615–22.
17. Anderson KM, Odell PM, Wilson PWF, Kannel WB. Cardiovascular disease risk profiles. Am Heart J 1991; 121:293–8[ISI][Medline]
18. Dyslipidaemia Advisory Group on behalf of the scientific committee of The National Heart Foundation of New Zealand. 1996 National Heart Foundation clinical guidelines for the assessment and management of dyslipidaemia. N Z Med J 1996; 109:224–32.[ISI][Medline]
19. Haq IU, Jackson PR, Yeo WW, Ramsay LE. A comparison of methods for targeting CHD risk for primary prevention. Heart 1997; 77(Suppl 1):36.
20. Assman G, ed. Results and conclusions of the prospective cardiovascular Münster (PROCAM) Study. In: Lipid metabolism disorders and coronary heart disease: primary prevention, diagnosis and therapy guidelines for general practice. München, MMV-Medizin-Verlag, 1993.
21. Armitage P, Berry G. Statistical methods in medical research, 2nd edn. Oxford, Blackwell Scientific, 1987.
22.
Raine R, Streetly A, Davis AM. Variation in local policies and guidelines for cholesterol management: national survey. Br Med J 1996; 313:1368–9.
23. Grover SA, Coupal L. Hu X-P. Identifying adults at increased risk of coronary disease. How well do the current cholesterol guidelines work? JAMA 1995; 274:801–6.[Abstract]
24.
Grover SA, Lowensteyn I, Esrey KL, Steinart Y, Joseoph L, Abrahamowicz M. Do doctors accurately assess coronary risk in their patients? Preliminary results of the coronary health assessment study. Br Med J 1995; 310:975–8
25. Evans J ST BT, Harries C, Dennis I, Dean J. General practitioners' tacit and stated policies in the prescription of lipid lowering agents. Br Med J 1994; 45:15–18.
26. Monthly Index of Medical Specialities. December 1998.
27.
Avins AL, Browner WS. Improving the prediction of coronary heart disease to aid in the management of high cholesterol levels. What a difference a decade makes. JAMA 1998; 279:445–9.
28. Durrington PN, Prais H, Bhatnagar D, France M, Cowley V, Khan J, Morgan J. Indication for cholesterol-lowering medication: comparison of risk-assessment methods. Lancet 1999; 353: 278–81.[ISI][Medline]
29. Hoerger TJ, Bala MV, Bray JW, Wilcosky TC, LaRosa J. Treatment patterns and distribution of low-density lipoprotein cholesterol levels in treatment-eligible United States adults. Am J Cardiol 1998; 82:61–5.[ISI][Medline]
30.
Greenland P, Grundy S, Pasternak RC, Lenfant C. Problems on the pathway from risk assessment to risk reduction. Circulation 1998; 97:1761–2.
31.
Haq IU, Ramsay LE, Yeo WW, Jackson PR, Wallis EJ. Is the Framingham risk function valid for northern European populations? A comparison of methods for estimating absolute coronary risk in high risk men. Heart 1999; 81: 40–6.
32. Kinosian B, Glick H, Preiss L, Puder KL. Cholesterol and coronary heart disease: predicting risks in men by changes in levels and ratios. J Invest Med 1995; 43:443–50.[ISI][Medline]
33. Bayly GR, Haddon A, Bartlett WA, Jones AF. Laboratory based calculation of coronary heart disease risk—a simple and accurate method of targeting primary prevention. Proc Roy Coll Phys Edin 1999; in press.
34.
Freemantle N, Barbour R, Johnson R, Marchment M, Kennedy A. The use of statins: a case of misleading priorities? National guidance that does not link costs and benefits is worthless. Br Med J 1997; 315:826–8.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. Ballantyne, B. Arroll, and J. Shepherd Lipids and CVD management: towards a global consensus Eur. Heart J., November 1, 2005; 26(21): 2224 - 2231. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. S. Ford, W. H. Giles, and A. H. Mokdad The distribution of 10-Year risk for coronary heart disease among U.S. adults: Findings from the National Health and Nutrition Examination Survey III J. Am. Coll. Cardiol., May 19, 2004; 43(10): 1791 - 1796. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. J. Pashkow The prudent person's paradox J. Am. Coll. Cardiol., May 7, 2003; 41(9): 1480 - 1481. [Full Text] [PDF]
|
|
|
|
|
|
H.-W. Hense, H. Schulte, H. Lowel, G. Assmann, and U. Keil Framingham risk function overestimates risk of coronary heart disease in men and women from Germany--results from the MONICA Augsburg and the PROCAM cohorts Eur. Heart J., May 2, 2003; 24(10): 937 - 945. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K S Rabindranath, N R Anderson, R Gama, and M R Holland Comparative evaluation of the new Sheffield table and the modified joint British societies coronary risk prediction chart against a laboratory based risk score calculation Postgrad. Med. J., May 1, 2002; 78(919): 269 - 272. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I U Haq, L E Ramsay, E J Wallis, C G Isles, L D Ritchie, and P R Jackson Population implications of lipid lowering for prevention of coronary heart disease: data from the 1995 Scottish health survey Heart, September 1, 2001; 86(3): 289 - 295. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P S Sanmuganathan, P Ghahramani, P R Jackson, E J Wallis, and L E Ramsay Aspirin for primary prevention of coronary heart disease: safety and absolute benefit related to coronary risk derived from meta-analysis of randomised trials Heart, March 1, 2001; 85(3): 265 - 271. [Abstract] [Full Text]
|
|
|
|
 |
|
 E. J Wallis, L. E Ramsay, I. U. Haq, P. Ghahramani, P. R Jackson, K. Rowland-Yeo, and W. W Yeo Coronary and cardiovascular risk estimation for primary prevention: validation of a new Sheffield table in the 1995 Scottish health survey population BMJ, March 11, 2000; 320(7236): 671 - 676. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||