Sheffield risk and treatment table for cholesterol lowering for primary prevention of coronary heart disease

Sheffield risk and treatment table for cholesterol primary prevention of coronary heart disease

Summary

Introduction

When used for the secondary prevention of coronary heart an inhibitor of disease, treatment with reductase in results hydroxymethylglutaryl-coenzyme-A risk worthwhile benefit that clearly exceeds any in patients

High serum contributing

whose risk of coronary death is 1·5% or more per year. This evidence can be extrapolated logically to primary prevention of coronary disease provided that treatment is targeted at those with similar or higher risk. We present a table that refines previously proposed methods of risk prediction. The table identifies subjects who have the specified degree of coronary risk; shows the serum cholesterol concentration that confers that degree of risk in the individual; and identifies subjects who will not have this degree of risk, irrespective of their cholesterol concentration. It is simple enough for use in ordinary practice. The table highlights the predominant effect of age on coronary risk; a person who is free of vascular disease and younger than 52 years is unlikely to have the specified degree of risk. Even in older people (60-70 years) several risk factors are generally required to attain this degree of risk. Some people are candidates for lipidlowering drug treatment with serum cholesterol as low as 5·5 mmol/L, whereas others with cholesterol as high as 9·0 mmol/L are not. Although cholesterol lowering is a powerful method for preventing coronary events in people at high risk, cholesterol measurement by itself is not a good way to identify those with high coronary risk. The method can be adapted readily to target a different level of coronary risk as new evidence on the benefit and risk of treatment becomes available.

Section of Clinical Pharmacology and Therapeutics, Department of Medicine and Pharmacology, Royal Hallamshire Hospital, Sheffield S10 2JF, UK (I U Haq MRCP, P R Jackson FRCP, W W Yeo MRCP, Prof L E Ramsay FRCP)

Correspondence to:

Prof L E

Ramsay

lowering for

cholesterol to

coronary

is

one I

risk,’

factor reduction of

important and

cholesterol concentration prevents coronary events.2-5 Implementation of cholesterol lowering in ordinary practice has, however, been hampered by the small effect of dietary change on serum cholesterol and concerns that the benefit from some classes of lipid-lowering drugs might be offset by an increase in non-coronary

mortality. 7-9 The findings of the Scandinavian Simvastatin Survival (4S) study4 substantially shifted the balance of benefit and risk for lipid-lowering drug therapy. In that trial4 survivors of myocardial infarction with total cholesterol concentrations higher than 5-5 mmol/L who were treated with simvastatin, an inhibitor of hydroxymethylglutarylcoenzyme-A (HMG-Co-A) reductase had clinically significant reductions in coronary events, cardiovascular events, and all-cause mortality. 16 patients had to be treated with simvastatin for 5 years to prevent one major coronary event. There was no apparent increase in noncoronary mortality to offset the beneficial effects, and no evidence that aggressive lipid lowering cerebral promoted haemorrhage.4 Patients who survive myocardial infarction and have serum cholesterol of 5-5 mmol/L or higher should now be treated with an HMGCo-A reductase inhibitor.1O These findings will probably be extrapolated to all forms of coronary disease,lo for example stable angina. The evidence is less strong for such patients, who have a lower coronary risk, but extrapolation seems justifiable because the margin of benefit over risk in the 4S study was large, and because the HMG-Co-A reductase inhibitors are safe, effective, and well-tolerated.4,11 Patients with symptomatic carotid disease or peripheral vascular disease have a coronary risk at least equal to that of patients with stable angina,"-’3 and a case can therefore be made for cholesterol reduction by an HMG-Co-A reductase inhibitor for secondary prevention in any patient with clinically overt atherosclerotic vascular disease. An important question is whether and how these new findings can be extrapolated to primary prevention of coronary disease. Understandably, caution has been advised in the use of lipid-lowering drugs for primary prevention,’" in agreement with guidelines for managing hyperlipidaemia.14,15 HMG-Co-A reductase inhibitor treatment might logically be extended to primary prevention, however, provided that it is targeted strictly at patients with a risk of coronary disease similar to or higher than that of patients in the 4S study. Placebotreated patients in that study4had a risk of coronary death of about 1-5% per year, and the balance of benefit and risk of treatment is likely to be equally favourable in subjects with that degree of coronary risk irrespective of the reasons for the high risk. Subjects without vascular disease may have a coronary risk higher than that of the

there

was

1467

Table: Sheffield risk and treatment table for primary prevention of coronary heart disease A patient whose value falls in the unshaded area has an estimated risk of coronary death of less than 1.5%.

survivors of myocardial infarction entered in the 4S study. Cholesterol reduction for primary prevention does prevent coronary eventsand any risk associated with HMG-Co-A reductase inhibitors is likely to be the same for primary or secondary prevention. A favourable benefit-risk balance should therefore result from primary prevention in appropriately selected high-risk subjects. The key to rational use of HMG-Co-A reductase inhibitors for primary prevention will be accurate targeting of treatment at subjects with a specified high risk of coronary events. This concept is embraced in the latest management guidelines,14,15 but in practice doctors are unable to assess coronary risk accurately.16 The latest European guidelines14 include a table based on the 1468

Framingham risk function" to help doctors estimate coronary risk more precisely. We propose here a refinement of this approach. The method of risk estimation set out in the European guidelines 14 requires cholesterol measurement in all adults. This approach implies that cholesterol’s predictive value for coronary disease is much greater in relation to other risk factors than it actually is. Let us consider two

examples. A 28-year-old woman has systolic blood pressure 160 mm Hg, no diabetes, and no left ventricular hypertrophy, and is a non-smoker. Serum cholesterol, measured as advised in the guidelines, is 8-6 mmol/L. After dietary advice the concentration falls to 8-4

mmol/L." Despite the caution advised in the guidelines,14,15 this woman is likely to be treated with a lipid-lowering drug. However, her estimated risk of a coronary event is only 0-3% in 10 years." To prevent one coronary event 670 patients like this would need treatment for 10 years. Such treatment is not justified by the evidence available. By contrast let us look at a 69-year-old man with systolic blood pressure 160 mm Hg, electrocardiographic (ECG) evidence of left ventricular hypertrophy, and diabetes, who persists in smoking cigarettes. His serum cholesterol is consistently 5-5mmol/L. This man would not be a candidate for lipid-lowering drug therapy by current guidelines,14,15 yet his estimated risk of a coronary event is 58-1% in 10 years." One coronary event would be prevented by treating only three such men with an HMGCo-A reductase inhibitor for 10 years. Such treatment is

justifiable. The first case illustrates the important point that some individuals cannot possibly reach a prespecified high risk of coronary disease irrespective of their cholesterol concentration. The second case highlights a further feature of the guidelines’"’’—they cite fixed cut-off levels for cholesterol’4 or low-density-lipoprotein (LDL) cholesterol’S above which lipid-lowering drug therapy should be considered. However, the cholesterol concentration that confers a specified risk of coronary disease varies among individuals. It is a function of other risk factors for coronary heart disease and can be calculated readily. Our aims in producing the Sheffield risk and treatment table were to identify for primary prevention individuals with a specified coronary risk, in whom treatment with an HMG-Co-A reductase inhibitor could be justified; to find the cholesterol concentration for such individuals that confers the specified degree of coronary risk; and to show those individuals who, whatever their cholesterol concentration, will not have a coronary risk that justifies

lipid-lowering drug therapy. Depending on the policy adopted, cholesterol measurement might be avoided in the latter subjects. Our other aim was to keep the table simple for use in ordinary practice. Methods The Sheffield table is based on a logistic regression equation predicting coronary risk derived from the Framingham population." Coronary deaths rather than coronary events were targeted because coronary death is defined more uniformly than coronary events in epidemiological and intervention studies. The target risk selected was a coronary death risk of 1-5% per year, with the risk of placebo-treated patients in the 4S study4 as a yardstick. For this degree of risk, treatment with an HMG-Co-A reductase inhibitor seems justifiable even from a conservative viewpoint. The Sheffield table does not indicate a degree of risk that would justify treatment with other classes of lipid-lowering drugs, since benefit may exceed risk with other drug classes only when the risk of coronary death is 3-0% or more per year.9 Certain assumptions were made to simplify the table. Hypertension was reduced to a dichotomous variable, present or absent, with the assumption that if present systolic blood pressure will be controlled to 160 mm Hg’8 and if absent systolic blood pressure will be 139 mm Hg, the average for middle-aged British men.19 High-density-lipoprotein (HDL) cholesterol was not included in the table, and population mean values of 1-15 mmol/L for men20 and 1.40 mmol/L for women were ascribed.2O Left ventricular hypertrophy was defined according to the ECG criteria used by the Framingham investigators-an increased Rwave potential and flattened or inverted T waves in the left precordial leads.

Two boundaries for cholesterol had to be set. The lower boundary is the cholesterol concentration below which lipidlowering drug therapy is not currently indicated. This was set at 5-5 mmol/L on the basis of the 4S study.4 The upper boundary for cholesterol indicates the highest concentration that is likely to be found in the relevant population. We set this boundary at three standard deviations above the mean (6-3 mmol/L for men, 6-7 mmol/L for women) for a middle-aged hypertensive population in the UK21 (ie, 9-3 mmol/L for men, 10-1mmol/L for women). Cholesterol values higher than this would be missed in the table for 1 in 740 hypertensive subjects. Since the mean cholesterol value chosen is higher than those of the UK population (5-9 mmol/L22), UK diabetic patients (5-7 mmol/U3), younger subjects,22 and the populations of most other countries,>4 the chance of missing cholesterol values above the upper boundary in general use should be much smaller than 1 in 740. The upper limit for age was set at 70 years, as in the 4S study,’ and because evidence is lacking for benefit from cholesterol reduction in older subjects.25 We calculated, using the Framingham risk function,"the cholesterol concentration that would confer a risk of coronary death of 1-5% per year, from the variables age, sex, smoking (present/absent), hypertension (present/absent), diabetes (present/absent), left ventricular hypertrophy (present/absent), and the population mean HDL cholesterol values. Since the Framingham equation is accurate for prediction between 4 and 12 years," a risk of coronary death of 7-5% over 5 years was calculated and divided by 5 to obtain the coronary death risk of 1-5% per year. This approach assumes that the distribution of coronary deaths over 5 years is uniform. When the cholesterol value predicting this degree of risk is lower than 5-5 mmol/L, it is entered in the Sheffield table as 5-5 mmol/L. When it exceeds the upper boundary for cholesterol (93mmol/L for men, 10-mmol/L for women), the table shows no

entry.

Results To

table, the correct column for hypertension, smoking, diabetes, and left ventricular hypertrophy and the row showing the age of the individual are identified (table). The cholesterol concentration at the intersection of the appropriate column and row is read. If there is a number entered, serum cholesterol should be measured. When the cholesterol value is repeatedly at or above the concentration shown, the risk of coronary death is predicted to be 1-5% or more per year and treatment with an HMG-Co-A reductase inhibitor may be justified. If the intersection of the appropriate column and row shows no entry for cholesterol, the estimated risk of coronary death use

the

is lower than 1-5% per year for any cholesterol concentration likely to be found in the population. Such individuals are unlikely to be candidates for treatment with an HMG-Co-A reductase inhibitor and, depending on the strategy adopted, cholesterol measurement may be

avoided. Use of the table Inspection of the table shows that men are at higher risk than women, as expected. Multiple risk factors in addition to high cholesterol are generally required to reach the specified degree of coronary risk. Of particular note is the profound effect of age in determining coronary risk. Men younger than 52 and women younger than 54 who are free of vascular disease are unlikely to reach a risk of coronary death of 1-5% or more per year at any realistic cholesterol concentration. This feature underlines the caution expressed in the latest guidelines 14, 15 against the use of lipid-lowering agents in younger subjects, and sets out the reason explicitly. There might be concern that 1469

failure to screen young people might result in irreversible vascular damage before they reach the age range targeted in the table. Somewhat surprisingly, this appears not to be the case. In the 4S study and previous intervention trials5 the coronary risk attributable to high cholesterol, in epidemiological terms, was reversed entirely by lipidlowering therapy after only 2 years of treatment. The implication is that 10 or 20 years of treatment is not needed to prevent coronary events. The table can be used readily to look forward and advise patients at what age they should have a cholesterol measurement and, if necessary, treatment. In contrast to the low coronary risk in the young, people aged 60-70 years who have several risk factors not uncommonly have a coronary risk that may justify treatment even with cholesterol concentrations as low as 5-5 mmol/L. The implications for coronary risk and its management if the age range was extended above 70 years

are

sobering.

We must emphasise strongly that the table estimates risk only for primary prevention. Most patients who already have clinically overt vascular disease have a much higher coronary risk, and will often be candidates for

secondary prevention with lipid-lowering drugs even in the absence of multiple risk factors. In the setting of primary prevention, other remedial risk factors should be managed appropriately. In particular giving up cigarette smoking has a major impact on coronary risk. Left ventricular hypertrophy can be assumed to be absent if there is no hypertension. Secondary causes of hyperlipidaemia should be borne in mind, although routine investigation to exclude these factors is probably not indicated.’-6 The serum cholesterol value used for decision-making should not be a single value, but the average of two4 or preferably morel1 measurements. Risk prediction is not an exact science and we emphasise that use of the Sheffield table and final treatment decisions may and indeed should be influenced by individual considerations. For example, it is not our intention that a 25-year-old man with a family history of coronary heart disease and tendon xanthomata should not have cholesterol measured because he does not the table. People from families with familial hyperlipidaemias will remain candidates for screening. Factors to be taken into account for final decisions on screening or treatment should include family history, an HDL-cholesterol concentration substantially different from the average, uncontrollable hypertension, and the personal wishes of the individual. In cases of difficulty an accurate estimate of coronary risk can be obtained by use of the full Framingham risk function. 17 The table can be used either as a guide to decisions on treatment alone, or as a guide to decisions on screening and treatment, according to the policy that individual doctors choose to adopt. This decision will depend on their view of the efficacy of diet change for lowering cholesterol. The latest guidelines 14,11 still hold that diet change is the cornerstone of management. These guidelines therefore advise cholesterol measurement in all adults and advise rigorous dietary intervention. If this policy is adopted, the table can be used to assess whether lipid-lowering drug therapy is indicated by taking the average cholesterol value after a trial of dietary treatment. We believe, based on an overview of long-term controlled trials of dietary reduction of cholesterol," that acceptable dietary change has very little effect on serum cholesterol, whereas diets that lower cholesterol effectively are unpalatable. Those who share our view can use the table appear

1470

on

avoid measuring cholesterol when they expect no benefit. One potential difficulty with use of the table concerns decisions to treat relatively low cholesterol concentrations in some subjects and not to treat high concentrations in others. With the table it is quite possible to find a cholesterol concentration of 8-0 mmol/L, for example, and yet conclude that lipid-lowering drug therapy is not indicated. This feature will require a major change in thinking on the part of both doctors and patients, to think only of coronary risk and not of cholesterol. "Your cholesterol is high, but your coronary risk is not" is often a correct summary of the clinical circumstances. to

Discussion

approach is an extension and refinement of current thinking that the management of hyperlipidaemia 14 and hypertension27 should be targeted at overall risk and not simply at cholesterol or blood pressure. Two points deserve particular discussion-the accuracy of risk prediction and the level of coronary risk to be targeted. The Sheffield table is based on the Framingham risk function, as are many other methods of risk prediction,14,27 partly because the Framingham function is geared specifically to primary prevention, and because it includes women, unlike many other risk prediction methods.28-30 However, the Framingham risk function will not predict risk precisely for individuals, even in the original population. Rather it predicts the average risk associated with a set of risk factors. Furthermore, it is not Our

established

that the risk function derived from Framingham will be applicable to other populations. When we calculated coronary risk for men from risk functions derived from British28 and German30 populations, and compared these estimates with the coronary risk calculated from the Framingham function, there was close agreement (unpublished); this finding suggests that prediction from Framingham is valid for these European populations, at least in men. In the interests of simplicity we have reduced the accuracy of prediction further by adopting approximations for systolic blood pressure and HDL cholesterol. The approximation for blood pressure is likely to overestimate risk slightly on average. We are investigating whether these approximations will result in important misclassification of individuals’ risk. Meanwhile we re-emphasise that individual considerations should override the table whenever that seems appropriate. The degree of risk targeted is open to debate. The table identifies subjects who have an average risk of coronary death higher than 1-5% per year, so it is conservative in comparison with the placebo-treated patients in the 4S study, who were used as the yardstick. The benefit of treatment with an HMG-Co-A reductase inhibitor is therefore expected to exceed any risk substantially in such individuals. The degree of risk targeted is broadly similar to that used in the latest European guidelines,14 which advise that lipid-lowering drug treatment should be considered when the risk of a coronary event is predicted at 20-40% over 10 years. Assuming that about half of all coronary events are fatal, this risk equates to a risk of coronary death of 1-2% per year. The risk targeted in our table is in the middle of this range, and it is therefore also slightly more conservative than the European guidelines. Doctors with a more optimistic view of the evidence available will doubtless continue to screen and treat more

than the table would advise. Furthermore, evidence from new trials of primary prevention10 may indicate that a different degree of coronary risk should be

aggressively

targeted. The precise

risk targeted is unimportant in the sense that the table can readily be recast to target any specified degree of risk as new evidence becomes available. What does matter is the principle that such a simple table can be produced. We believe that this approach is a useful advance on other proposed methods. 14 It will require doctors and the public to think in terms of measuring coronary risk and not measuring cholesterol-but, of course, coronary risk and the prevention of coronary disease are the variables of interest. References 1 Kannel WB, Castelli WP, Gordon T, McNamara PM. Serum cholesterol, lipoproteins, and the risk of coronary heart disease. Ann Intern Med 1971; 74: 1-12. 2 Lipid Research Clinics Programme. The lipid research clinics coronary primary prevention trial results 1: reduction in incidence of coronary heart disease. JAMA 1984; 251: 351-64. 3 Manninen V, Elo MO, Frick MH, et al. Lipid alterations and decline in the incidence of coronary heart disease in the Helsinki Heart Study. JAMA 1988; 260: 641-51. 4 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: 1383-89. 5 Law MR, Wald NJ, Thompson SG. By how much and how quickly does reduction in serum cholesterol concentration lower risk of ischaemic heart disease? BMJ 1994; 308: 367-72. 6 Ramsay LE, Yeo WW, Jackson PR. Dietary reduction of serum cholesterol concentration: time to think again. BMJ 1991; 303: 953-57. 7 Muldoon MF, Manuck

8

9

10 11

12

SB, Matthews KA. Lowering cholesterol concentrations and mortality: a quantitative review of primary prevention trials. BMJ 1990; 301: 309-14. Holme I. An analysis of randomized trials evaluating the effect of cholesterol reduction on total mortality and coronary heart disease incidence. Circulation 1990; 82: 1916-24. Davey Smith G, Song F, Sheldon TA. Cholesterol lowering and mortality: the importance of considering initial level of risk. BMJ 1993; 306: 1367-73. Oliver M, Poole-Wilson P, Shepherd J, Tikkanen MJ. Lower patient’s cholesterol now. BMJ 1995; 310: 1280-81. Andrade SE, Walker AM, Gottlieb LK, et al. Discontinuation of antihyperlipidaemic drugs-do rates reported in clinical trials reflect rates in primary care settings? N Engl J Med 1995; 332: 1125-31. UK-TIA Study Group. The UK-TIA Aspirin Trial: the interim results. BMJ 1988; 296: 316-20.

13

Criqui MH, Langer RD, Fronek A, years in patients with 326: 381-86.

peripheral

et al. Mortality over a period of 10 arterial disease. N Engl J Med 1992;

De Backer G, Graham I, et al. 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. 15 Anonymous. National Cholesterol Education Program. Second report of the expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult treatment panel II). Circulation 1994; 89: 1333-45. 16 Grover SA, Lowensteyn I, Esrey KL, Steinert Y, Joseph L, Abrahamowicz M. Do doctors accurately assess coronary risk in their patients? Preliminary results of the coronary health assessment study. 14

Pyörälä K,

BMJ 1995; 310: 975-78. 17 Anderson KM, Odell PM, Wilson PWF, Kannel WB. Cardiovascular disease risk profiles. Am Heart J 1991; 121: 293-98.

18 Sever P, Beevers G, Bulpitt C, et al. Management guidelines in essential hypertension: report of the second working party of the British Hypertension Society. BMJ 1993; 306: 983-87. 19 Whincup PH, Perry IJ, Shaper AG. Regional differences in blood pressure in developed countries: fetal influences. In: Swales JD, ed. Textbook of hypertension. Oxford: Blackwell Scientific Publications, 1994: 38. 20 Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR. High density lipoprotein as a protective factor against coronary heart disease. Am J Med 1977; 62: 707-14. 21 Thompson SG, Pocock SJ. The variability of serum cholesterol measurements: implications for screening and monitoring. J Clin Epidemiol 1990; 43: 783-89. 22 Mann JI, Lewis B, Shepherd J, et al. Blood lipid concentrations and other cardiovascular risk factors: distribution, prevalence, and detection in Britain. BMJ 1988; 296: 1702-06. 23 Hypertension in Diabetes Study Group. Hypertension in Diabetes Study (HDS) 1: prevalence of hypertension in newly presenting type 2 diabetic patients and the association with risk factors for cardiovascular and diabetic complications. J Hypertens 1993; 11: 309-17. 24 World Health Organization. World health statistics annual. Geneva: World Health Organization, 1989. 25 Krumholz HM, Seeman TE, Merrill SS, et al. Lack of association between cholesterol and coronary heart disease mortality and morbidity and all-cause mortality in persons older than 70 years. JAMA 1994; 272: 1335-40. 26 Evans P, Pereira Gray D. Value of screening for secondary causes of hyperlipidaemia in general practice. BMJ 1994; 309: 509-10. 27 Jackson R, Barham P, Bills J, et al. Management of raised blood pressure in New Zealand: a discussion document. BMJ 1993; 307: 107-10. 28 Tunstall-Pedoe H. The Dundee coronary risk-disk for management of change in risk factors. BMJ 1991; 303: 744-47. 29 Shaper AG, Pocock SJ, Phillips AN, Walker M. Identifying men at high risk of heart attacks: strategy for use in general practice. BMJ 1986; 293: 474-79. 30 Assmann G, Schulte H. Procam trial. Panscienta Verlag, HedingenZürich, 1986.

1471