Relevant BNF section: 2.12
In England, only 30% of patients with established coronary heart disease (CHD) and raised serum lipids, and fewer than 4% of individuals eligible for primary prevention, receive lipid-lowering therapy.1 Target total cholesterol concentrations are achieved in fewer than 50% of patients who do receive such treatment.1 Here, we review the use of statin therapy in the prevention of CHD events.
Statistics from Altmetric.com
Relevant BNF section: 2.12
Elevated serum concentrations of total cholesterol, low-density-lipoprotein (LDL) cholesterol and total triglycerides are associated with increased CHD risk. High-density lipoproteins (HDLs), however, confer protection, and in general, CHD risk rises as serum HDL-cholesterol concentration falls or as the ratio of total cholesterol to HDL-cholesterol (TC:HDL-C) rises. Serum lipids are, however, only one of several influences crucial in assessing a patient's absolute risk of developing complications of, or dying from, CHD within a given period. Others include age, gender, family or personal history of atherosclerotic disease, hypertension, presence of diabetes mellitus, ethnic group, smoking status and other lifestyle factors, socio-economic status, and obesity. The risk is particularly high (above 30% over 10 years) in patients who already have clinically obvious atherosclerotic disease (e.g. history of myocardial infarction, angina or coronary revascularisation, peripheral vasular disease, transient ischaemic attack or ischaemic stroke).
Benefits of statins
The efficacy of statins in the prevention of CHD events has been demonstrated in five randomised placebo-controlled trials,2–6 involving a total of 30,817 patients (87% men). In these studies, daily treatment for around 5 years with pravastatin 40mg,3,5,6 simvastatin 10-40mg4 or lovastatin 20-40mg2 reduced the risk of developing major coronary events by 34% (95% CI 23-43%)7 in primary prevention2,3 and by 30% (95% CI 24-35%)7 in secondary prevention,4–6 with benefits becoming obvious by 6 months.2,3 Overall mortality fell in secondary prevention (odds ratio 0.77, 95% CI 0.70-0.85), but not in primary prevention studies.7 Statin therapy was associated with mean reductions in serum concentrations of total cholesterol (by 20%), LDL-cholesterol (by 28%) and triglycerides (by 13%), and a 5% increase in HDL-cholesterol concentration.7 Statin therapy, taken for up to around 5 years, also seems to reduce the risk of stroke by 19-32% in patients with or without evidence of CHD.4–6,8,9 Treatment benefits with statins appear similar in men and women, and are independent of age up to 75 years,7,10 but data are lacking for patients above this age.
Who should receive statins?
Statins provide useful preventative benefit for patients with clinically obvious atherosclerotic disease, such as CHD, and for those without such features but with an absolute risk of 15% or more over 10 years.2–6,11 However, treating all those with a risk of 15% or more over 10 years would in effect include around 25% of UK adults12 and is not achievable with current NHS funding. A more realistic approach, therefore, is to ensure that all those with an absolute risk above 30% over 10 years receive optimum statin therapy, with appropriate monitoring of lipid concentrations and advice on non-drug measures (see pages 21-24, this issue), and extending statin therapy to remaining individuals with a risk level of at least 15% over 10 years, as resources permit.11
These objectives involve, as a first priority, identifying all patients with clinically obvious atherosclerotic disease, who do not need formal assessment of absolute coronary risk before starting secondary prevention interventions.11,13 It also involves identifying people without such clinical features, but who nonetheless are likely to be at increased CHD risk, such as those with a family history of premature CHD, clinical signs of hyperlipidaemia, or hypertension, or who smoke. These individuals require formal assessment of absolute risk.
Special consideration needs to be given to certain groups. These include: patients with diabetes mellitus (whose risk of developing and dying from CHD is around 2-5 times that of people without diabetes); those of South Asian descent (who typically have a 40% greater risk of developing CHD than white UK populations); and patients with familial hypercholesterolaemia (who are at particularly high risk of dying from CHD).13 Families of patients with inherited dyslipidaemia, particularly familial hypercholesterolaemia, including children, deserve similar attention.14
Patients with clinically overt atherosclerotic disease should be started on a statin. The drug should be given, with dietary advice, to lower serum total cholesterol concentrations to below 5mmol/L or by 20-25% (or LDL-cholesterol concentrations to below 3mmol/L or by 30%), whichever results in the lower concentration.13,15
In patients without clinically overt atherosclerotic disease, but with an estimated CHD risk of above 30% over 10 years, statins should be used to lower serum total cholesterol concentrations to below 5mmol/L or by 20-25% (or LDL-cholesterol concentrations to below 3mmol/L or by 30%), whichever results in the lower concentration.
Whether or not treatment with a statin is indicated, appropriate lifestyle modifications should be identified and continued indefinitely in all patients worthy of risk assessment, even if their risk is below 30% over 10 years.16 In primary prevention, smoking cessation alone may reduce absolute CHD risk sufficiently to eliminate the need for statin treatment.17 Other appropriate measures include weight reduction, reducing intake of saturated fat and increasing intake of polyunsaturated fat, regular physical activity, and, when appropriate, additional measures to reduce blood pressure and to control blood glucose.11,14,16 In patients who have clinical features of atherosclerotic disease, statin therapy should be given with other appropriate secondary prevention therapy, such as low-dose aspirin, ACE-inhibitors or beta-blockers.13
Tools for estimating risk
The most accurate way of estimating absolute CHD risk and, therefore, potential benefits from primary prevention interventions, is by weighting and collating the influence of all major risk factors, using a risk function derived from epid-emiological data.18 Several methods have been developed for assessing risk in the UK population, and are generally available as printed charts or computer programmes. These usually assess 'CHD' risk (based on fatal or non-fatal myocardial infarction plus angina), rather than the closely associated 'cardiovascular' (CV) risk (based on CHD, stroke, peripheral vascular disease and heart failure). (A 10-year CHD risk of 30% approximates to a CV risk of 40%).19 All of the methods discussed here are deemed acceptable by the National Service Framework on CHD.13
Accurate assessment of CHD risk in primary prevention requires measurement of TC:HDL-C ratios, rather than total cholesterol alone.20
The Sheffield table (distributed by DTB in 199621) can collate the influence of a patient's CHD risk factors, without prior knowledge of blood lipid concentrations. It can thereby predict whether an individual's absolute CHD risk might be high enough to justify starting a statin and, therefore, whether measurement of lipid concentrations is needed to confirm or exclude this level of risk. This obviates the need to measure lipid concentrations in everyone being assessed, an advantage over other risk-assessment methods. The table detects when a given CHD risk level (e.g. at least 30% over 10 years) has been crossed, but does not allow further quantification of risk. A newer version of the table improves on the original22–24 and should be used in its place. Unlike the original, the newer table uses ratios of TC:HDL-C, identifies two CHD risk levels (15% and 30% over 10 years), and attempts to quantify the influences of family history of premature CHD, and left-ventricular hypertrophy on ECG. The Sheffield table treats hypertension as a dichotomous variable (i.e. either 'present' or 'absent') and assumes an 'average' HDL-cholesterol concentration of 1.2mmol/L (when an actual value is unavailable).25 In one study, the newer table had 97% sensitivity and 95% specificity for a CHD event risk of 15% or more over 10 years, and 82% sensitivity and 99% specificity for a CHD event risk of 30% or more over 10 years.25
Various coloured charts are available for estimating risk, including, for example, the European Coronary Risk chart and the New Zealand risk assessment tool. Another such chart, the Joint British Societies chart, estimates 10-year CHD risk.11 A version of this chart can be found in the British National Formulary.26 This chart uses TC:HDL-C ratios, and considers systolic blood pressure as a continuous variable over the range 110-210mmHg. If no HDL-cholesterol concentrations are available, risk can still be estimated by assuming a concentration of 1.0mmol/L. The chart, however, averages the influence of age into 10-year blocks. In a comparative study, the chart had a sensitivity of 85% (vs. 91% for the revised Sheffield table) and specificity of 99% (vs. 96% for the revised Sheffield table) for identifying 10-year CHD risk levels of 30% or higher.24 A recent survey found that GPs and practice nurses preferred the British Societies chart to the Sheffield table and that nurses (but not GPs) interpreted it more accurately.27 These potential advantages must be weighed against the need to measure blood lipids in all patients assessed with the charts, unlike with the Sheffield table.
Two computer programmes recommended by the National Service Framework on CHD calculate absolute 10-year CHD and stroke risk.11,28 These programmes (based on Microsoft Excel 5.0 [CardioRisk Manager28 and Cardiac Risk Assessor11] or MS DOS [Cardiac Risk Assessor11]) run on personal computers and use the same risk equation. However, each programme presents risk data differently, that is graphically and numerically28 or just numerically.11 In general, such programmes should, in theory, generate more precise risk estimates than paper-based tables or charts,13 but they require knowledge of all standard risk factors (including lipid concentrations), and therefore, unlike the Sheffield table, cannot function as a screening tool. Also, they may be inconvenient to use in some clinical settings and data entry may take longer than paper-based methods. On the other hand, using these programmes interactively during consultations to demonstrate the potential impact of an intervention might encourage adherence to therapy.
Limitations of risk-assessment methods
The methods discussed above all use the Framingham risk equation (based on high-risk, middle-aged, largely white North American populations) to determine the risk of developing CHD events. This equation does not accurately predict risk in Southern European populations (generally at lower risk of CHD) and may underestimate risk in the following groups: British Asians; women with a premature menopause; patients with hypertriglyceridaemia, familial hyperlipidaemia or a strong family history of premature CHD; and patients with chronic renal disease.14 No allowances are made for heaviness of smoking or socio-economic status in risk calculations; moreover, printed charts do not allow CHD risk estimates to be determined in those aged over 70 or 75 years. All this emphasises the need to regard risk-assessment methods as aids to decision-making, which should be used in conjunction with clinical judgment.
Using statins in practice
Choosing between statins
The five statins licensed in the UK (atorvastatin, cerivastatin, fluvastatin, pravastatin and simvastatin) differ in important respects, including the strength of outcome evidence in clinical trials, maximum efficacy in lowering cholesterol concentrations at licensed dosages, and cost.
Pravastatin and simvastatin have a clear advantage over the other statins in terms of weight of clinical evidence for long-term efficacy and safety3–6 and are the only statins licensed in the UK for secondary prevention of CHD events. Pravastatin is also licensed for primary prevention of CHD events.
Efficacy in lowering lipid concentrations
The dosage required to lower serum LDL-cholesterol concentrations by a given amount varies substantially between drugs. At current maximum licensed dosages, atorvastatin appears the most effective in lowering LDL-cholesterol concentrations, followed by simvastatin, then cerivastatin, pravastatin and fluvastatin.29–32 All the statins lower plasma triglyceride concentrations; at maximum licensed dosages, atorvastatin and simvastatin appear to have the greatest effect,29–32 and may be useful in treating moderate hypertriglyceridaemia in patients with familial combined hyperlipidaemia.
So, which drug?
On current evidence, it seems reasonable to use simvastatin as routine first-line therapy for the prevention of CHD events, on the basis of long-term clinical outcome data and efficacy in reducing serum cholesterol concentrations.
Licensed dosages for the available statins are given in the table. It has been suggested that rather than starting with the lowest licensed dosage, initial dosages of simvastatin (20mg) or pravastatin (40mg) should be used once daily as in published prevention studies. Lower statin dosages will be required when these drugs are used in combination with other lipid-lowering drugs. Statins should be taken in the evening for maximal effect, and require at least 4 weeks to exert their full effect on lipid concentrations. Adequate monitoring is essential to ensure that all patients receive optimal dosages. Measurements of total or LDL-cholesterol concentrations are usually sufficient for monitoring the effects of statins and for titrating dosages. Such measurements should be repeated 6 weeks after dosage adjustments are made until the desired lipid concentrations are achieved, and then annually.14
Unwanted effects and precautions
Gastrointestinal upset, altered liver function tests and muscle aches are the commonest unwanted effects with statins. Myopathy, hepatitis, rash, peripheral neuropathy, insomnia, nightmares or vivid dreams, and difficulty in concentrating have been reported more rarely. Patients reporting adverse central nervous system effects on a statin should be tried on a drug with less penetration into the central nervous system, such as pravastatin or atorvastatin.33
Statins should be used cautiously in patients with a history of liver disease or with a heavy alcohol intake, and avoided in active liver disease. Liver function tests should be performed in all patients before and within 1-3 months of starting treatment, and thereafter, 6-monthly for 1 year, unless indicated sooner by clinical suspicion. Treatment should be discontinued if serum transaminase concentrations rise to, and persist at, three times the upper limit of normal. These elevations are generally not associated with clinical features of liver disease and usually decline to normal once therapy is stopped.
If myopathy is suspected or diagnosed clinically and if creatine kinase concentrations become markedly elevated (over 10 times the upper limit of normal), treatment should be stopped. Patients should be advised that if they develop unexplained persistent generalised muscle pain they should discontinue statin treatment and seek medical advice. Rhabdomyolysis is a rare, but potentially serious unwanted effect; the incidence may be increased in those with renal impairment. Combining a statin with fibrate therapy increases the risk of muscle complications.
Atorvastatin, cerivastatin and simvastatin are all metabolised by cytochrome P450 3A4. Cerivastatin is also metabolised by cytochrome P450 2C8. Pravastatin is primarily metabolised by sulphation, whereas fluvastatin is metabolised by cytochrome P450 2C9. Antibiotics (e.g. macrolides), antifungals, cyclosporin and HIV-protease inhibitors commonly inhibit cytochrome P450 3A4 and 2C9, leading to raised serum statin concentrations33 and an increased risk of myopathy. In contrast, barbiturates, carbamazepine, phenytoin and rifampicin induce cytochrome P450 3A4 and 2C9,33 so reducing serum statin concentrations.
Changes in prothrombin times have been reported with atorvastatin, fluvastatin and simvastatin when these drugs have been taken concomitantly with warfarin. Atorvastatin and simvastatin may increase plasma digoxin concentrations.
Serum lipid concentrations are only one of several factors that determine a patient's absolute risk of developing complications of coronary heart disease (CHD) and, therefore, whether the individual is likely to benefit clinically from statin therapy. The risk is particularly high in patients with clinically overt atherosclerotic disease and this justifies starting a statin (with appropriate dietary advice) in all such individuals. Given current NHS funding, it is practicable to treat patients without clinically overt atherosclerotic disease only if their absolute risk of CHD events exceeds 30% over 10 years. These individuals should be identified by specifically targeting patients whose personal or family medical history suggests they could have an elevated CHD risk. They should have their risk measured using a risk-assessment tool and given statin therapy, with appropriate dietary and lifestyle advice (with smoking cessation being a particular priority). Computer-based risk-assessment methods are in theory more precise than paper-based methods, since they allow more risk factors to be taken into account. But all available methods tend to underestimate risk in certain groups, such as British Asians and those with inherited dyslipidaemias.
Of the five statins available in the UK, it seems reasonable to use simvastatin as routine first-line therapy for the prevention of CHD events, on the basis of long-term clinical outcome data and cholesterol-lowering efficacy. Once a statin is started, the patient must have appropriate monitoring of their lipid concentrations, with titration of the dosage to achieve the required serum total cholesterol concentration (a fall to below 5mmol/L or by 20-25%, whichever results in a lower concentration [or a fall in low-density-lipoprotein cholesterol to below 3mmol/L or by 30%]).
[M=meta-analysis; R=randomised controlled trial]
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.