NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Primary Care (UK). Lipid Modification: Cardiovascular Risk Assessment and the Modification of Blood Lipids for the Primary and Secondary Prevention of Cardiovascular Disease [Internet]. London: Royal College of General Practitioners (UK); 2008 May. (NICE Clinical Guidelines, No. 67.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Lipid Modification

Lipid Modification: Cardiovascular Risk Assessment and the Modification of Blood Lipids for the Primary and Secondary Prevention of Cardiovascular Disease [Internet].

Show details

7Drug therapy for the secondary prevention of cardiovascular disease (CVD)

7.1. Recommendations

[Hyperlink to Introduction]

7.1.1.

When considering lipid modification therapy in primary and secondary prevention, drugs are preferred for which there is evidence in clinical trials of a beneficial effect on CVD morbidity and mortality.

Drug therapy for secondary prevention

7.1.2.

For secondary prevention, lipid modification therapy should be offered and should not be delayed by management of modifiable risk factors. Blood tests and clinical assessment should be performed, and comordbidities and secondary causes of dyslipidaemia should be treated. Assessment should include:

  • smoking status
  • alcohol consumption
  • blood pressure (see ‘Hypertension’, NICE clinical guideline 34)
  • body mass index or other measure of obesity (see ‘Obesity’, NICE clinical guideline 43)
  • fasting total cholesterol, LDL cholesterol, HDL cholesterol and triglycerides (if fasting levels are not already available)
  • fasting blood glucose
  • renal function
  • liver function (transaminases)
  • thyroid-stimulating hormone (TSH) if dyslipidaemia is present.
7.1.3.

If a person has acute coronary syndrome, statin treatment should not be delayed until lipid levels are available. A fasting lipid sample should be taken about 3 months after the start of treatment.

Statins for secondary prevention

7.1.4.

Statin therapy is recommended for adults with clinical evidence of CVD.15

7.1.5.

The decision whether to initiate statin therapy should be made after an informed discussion between the responsible clinician and the person about the risks and benefits of statin treatment, taking into account additional factors such as comorbidities and life expectancy. 16

7.1.6.

When the decision has been made to prescribe a statin, it is recommended that therapy should usually be initiated with a drug with a low acquisition cost (taking into account required daily dose and product price per dose).20

7.1.7.

People with acute coronary syndrome should be treated with a higher intensity statin17. Any decision to offer a higher intensity statin should take into account the patient’s informed preference, comorbidities, multiple drug therapy, and the benefits and risks of treatment.

7.1.8.

Treatment for the secondary prevention of CVD should be initiated with simvastatin 40 mg. If there are potential drug interactions, or simvastatin 40 mg is contraindicated, a lower dose or alternative preparation such as pravastatin may be chosen.

7.1.9.

In people taking statins for secondary prevention, consider increasing to simvastatin 80 mg or a drug of similar efficacy and acquisition cost if a total cholesterol of less than 4 mmol/litre or an LDL cholesterol of less than 2 mmol/litre is not attained. Any decision to offer a higher intensity statin21 should take into account informed preference, comorbidities, multiple drug therapy, and the benefit and risks of treatment.

7.1.10.

An ‘audit’ level of total cholesterol of 5 mmol/litre should be used to assess progress in populations or groups of people with CVD, in recognition that more than a half of patients will not achieve a total cholesterol of less than 4 mmol/litre or an LDL cholesterol of less than 2 mmol/litre.

Fibrates for secondary prevention

7.1.11.

Fibrates may be considered for secondary prevention in people with CVD who are not able to tolerate statins.

Nicotinic acid for secondary prevention

7.1.12.

Nicotinic acid may be considered for secondary prevention in people with CVD who are not able to tolerate statins.

Anion exchange resins for secondary prevention

7.1.13.

Anion exchange resins may be considered for secondary prevention in people with CVD who are not able to tolerate statins.

Ezetimibe for secondary prevention

7.1.14.

People with primary hypercholesterolaemia should be considered for ezetimibe treatment in line with ‘Ezetimibe for the treatment of primary (heterozygous-familial and non-familial) hypercholesterolaemia’ (NICE technology appraisal guidance 132).

Monitoring of statin treatment for primary and secondary prevention

7.1.15.

If a person taking a statin starts taking additional drugs, or needs treatment for a concomitant illness that interferes with metabolic pathways or increases the propensity for drug and food interactions, consider reducing the dose of the statin, or temporarily or permanently stopping it.

7.1.16.

People who are being treated with a statin should be advised to seek medical advice if they develop muscle symptoms (pain, tenderness or weakness). If this occurs, creatine kinase should be measured.

7.1.17.

Creatine kinase should not be routinely monitored in asymptomatic people who are being treated with a statin.

7.1.18.

Baseline liver enzymes should be measured before starting a statin. Liver function (transaminases) should be measured within 3 months of starting treatment and at 12 months, but not again unless clinically indicated.

7.1.19.

People who have liver enzymes (transaminases) that are raised but are less than 3 times the upper limit of normal should not be routinely excluded from statin therapy.

7.1.20.

If a person develops an unexplained peripheral neuropathy, statins should be discontinued and specialist advice sought.

7.2. Introduction to drug therapy for secondary prevention

7.2.1. The effectiveness of lipid modifying drugs

The GDG based recommendations to use lipid modifying drugs on trial evidence of improvement in cardiovascular outcomes and where available, total mortality. For people with established CVD there is substantive trial evidence that statins reduce total mortality, cardiovascular mortality and morbidity and total mortality, and are cost-effective. This evidence is strongest for people with coronary heart disease (CHD) (Baigent, C., Keech, A., Kearney, P. M. et al, 2005; National Institute for Health and Clinical Excellence, 2006).

Among people with CHD treated with statins there is a reduction in recurrent CHD events of about 23%, (rate ratio (RR) 95% CI 0.74 to 0.80) and a reduction in stroke events by 17% (0.78 to 0.88) (Baigent, C., Keech, A., Kearney, P. M. et al, 2005). For people with stroke there is a reduction in stroke and cardiovascular events using higher intensity statins (Amarenco, P., Bogousslavsky, J., Callahan, A. S. et al, 2003). No trials have compared the effectiveness of higher intensity statin therapy with standard intensity statin therapy in people following a stroke.

Although there have been no statin trials specifically in people with peripheral arterial disease (PAD), the Heart Protection Study demonstrated the benefits of statin therapy in patients with PAD. Allocation to simvastatin 40 mg daily reduced the rate of first major vascular events by about one-quarter, and that of peripheral arterial events by about one-sixth, with large absolute benefits seen in participants with PAD because of their high vascular risk (Heart Protection Study Collaborative Group., 2007).

Fibrates have been shown to reduce some cardiovascular events in people with CHD though in comparison to statins their lower efficacy and adverse event profile has meant that statins are the drug of first choice for most people. Nicotinic acid and anion-exchange resins have also shown evidence of cardiovascular benefit.

The NICE Statin Technology Appraisal ‘Statins for the prevention of cardiovascular events’ 2006 has thoroughly and comprehensively reviewed the evidence on the effectiveness and cost-effectiveness of statins, and our recommendations on the initiation of statin therapy are based upon this report which states that:

  • Statin therapy is recommended for adults with clinical evidence of CVD
  • The decision to initiate statin therapy should be made after an informed discussion between the responsible clinician and the individual about the risks and benefits of statin treatment, and taking into account additional factors such as comorbidity and life expectancy
  • When the decision has been made to prescribe a statin, it is recommended that therapy should be initiated with a drug with a low acquisition cost (taking into account required daily dose and product price per dose).

7.2.2. The association between lipid modification using drugs and cardiovascular events

The epidemiological relationship between cholesterol as a risk factor in populations and groups and cardiovascular events is well established. As cholesterol increases, so does the risk of CVD. This relationship is such that each 1mmol/l rise in total cholesterol is associated with a 72% increase in the risk of a major coronary event (Emberson, J. R., Whincup, P. H., Morris, R. W. et al, 2003).

There is now compelling randomised controlled trial evidence in people with established CVD, that lowering cholesterol with statins reduces total mortality, cardiovascular mortality and morbidity. For the statin class at lower and moderate intensity each 1 mmol/l reduction in LDL cholesterol will produce a proportional reduction in major vascular events of 23% (at least down to an LDL cholesterol of 2 mmol/l) (Baigent, C., Keech, A., Kearney, P. M. et al, 2005).

Statins are highly cost-effective with a good record of safety. There is also good evidence that higher intensity statins are associated with additional cost-effective reductions in cardiovascular events for people after recent myocardial infarction (MI) and acute coronary syndrome (ACS).

However the benefits of cholesterol lowering and safety cannot be assumed for all drug classes or for all drugs within the same class (Psaty, B. M., Weiss, N. S., Furberg, C. D. et al, 1999) and cardiovascular outcome and adverse event data should be available for every drug from clinical trials. The withdrawal of the statin cerivastatin because of adverse events is a salutary reminder that all drugs within a class are not the same and that there may be specific drug effects within a drug class.

The same strength of evidence that exists for statins does not exist for other classes of lipid lowering drugs (fibrates, anion exchange resins, nicotinic acid) where the trials are fewer in number, the total patient population studied can be small, and trials have shown variable benefits on cardiovascular events despite reduction in cholesterol.

Other classes of drug have either failed to improve cardiovascular outcomes or even increased mortality. Torcetrapib, one of a new class of lipid modifying drug therapies (CETP inhibitor) which raises HDL cholesterol, was being evaluated in a clinical trial which was stopped prematurely because of excess mortality (Jensen, G. B. and Hampton, J., 2007; Nissen, S. E., Tardif, J. C., Nicholls, S. J. et al, 2007).

The potential advantages of drug combinations from different classes cannot be assumed as there are no cardiovascular outcome data for any drug combination in lipid management. There is a greater propensity for major adverse events when statins are combined with fibrates or other drugs particularly when statins are used at higher doses.

7.2.3. The use of statins in clinical practice

In the period 1981–2000, CHD mortality under age 84 years in England and Wales fell by 54%; 68 230 fewer deaths. Modelling of the effects of changes in the three major risk factors, smoking, blood pressure and serum cholesterol suggests that these changes are associated with 45 370 fewer deaths. The biggest single contribution to reduction in mortality was estimated to be a decrease in smoking. Approximately 2135 fewer deaths were attributed to statin treatment: 1990 in CHD patients and 145 in people without established disease (Unal, B., Critchley, J. A., and Capewell, S., 2005).

Prescription of statins and other drugs to improve risk factors remains suboptimal despite the fact that half the survivors of hospital admission for acute MI or angina experience a further major coronary event or death within 5 years of discharge (Capewell, S., Unal, B., Critchley, J. A. et al, 2006).

Statin prescription has increased dramatically in the last 10 years particularly for people with established CVD. In 1997 Brady et al reported 18% of people with CHD in primary care were on statins (Brady, A. J., Oliver, M. A., and Pittard, J. B., 2001). In 2006, among 150 general practices in East London, statin prescription for people with CHD was 81% (Report: East London Clinical Effectiveness Group Queen Mary University of London 2007).

There is still considerable variation in prescribing and under-dosing by practice and evidence of inequity in prescribing by age and also by sex. Statins are less likely to be prescribed to people over 75 years and women (de Lusignan, S., Belsey, J., Hague, N. et al, 2006; DeWilde, S., Carey, I. M., Bremner, S. A. et al, 2003).

Patient adherence to treatment with statins remains a major challenge and only half the patients at highest risk after MI continue to take their statins at 2 years (Penning-van Beest, F. J., Termorshuizen, F., Goettsch, W. G. et al, 2007; Wei, L., Ebrahim, S., Bartlett, C. et al, 2005).

7.3. Statins

[Return to Recommendations]

7.3.1. Evidence statements for statins

NICE Technology Appraisal evidence statement for statins

7.3.1.1.

In a meta-analysis of 14 randomised controlled trials of secondary prevention in CHD, statin therapy was associated with a reduction in all-cause mortality, CVD mortality, CHD mortality, fatal MI, and coronary revascularisation compared with placebo. (NICE technology appraisal 94, ‘Statins for the prevention of cardiovascular events’ 2007).

7.3.2. Evidence statements for higher intensity statin therapy

7.3.2.1.

Meta-analysis of four randomised controlled trials in patients with CHD found that higher intensity statin therapy compared with lower intensity statin therapy was associated with a reduction in the composite outcome of coronary death or MI, and with a reduction in the composite outcome of coronary death or any cardiovascular event (MI, stroke, hospitalization for unstable angina or any revascularisation).

7.3.2.2.

Higher intensity statin therapy was not associated with a reduction in all cause mortality but there was a trend for significance in cardiovascular mortality compared with lower intensity statin therapy. Higher intensity statins reduced coronary death or any cardiovascular event compared with lower intensity statins.

7.3.2.3.

No randomised controlled trials were identified that compared higher intensity statin therapy with lower intensity statin therapy in patients with peripheral arterial disease or following stroke.

7.3.2.4.

One randomised controlled trial in patients following stroke or transient ischaemic attack found that higher intensity statin therapy with atorvastatin 80 mg was associated with a reduction in fatal stroke, the composite of fatal and non-fatal stroke and any cardiovascular event compared with placebo. Post-hoc analysis found this beneficial effect to be restricted to patients after ischaemic stroke whereas a harmful effect was found for those patients after hemorrhagic stroke.

7.3.2.5.

Higher intensity statin therapy did not confer any benefit over placebo for the outcome of non-fatal stroke compared with placebo.

7.3.2.6.

Using a model developed for the guideline, higher intensity statin therapy compared to low intensity statin therapy was found to be cost-effective in the base case in patients following acute coronary syndrome. Treatment is most cost-effective using drugs with lowest acquisition costs

7.3.2.7.

Using a model developed for the guideline, higher intensity statin therapy is not cost-effective in the base case compared to low intensity statin therapy in patients with stable coronary artery disease (£27,840/QALY). However if generic drug prices are assumed high intensity statins will dominate lower intensity statins (they will result in more QALYs and cost savings) in patients with stable CAD.

7.3.2.8.

Using a model developed for the guideline, a titration strategy based on a target total cholesterol of 4mmol/l was found to be cost- effective compared to a fixed dose strategy of low intensity statins, but only if titrating using generic drugs.

Adverse events associated with higher intensity statin therapy

7.3.2.9.

Four randomised controlled trials in patients with CHD found that higher intensity statin therapy was associated with a greater persistent elevation in alanine aminotransferase and/or aspartate aminotransferase levels compared with lower intensity therapy. This was not found to be associated with a significant increase in clinical liver disease.

7.3.2.10.

Three of the four trials found higher intensity statin therapy was not associated with an increase in myalgia compared with lower intensity therapy and one found an excess of myalgia but no increase in the incidence of myopathy.

7.3.2.11.

Three of the four trials found that higher intensity statin therapy was not associated with an increase in rhabdomyolysis compared with lower intensity therapy and one found an excess of rhabdomyolysis in the higher intensity group which was found to be associated with identifiable secondary causes.

7.3.2.12.

A retrospective analysis of pooled data from 49 clinical trials found higher intensity statin therapy with atorvastatin 80 mg to be associated with a greater incidence of persistent elevations in alanine aminotransferase and/or aspartate aminotransferase > 3 x ULN compared to standard intensity therapy with atorvastatin 10 mg or placebo.

No incidences of myopathy or rhabdomyolysis were reported and serious hepatic adverse events were rare although a small number of patients receiving high intensity statin therapy developed hepatitis which resolved after discontinuation of drug therapy.

7.3.3. Clinical effectiveness of statins

Throughout the guideline, we have reported 95% confidence intervals for relative risks (RR) and odds ratios (OR). Where the 95% confidence interval crosses the ‘line of no effect’ i.e., when the confidence intervals included 1, we have interpreted this as being non-significant. This interpretation holds even when the upper or lower limit of the confidence interval is 1.00.

The NICE Technology Appraisal 94 (NICE technology appraisal guidance 94, ‘Statins for the prevention of cardiovascular events’ 2006) states that:

  • Statin therapy is recommended for adults with clinical evidence of CVD.

The recommendation was based on the meta-analysis of 14 randomised controlled trials of secondary prevention in CHD. Of these, four were conducted in MI and/or angina patients (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., 1998; Liem, A. H., van Boven, A. J., Veeger, N. J. et al, 2002; Pedersen, T. R., Kjekshus, J., Berg, K. et al, 2004; Sacks, F. M., Tonkin, A. M., Shepherd, J. et al, 2000). Four studies recruited patients with CAD (Crouse, J. R., Byington, R. P., Bond, M. G. et al, 1995; Jukema, J. W., Bruschke, A. V., van Boven, A. J. et al, 1995; Pitt, B., Mancini, G. B., Ellis, S. G. et al, 1995; Teo, K. K., Burton, J. R., Buller, C. E. et al, 2000) two studies recruited patients with CAD and hypercholesterolaemia (Bestehorn, H. P., Rensing, U. F., Roskamm, H. et al, 1997; Riegger, G., Abletshauser, C., Ludwig, M. et al, 1999) one study recruited patients with mild CAD (Oliver, M. F., de Feyter, P. J., Lubsen, J. et al, 1994) two studies enrolled patients after coronary balloon angioplasty (Serruys, P. W., Foley, D. P., Jackson, G. et al, 1999) and (Bertrand, M. E., McFadden, E. P., Fruchart, J. C. et al, 1997), and one study enrolled patients after percutaneous coronary intervention (Serruys, P. W., de Feyter, P., Macaya, C. et al, 2002). Statin therapy was associated with a reduction in the following clinical outcomes compared with placebo: all-cause mortality (RR 0.79, 95% CI 0.70 to 0.90), CVD mortality (RR 0.75, 95% CI 0.68 to 0.83), CHD mortality (RR 0.72, 95% CI 0.64 to 0.80), fatal MI (RR 0.57, 95% CI 0.45 to 0.72), unstable angina (RR 0.82, 95% CI 0.72 to 0.94), hospitalisation for unstable angina (RR 0.90, 95% CI 0.70 to 0.90), nonfatal stroke (RR 0.75, 95% CI 0.59 to 0.95), new or worse intermittent claudication (RR 0.64, 95% CI 0.46 to 0.91) and coronary revascularisation (RR 0.77, 95% CI 0.69 to 0.85).

The NICE Technology Appraisal 94 (NICE technology appraisal guidance 94, Statins for the prevention of cardiovascular events’ 2006) further states that:

  • The decision to initiate statin therapy should be made after an informed discussion between the responsible clinician and the individual about the risks and benefits of statin treatment, and taking into account additional factors such as comorbidity and life expectancy.
  • When the decision has been made to prescribe a statin, it is recommended that therapy should be initiated with a drug with a low acquisition cost (taking into account required daily dose and product price per dose).

7.3.4. Clinical effectiveness of higher intensity versus lower intensity statin therapy

No randomised controlled trials were identified that compared higher intensity statin therapy with lower intensity therapy in patients with angina alone, stroke or peripheral arterial disease. In addition, no randomised controlled trials were identified on the effectiveness of up-titrating statin dose compared with giving a fixed dose.

Three randomised controlled trials compared higher intensity statin therapy with lower intensity statin therapy in patients with CHD: one in patients after ACS (PROVE-IT-TIMI-22) (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004), one in patients with previous MI (IDEAL) (Pedersen, T. R., Faergeman, O., Kastelein, J. J. et al, 2005) and one which included previous MI 58% and/or angina/revascularization (TNT) (LaRosa, J. C., Grundy, S. M., Waters, D. D. et al, 2005)). None of these trials treated to a pre-specified target total or LDL cholesterol, although the achieved levels were lower in each of the higher intensity statin groups, compared with the respective lower intensity statin groups. A fourth trial in patients after ACS, compared early intensive statin therapy with delayed conservative statin therapy (A to Z) (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004).

The first randomised controlled trial (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004) recruited patients within 10 days of an ACS event (29% had unstable angina, 36% non-ST elevation MI and 35% ST elevation MI). A high proportion of trial participants were taking other secondary prevention drugs and over two thirds were revascularised for treatment of the index event. At recruitment patients had to have a total cholesterol of 6.21 mmol/l or less. Patients were randomised to receive either higher intensity statin therapy with atorvastatin (80 mg once daily) or lower intensity statin therapy with pravastatin (40 mg once daily). Lipid values at the start of the study were similar in both groups. At follow up, patients in the atorvastatin group achieved lower levels of LDL cholesterol compared with the pravastatin group (1.60 mmol/l versus 2.46 mmol/l) and patients in the pravastatin group achieved higher HDL cholesterol levels.

During a mean follow up of 24 months, there was a reduction in the primary outcome (a composite of death from any cause, MI, documented unstable angina requiring rehospitalisation, revascularisation or stroke) with higher intensity therapy compared with lower intensity (HR 0.84, 95% CI 0.74 to 0.95). Similarly, higher intensity therapy was associated with a risk reduction of 14% (P = 0.029) for the secondary outcome of a composite of death from CHD, nonfatal MI or revascularisation. There was no significant reduction in death from any cause or reinfarction with higher intensity therapy compared with lower intensity (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004).

The second study was an open label randomised trial in patients with prior MI (median time since last MI was 22 months) (Pedersen, T. R., Faergeman, O., Kastelein, J. J. et al, 2005). Most trial participants were taking aspirin and beta blockers, but almost 2/3 were not taking ACE inhibitors or ARBs. Patients were assigned to higher intensity atorvastatin 80 mg once daily or lower intensity simvastatin (20 mg once daily). Further drug titration could be undertaken at 24 weeks within the study protocol, based on achieved total cholesterol levels. Twenty one percent of patients in the simvastatin group had their dose increased to 40 mg daily, and 6% of patients in the atorvastatin group had their dose reduced to 40 mg daily. At the end of the study, 23% were treated with simvastatin 40 mg daily and 13% with atorvastatin 40 mg daily. During treatment, patients in the atorvastatin group had lower levels of LDL cholesterol, total cholesterol, triglycerides and apolipoprotein B compared with the simvastatin group. HDL cholesterol and apolipoprotein A1 levels were higher in the simvastatin group compared with the atorvastatin group. Mean LDL cholesterol levels were 2.7 mmol/l in the simvastatin group and 2.1 mmol/l in the atorvastatin group.

For the primary endpoint of major coronary event (defined as coronary death, hospitalisation for nonfatal acute MI, or cardiac arrest with resuscitation) there was no significant difference in event rates between the two treatment groups during a median follow up of 4.8 years. There was a reduction in the nonfatal MI component of this primary endpoint with atorvastatin therapy compared with simvastatin treatment (HR 0.83, 95% CI 0.71 to 0.98). Atorvastatin treatment was associated with a reduction in the secondary endpoint of any CHD event (HR 0.84, 95% CI 0.76 to 0.91) and also a reduction in any major cardiovascular event (HR 0.87, 95% CI 0.78 to 0.98) compared with simvastatin treatment. There were no differences in cardiovascular or all cause mortality (Pedersen, T. R., Faergeman, O., Kastelein, J. J. et al, 2005).

The third randomised controlled trial recruited patients with clinically evident stable CHD (59% had a prior MI, 82% angina) (LaRosa, J. C., Grundy, S. M., Waters, D. D. et al, 2005). To ensure that, at baseline, all patients had LDL cholesterol levels consistent with the then current guidelines for the treatment of stable CHD, patients with LDL cholesterol levels between 3.4 and 6.5 mmol/l entered an eight week run in period of open-label treatment with 10 mg of atorvastatin per day. At the end of the run in phase, those patients with a mean LDL cholesterol of less than 3.4 mmmo/l were randomised. Patients were assigned to either higher intensity atorvastatin (80 mg once daily) or lower intensity atorvastatin (10 mg once daily). The trial follow up was for a median of 4.9 years. No information was given on concomitant medications at baseline or during the trial but it was stated that medication usage was similar in the two groups at the start of the trial. Mean LDL cholesterol levels during the study were 2.0 mmol/l in the group treated with atorvastatin 80 mg once daily and 2.6 mmol/l in the group treated with atorvastatin 10 mg once daily. There was a 22% reduction (95% CI 11% to 31%) in the primary end point (defined as the combination of death from CHD, nonfatal non-procedural MI, resuscitation after cardiac arrest, or fatal or nonfatal stroke) in patients treated with atorvastatin 80 mg daily compared with patients treated with atorvastatin10 mg daily. Patients treated with high dose atorvastatin had a decreased incidence of the following components of this primary endpoint: nonfatal MI (HR 0.78, 95% CI 0.66 to 0.93), and fatal or nonfatal stroke (HR 0.75, 95% CI 0.59 to 0.96). Higher intensity treatment was also associated with a lower incidence of the following secondary outcomes: major coronary event (HR 0.80, 95% CI 0.69 to 0.92), cerebrovascular event (HR 0.77, 95% CI 0.64 to 0.93), hospitalisation for congestive heart failure (HR 0.75, 95% CI 0.59 to 0.93), any cardiovascular event (HR 0.81, 95% CI 0.75 to 0.87) and any coronary event (HR 0.79, 95% CI 0.73 to 0.86). There was no difference in all cause mortality between higher and lower intensity atorvastatin treatment (LaRosa, J. C., Grundy, S. M., Waters, D. D. et al, 2005).

A fourth trial compared early intensive statin therapy with delayed lower intensity statin therapy (A to Z) (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004). This trial consisted of 2 overlapping phases. The first phase was an open labelled trial comparing enoxaprin with unfractionated heparin in patients with non ST elevation ACS who were treated with tirofiban and aspirin. The second phase recruited patients initially from the first phase who had stabilised (for at least 12 consecutive hours within 5 days after symptom onset). In addition, recruits had at least one of the following characteristics: age older than 70 years, diabetes mellitus, prior history of coronary artery disease, peripheral arterial disease or stroke. Subsequently, the protocol was amended to allow patients with non ST elevation ACS who were not enrolled in the first phase, and also patients with ST elevation MI to enter into the second phase directly (overall non ST-segment elevation ACS: 60%, ST elevation MI: 40%).

At baseline almost all the participants were taking aspirin and beta blockers, three quarters were taking ACE inhibitors and almost half were revascularised for treatment of the index event. Patients were randomised to either simvastatin 40 mg once daily for 1 month followed by 80 mg once daily thereafter (early higher intensive therapy) or placebo for 4 months followed by simvastatin 20 mg once daily thereafter (delayed conservative therapy) (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004).

Early high intensity statin therapy decreased LDL cholesterol levels by 39% compared with baseline levels during the first month of therapy with simvastatin 40 mg, and then by a further 6% following an increase in simvastatin dosage to 80 mg. For the delayed conservative statin treatment group, LDL cholesterol levels increased by 11% during the 4 month placebo period, then decreased from baseline by 31% after 4 months of therapy with simvastatin 20 mg (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004).

For the primary endpoint of the combination of cardiovascular death, nonfatal MI, readmission for ACS or stroke, early higher intensity statin therapy did not confer benefit compared with delayed lower intensity therapy. There was also no benefit found in any of the individual components of the primary endpoint. Likewise no benefit was observed in the secondary endpoints of all cause mortality and coronary revascularisation due to documented ischaemia. There was a reduction in the incidence of new onset congestive heart failure in the early intensive statin treatment group compared with the delayed conservative treatment group (HR 0.72, 95% CI 0.53 to 0.98) but not a reduction in cardiovascular related death (HR 0.75, 95% CI 0.51 to 1.00) (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004).

A meta-analysis of these four studies has been conducted by Cannon et al (Cannon, C. P., Steinberg, B. A., Murphy, S. A. et al, 2006) using a fixed-effects model. Higher intensity statin therapy did not confer any significant benefit over lower intensity statin therapy for the outcomes of all cause mortality (OR 0.94, 95 % CI 0.85 to 1.04), cardiovascular mortality (OR 0.88, 95 % CI 0.78 to 1.00) or non-cardiovascular mortality (OR 1.03, 95 % CI 0.88 to 1.20). Higher intensity statin therapy was associated with a reduction in the combination of coronary death or MI (OR 0.84, 95 % CI 0.77 to 0.91), stroke (OR 0.82, 95 % CI 0.71 to 0.96) and coronary death or any cardiovascular event (OR 0.84, 95 % CI 0.80 to 0.89).

In addition to the four trials comparing higher intensity therapy with lower intensity therapy, two randomised controlled trials were identified that compared higher intensity statin therapy with placebo. The first trial recruited patients with ACS (Schwartz, G. G., Olsson, A. G., Ezekowitz, M. D. et al, 2001) and the second recruited patients with a history of stroke or transient ischaemic attack (Amarenco, P., Bogousslavsky, J., Callahan, A., III et al, 2006).

The trial in patients with ACS (Schwartz, G. G., Olsson, A. G., Ezekowitz, M. D. et al, 2001) randomised a total of 3086 patients with unstable angina or non-Q-wave acute MI to receive either atorvastatin 80 mg daily or placebo. Patients were hospitalised within 24 hours of the index event and randomised after a mean of 63 hours of hospitalisation. During or after hospitalisation for the index event, most were treated with aspirin, three quarters with beta blockers and half with ACE inhibitors or ARBs.

The study period was for 16 weeks and during this period the primary end point (combination of death, nonfatal acute MI, cardiac arrest with resuscitation, or recurrent symptomatic myocardial ischemia with objective evidence requiring emergency rehospitalisation) was not significantly reduced in patients randomised to atorvastatin compared with those who received placebo (RR 0.84, 95% CI 0.70 to 1.00). Atorvastatin therapy was not associated with a reduction in the following individual components of the primary outcome: death, non-fatal MI or cardiac arrest with resuscitation but was associated with a lower risk of recurrent myocardial ischaemia requiring rehospitalisation compared with placebo (RR 0.74, 95% CI 0.57 to 0.95). However, it should be noted that the study was only powered to detect differences between groups in the primary outcome. At the end of the study, compared to baseline, LDL cholesterol had increased by an adjusted mean of 12% in the placebo group and had decreased by an adjusted mean of 40% in the atorvastatin group (Schwartz, G. G., Olsson, A. G., Ezekowitz, M. D. et al, 2001).

Incidences of the following secondary outcomes were not different in the atorvastatin group compared with placebo: coronary revascularisation procedures, worsening congestive heart failure or worsening angina. Non-fatal stroke was reduced in the atorvastatin group compared with placebo (RR 0.41, 95% CI 0.20 to 0.87) as was the composite outcome of fatal and non-fatal stroke (RR 0.50, 95% CI 0.26 to 0.99) (Schwartz, G. G., Olsson, A. G., Ezekowitz, M. D. et al, 2001).

The second randomised controlled trial (Amarenco, P., Bogousslavsky, J., Callahan, A., III et al, 2006) recruited patients without known CHD and with previously documented stroke (69%) (66.5% ischaemic and 2.5% haemorrhagic) or transient ischaemic attack (31%), 1 to 6 months prior to randomisation. A total of 4731 participants were randomised to receive either 80 mg atorvastatin or placebo and were followed up for a mean duration of 4.9 years. Most patients were taking aspirin or other antiplatelets (not heparin) although only 29% were taking ACE inhibitors and 18% beta blockers. For the primary endpoints, high dose atorvastatin decreased the risk of fatal stroke (HR 0.57, 95 % CI 0.35 to 0.95) and the composite of fatal and non-fatal stroke (HR 0.84, 95 % CI 0.71 to 0.99) compared with placebo. High dose atorvastatin also reduced the risk of any cardiovascular event (stroke plus any major coronary event) (HR 0.80, 95 % CI 0.69 to 0.92) compared with placebo. No benefit was found for the outcome of non-fatal stroke. Post hoc analysis indicated significant differences in hazard ratios based on the type of stroke occurring during the trial; the cause specific adjusted hazard ratios compared to placebo showed a beneficial effect in those experiencing ischaemic stroke during the trial (HR 0.78, 95 % CI 0.66 to 0.94), but a harmful effect on those experiencing hemorrhagic stroke (HR 1.66, 95 % CI 1.08 to 2.55). Atorvastatin conferred benefit compared with placebo for the following secondary outcomes: major coronary event (HR 0.65, 95 % CI 0.49 to 0.87), major cardiovascular event (HR 0.80, 95 % CI 0.69 to 0.92), any cardiovascular event (HR 0.74, 95 % CI 0.66 to 0.83), acute coronary event (HR 0.65, 95 % CI 0.50 to 0.84), any coronary event (HR 0.58, 95 % CI 0.46 to 0.73), non-fatal MI (HR 0.51, 95 % CI 0.35 to 0.74), revascularisation (HR 0.55, 95 % CI 0.43 to 0.72), transient ischaemic attack (HR 0.74, 95 % CI 0.60 to 0.91), the composite of stroke or transient ischaemic attack (HR 0.77, 95 % CI 0.67 to 0.88). No benefit was seen for the outcomes of cardiovascular mortality or all cause mortality but the trial was not statistically powered for this endpoint (Amarenco, P., Bogousslavsky, J., Callahan, A., III et al, 2006).

7.3.5. Cost-effectiveness of statins

The NICE Technology Appraisal (NICE technology appraisal guidance 94, Statins for the prevention of cardiovascular events’ 2006) states that:

  • When the decision has been made to prescribe a statin, it is recommended that therapy should usually be initiated with a drug of low acquisition cost (taking into account required daily dose and product price per dose).

7.3.6. Cost-effectiveness of higher intensity statin therapy compared with lower intensity statin therapy

When initial searches were undertaken, no studies were found which compared cost-effectiveness of higher intensity statins with lower intensity statins in patients with coronary artery disease (CAD). Consequently, the GDG requested the development of an economic model to help inform the guideline.

A Markov model was developed to estimate the incremental cost per quality adjusted life year (QALY) of lifetime treatment with high intensity statins (atorvastatin 80 mg and simvastatin 80 mg) compared with low intensity statins (simvastatin 40 mg) from a UK NHS perspective. The base case assumptions model two cohorts of hypothetical patients aged 65 years of age:

  1. Patients with acute ACS, and;
  2. Patients with stable coronary artery disease (CAD).

Intermediate outcomes included in the model include the numbers of MI, stroke, TIA, PAD, heart failure, revascularisation, and angina events, and deaths from CVD and other causes. Effectiveness data for ACS patients were drawn from two studies which were meta-analysed; A to Z (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004), in which patients were randomised to either simvastatin 40 mg once daily for 1 month followed by 80 mg once daily thereafter (early intensive therapy) or placebo for 4 months followed by simvastatin 20 mg once daily thereafter (delayed conservative therapy) and PROVE-IT where patients were randomised to receive either higher intensity statin therapy with atorvastatin (80 mg once daily) or lower intensity statin therapy with pravastatin (40 mg once daily) (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004) For the stable CAD patient model, effectiveness data were drawn from the TNT where patients were assigned to either higher intensity atorvastatin (80 mg once daily) or lower intensity atorvastatin (10 mg once daily) (LaRosa, J. C., Grundy, S. M., Waters, D. D. et al, 2005) and IDEAL where patients were assigned to higher intensity atorvastatin 80 mg once daily or lower intensity simvastatin (20 mg once daily) (Pedersen, T. R., Faergeman, O., Kastelein, J. J. et al, 2005) trials. Again, these were meta-analysed.

The models make the conservative assumption that the all cause mortality rate in the modelled population is twice that of the general population. Health state utility values were taken from published sources (see Appendix C for details). All cause mortality rates were taken from the Government Actuarial Department.(Government Actuaries Department, 2006), The model makes the conservative assumption of no adverse events from treatment using high intensity statins. Cost of drugs were taken from the Prescription Pricing Authority Drug Tariff Feb 27th 2008 (atorvastatin 80 mg £367.74/year, simvastatin 80 mg £64.53/year, simvastatin 40 mg, £18.12/year) (NHS Prescription Pricing Authority, 2008). Costs of cardiovascular events were taken from the statins TA94 (National Institute for Health and Clinical Excellence, 2006). In order to reflect social values for time preference, as is standard in economic models, costs and QALYs have been discounted at 3.5% as recommended by NICE.(National Institute for Health & Clinical Excellence., 2006) All of these and other model assumptions have been tested in sensitivity analyses.

The base case results are presented below, and cost-effectiveness is assessed against a threshold of £20,000/QALY.

7.3.6.1. Results for patients with ACS

Table 2 indicates the modelled number of events for a hypothetical population of 1,000 ACS patients treated with either high intensity or low intensity statins. The table indicates that fewer cardiovascular events occur in the population treated with high intensity statins. This translates to a gain of 0.32 discounted QALYs when compared with low intensity statins.

Table 2. Lifetime modelled events for a cohort of 1,000 ACS patients treated with either low or high intensity statins.

Table 2

Lifetime modelled events for a cohort of 1,000 ACS patients treated with either low or high intensity statins.

a. Cost-effectiveness results for ACS patients

The model estimates the life-time incremental cost per QALY of using high intensity statins (both simvastatin and atorvastatin 80mg) compared with low intensity statins both simvastatin and pravastatin is about £4,700, indicating that high intensity statins are cost-effective in ACS patients. The probability that high intensity statins is cost-effective is about 94% when compared with low intensity statins. .

7.3.6.2. Results for patients with stable coronary artery disease (CAD)

Table 3 indicates the modelled number of lifetime events for a hypothetical 1000 patients treated with either high or low intensity statins. The table indicates that fewer cardiovascular events occur in the population treated high intensity statins. This translates to a gain of 0.08 discounted QALYs per patient when compared with low intensity statins.

Table 3. Lifetime modelled events for a cohort of 1000 CAD patients treated with either low or high intensity statins.

Table 3

Lifetime modelled events for a cohort of 1000 CAD patients treated with either low or high intensity statins.

a. Cost-effectiveness results

The model estimates the life-time incremental cost per QALY of using high intensity statins (atorvastatin 80mg) compared with low intensity statins (simvastatin 40mg) is about £27,840 indicating that high intensity statins are not cost-effective in patients with stable CAD. The probability that high intensity statins is cost-effective is about 42% when compared with low intensity statins. .

Updated Economic Publication Searches

Subsequent to this model being built, updated searches retrieved one publication which compared higher intensity statins with lower intensity statins in patients with ACS and stable CAD in North America.(Chan, P. S., Nallamothu, B. K., Gurm, H. S. et al, 2007) The study is a cost-utility analysis conducted from a third payer’s perspective, using a Markov model for a hypothetical population of 60 year old patients. Effectiveness data were drawn from the A to Z (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004) and PROVE-IT (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004) trials for the ACS model, and from the TNT (LaRosa, J. C., Grundy, S. M., Waters, D. D. et al, 2005) and IDEAL (Pedersen, T. R., Faergeman, O., Kastelein, J. J. et al, 2005) trials for the stable CAD model. Utility data were derived from published literature. The estimated ICER for the ACS population was below US$30,000/QALY and is stable in sensitivity analysis. The ICER for the stable CAD population was reported as US$33,400/QALY but the ICER is very sensitive to assumptions about statin efficacy (ICER range from $10,300/QALY to dominated) and cost of statins. ICERs range from dominant using the lower price of atorvastatin to $84,000/QALY when the higher price is used. The results of this study are similar to those found as a result of our modelling work.

Summary of cost-effectiveness of higher versus lower intensity statins

In conclusion, compared with low intensity statins, high intensity statins in patients with ACS are cost-effective when compared with low intensity statins. In patients with stable CAD, atorvastatin 80 mg is not cost-effective using a £20,000/QALY threshold. However, assuming the use of generic simvastatin 80 mg is makes the model highly cost-effective. Thus cheaper generic high intensity statins may be used in patients with stable CAD.

Cost-effectiveness of treating to target (titration threshold) compared with fixed doses of statins

A systematic literature search identified 408 papers. Eighteen papers were assessed in full. None of them met the inclusion criteria. In light of the lack of published evidence, the GDG requested the development of an economic model in order to generate cost-effectiveness estimates.

Model Structure and Assumptions

The population modelled is a hypothetical cohort of 1000 adults with hyperlypidemia and with a history of CHD/CVD, and who are free from diabetes. The population modelled was based on a distribution of patients taken from The Health Improvement Network (THIN) database, having an average untreated total cholesterol level of 6.1 mmol/l and an average age of 61 years.

The model estimates lifetime costs and quality adjusted life years (QALYs) of statin treatment using a target titration treatment strategy versus a fixed dose treatment strategy. The model has been used to estimate the cost-effectiveness of both 4 mmol/l and 5 mmol/l targets using 1 and 2 step titrations.

In the fixed dose strategy, all patients are assumed to be given simvastatin 40 mg daily, with no further consultations, or measurements performed. This treatment strategy was initially compared with a two-stage titration strategy, in which patients are initially given simvastatin 40 mg daily, with those failing to reach the pre-specified target then being titrated to the next therapy (simvastatin 80 mg). Measurements are again taken for the latter group of patients, and anyone still not achieving the pre-specified target is then assumed to be titrated up to atorvastatin 80 mg. In the one-step titration model, patients not achieving target on simvastatin 40 mg are titrated once only up to simvastatin 80 mg, with no further up-titration.

For both treatment arms, the modelled percentage reductions in cholesterol levels are estimated using the results of the STELLAR trial (Jones, P. H., Hunninghake, D. B., Ferdinand, K. C. et al, 2004). Subsequent reductions in CVD event and mortality outcomes were estimated using equations derived from a meta-analysis by Law et al (Law, M. R., Wald, N. J., Rudnicka, A. R. et al, 2003).

Costs of drugs are based on prices quoted by the PPA as at February 27th 2008.

Table 4. Costs of modelled Statins as at Feb 27th 2008.

Table 4

Costs of modelled Statins as at Feb 27th 2008.

Each titration step is assumed to cost £26 based on the cost of a GP consultation and a blood test.(Netten, A. & Curtis L., 2007) Cost of health states including treatment for MI, stroke, TIA, PAD, HF, and angina were estimated using various published sources (details in Appendix C). Health state utility values were taken from published sources (Appendix C). All cause mortality rates are from the Government Actuarial Department.(Government Actuaries Department, 2006) The model makes the conservative assumption that the all cause mortality rate in the modelled population is twice that of the general population. Also, the model assumes no adverse events from treatment using high dose statins.

As recommended by NICE (National Institute for Health & Clinical Excellence., 2006) and to reflect social values, future costs and QALYs are both discounted at a rate of 3.5% in the model. These and other model assumptions have been tested in sensitivity analyses.

Results

Table 5 indicates that with a target of 5 mmol/l total cholesterol, the majority of patients (69%) are modelled to reach target on simvastatin 40 mg. This is true of both the fixed and the titration population groups in the model. With a target of 4 mmol/l, only 31% of patients will reach target on simvastatin 40 mg. In the 2 step titration model an additional 15% of patients reach target on simvastatin 80 mg, if the target is 5 mmol/l and an additional 6% reach target using 4 mmol/l.

Table 5. Proportion of patients modelled to be on each of the three included drugs under four treatment strategies.

Table 5

Proportion of patients modelled to be on each of the three included drugs under four treatment strategies.

Table 6 indicates the modelled number of events for the hypothetical 1000 patient cohorts having assumed a 2-step titration and a target total cholesterol of 5 mmol/l for illustrative purposes. The table indicates that fewer CVD events occur in the population treated using the titration strategy.

Table 6. Lifetime event outputs modelled for a cohort of 1,000 patients using a 2-stage titration treatment strategy with a target of 5 mmol/l total cholesterol compared with a fixed low dose treatment strategy.

Table 6

Lifetime event outputs modelled for a cohort of 1,000 patients using a 2-stage titration treatment strategy with a target of 5 mmol/l total cholesterol compared with a fixed low dose treatment strategy.

The incremental cost-effectiveness analysis indicates that compared to a fixed dose treatment strategy, a 1-step titration to simvastatin 80mg treatment strategy using a target of 4mmol/l has an ICER of £14,089 per QALY. One step titration to 5mmol/l is ruled out by extended dominance and 2 –step titration to 5 is dominated by I step titration to 4mmol/l. Two step-titration to 4mmol/l is not cost-effective and has an ICER of £66,819/QALY when compared to 1 step-titration to 4mmol/l. Our model indicates that with the 1 step titration to a target of 4 mmol/l (simvastatin 80mg) 63% of patients would not achieve this target, however the analysis indicates that it would not be cost-effective to try to get more patients to target.

Conclusion

In conclusion, the result of modelling suggest that titration using a threshold target of 4 mmol/l total cholesterol is cost-effective so long as titration stops at simvastatin 80 mg. Most patients would not achieve a target of 4mmol/l total cholesterol and modelling suggests that it is not cost-effective to try to take more patients to target using higher cost statins such as atorvastatin. Details of the economic model and the analyses are available in Appendix C.

7.3.7. Adverse events associated with lower intensity statin therapy

Adverse events associated with lower intensity statin therapy are discussed in the primary prevention drug therapy chapter (Section 6.3.2.3).

7.3.8. Adverse events associated with higher intensity statin therapy

Four randomised controlled trials were identified that compared higher intensity statin therapy with lower intensity statin therapy, the details and results of which have been described in section 1.3.3 (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004) (Pedersen, T. R., Faergeman, O., Kastelein, J. J. et al, 2005) (LaRosa, J. C., Grundy, S. M., Waters, D. D. et al, 2005) (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004).

The first trial (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004) found elevations in alanine aminotransferase levels to be greater in patients who received atorvastatin 80 mg compared with those receiving pravastatin 40 mg. Discontinuation of study medication due to myalgia, muscle aches or elevations in creatine kinase levels were similar in the two treatment groups. No cases of rhabdomyolysis were reported in either group (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004).

The second trial (Pedersen, T. R., Faergeman, O., Kastelein, J. J. et al, 2005) found that patients who received atorvastatin 80 mg had higher rates of discontinuation due to non-serious adverse events than those allocated to simvastatin 20 mg. There were no differences in the frequency of serious adverse events between the two treatment groups. Serious myopathy and rhabdomyolysis were rare in both groups (Pedersen, T. R., Faergeman, O., Kastelein, J. J. et al, 2005).

The third trial (LaRosa, J. C., Grundy, S. M., Waters, D. D. et al, 2005)) found therapy with atorvastatin 80 mg to be associated with an increase in adverse events, with a higher rate of treatment discontinuation compared with the atorvastatin 10 mg group. Treatment related myalgia was similar in the two groups and there were no persistent elevations in creatine kinase. Five cases of rhabdomyolysis were reported (2 in the high dose group, 3 in the low dose group). More patients in the high dose group had persistent elevation in alanine aminotransferase, aspartate aminotransferase or both, compared with the low dose group (LaRosa, J. C., Grundy, S. M., Waters, D. D. et al, 2005).

The fourth trial (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004) compared early intensive therapy (simvastatin 40 mg once daily for 1 month followed by 80 mg once daily thereafter) with delayed conservative therapy (placebo for 4 months followed by simvastatin 20 mg once daily thereafter). Incidences of elevated alanine aminotransferase or aspartate transaminase levels (greater than 3 X ULN) were found to be similar in the two treatment groups. Discontinuation of study medication due to muscle-related adverse events was also comparable between the two groups. A total of 10 patients developed myopathy (creatine kinase > 10 X ULN on 2 consecutive measurements). Of the nine patients treated with simvastatin 80 mg, three patients had creatine kinase levels > 10 000 units/l and met the criteria for rhabdomyolosis. Of these 3 patients, 1 had contrast media renal failure and 1 patient was receiving concomitant verapamil (inhibitor of cytochrome P450 3A4 (CYP3A4)). In addition, 1 patient receiving 80 mg simvastatin had a creatine kinase level 10 X ULN without muscle symptoms, which was associated with alcohol abuse (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004).

Two randomised controlled trials were identified that compared higher intensity statin therapy with placebo (Amarenco, P., Bogousslavsky, J., Callahan, A., III et al, 2006; Schwartz, G. G., Olsson, A. G., Ezekowitz, M. D. et al, 2001), the details and results of which have also been described in section 9.3.3.

The first trial (Schwartz, G. G., Olsson, A. G., Ezekowitz, M. D. et al, 2001) found that more patients in the atorvastatin 80 mg group developed liver transaminase levels > 3 X ULN compared with those allocated placebo. There were no cases of myositis.

The second trial (Amarenco, P., Bogousslavsky, J., Callahan, A., III et al, 2006) compared treatment with atorvastatin 80 mg to placebo and found no significant difference in the incidence of serious adverse events between groups, although persistent elevation of alanine or aspartate aminotransferase (> 3 ULN on two consecutive occasions) was more frequent in the atorvastatin group (2.2 %) versus placebo (0.5 %), P < 0.001.

A retrospective analysis of pooled data from 49 clinical trials of atorvastatin was identified which compared the relative safety of lower intensity atorvastatin 10 mg with higher intensity atorvastatin 80 mg (Newman, C., Tsai, J., Szarek, M. et al, 2006). Data were pooled from 49 clinical trials (n = 14 236 participants) in which patients were randomised to receive active treatment for a period ranging from 2 weeks to 52 months (atorvastatin 10 mg: n = 7258, atorvastatin 80 mg: n = 4798 and placebo: n = 2180). The incidence rate (per 1000 patient-years of exposure) of various safety parameters and adverse events was calculated for each of the three groups. The overall safety profile was comparable between atorvastatin 80 mg, 10 mg and placebo in terms of incidence rate of patients experiencing ≥ 1 adverse event, withdrawals due to adverse events and serious, nonfatal adverse events. Musculoskeletal safety parameters were also similar across groups and there were no incidences of myopathy or rhabdomyolysis reported. In this analysis, a greater incidence of persistent alanine aminotransferase and/or aspartate aminotransferase > 3 X ULN was observed in the atorvastatin 80 mg group compared with the other two groups. Serious hepatic adverse events were rare although five patients in the atorvastatin 80 mg group developed hepatitis, which resolved after discontinuation of atorvastatin. The adverse events of haematuria and albuminuria were also examined but the incidence in each atorvastatin group was low compared to placebo. Incidence of death was low in all groups and none were considered to be related to treatment.

A number of cohort studies have examined the safety of rosuvastatin used in clinical practice.

The first was a Dutch study that followed three separate cohorts, namely incident rosuvastatin users, other incident cohort users and non-statin exposed controls for cases of myopathy, rhabdomyolysis, acute renal failure and liver impairment/failure (Goettsch, W. G., Heintjes, E. M., Kastelein, J. J. et al, 2006). Exclusion criteria for the two statin cohorts were as follows; not incident users, statin use < 12 months, age < 20 or > 84 years, missing information in the PHARMO system, serious adverse event in history (e.g. of myopathy, rhabdomyolysis). The control cohort had to be aged between 20 and 84 and have no history of statin usage (≥ 12 months), and individuals were excluded if they had a history of a serious adverse event (e.g. of myopathy, rhabdomyolysis). Data were obtained from the PHARMO medical record linkage system that included drug-dispensing records from community pharmacies and hospital discharge records of more than 2 million residents throughout the Netherlands. Potential cases of hospitalisation for myopathy, rhabdomyolysis, acute renal failure or hepatic impairment for each of the three cohorts were validated through a multi-step process using data obtained from hospital records. Cases of all cause mortality were obtained from notifications in the hospital and pharmacy databases and were not validated (Goettsch, W. G., Heintjes, E. M., Kastelein, J. J. et al, 2006).

In 2002 and 2004, of 119 681 statin users 47 543 incident statin users met the inclusion criteria. More than 20% of those patients started with rosuvastatin (10 147), 15 091 patients with atorvastatin, 14 198 with simvastatin, 7290 with pravastatin and 817 with floatation. There were 99 935 controls selected from the PHARMO system. In total, 102 events (excluding death) were identified in 96 patients, 21 in the category myopathy/rhabdomyolysis, 48 in acute renal failure, and 33 events as hepatic Impairment. Only 81% of cases could be validated (79.4%) because some hospitals did not cooperate for several not medical reasons. The validation process resulted in 1 case of myopathy, 1 case of rhabdomyolysis, 13 cases of renal impairment and 11 cases of hepatic impairment. The total number of deaths identified was 1388, and after adjustment for age and gender in the three cohorts, all cause mortality was not increased in the statin user groups compared with the control group (Goettsch, W. G., Heintjes, E. M., Kastelein, J. J. et al, 2006).

The total incidence of serious adverse event was very low, in the users of statins only 15 validated events were identified in more than 45 000 years of follow up (> 1 per 3000 person years). Only one case of myopathy could be identified among the users of other statins cohort, and one case of rhabdomyolysis in the non statin control cohort. The number of validated cases of acute renal failure was higher, and the incidence in both statin cohorts was increased compared with controls (rosuvastatin RR 5.91, 95%CI 1.19 to 29.36, other statins RR 3.27 95%CI 0.84 to 12.75). No significant difference was observed in the incidence of acute renal failure between the rosuvastatin and other statin cohorts (RR 1.81, 95%CI 0.47 to 7.02). Hepatic impairment incidences’ were comparable in the other statin and control cohorts, while no incidences of hepatic impairment were found in the rosuvastatin cohort (Goettsch, W. G., Heintjes, E. M., Kastelein, J. J. et al, 2006).

The second study was an observational cohort study in which patients were identified from dispensed prescriptions issued by primary care physicians/general practitioners between August and December in the England (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007). At least 6 months after the initial prescription, questionnaires known as Green forms were sent to the general practitioners requesting information regarding any event that occurred since initiation of rosuvastatin. The term event was defined as ‘any new diagnosis, any reason for referral to a consultant or hospital admission, any unexpected deterioration (or improvement in concurrent illness, and suspected drug reaction, any alteration of clinical importance in laboratory values, or any other significant event requiring documentation. All returned forms were reviewed by medically qualified staff, and events that required further assessment were followed up. These included muscular, hepatic and renal events, suspected adverse drug events, and events with unknown aetiology for example jaundice (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

Of 31 228 Green forms sent, 12 543 (40.2%) were returned, and 863 (6.9%) were classified as void and excluded from the study. The study cohort comprised of 11 680 patients, of which 50.3% were male (5880), 49.2% (5745) were female, and for 0.5% (55) the sex was not specified. The median age was 64 years (interquartile range 56 to 72 years), and the age range was 17 to 101 years. The median treatment period was 9.8 months (interquartile range 4.6 to 11.7 months) (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

Data derived from the Green forms were used in an incident density analysis of all events reported during treatment within specified time periods and also provided information on clinical events reported as the reason for discontinuation of rosuvastatin (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

A total of 2047 (17.5%) patients were reported to have stopped treatment with rosuvastatin. Musculoskeletal events accounted for 20.3% (414 of 2037) of the reasons for discontinuation. Myalgia was the most frequents cause (277 cases, 13.6% of all reasons specified), followed by patient request (144 of 2037), drug information including adverse publicity/reports in the media (123 of 2037), non formulary reasons such as change in general practitioner, prescribing policy (91 of 2037). Abnormal liver function tests and elevated creatine kinase levels accounted for 57 and 33 cases of discontinuation, respectively (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

Incident densities (ID) were calculated for events occurring in the first month (ID1) of treatment, during months 2–6 (ID2–6) of treatment and for events occurring during the overall treatment period. The ten most common adverse events in order of first month IDs were: Myalgia, malaise, dizziness, nausea/vomiting, intolerance, headache/migraine, abdominal pain, dyspepsia, abnormal LFTs and joint pain. Myalgia was the adverse event with the highest incident density during month 1 (ID1 = 7.70 events per 1000 patient-months of treatment) and it also had the highest ID for the whole treatment period. The difference between IDs for the first month and during months 2–6 were calculated to establish which events may have been early-onset events with rosuvastatin. There were six clinical events for which the rate of event in month 1 was significantly greater than the rate of event in months 2–6: Myalgia (ID1-ID2–6 = 4.0 (99% CI 1.67 to 6.33)), malaise (ID1-ID2–6 = 2.28 (99% CI 0.64 to 3.91)), dizziness (ID1-ID2–6 = 1.90 (99% CI 0.49 to 3.30)), nausea/vomiting (ID1-ID2–6 = 1.54 (99% CI 0.17 to 2.91)), intolerance (ID1-ID2–6 = 1.71 (99% CI 0.38 to 3.04)), and headache/migraine (ID1-ID2–6 = 1.43 (99% CI 0.11 to 2.75)) (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007),

IDs were also stratified by starting dose of rosuvastatin: the IDs for the 20 mg/day and 40 mg/day dosages were compared with the 10 mg/day dose. A 2.5 fold increase in the rate of abnormal LFT results was found for patients started on the rosuvastatin 40 mg/day dose compared with those started on the 10 mg/day dose (Incidence density ratio = 2.71 (95% CI 1.53 to 4.53)). Although there was an increase in the incidence density ratio for the 40 mg/day dose compared with the 10 mg/day dose for elevated CK, raised urea/creatinine, haematuria and proteinuria, these differences were not significant. No differences were found between dosage groups in the rates of myalgia, limb pain or cramps (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

Where events described on the Green forms required further assessment, follow-up questionnaires were sent to the GPs. A total of 685 questionnaires were posted to prescribing GPs of which 585 (85%) were returned. Data from these questionnaires were used in a causality assessment for adverse events relating to the muscular, hepatic and renal system-organ classes. Events were assessed as ‘probably’ or ‘possibly’ related to rosuvastatin depending upon various factors including whether the adverse events were clinically and/or pathologically well-defined with reasonable time-sequence in relation to administration of rosuvastatin and whether they were more likely to be attributed to rosuvastatin than to concurrent disease or other drugs and whether dechallenge or rechallenge was positive (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

Regarding musculoskeletal events, there were no cases of rhabdomyolysis reported in this cohort; there were 2 cases of myopathy reported however follow-up data was not available and thus causality assessment was not performed. Of the 229 cases of myalgia that were followed up, 128 were assessed as probably related to rosuvastatin and 69 possibly related to rosuvastatin. Overall, musculoskeletal events were the most frequently reported adverse event. Where causality assessment was conducted, a high proportion of musculoskeletal events were assessed as probably or possibly related to rosuvastatin (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

Regarding hepatic events, follow-up data was available for 101 cases of abnormal LFTs, 19 and 48 of these were assessed as probably or possibly related to rosuvastatin respectively. In addition, one case of autoimmune hepatitis and another case of jaundice, raised alkaline phosphatise and ALT were assessed as possibly related to rosuvastatin (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

Regarding renal events, there were 25 cases of raised urea/creatinine, 5 of which were assessed as possibly related to rosuvastatin; there were 7 cases of haematuria, 3 of which were assessed as possibly related to rosuvastatin; 9 cases of proteinuria, one of which were assessed as possibly related to rosuvastatin and another was assessed as probably related to rosuvastatin. Two cases of renal failure were reported although follow-up data was not available for either of these cases (Kasliwal, R., Wilton, L. V., Cornelius, V. et al, 2007).

The fourth study was a retrospective matched cohort study with a follow-up duration of up to 18 months in patients initiating treatment with rosuvastatin compared with other statins (McAfee, A. T., Ming, E. E., Seeger, J. D. et al, 2006). All patients receiving a statin were identified from the administrative database of a large health insurer in the U.S. for the period 1st September 2003 to 29th February 2004. Patients were included in the cohort if they had no prescription for a statin (naïve initiators) or if they had been prescribed a different statin than the index prescription (switcher initiators) during the baseline period defined as 183 days prior to the index date. Only patients who were at least 18 years of age with complete demographic and enrolment information and at least 183 days of complete enrolment before the index date were included. Patients were excluded if they had claims-based diagnoses of myopathy, rhabdomyolysis, renal dysfunction or hepatic dysfunction associated with a hospitalization during the baseline period (McAfee, A. T., Ming, E. E., Seeger, J. D. et al, 2006).

A total of 194 320 patients were identified as having at least one prescription claim for a statin during the defined time period who were either naïve or switcher initiators of a particular statin. Of these patients, 106 926 met the inclusion criteria, 12 217 of which were rosuvastatin initiators and 94 709 were initiated on other statins. Rosuvastatin initiators were matched to other statin initiators by a multivariate technique (propensity score analysis and matching) in order to balance covariate patterns and account for any baseline characteristics of rosuvastatin initiators that differed from other statin initiators in that time period. All analyses were also adjusted by the number of matched comparators. Thus, 11 249 rosuvastatin initiators were matched to 37 282 other statin initiators (statin used: 54.2% atorvastatin, 21.2% simvastatin, 11.0% pravastatin, 10.6% lovastatin and 3.1% fluvastatin) (McAfee, A. T., Ming, E. E., Seeger, J. D. et al, 2006).

Potential incident cases associated with hospitalization for myopathy, rhabdomyolysis, renal dysfunction, or hepatic dysfunction and in-hospital death were identified from health insurance claims and data on 403 (81%) of these potential outcomes were successfully abstracted from written medical records with 125 (31%) cases of outcome incidence being confirmed (McAfee, A. T., Ming, E. E., Seeger, J. D. et al, 2006).

Incidences of adverse events were low. Five cases of rhabdomyolysis or myopathy were found among 43 585 person-years for the entire study cohort (Incidence Rate = 1.15 per 10 000 person-years (95% CI 0.37 to 2.68)). Adjusted Hazard Ratios were calculated and it was found that there were no significant differences between those initiated on rosuvastatin compared with those initiated on other statins for any outcome measure (HR = 1.98 (95% CI 0.18 to 21.90) for rhabdomyolysis, HR= 0.90 (95% CI 0.47 to 1.73) for renal dysfunction, HR not calculable for myopathy, HR=0.87 (95% CI 0.18 to 4.14) for hepatic dysfunction and HR=0.51 (95% CI 0.24 to 1.10) for in-hospital death) (McAfee, A. T., Ming, E. E., Seeger, J. D. et al, 2006).

The fifth study reviewed adverse event reports (AERs) to the Food and Drug Administration USA (FDA) to determine the frequency of rosuvastatin-associated events relative to other commonly used statins, namely; atorvastatin, simvastatin, pravastatin and cerivastatin (for cerivastatin during the time it was available). Two comparative primary analyses were performed. For the first analysis, AERs were determined for the first year during which rosuvastatin was available in the USA (October 2003 to September 2004) and these AERs were compared with the concomitant time period for the other statins (defined as ‘concurrent time period analysis’). The mean doses of statins during this time period was as follows; rosuvastatin 16.7±1.2 mg, simvastatin 53±2.8 mg, pravastatin 18.8±2.0 mg and atorvastatin 21.8±1.4 mg. The second analysis was performed to address the potential of preferential reporting of adverse events with newly marketed drugs. Thus rates of rosuvastatin-associated AERs were compared with those during the first year of marketing for atorvastatin (1997), simvastatin (1992), pravastatin (1992) and cerivastatin (1998). This was defined as ‘first year of marketing analysis’. The rates of AERs were calculated as AERs per million prescriptions for various AERs associated with each of the statins (Alsheikh-Ali, A. A., Ambrose, M. S., Kuvin, J. T. et al, 2005).

For the concurrent time period analysis, the rate of rosuvastatin AERs (a composite of rhabdomyolysis, proteinuria/nephropathy, or renal failure) was higher than AERs for simvastatin (P < 0.001), pravastatin (P < 0.001) and atorvastatin (P < 0.001). For the first year of marketing analysis the rate of rosuvastatin-associated composite AERs was not significantly different than simvastatin AERs, but was significantly higher compared with pravastatin (P < 0.001) and atorvastatin (P < 0.001). Compared with AERs for cerivastatin during its first post marketing year, rosuvastatin composite AERs were less frequent (P < 0.001). Sixty two percent of rosuvastatin-associated AERs occurred at doses of 10 mg/day, and occurred earlier after the initiation of therapy (within the first 12 weeks) compared to other statins. There was no gender predominance. While fatalities were rare, most composite AERs listed hospitalisation as an outcome (Alsheikh-Ali, A. A., Ambrose, M. S., Kuvin, J. T. et al, 2005).

The increased rate of rosuvastatin-associated AERs relative to the other statins was also observed in secondary analysis.

For the concurrent time period analysis, the rate of rosuvastatin-associated AERs for any adverse event was higher than that observed for simvastatin, pravastatin and atorvastatin (P < 0.001 all statins versus rosuvastatin). Likewise for serious AERs (life threatening or requiring hospitalisation), liver AERs, muscle AERs without rhabdomyolysis and also renal failure AERs, rosuvastatin had higher rates of adverse events (P < 0.001 all statins versus rosuvastatin). Furthermore, rhabdomyolysis AERs, although rare, were also higher for rosuvastatin (simvastatin; P < 0.01, pravastatin and atorvastatin; P < 0.001) (Alsheikh-Ali, A. A., Ambrose, M. S., Kuvin, J. T. et al, 2005).

For the first year of marketing analysis the rate of rosuvastatin-associated AERs was similarly higher for the following AERs compared with other statins; all AERs (simvastatin, pravastatin atorvastatin, cerivastatin P < 0.001 all statins versus rosuvastatin), muscle AERs without rhabdomyolysis (simvastatin, pravastatin atorvastatin, cerivastatin P < 0.001 all statins versus rosuvastatin). Liver AERs were higher for rosuvastatin compared with simvastatin, pravastatin and atorvastatin, but were not significantly different with the rate observed with cerivastatin. Serious AERs were higher for rosuvastatin compared with pravastatin and atorvastatin (P < 0.001 for both); however, the rosuvastatin rate was lower than that observed for simvastatin (P < 0.001) and cerivastatin (P < 0.01). Rosuvastatin was also significantly more likely than simvastatin, pravastatin and atorvastatin to be associated with reports of rhabdomyolysis (P < 0.001 all statins versus rosuvastatin), but compared with the first year of cerivastatin, the rate of rosuvastatin rhabdomyolysis events was significantly less (P < 0.001). Finally, the rate of rosuvastatin-associated renal failure AERs was higher compared with pravastatin and atorvastatin (P < 0.001 for both), but similar to that observed with simvastatin and cerivastatin (Alsheikh-Ali, A. A., Ambrose, M. S., Kuvin, J. T. et al, 2005).

There are a number of intrinsic limitations of post marketing adverse event analysis. The analysis is based on reporting rates, not on actual adverse event rates. In clinical practice, adverse events are under reported, and serious adverse events are more likely to be reported than less serious events. The retrospective nature of the analysis does not allow confirmation of causality, or control of potential confounders. For example, providers tend to report preferentially adverse events with newly marketed drugs. In addition, certain adverse events may not be recognised as related to a particular class of drug. Post marketing analysis can also be influenced by publicity, favourably or unfavourably. Another time dependent post marketing variable could be related to the availability of drug dosage. In this context, the relatively low rate of atorvastatin-associated AERs during its first year of marketing may be partially attributable to the fact that only the 10 mg dose was available in the first year (Alsheikh-Ali, A. A., Ambrose, M. S., Kuvin, J. T. et al, 2005).

Not with standing these limitations, the review found that rosuvastatin had a higher rate of AERs compared with other commonly prescribed statins based upon adverse event reports to the FDA. The authors of the review stated that the reported occurrence of these AERs early after initiation of therapy (within 12 weeks on average) suggests that vigilant monitoring for adverse events may ameliorate the risk of toxicity when rosuvastatin is used. They also stated that it would seem prudent for healthcare providers to consider other statins as first line therapy, to initiate rosuvastatin therapy in appropriate patients at lower doses as well as careful monitoring for adverse events (Alsheikh-Ali, A. A., Ambrose, M. S., Kuvin, J. T. et al, 2005).

7.3.9. Evidence to recommendations – statins

The NICE technology appraisal on statins (NICE technology appraisal guidance 94, Statins for the prevention of cardiovascular events’ 2006) considered twenty-eight randomised controlled trials of statins in adults with or at risk of CVD.

No studies that reported cardiovascular events as outcomes were identified for rosuvastatin. Fourteen placebo-controlled studies in which all participants had CHD at study entry were identified for inclusion in a meta-analysis. There were significant reductions in all cause mortality (RR 0.79, 95% CI 0.70 to 0.90), CVD mortality (RR 0.75, 95% CI 0.68 to 0.83), CHD mortality (RR 0.72, 95% CI 0.64 to 0.80), fatal MI (RR 0.57, 95% CI 0.45 to 0.72), nonfatal MI (RR 0.69, 95% CI 0.59 to 0.95), new or worsening intermittent claudication (RR 0.64, 95% CI 0.46 to 0.91). There was no significant reduction in stroke mortality (RR 1.07, 95% CI 0.67 to 1.71) or TIA (RR 0.66 95% CI 0.37 to 1.17). The relative effectiveness of statins did not differ by sex, in people with and without diabetes, or in people over 65 years compared with younger people. For secondary CHD prevention the incremental cost per QALY ranged from £10,000 to £16,000 for all age groups with little difference for men and women.

The NICE technology appraisal (NICE technology appraisal guidance 94, ‘Statins for the prevention of cardiovascular events’ 2006) recommended statin therapy for all adults with clinical evidence of CVD and that when the decision has been made to prescribe a statin, it is recommended that therapy should usually be initiated with a drug with a low acquisition cost (taking into account required daily dose and product price per dose). The GDG considered that for initiation of treatment, simvastatin 40 mg was the most effective drug with a low acquisition cost in secondary prevention.

7.3.10. The use of higher intensity statins and cholesterol targets

International and national guidelines on lipid lowering for CVD prevention have all defined goals or targets of therapy. These target levels have become progressively lower over time and differ between guidelines. The Joint British Societies first recommended in 1998 a total cholesterol target of less than 5.0 mmol/l and an LDL cholesterol target of less than 3.0 mmol/l, or a 25% total cholesterol reduction or a 30% LDL cholesterol reduction, whichever is greater (Joint British recommendations on prevention of coronary heart disease in clinical practice. British Cardiac Society, British Hyperlipidaemia Association, British Hypertension Society, endorsed by the British Diabetic Association, 1998). The National Service Framework for CHD in 2000 recommended levels less than total cholesterol 5 mmol/l or LDL cholesterol 3 mmol/l (or a 25% TC reduction or 30% LDL cholesterol reduction whichever is greater) and these remain the current national advice (DoH March 2000 website). In 2003 the Joint European Societies Task Force on CVD Prevention recommended a total cholesterol level less than 4.5 mmol/l and LDL cholesterol levels below 2.5 mmol/l. Since 2004 in the USA high risk CVD patients are advised to achieve LDL cholesterol levels below 1.81 mmol/l (http://www.nhlbi.nih.gov/guidelines/cholesterol/atp3_rpt.htm). The most recent Joint British Societies 2005 guideline recommended target levels below total cholesterol 4 mmol/l and LDL cholesterol 2 mmol/l (or a 25% reduction in total cholesterol and a 30% reduction in cholesterol if that yields a lower value) (Wood, D., Wray, R., Poulter, N. et al, 2005). More recently the Scottish Sign Guideline 2007 considered total cholesterol targets of 4 mmol/l or 4.5 mmol/l would have major resource implications for NHS Scotland (Scottish Intercollegiate Guidelines Network., 2007), but this was not based on a formal cost-effectiveness analysis. SIGN recommended that pending further studies on mortality, safety, and cost-effectiveness, a total cholesterol target of less than 5 mmol/l in individuals with CVD should be a minimum standard of care (Scottish Intercollegiate Guidelines Network., 2007).

The Cholesterol Trialists Collaboration (Baigent, C., Keech, A., Kearney, P. M. et al, 2005) reported an approximately linear relationship between the absolute reductions in LDL cholesterol achieved 14 statin trials and the proportional reductions in the incidence of coronary and other events. The authors of the Cholesterol Trialists Collaboration state that there is a significant trend towards greater proportional reductions in major coronary events being associated with greater mean absolute LDL cholesterol reductions in the different trials (Baigent, C., Keech, A., Kearney, P. M. et al, 2005). There was no significant heterogeneity between the relative effects after weighting for the absolute LDL cholesterol reduction (Baigent, C., Keech, A., Kearney, P. M. et al, 2005). They found that the proportional reduction in the event rate per mmol/l reduction in LDL cholesterol was largely independent of the presenting cholesterol level. So, lowering the LDL cholesterol level from 4 mmol/l to 3 mmol/l reduced the risk of vascular events by about 23% and lowering LDL cholesterol from 3 mmol/l to 2 mmol/l also reduced residual risk by about 23%. There is a near linear relationship between the log of the risk and cholesterol reduction, but it is important to appreciate that although the relative risk reduction remains constant, at lower cholesterol levels there is a smaller absolute reduction in cardiovascular events, and it is absolute risk reduction that determines cost-effectiveness.

This log linear relationship describes the effect of cholesterol lowering with statins, at least down to a LDL cholesterol of 2 mmol/l. A meta-analysis of higher intensity statins (Cannon, C. P., Steinberg, B. A., Murphy, S. A. et al, 2006) confirmed that the observed 0.67 mmol/l reduction in LDL cholesterol would be expected to lead to a 14% reduction in cardiovascular events on the basis of the log linear hypothesis and the observed reduction of 16% was consistent with this.

The majority of randomised controlled trials to date have not shown a reduction in LDL cholesterol below 2 mmol/l with statin therapy (Figure 1, JBS2 (Wood, D., Wray, R., Poulter, N. et al, 2005)). LDL cholesterol was reduced below an average value of 2 mmol/l in only three of the twenty trials shown; PROVE-IT 1.6 mmol/l (Cannon, C. P., Braunwald, E., McCabe, C. H. et al, 2004), A-Z 1.7 mmol/l (de Lemos, J. A., Blazing, M. A., Wiviott, S. D. et al, 2004), MIRACL 1.9 mmol/l (Schwartz, G. G., Olsson, A. G., Ezekowitz, M.D. et al, 2001). These are all recent randomised controlled trials at maximal licensed statin dosage. These trials had strict recruitment criteria and patients with higher levels of LDL cholesterol tended to be excluded, and are not representative of the general population with CVD. Moreover, the reported LDL cholesterol reductions were median values of the trial participants.

Figure 1. Statin trials showing % reduction in major cardiac events and LDL cholesterol (mmol/l).

Figure 1

Statin trials showing % reduction in major cardiac events and LDL cholesterol (mmol/l). (Figure from JBS2 (Wood, D., Wray, R., Poulter, N. et al, 2005))

GDG discussion on use of targets

Within the GDG, there were differing views on the use of cholesterol “targets” i.e. levels of total and LDL cholesterol that patients on lipid lowering therapy should either aim to be below or should achieve. Proponents of targets considered that the log linear hypothesis from the Cholesterol Trialists Collaboration (Baigent, C., Keech, A., Kearney, P. M. et al, 2005) supported the use of targets because it confirmed that for LDL cholesterol “lower is better”. GDG members were concerned that patients could be potentially under treated if no goal or target were specified. As a proportion of patients can reach cholesterol targets of total cholesterol of less than 4 mmol/l or LDL cholesterol of less than 2 mmol/l on standard doses of statins such as simvastatin 40mg the use of a target would reduce the likelihood that patients would be under-treated with suboptimal doses of statins such as simvastatin 10mg.

Opponents of setting targets raised a number of concerns. There was a minority view within the GDG that any targets are essentially misleading as trials have not treated to target but have used specific drugs to treat patients. For other members of the GDG there was concern as to how targets may be interpreted. Firstly, in practice targets can be interpreted to mean that all patients on treatment should attain the recommended level, irrespective of their starting cholesterol level. This takes no account of the distribution of cholesterol levels in the population prior to commencement of treatment, nor of differing responses to treatment and differing adherence to treatment. It is also important to note that the majority of randomised controlled trials which recruited selected populations did not find statin therapy reduced LDL cholesterol below 2 mmol/l (Figure 1). Opponents of setting targets considered it misleading for both professionals and patients, to set a target that is interpreted as ‘should be achieved’, knowing that many patients will not achieve this.

Secondly, two-thirds of the gain from a statin is realised by the initial dose. Lower cholesterol levels for individual patients may be achieved by using higher intensity statins but for each doubling of dose there is a smaller absolute reduction in cardiovascular events. There was concern that the adoption of targets may encourage the indiscriminate use of either high dose statins or combination lipid therapy.

Finally, there is no trial evidence that drug combinations such as a statin plus a fibrate, will produce additional cost-effective absolute reductions in cardiovascular events.

The GDG concluded by majority that the use of higher intensity statins or drug combinations should be driven by trial evidence of absolute benefit in clinical outcomes and cost effectiveness, and less by targets and relative risk. The GDG accepted again by a majority that the use of a target figure can be helpful in guiding increases of lipid lowering drugs as long as it is clear that this figure is intended to guide treatment rather than be a figure patients are expected to achieve. The wording of the recommendations was agreed to reflect this.

The GDG agreed using the clinical and cost effectiveness evidence that patients with ACS benefit from immediate high intensity statins. Health economic analyses for this guideline and published literature indicate that high intensity statins are less cost effective for patients with CAD. These patients should start on a standard dose of statin and the target figure used to inform increases in treatment.

The GDG recognised from the health economic modelling that over half of patients with stable CAD will not achieve total cholesterol level of 4 mmol/l and LDL cholesterol of 2 mmol/l when given 80 mg simvastatin.. An audit level of total cholesterol 5 mmol/l may help to assess progress in populations and groups.

Table 7 and Table 8 show absolute total and LDL cholesterol reduction and percentage reductions in serum concentrations according to statin and daily dose

Table 7. Absolute LDL cholesterol reduction and percentage reductions in serum LDL cholesterol concentration according to statin and daily dose (summary estimates from 164 randomised controlled trials).

Table 7

Absolute LDL cholesterol reduction and percentage reductions in serum LDL cholesterol concentration according to statin and daily dose (summary estimates from 164 randomised controlled trials).

Table 8. Absolute cholesterol reduction and percentage reductions in serum total cholesterol concentration according to statin and daily dose (summary estimates from 164 randomised controlled trials).

Table 8

Absolute cholesterol reduction and percentage reductions in serum total cholesterol concentration according to statin and daily dose (summary estimates from 164 randomised controlled trials).

7.4. Fibrates

[Return to Recommendations]

7.4.1. Evidence statements for fibrates

7.4.1.1.

Two randomised controlled trials in patients after an MI and/or with angina found that clofibrate therapy was not associated with a reduction in fatal MI or sudden death in people with angina compared with placebo. One trial found that clofibrate therapy was not associated with a reduction in cardiovascular morbidity compared with placebo while the other found that clofibrate therapy was associated with a reduction in the rate of first non-fatal infarct in women with a history of angina compared with placebo.

7.4.1.2.

One randomised controlled in patients after an MI and/or with angina found that bezafibrate therapy was not associated with a reduction in the composite of fatal MI, non-fatal MI and sudden death compared with placebo. In addition, no benefit was seen for cardiovascular morbidity.

7.4.1.3.

One randomised controlled trial in men after an MI and/or with angina found that gemfibrozil therapy was associated with a reduction in the composite of fatal MI, sudden death, death due to congestive heart failure and death as a complication of invasive cardiac procedures compared with placebo.

7.4.1.4.

Two randomised controlled trials in patients following stroke or TIA found that clofibrate therapy was not associated with a reduction in all cause mortality or cardiovascular morbidity compared with placebo.

7.4.1.5.

One randomised controlled trial in patients with peripheral arterial disease showed that bezafibrate therapy was not associated with a reduction in the combination outcome of fatal and nonfatal CHD events and stroke compared with placebo although bezafibrate therapy was associated with a reduction in the incidence of non- fatal coronary heart disease.

7.4.2. Clinical effectiveness of fibrates

Seven randomised controlled trials were identified that compared fibrate therapy with placebo in patients with a history of CVD. Four of these were in patients after an MI and/or with angina, two were in patients following a stroke or transient ischaemic attack and one was in patients with peripheral arterial disease.

Four randomised controlled trials were identified in patients after an MI and/or with angina (Behar, S., Brunner, D., Kaplinsky, E. et al, 2000) (Rubins, H. B., Robins, S. J., Collins, D. et al, 1999) (Research Committee of the Scottish Society of Physicians., 1971), (Group of Physicians of the Newcastle Upon Tyne Region., 1971).

The first randomised controlled trial (Research Committee of the Scottish Society of Physicians., 1971) recruited patients aged 40–69 years with a history of angina, MI or both (27% had angina only). A total of 717 patients were randomised to receive either clofibrate or placebo (olive oil) and were followed up for a mean duration of 4 years. In patients with a history of angina only, treatment with clofibrate did not decrease the rates of sudden death, fatal MI or first non-fatal MI compared to placebo.

The second randomised controlled trial (Group of Physicians of the Newcastle Upon Tyne Region., 1971) recruited patients under 65 years with a history of angina, MI or both (40% had angina only). A total of 497 patients were randomised to receive either clofibrate or placebo (corn oil) and were followed up for 5 years. In patients with a history of angina only, treatment with clofibrate did not decrease the rates of sudden death or fatal MI compared to placebo but was found to decrease the rate of first non-fatal infarct compared to placebo in women with a history of angina (P < 0.05) but not men.

Both of these studies used the drug clofibrate which has now been withdrawn from the British National Formulary.

The third randomised controlled trial (Rubins, H. B., Robins, S. J., Collins, D. et al, 1999) recruited men with an HDL cholesterol of 1.0 mmol/l or less, LDL cholesterol 3.6 mmol/l or less and triglycerides less than 3.4 mmol/l with documented coronary artery disease defined as a history of MI, angina, having undergone coronary revascularization, or angiographic evidence of coronary stenosis. Of these, 61% had a prior history of MI. Concomitant drug therapy at the start of the trial was as follows; aspirin 82%, beta blockers 43%, nitrates 46%, ACE inhibitors 21%, calcium channel blockers 53%. Patients were randomised to either gemfibrozil or placebo. Patients were followed for a mean 5.1 years. Gemfibrozil therapy was associated with a reduction in the primary endpoint of a combination of nonfatal MI and death from CHD compared with placebo. The incidence of the secondary outcome of a combination of nonfatal MI, death from CHD and confirmed stroke was also reduced in the gemfibrozil treatment group compared with the placebo. In addition, gemfibrozil therapy was associated with a reduction in the following outcomes compared with placebo: nonfatal MI, investigator-designated stroke, transient ischaemic attack, carotid endarterectomy and hospitalisation for congestive heart failure. Treatment with gemfibrozil was not associated with any benefit for the following outcomes: death due to coronary heart disease, death from any cause, confirmed stroke, revascularisation, coronary artery bypass graft, percutaneous transluminal coronary angioplasty, peripheral vascular surgery and hospitalisation for unstable angina.

Patients assigned to gemfibrozil had lower total cholesterol and triglycerides levels and higher HDL cholesterol levels compared to patients in the placebo group. LDL cholesterol levels were the same in both groups. Gemfibrozil treatment was associated with a greater incidence of dyspepsia (Rubins, H. B., Robins, S. J., Collins, D. et al, 1999).

The fourth randomised controlled trial (Behar, S., Brunner, D., Kaplinsky, E. et al, 2000) recruited patients with a history stable angina pectoris and/or MI. Of these, 57% had prior angina (and 78% had a history of MI). A total of 3090 patients were randomised to receive either bezafibrate (retard) or placebo and were followed up for a mean duration of 6.2 years. Treatment with bezafibrate did not confer any benefit over placebo for the primary endpoint of a composite of fatal MI, nonfatal MI and sudden death. There was also no benefit observed for any of the individual components of this endpoint. Bezafibrate had no benefit over placebo for the following secondary endpoints: combination of hospitalisation for unstable angina, percutaneous transluminal coronary angioplasty or coronary artery bypass graft, hospitalisation for unstable angina, percutaneous transluminal coronary angioplasty, coronary artery bypass graft, mortality, cardiac mortality, noncardiac mortality, stroke or ischemic stroke.

Compared with the placebo group, triglyceride levels were lower in the bezafibrate subgroup that had triglyceride levels ≥ 2.26 mmol/l. The overall incidence of any adverse event was 69% in both groups, and the frequency of each type adverse event was similar in both groups (Behar, S., Brunner, D., Kaplinsky, E. et al, 2000).

Two randomised controlled trials were identified that compared fibrate therapy with placebo in patients with a history of stroke or transient ischaemic attack (Acheson, J. and Hutchinson, E. C., 1972) (Noble, J. D., Feringa, E. R., Greenhouse, A. H. et al, 1973). Both of these trials used clofibrate.

The first randomised controlled trial (Acheson, J. and Hutchinson, E. C., 1972) recruited patients with focal cerebral vascular disease (those with one stroke, multiple strokes or transient cerebral ischaemia) who had a serum cholesterol level of 250 mg/100ml or higher. A total of 95 patients were randomised to receive either clofibrate or placebo and the period of observation was from 4 months to 4 years. Compared with placebo, clofibrate therapy was not associated with a decrease in all cause mortality. Patients assigned to clofibrate had lower levels of serum cholesterol compared to those who received placebo; mean proportional change in serum cholesterol level was −12.69% for control and −21.41% for clofibrate (P < 0.05). It should be noted that this was a small study and is likely to be underpowered for the outcomes described.

The second randomised controlled trial (Noble, J. D., Feringa, E. R., Greenhouse, A. H. et al, 1973) recruited male veterans with one or more cerebral infarctions or transient ischaemic attack within the past 12 months. A total of 532 men were randomised to receive either clofibrate or placebo and were followed up for an average duration of 21 months. Compared with placebo, clofibrate therapy was associated with a non significant decrease in all cause mortality: 30/264 deaths occurred in the placebo group versus 22/268 in the group allocated to receive clofibrate. For the outcome of vascular morbidity, there was no difference between the groups in the incidence of MI, TIA or angina. There was an increase in recurrence of cerebral infarction (23/264 placebo versus 37/268 clofibrate) and an increase in the incidence of congestive heart failure (4/264 placebo versus 15/268 clofibrate) in the clofibrate group compared to those receiving placebo but these differences were not tested for statistical significance. All other side effects were similar between groups. Regarding blood lipids, clofibrate decreased triglycerides compared to the control group (29% decrease clofibrate versus a 4% increase control) but had a negligible effect on cholesterol levels. Again, no statistical analysis was performed so the significance of these results is unknown It should be noted that this was a small study and is likely to be underpowered for the outcomes described.

One randomised controlled trial was identified that compared fibrate therapy with placebo in patients with a history of peripheral arterial disease (Meade, T., Zuhrie, R., Cook, C. et al, 2002). This trial recruited men with lower extremity arterial disease, 24% had stable angina, 21% had a previous MI and 12% had a history of stroke. A total of 1568 men were randomised to receive either bezafibrate (as Bezalip mono) or placebo and were followed up for a mean of 4.6 years. Bezafibrate therapy did not confer any benefit over placebo for the primary endpoint of a composite of CHD events (both fatal and non-fatal) and all strokes. When the individual endpoints were analysed separately, bezafibrate had no benefit over placebo for the primary outcome of a composite of CHD events and all strokes, but was associated with a reduction in the incidence of non-fatal CHD events (RR 0.60, 95% CI 0.36 to 0.99).

7.4.3. Cost-effectiveness of fibrates

There were no cost-effectiveness studies found on the use of fibrates compared with placebo in secondary prevention of CVD.

7.4.4. Evidence into recommendations

The GDG considered that there was insufficient evidence to routinely recommend the use of fibrates as a first line treatment for patients with CVD. It was decided however, that they may be offered as an alternative for those who are intolerant of statin therapy.

7.5. Nicotinic acids

[Return to Recommendations]

7.5.1. Evidence statements for nicotinic acids

7.5.1.1.

No randomised controlled trials were identified that compared nicotinic acid therapy with placebo in patients with angina, peripheral arterial disease or following stroke.

7.5.1.2.

One randomised controlled trial in patients after MI found that nicotinic acid therapy was associated with a reduction in non-fatal MI and the combination of coronary death or non-fatal MI compared with placebo. Nicotinic acid therapy was not associated with a reduction in all cause mortality, cardiovascular mortality or cardiovascular morbidity compared with placebo.

7.5.2. Clinical effectiveness of nicotinic acids

No randomised controlled trials were identified that compared nicotinic acid therapy with placebo in patients with angina, peripheral arterial disease or following stroke. Due to the lack of trial evidence, it was decided by the GDG to consider evidence used in the NICE Myocardial Infarction guidance (Myocardial infarction - Secondary prevention in primary and secondary care for patients following a myocardial infarction, CG48, 2007)

One paper was identified that compared niacin treatment with placebo in patients after an MI (Wilkins, R. W., Bearman, J. E., Boyle, E. et al, 1975). The Coronary Drug Project Research Group randomly assigned post MI patients to six treatment groups: low and high conjugated oestrogen therapy, clofibrate, dextrothyroxine sodium, niacin and a placebo. The oestrogen and dextrothyroxine arms were stopped early because of an excess of nonfatal cardiovascular events and death, respectively. Patients were followed for 5 years.

Compared with placebo, niacin was not associated with a reduction in the incidence of the following outcomes: all cause mortality, the individual components of all cause mortality, definite pulmonary embolism (fatal or nonfatal), fatal or nonfatal stroke or intermittent cerebral ischaemic attack, definite or suspected fatal or nonfatal pulmonary embolism or thrombophlebitis and also any definite or suspected fatal or nonfatal cardiovascular event. Niacin therapy reduced the incidence of nonfatal MI and also the combination of coronary death or nonfatal MI, compared with placebo. Cholesterol and triglycerides levels decreased in the niacin group compared with the placebo group.

Patients in the niacin group had a greater incidence of the following side effects compared with the placebo group: the combination of diarrhea, nausea, vomiting, black tarry stools, stomach pain, flushing, itching of skin, urticaria, other type of rash, pain or burning when urinating, decrease in appetite, unexpected weight loss, and excessive sweating (Wilkins, R. W., Bearman, J. E., Boyle, E. et al, 1975).

7.5.3. Cost-effectiveness of nicotinic acids

There were no cost-effectiveness studies found on the use of nicotinic acids compared with placebo in secondary prevention of CVD.

7.5.4. Evidence into recommendations

The GDG considered that there was insufficient evidence to routinely recommend the use of nicotinic acids as a first line treatment for patients with CVD. It was decided however, that they may be offered as an alternative for those who are intolerant of statin therapy.

7.6. Anion exchange resins

[Return to Recommendations]

7.6.1. Evidence statements for anion exchange resins

7.6.1.1.

No randomised controlled trials were identified in patients with CVD that compared anion exchange resin therapy with placebo for the outcomes mortality or morbidity.

7.6.1.2.

One small randomised controlled trial in patients with a history of CVD found that cholestyramine therapy was associated with a reduction in total cholesterol and LDL cholesterol compared with placebo.

7.6.2. Clinical effectiveness of anion exchange resins

No randomised controlled trials were identified in patients with CVD that compared anion exchange resin therapy with placebo for the outcomes mortality or morbidity.

One small randomised controlled trial was identified on the clinical effectiveness of anion exchange resins compared with placebo to improve lipid level profiles in patients with coronary artery disease (Brensike, JF., 1984). This trial recruited people with elevated LDL cholesterol and angiographic evidence of coronary artery disease (50% of whom had symptomatic angina and/or MI). A total of 143 patients were randomised to receive either cholestyramine 24 g per day or placebo and were followed up for five years. Treatment with cholestyramine resulted in decreases in total and LDL cholesterol compared with placebo (5 year mean lipid level differences were - 0.1 mmol/l placebo versus - 1.4 mmol/l cholestyramine (P < 0.001) for total cholesterol and - 0.26 mmol/l placebo versus - 1.66 mmol/l cholestyramine (P < 0.001) for LDL cholesterol). Cholestyramine therapy did not have an effect on triglycerides or HDL cholesterol. There were negligible differences between groups for the ancillary outcomes of mortality and morbidity.

7.6.3. Cost-effectiveness of anion exchange Resins

There were no cost-effectiveness studies found on the use of anion exchange resins compared with placebo in secondary prevention of CVD.

7.6.4. Evidence into recommendations

The GDG considered that there was insufficient evidence to routinely recommend the use of anion exchange resins as a first line treatment for patients with CVD. It was decided however, that they may be offered as an alternative for those who are intolerant of statin therapy.

7.7. Ezetimibe

[Return to Recommendations]

7.7.1. Evidence statements for ezetimibe

7.7.1.1.

Please refer to NICE Technology Appraisal No. TA132 ‘Ezetimibe for the treatment of primary (heterozygous familial and non-familial) hypercholesterolaemia’, (National Institute for Health and Clinical Excellence., 2007)

7.7.2. Clinical effectiveness of ezetimibe

The NICE Technology Appraisal TA132 is entitled ‘Ezetimibe for the treatment of primary (heterozygous familial and non-familial) hypercholesterolaemia’, (National Institute for Health and Clinical Excellence., 2007). The guidance recommends ezetimibe as a treatment option for primary (heterozygous familial and non-familial) hypercholesterolaemia and states that its recommendations should be read in the context of the lipid modification clinical guideline (this guidance).

The population groups covered by the ezetimibe Technology Appraisal TA132 (National Institute for Health and Clinical Excellence., 2007) are:

  • Adults with primary (heterozygous familial and non-familial) hypercholesterolaemia who are candidates for treatment with statins on the basis of their CVD status or risk and;
  • whose condition is not appropriately controlled with a statin alone or;
  • in whom a statin is considered inappropriate or is not tolerated.

The term “not appropriately controlled with a statin alone” is defined as failure to achieve a target lipid level that is appropriate for a particular group or individual. It also assumes that statin therapy is optimised.

The NICE Technology Appraisal TA132 (National Institute for Health and Clinical Excellence., 2007) ‘Ezetimibe for the treatment of primary (heterozygous familial and non-familial) hypercholesterolaemia’ did not identify any randomised controlled trials that reported health-related quality of life or clinical endpoints such as cardiovascular morbidity and mortality; in the trials identified, surrogate outcomes such as total cholesterol, LDL cholesterol, HDL cholesterol and triglyceride levels were used as indicators of clinical outcomes.

To represent the population of people with hypercholesterolaemia that is not appropriately controlled with statin therapy, six 12-week fixed-dose randomised controlled trials (n = 3610) were identified that compared ezetimibe plus statin therapy with statin therapy alone.

Seven randomised controlled trials (n = 2577) comparing ezetimibe monotherapy with placebo represented the population where statin therapy is considered inappropriate or is not tolerated. All were 12-week studies and were included in a meta-analysis performed by the Assessment Group.

All trials involved people with primary hypercholesterolaemia with average baseline LDL cholesterol levels ranging from 3.4 mmol/litre to 6.5 mmol/litre and included mixed populations of people with and without a history of CVD.

7.7.3. Cost-effectiveness of ezetimibe

Please refer to results of the cost-effectiveness analysis carried out by the NICE Technology Appraisal 132 (National Institute for Health and Clinical Excellence, 2007).

7.7.4. Evidence into recommendations

Please refer to the NICE Technology Appraisal 132 entitled ‘Ezetimibe for the treatment of primary (heterozygous familial and non-familial) hypercholesterolaemia’.

Footnotes

15

This recommendation has been taken from ‘Statins for the prevention of cardiovascular events’, NICE technology appraisal 94. See www​.nice.org.uk/TA094

16

This recommendation has been taken from ‘Statins for the prevention of cardiovascular events’, NICE technology appraisal 94. See www​.nice.org.uk/TA094

17

‘Higher intensity statins’ are statins used in doses that produce greater cholesterol lowering than simvastatin 40 mg, for example simvastatin 80 mg.

Copyright © 2008, Royal College of General Practitioners.

All rights reserved. No part of this publication may be reproduced in any form (including photocopying or storing it in any medium by electronic means and whether or not transiently or incidentally to some other use of this publication) without the written permission of the copyright owner. Applications for the copyright owner’s written permission to reproduce anypart of this publication should be addressed to the publisher.

Bookshelf ID: NBK55506

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (749K)
  • Disable Glossary Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...