How should we balance individual and population benefits of statins for preventing cardiovascular disease?BMJ 2011; 342 doi: http://dx.doi.org/10.1136/bmj.c6244 (Published 26 January 2011) Cite this as: BMJ 2011;342:c6244
- Aroon D Hingorani, professor of genetic epidemiology1,
- Harry Hemingway, professor of clinical epidemiology2
- 1Genetic Epidemiology Group, Department of Epidemiology and Public Health, University College London, London WC1E 6BT, UK
- 2Clinical Epidemiology Group, Department of Epidemiology and Public Health, University College London, UK
- Correspondence to: A D Hingorani
- Accepted 18 October 2010
Guideline groups in the United Kingdom and the United States are reviewing recommendations for the primary prevention of cardiovascular disease. At their disposal will be good quality evidence on the quantitative relation between the major risk factors and the probability of a first cardiovascular event1; the extent of the current population exposure to risk factors2 3; and the safety and efficacy of drugs to lower blood pressure and cholesterol concentrations.4 5 The decisions of these expert groups will have far reaching consequences for the millions of adults in both countries, where cardiovascular disease is the biggest cause of mortality and morbidity; around half of men and a third of women have a cardiovascular event during their life.6
There are new drivers to modify the existing guidance (table 1⇓; box), and two extreme positions could be envisaged (table 2⇓). The first is to redouble efforts to identify high risk individuals by enhancing currently used risk prediction tools with new information from blood biomarkers or non-invasive vascular imaging; to treat those at high risk with newer more expensive statins that achieve the greatest cholesterol reduction; and to tailor treatment for each individual to achieve target cholesterol levels. The interest in C reactive protein as a new biomarker of cardiovascular risk7 and the recent US Food and Drug Administration licence extension of the patent for rosuvastatin, with C reactive protein as a companion test, suggests that the US could follow this course. The diametric alternative is to use generic versions of the older statins in a wider population by including people whose risks fall below the current absolute risk thresholds for drug intervention and to dispense with a target cholesterol level. In the UK, eligibility criteria for statins in primary prevention have been relaxed over the years and a switch to generic statins, where possible, is already saving substantial sums.8
Pressures for changing primary prevention strategy
Drivers for widening statin eligibility
Reduced cost through patent expiry and availability of inexpensive generic formulations with excellent long term safety profile
Policy emphasis on disease prevention
Development of polypill concept
Drivers for more individualised primary prevention
Emphasis on personalised medicine
Claimed predictive value of new biomarkers and imaging tools such as C reactive protein, carotid intima-media thickness, and coronary artery calcification
But is the first approach an unacceptably expensive strategy that fails to exploit increased opportunity for disease prevention from wider access to effective, safe, inexpensive generic statins? Is the second insufficiently refined for an era of personalised or stratified medicines where the aim is to maximise individual benefit and minimise harm?
High risk individual v population based approaches
Geoffrey Rose developed the important concept of the “prevention paradox”: that more cases of cardiovascular disease occur among the majority at average risk than among the minority at high risk.9 The paradox arises because risk factors such as cholesterol and blood pressure are a continuum. Each exhibits a log-linear association with risk of coronary heart disease with no safe threshold, while the population risk factor exposure follows a normal distribution.10 A consequence is that the distribution of blood pressure and cholesterol values overlaps substantially among those who do and do not develop coronary events later in life, leading to a seemingly counterintuitive observation that risk factors such as low density lipoprotein (LDL) cholesterol are poor predictors of clinical events despite being causally related to coronary heart disease.
The prevention paradox leads to a tension. Focusing exclusively on people with high levels of one risk factor, or at high absolute risk of a clinical event calculated from multiple risk factors, overlooks the burden of events that will occur among the average majority. Conversely, a population based strategy, which seeks to address this, leads to people at low individual risk being exposed to an intervention with less personal gain. The merits of each approach are intimately linked to the efficacy, safety, cost, and convenience of the available interventions. In Rose’s era, the lipid lowering drugs were poorly tolerated and only modestly effective. For this reason, dietary and lifestyle interventions became aligned with the population approach, the aim being to achieve a large overall benefit by even modest shifts in the risk factor distribution in the whole population.
Primary prevention in the early statin era
The first statins to market were better tolerated and lowered cholesterol more effectively than the preceding drugs.11 However, their broader use in primary prevention was constrained initially by the high costs and uncertainty about long term safety. An individualised approach to primary prevention therefore persisted in the UK based on absolute risk. Absolute risk was chosen rather than LDL cholesterol because LDL cholesterol on its own poorly differentiates those who will have events and two people with the same LDL cholesterol concentration can have widely differing risks of coronary heart disease depending on other risk factors such as age, sex, smoking habits, and blood pressure.12 Targeting statins on the basis of absolute risk makes the justifiable assumption that the relative risk reduction from statin treatment is constant (such that the absolute benefit and number needed to treat are proportional to absolute risk) and that the particular constellation of risk factors in an individual does not modify the treatment effect.13 Statins are as effective in people whose cardiovascular risk is mainly influenced by high blood pressure or diabetes as among people whose cholesterol concentration is the main determinant of risk.
Europe and Australasia have broadly adopted a similar absolute risk based approach to intervention, which has encouraged use of computerised, point of care risk assessment tools based on results from observational studies like the US Framingham Heart Study or routinely collected clinical data such as QRISK in the UK.14 However, in the US, where risk assessment was initially based on LDL cholesterol alone, guidelines continue to recommend consideration of both LDL cholesterol and absolute risk when prescribing statins for primary prevention.
Why change guidance?
The absolute risk based approach has limitations. Risk assessment models assign individuals to groups with observed event rates close to those predicted. But they perform little better than their constituent variables (for example, age, blood pressure, and cholesterol) in differentiating individuals who eventually have events.15 Just as with cholesterol, many cardiovascular events occur in people at intermediate risk, below current thresholds for statin eligibility.
Raised concentrations of blood biomarkers such as C reactive protein and subclinical atherosclerosis in the carotid or coronary artery, detected by ultrasonography or computed tomography, have been associated with a higher risk of cardiovascular events. And at the same time, the cost of the first statins to market (simvastatin, pravastatin, etc) is dropping abruptly because their patents have expired. For example, generic simvastatin 40 mg daily now costs less than £1.40 (€1.60; $2.30) a month compared with £24.64 a month for 20 mg atorvastatin, the patent on which expires next year. Evidence of long term safety is available for established statins but not yet for the newer drugs. These developments are opening up two approaches to deal with the current limitations of absolute risk assessment based on established risk factors.
The first seeks to apply new technology, whether biomarkers or imaging tools, to identify more accurately people at intermediate risk who will have events. The second seeks to circumvent the inherent difficulties in prediction altogether by simply offering statins to a wider range of adults than would currently be treated. In effect, this represents a shift towards a population based approach to prevention that includes use of cholesterol lowering drugs as well as dietary and lifestyle measures. How do the options compare?
New tools for risk assessment
The European view is that C reactive protein is little better than cholesterol at predicting risk (for similar reasons) and adds little to existing risk models.16 However, C reactive protein is already considered an option for risk assessment in the US, based in part on a recommendation from the American Heart Association and Centers for Disease Control and Prevention17 and is now established in Canadian guidelines.18
A recent meta-analysis found that thickness of the carotid intima-media and identification of plaque were insufficiently useful for screening people at intermediate risk,19 and comparison trials have not been conducted for primary prevention. It is therefore unsurprising that neither carotid nor coronary imaging has yet been adopted by the NHS for primary prevention and that the American Heart Association, the American College of Cardiology, and the US Preventative Services Task force have been cautious about their role in predicting risk.20 21 Despite this, the newly enacted Texas Heart Attack Prevention Bill requires health insurers to cover up to $200 towards the cost of measuring coronary calcium or carotid intima-media thickness every five years. The bill followed publication of guidelines from the Society for Heart Attack Prevention and Eradication (SHAPE) task force,22 which were not part of the established guideline development framework of the American College of Cardiology or American Heart Association.23
Wider use of generic statins
The most recent guidelines from the Scottish Intercollegiate Guidelines Network and National Institute for Health and Clinical Excellence (NICE) lowered the risk threshold for prescribing statins from a 10 year coronary heart disease risk of 30% (roughly equivalent to a 10 year cardiovascular disease risk of 45%) adopted in the UK soon after the introduction of statins, to a 10 year cardiovascular disease risk of 20% (roughly equivalent to a 10 year coronary heart disease risk of 15%). The proposed Department of Health vascular health checks for all 40-74 year olds would also move from opportunistic testing of cardiovascular risk factors in primary care to systematic identification of all eligible adults.24
But are the new risk thresholds and vascular health checks an equitable means of expanding the use of statins? Around half of men aged over 50 years in England and Wales will now be eligible for statins, but a substantial proportion of all events would be expected in men of this age group whose risk falls below the threshold of 20%, arguably leading to inequity of access to statins. Furthermore there are unresolved concerns about the extent to which risk thresholds may address, or exacerbate, social inequalities in vascular risk.25 The number needed to screen in the vascular checks to prevent one event is also estimated to be high: 449 for men and 1638 for women.26
An alternative is to offer statins on the basis of age with no risk factor screening. Over 95% of cardiovascular events occur after the age of 50, and age is the overarching determinant of absolute risk.27 Age on its own may be nearly as effective in discriminating cardiovascular events as risk equations that incorporate additional variables. Age based eligibility for statins would obviate the need for risk factor screening and reduce potential inequity of access to statins. However, it would result in large numbers of adults at low risk taking drugs for many years, which may make it difficult to implement.
Is there any role for newer more expensive statins?
A lower “on-treatment” LDL cholesterol level within trials and a larger average LDL cholesterol reduction across trials have been associated with greater reductions in cardiovascular risk. As a result, some guidelines on primary prevention have proposed treatment to target cholesterol concentrations.28 However, the cost effectiveness of more intensive lowering of cholesterol with expensive patented statins (such as rosuvastatin) versus less intensive lowering with cheaper generic drugs has not been evaluated in primary prevention. Moreover, someone taking simvastatin but failing to reach an arbitrary cholesterol target may stand to gain a similar or greater reduction in cardiovascular risk from the addition of an inexpensive generic blood pressure lowering drug as from a switch to a different statin. The principle of targeting multiple risk factors simultaneously to maximise risk reduction at low cost is being evaluated in trials of combination tablets containing generic blood pressure lowering drugs and statins (polypills).29
From guidelines to health policy
Rose recognised that strategies for preventing cardiovascular disease have sociopolitical repercussions and their development could benefit from involvement not only of medical experts but also policy makers and patients. Studies are now required to evaluate the preferences of people being targeted for primary prevention, who have yet to be properly invited to the debate and to formally model the cost effectiveness of the different screening options. Prevention is now high on the health agenda. Recent guidance from NICE on prevention of cardiovascular disease at population level30 is aimed at “government, the NHS, local authorities, industry and all those whose actions influence the population’s cardiovascular health” and focuses on “legislative, regulatory, and voluntary changes” relating to salt, saturated fat and trans fat consumption, food marketing and labelling, public sector catering, and increasing physical activity.
In an era of safe, inexpensive generic statins where new methods for risk assessment poorly discriminate cases of cardiovascular disease, the balance of evidence appears currently to favour wider eligibility for statins, as part of a broader population based effort to reduce cardiovascular risk. However, should new preventive treatments of uncertain long term safety emerge from ongoing clinical trials, these would again need to be targeted at people at highest risk, applying the most cost effective screening tools available.
Cite this as: BMJ 2010;341:c6244
We thank Bruce Psaty for previous discussions that helped to formulate several of the ideas included in this article.
Contributors and sources: ADH and HH are engaged in research on the primary and secondary prevention of cardiovascular disease. Both have an interest in the critical appraisal of new biomarkers. ADH is an honorary consultant in the University College London Hospital’s cardiovascular risk clinic.
Funding: ADH is supported by a British Heart Foundation senior research fellowship FS05/125.
Competing interests: All authors have completed the unified competing interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no support from any organisation for the submitted work; ADH has received honorariums for speaking at courses and meetings on cardiovascular risk prediction, which were donated in part to charity, and is the lead investigator of an MRC biomarkers award cofounded by Pfizer; ADH is on the editorial board of Drug and Therapeutics Bulletin, has provided non-remunerated advice to GlaxoSmithKline and London Genetics, and is a member of the JBS3 Guidelines development group.
Provenance and peer review: Commissioned; externally peer reviewed.