I write to contribute to the ongoing debate about the use of Framingham risk calculations in the British population. In his letter (Guidelines in Practice January 2004), Dr Lockyer cites the paper by Brindle et al.1 which demonstrates an overestimation of the 10-year CHD risk in British men.
As you point out in your news story (Guidelines in Practice December 2003), the authors say that the accuracy can be improved by a simple calculation. After making this adjustment, the predicted risk became close to the observed rate at all levels of risk.2
When determining who should receive primary prevention we want to be able to select those at the highest absolute risk of coronary events. We could have a ‘modified-for-Britain’ Framingham equation which would produce proportionately lower absolute 10-year risks; those who currently fall into the high risk (>=30%) group would have the highest values with the new equation.
Based on Framingham scores, an arbitrary cut-off of >=30% 10-year risk was chosen to signify high risk. This value was selected mainly to limit statin prescribing. With our new ‘modified’ equation we could reduce the arbitrary cut-off to, say, >=25% 10-year risk.
The European Society of Cardiology is in the advanced stages of developing a risk calculator that will give us a European evidence-based prediction tool (see http://www.escardio.org/prevention/scorecard.htm).
The society’s recommendation for the high-risk cut-off point is likely to be well below the >=30% Framingham equivalent and it is likely to urge us to provide intervention, including statins, for ever larger numbers of patients.
I fully recognise that we should never lose sight of the major benefits of lifestyle changes, and that we are limited to providing only the prevention that our system can afford. However, risk prediction can be a useful clinical management tool.
It would be a great shame if GPs were discouraged from carrying out risk prediction and consequently primary prevention whether through lifestyle changes or medication.
Dr David F. Law,
GP Registrar, Malvern
Dr Lockyer replies:
I am grateful to Dr Law for his thoughtful and stimulating letter.
I am not sure that an adjustment will be as simple as a ‘bolt on’ formula. Framingham risk tables already fail to take account of other significant features, especially family history. An estimated adjustment is usually made for this. There are no other lifestyle factor adjustments apart from that for smoking.
This is not a question for a GP to answer, but I wonder if we have reached the limit of extrapolation of Framingham data? European data clearly exist, so perhaps we should attempt more accurate modelling for our population.
We often ask asymptomatic patients to take medication for primary prevention for a protracted period of time on the basis of risk analysis. As a GP I wish to feel secure that I am giving the best advice.
One of the other important points made by the paper by Brindle et al. was that the majority of events did not occur in high risk groups. If the paper’s findings are accepted, it will add to the need for clarification of the issues of primary prevention prescribing.
I did not advocate withdrawal from primary prevention treatment with statins. I did, however, suggest that it would be important to assimilate this new information into the guidance that GPs are given.
There may be other issues that GPs have to face. Some patients may feel that they have been treated unnecessarily because of the over-estimated risk.
We are all aware of how anxious some patients can be over health issues such as this.
There is also good evidence at present that we do not prescribe as thoroughly for secondary prevention as we might.
We can argue a strong case for primary prevention in high-risk groups such as patients with diabetes or hypertension, whatever their lipid levels. For example, it has been argued that the Heart Protection Study3 supports generalised use of statins in patients over 40 years with diabetes.
I am not advocating ceasing to prescribe for primary prevention. I am advocating that important research like this should, if verified, be quickly acted upon as it may change the nature of the advice we give our patients.
Dr Matthew Lockyer, GP, Suffolk
- Brindle P, Emberson J, Lampe F et al. Predictive accuracy of the Framingham coronary risk score in British men: prospective cohort study. Br Med J 2003; 327: 1267-70.
- Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: a randomised placebo-controlled trial. Lancet 2002; 360: 7-22.