To the detriment of patients and doctors, tools for screening cardiovascular risk are still relatively crude. Risks are not tightly linked to the hypothetical outcome and treatment consists of lifelong drug treatment. So it is no surprise NICE’s proposal to reducing the current threshold from 20% to 10% has been highly controversial.1
I consider there are five main aspects of the new guidance, and I’d like to consider each in turn:
- calculating risk
- estimating NNTs
- reducing morbidity
- monitoring cholesterol
- making decisions with the patient.
Firstly, let’s look at the 10% threshold in context. Despite the furore surround its draft guideline NICE has essentially retained the 10% threshold for cardiovascular risk. A risk threshold serves one important function – identifying which people should be offered treatment.
Many GPs will have a sense of the meagre benefits of treatment at the 10% level of cardiovascular risk, as well as the increasing workload, poor rates of treatment uptake and increasing waits for appointments. Only time will tell how many GPs will embrace a 10% threshold as ‘high risk’, or regard this latest guideline as ‘the wrong medicine at the wrong time’.
NICE has responded to feedback and underlined the importance of first tackling lifestyle before offering statins. Drug therapy is and never should be the first option in low risk people – a point also emphasised in previous guidance (2008).
NICE also acknowledges that patients should be prioritised for risk assessment based on an assessment of their prior risk. This is logical, but the guideline then goes on to suggest that patients with a greater than 10% estimated risk should be offered a formal risk assessment – seemingly counter to the earlier suggestion to prioritise patients.
Secondly, analysis of data published within a Lancet meta-analysis suggests that if there is a benefit at all, the NNT to prevent one death over five years could be the order of 1,000-2,000 patients – very high by conventional standards.2 So for patients within the new catchment population of 10-20% who might take statins, it’s doubtful there is any mortality benefit at all.
Thirdly, the Lancet meta-analysis suggests that statins have an NNT for major cardiovascular events of approximately 216 (216 patients with a risk between 10-20% would need to be treated per annum to prevent one major cardiovascular event) and most of this benefit would be in patients with a risk in the 15-20% part of the range.
Over the last decade death rates from cardiovascular disease have more than halved as a result of both increases in medical treatment and improvements in risk factors, contributing in almost equal measure. This is an important point – we GPs are already routinely use drugs for treating common conditions that have no mortality benefit. However, the difference is that the morbidity of established diabetes is not the same as the morbidity of cardiovascular risk, the latter being only hypothetical.
Fourthly, NICE has dispensed with the need to obtain serial cholesterol measurements by recommending the reporting of non-LDL cholesterol, a major improvement for clinical practice. Another advance is the switch to recommending fixed doses of statins, notably atorvastatin 20 mg for primary prevention, removing any need for titrating treatment against cholesterol levels.
Lastly and perhaps most importantly, comes the patient’s view. Considerable research has been done on patient preferences in cardiovascular prevention but NICE has not considered it. We know patients are much more sceptical about prevention than doctors and need to be presented with much greater benefits to be convinced. And, interestingly, doctors who don’t recommend treatment have been shown to convey much more accurate estimates of the likely benefits of treatment than those that do.3 For cardiovascular prevention the evidence actually shows that most people (they are not quite patients yet) ‘opt in’ to treatment at the 30% threshold but only around half opt in at the 20% threshold. We don’t know how many would opt in at the 10% threshold – except that it will be considerably less.4
The pragmatism and scepticism of patients and GPs means this recommendation will probably have a very slight impact in practice. Guidelines can only suggest when treatment is offered – only patients can decide whether to take them. NICE really needs to update its methodology to formally consider patient preferences – it may be institutionally established but the environment it exists in is a dynamic one.
Of course the validity of any process lies in its output so it is important that guideline groups are constituted with thought, and legitimacy with their intended constituency. In this respect, there is a question whether NICE’s stakeholder consultation process sometimes appears to amount to little more than a defence of its own status quo.
Dr Rubin Minhas is a GP in Kent. He was a member of the 2008 NICE guideline group on lipid modification and a co-author of QRISK2
2 The Lancet. The effects of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: meta-analysis of individual data from 27 randomised trials. 17 May 2012
3 Lewis, D et al. Factors involved in deciding to start preventive treatment: qualitative study of clinicians’ and lay people’s attitudes. 2012. BMJ 327, 841.
4 Friedmann et al. Differences in generalists’ and cardiologists’ perceptions of cardiovascular risk in the outcomes of preventive therapy in cardiovascular disease. Ann Intern Med 1996; 124, 414–421.