GP and hospital practitioner in cardiology Dr Matt Hughes outlines five aspects of risk management that have been subject to recent debate
Framingham versus QRISK
The Framingham risk score is based on a massive dataset. The Framingham Heart Study is a long-term, ongoing cardiovascular study of the residents of one US town which began in 1948 and is now on its third generation of participants.
But even in 1998 – when the score was launched in its current form – the limitations of applying the risk score to European populations were recognised.
So when the QRESEARCH database – of over 10 million UK patients in 550 practices – was used to develop QRISK, many predicted it would quickly supersede Framingham. We should also not forget ASSIGN – a risk score based on Scottish data that has been endorsed by the Scottish Intercollegiate Guidelines Network.1
But Framingham has proved surprisingly tenacious, and not just because it’s firmly lodged in the back of the BNF. Many specialists have serious doubts that QRISK should automatically take its place.
For this article it would be useful to look at the advantages and disadvantages of the new kid on the block – QRISK – compared with the old hand, Framingham.
• QRISK scores are calibrated to a UK population, and a much more recent one than Framingham.
• It has been subject to UK validation using a separate research database.
• Risk is adjusted for variables not included in Framingham – like deprivation and current treatment for hypertension.
• Although the QRISK dataset is British and more recent, it is less complete than Framingham’s. QRISK relies on age and sex-based averages for missing data which are then interpolated – by as much as 60% for HDL-C, for example.
The one thing that will finally decide this issue is the unequivocal recommendation of one over the other in a major UK guideline – something we’re still waiting for. NICE was seen by some as ducking the issue when it announced in March 2010 that Framingham was no longer the tool of choice and that GPs should use whichever was available locally.
I’d say utility is a major factor, so you should probably choose the one you feel is more user-friendly. But I think the evidence suggests the balance is in favour of QRISK.
I suspect the switch from Framingham, although slower than many thought, is inevitable.
But there’s one very interesting and important message from the past few years of research into this issue. It’s glaringly clear from the development and validation of QRISK that far too many patients are still not getting a comprehensive cardiovascular risk assessment.
Getting all JBS2 recommendations into routine practice would probably make more of a difference to the population than choosing a newer, perhaps marginally more accurate, risk tool over an older, but still incredibly useful, one.
Co-morbidities and cardiovascular risk
The development of CVD is associated with risk factors that our patients can – in theory at least – change or improve: smoking, lipid levels, physical activity, diet. Other factors are fixed, like age and sex.
But certain co-morbidities can also increase cardiovascular risk – including diabetes, chronic kidney disease, rheumatoid arthritis and psoriasis.
Some studies suggest coronary mortality is as high in patients with diabetes without CHD as in patients with CHD, which led some to consider diabetes as a ‘coronary equivalent’. So people with diabetes are generally classified as at high risk, and the advice from NICE is:
• consider patients to be at high premature cardiovascular risk for their age unless they:
– are not overweight
– are normotensive – less than 140/80mmHg if not on antihypertensive therapy
– have no evidence of microalbuminuria
– are a non-smoker
– do not have a high-risk lipid profile
– have no history of CVD and no family history of CVD
• if the person is considered not to be at high cardiovascular risk, estimate cardiovascular risk annually using the UK Prospective Diabetes Study (UKPDS) risk engine (www.dtu.ox.ac.uk/riskengine)
• consider using cardiovascular risk estimates from the UKPDS risk engine when discussing cardiovascular complications with the individual
• order a full lipid profile when assessing cardiovascular risk after diagnosis and annually, and before starting lipid-modifying therapy.
Chronic kidney disease
People with CKD have a far greater likelihood of cardiovascular death than progression to established renal failure.
But the exact relationship between CVD and renal disease is not entirely clear – making it more difficult to estimate cardiovascular risk in these patients.
Currently, the recommendations are to use the usual risk tools. CKD patients with a 10-year risk greater than 20% should receive statins for primary prevention in the same way as in people without CKD.
But most of the studies on statins in cardiovascular prophylaxis have excluded patients with significant renal impairment – although some patients with CKD3 will have been included.
The ongoing Study of Heart and Renal Protection is looking at the role of lipid lowering in patients with CKD and should help shed some light on this issue.
Cardiovascular risk is greater in patients with inflammatory arthritis, compared with those without. The increase seems to be as high as that seen in type 2 diabetes, at 1.5- to 3-fold.
Any patient with inflammatory arthritis should now be undergoing routine monitoring for cardiovascular risk factors.
Last year, the European League Against Rheumatism updated its guidance on cardiovascular risk management in patients with rheumatoid arthritis, including a recommendation to multiply a patient’s risk score by 1.5. But in the UK, QRISK now incorporates an additional weighting for these patients.
Psoriasis seems to be an independent risk factor for CHD, with the greatest relative risk in young patients with severe disease.
This is an active area of research and unfortunately we do not have any clear guidance on whether we should manage these patients differently.
The firmest piece of evidence yet comes from a study presented at the 2010 American College of Cardiology meeting, which suggested that patients with severe psoriasis were at increased risk of an MI, atrial fibrillation (AF), stroke, and percutaneous coronary intervention (PCI). Even patients with mild disease were more likely to experience AF, stroke and PCI.2
What we do not know is whether patients with psoriasis should receive statin therapy earlier than predicted by traditional risk scores.
But it is clear that patients with severe psoriasis or psoriatic arthritis should have their BMI, blood pressure, lipid profile and blood glucose checked annually.
Lifetime versus 10-year risk
Ten-year estimation has been the basis of risk assessment for around 35 years, but that could soon change as the Joint British Societies considers whether to recommend lifetime risk estimation in their next set of guidelines. The change is being considered due to concerns that relying on 10-year risk scores means therapy is started too late for maximum benefit.
It is a persuasive argument, and it’s true that to be classified as being at a high cardiovascular risk under current guidance means you need to be pretty old. The years when heart disease is quietly developing are being ignored.
The JBS guidance is a scientifically rigorous set of standards and will be hugely influential, but the change would have profound implications. QOF and NICE guidance are, of course, based on 10-year risk, and they would both either have to change or risk years of confusion when JBS recommendations would once again be at odds with other guidelines.
It would also breathe new life into the debate on mass medication. We would be asking many more younger people to consider the risks and benefits of starting drug therapy for life – and adding to the problem of polypharmacy in those already on some sort of prescribed medication. The resource issue also needs serious debate.
The developers of QRISK have released a lifetime cardiovascular risk calculator available at www.pulsetoday.co.uk/downloads, but currently there are no recommended thresholds set for initiating measures to reduce risk based on lifetime risk.
Aspirin in primary prevention
The practice of advising people with no history of cardiovascular disease to take aspirin is now in serious doubt. The evidence suggests the risks do not justify the benefits – partly because long-term, low-dose aspirin substantially increases the likelihood of gastrointestinal haemorrhage. But also because the evidence has always been pretty equivocal; the latest studies have simply undermined its use further.
In 2009, a meta-analysis of six trials involving 95,000 patients found aspirin use reduced serious vascular events by about 0.07% per year – mainly due to an absolute reduction of about 0.05% in the likelihood of non-fatal MI. But it resulted in an absolute increase in major gastrointestinal or other extracranial bleeding of about 0.03% per year. Rates of all-cause mortality, death due to coronary heart disease and stroke did not differ significantly between the aspirin and the control groups.3
But three major UK guidelines published between 2005 and 2008 – SIGN, JBS and NICE – recommended aspirin for the primary prevention of cardiovascular disease in various groups of patients, including some with type 2 diabetes and those with a 10-year CVD risk of 20% or more. And last year the British Hypertension Society reaffirmed its guidance recommending aspirin for primary prevention.4
Of course, the aspirin naysayers will ask whether, if aspirin is so useful in primary prevention, why is it still unlicensed for that indication? All I can add to the debate is to set out what I do in practice.
I haven’t used low-dose aspirin for the primary prevention of cardiovascular disease for a few years, and when I get the chance I review patients currently taking it and try and involve them in the decision about whether to carry on with treatment.
New biomarkers for assessing risk
Finding a new biomarker that would help us more accurately assess cardiovascular risk is something of a holy grail for clinical biochemists. And a huge number have been proposed over just the past five years, including homocysteine, fibrinogen, various lipoproteins and apolipoproteins and a protease inhibitor called cystatin C.
Arguably, the most promising have been C-reactive protein and B-type natriuretic peptide. The body of evidence linking plasma levels of CRP with CVD risk is fairly impressive.
Last year a paper was published in The Lancet providing convincing evidence that the relationship between CRP and CVD risk is consistent and about as strong as that of major lipids and blood pressure.5
However, causality has yet to be demonstrated – and I suspect that is holding it back from being recommended for clinical use. Nonetheless, CRP is a strong and consistent risk marker for CVD risk, and the Framingham Heart Study as well as the Women’s Health Study suggested that including CRP levels in CVD risk prediction improves risk prediction.
As for BNP, investigators on the Framingham Heart Study found it to be the best of the panel of circulating biomarkers examined. But the association wasn’t yet thought to be strong enough to recommend routine measurement of BNP for cardiovascular risk assessment, although NTproBNP might show more promise.
Both these markers are subject to ongoing research, and naturally the fact they are both currently used routinely in primary care might well make their incorporation into future cardiovascular recommendations more likely.
Dr Matt Hughes is a GP and hospital practitioner in cardiology in Cardiff
Competing interests None declared
The use of asprin in primary prevention is now in serious doubt Asprin