This site is intended for health professionals only

At the heart of general practice since 1960

The cholesterol debate: the lower the better?

Continuing our series on key clinical controversies, Dr Tony Wierzbicki looks at the evidence for cutting cholesterol levels

Continuing our series on key clinical controversies, Dr Tony Wierzbicki looks at the evidence for cutting cholesterol levels

The epidemiology of hyperlipidaemia shows that cholesterol levels are associated with cardiovascular risk down to levels of 2.5mmol/l for total cholesterol (1.5mmol/l for LDL cholesterol). Statin trials have progressively shown benefits in lower risk populations and at lower final LDL levels.



The Heart Protection Study (HPS) validated the use of simvastatin 40mg in high-risk patients with established atherosclerosis (coronary heart disease [CHD], stroke, peripheral arterial disease) or those at high risk of developing occlusive arterial disease – patients with type 2 diabetes and some with type 1. The Cholesterol Treatment Triallists' (CCT) pooled analysis that showed a reduction of 12 per cent in mortality and 20-25 per cent in cardiovascular events, depending on the endpoint per 1mmol cholesterol reduction.

Indeed, further analysis of the HPS cohort shows that use of simvastatin is cost-effective over a lifetime in prevention of cardiovascular events.

The results of the secondary prevention studies were incorporated in the JBS-2 guidelines, and JBS-2 also placed diabetes as a CHD risk equivalent. For simplicity, the targets in primary prevention were made the same as secondary prevention at 4mmol/l for total cholesterol and 2mmol/l for LDL.

In contrast, the NSF for CHD and QOF use targets of 5mmol/l and 3mmol/l, but have not been reviewed recently. NICE is currently reviewing the evidence for cholesterol targets as part of its review of the post-myocardial infarction (MI) guidelines.

Parallel reviews in other areas are occurring as part of the NICE statins review and the hyperlipidaemia guideline. Yet, despite controversy, all groups agree the minimum target (the JBS audit standard) should be 5 and 3mmol/l. The question is, how much is to be gained from further incremental reductions in cholesterol and at what cost?

Although there is no mortality benefit with more aggressive treatment, there does seem to be a benefit on morbidity and this occurs down to LDL of 1.5mmol/l. While the initial NICE post-MI analysis concentrates on mortality and shows no benefit, data from morbidity assessed in the NICE statins review suggests that there may be benefits to the lower targets in secondary prevention. Thus, it remains possible that lower targets will be validated by NICE and later incorporated into QOF.

The fall-out

Statins are now the new aspirin – everybody with cardiovascular disease gets them. As these are retrospective analyses, controversy has begun to rage over whether lower is better when high-dose statin is compared with lower dose in prospective studies.

The NSF and QOF have led to more than 70 per cent of patients with atherosclerosis needing to be on treatment and achieve a total cholesterol less than 5mmol/l as a definition of adequate care. As the average cholesterol in post-MI patients is 5.6mmol/l, moderate statin therapy achieves the current targets easily, but at significant cost to the NHS (£1bn).

There are now strenuous efforts to promote the use of generic statins (simvastatin and pravastatin) as their cost differential allows five to six patients to be treated for each one on an on-patent brand.

Efficacy equivalence matrices are commonly used as the basis of switch policies (see box below). Most patients achieve better LDL control on 40mg simvastatin than the most commonly used 10mg atorvastatin dose. Some, however, get worse results although, given the variation in daily lipid levels, it is difficult to be certain until repeated measurements have been made.

To achieve lower targets both higher doses and substantial use of currently on-patent statins are necessary, as well as additional therapies to lower cholesterol further. The financial consequences are substantial.

Reflecting on the evidence

Acute coronary heart disease

Two trials, PROVE-IT and A-Z, have addressed the controversy in patients with unstable angina or non-Q wave MI, where risks of subsequent coronary events are extreme. The markedly different designs make it difficult to compare the studies, but it seems that a lower LDL (1.75mmol/l) is associated with benefits in terms of reduced morbidity especially in patients undergoing angioplasty.

Chronic coronary disease

Two trials have addressed the chronic benefits of lowering LDL aggressively in patients with established CHD. In TnT, 10,001 patients received 10mg or 80mg atorvastatin and achieved LDL levels of 2.6mmol/l (imputed 35 per cent reduction) and 2.0mmol/l (imputed 50 per cent reduction) respectively. The difference was associated with 22 per cent relative risk reduction in coronary events, although no effect was seen on the very low cardiovascular mortality rate. The benefits were achieved at the price of a 3 per cent rate of significant liver dysfunction (elevated transaminases). In the IDEAL trial, 8,888 patients received 20-40mg simvastatin in an analogue of the 4S study protocol, while the other group received 80mg atorvastatin. Lipid changes were similar to those in TnT and similar results were obtained.

Diabetes and primary prevention

The 'lower is better' hypothesis in primary care is based mostly on lowest quartile data from positive trials in type 2 diabetes (such as CARDS) using 10mg atorvastatin.

However, this data has been complicated by the results of the ASPEN trial where 10mg atorvastatin resulted in only a 10 per cent cut in cardiovascular events in a complicated mixed primary and secondary prevention group despite a 35 per cent (1.1mmol/l) LDL reduction. These results duplicated the extensive type 2 diabetes group (6,000 patients) in the ASCOT trial where similar results were obtained.

Data from the ASCOT study in patients without diabetes showed benefits in the lower quartile of lower LDL in post hoc analyses, but implies a high numbers needed to treat (NNT) even in patients with three to four cardiovascular risk factors.

Also in ASCOT a positive interaction was seen between statins therapy and calcium-channel blocker as opposed to ß-blocker-based antihypertensive therapy.

The recent MEGA trial using 10-20mg pravastatin and achieving a 15 per cent LDL (0.6mmol/l) reduction in primary prevention patients with a CHD event rate of 5 per cent per decade showed a 33 per cent reduction in events at five and 10 years with a NNT of 119. Intriguingly, the final LDL achieved was 3.30mmol/l and the benefits were mostly seen in the minority of men recruited to the study.

Practical implications for GPs


The evidence base exists for the lower 4 and 2mmol/l JBS-2 targets in patients with established CHD (secondary prevention) in terms of morbidity with a NNT of 40. A pooled analysis from CTT for diabetes is awaited, but again given their high cardiovascular risk it is likely that the NNT will be sufficiently favourable for their treatment to be cost-effective (NNT less than 100).

The NNT in the primary prevention meta-analyses is large (generally three times compared with secondary prevention) for 5 and 3mmol/l (for instance NNT AF/TexCAPS = 54) so it may not be cost-effective to go to 4 and 2mmol/l.

Which statin?

Simvastatin is a moderately potent agent. Atorvastatin and rosuvastatin are more potent on straight dose comparison. All statins lower LDL and triglycerides so apart from a differential ability to reach targets based on efficacy there is little between them.

What dose?

The only solid prospective evidence is for greater LDL cholesterol reduction with higher dose studies. Small-scale open follow-up studies suggest that cholestyramine added to statins reduces progression of atherosclerosis.

As yet there is no evidence for the use of ezetimibe, although studies suggest that any mechanism of reducing LDL, including ileal bypass surgery, will result in lower CHD event rates. Side-effectsHigher dose statins cause more side-effects, with discontinuation rates being 3-4 per cent for transaminase elevations in IDEAL and TnT with atorvastatin.

The elevation has to be consistent on repeat measurement. In terms of liver function, only AST or ALT matter. Gamma-GT is completely irrelevant. The discontinuation level is a rise to three times the upper limit of normal (about 150iu/l).

With myalgia, no rise in CK is seen. In myositis and rhabdomyolysis, CK levels rise and are associated with symptoms. NICE recommends CK is only measured in cases with symptoms of myopathy.

A consistent elevation to 10 x ULN (about 1,000iu/l) mandates cessation of statin therapy. A more conservative approach applies to African patients where normal CK levels of 500-1,000iu/l are seen and a rise to five times baseline is considered significant.

How to do it

The presumption is to start with 40mg simvastatin – most patients will be adequately treated including 50 per cent of those who need secondary prevention. Thereafter, a protocol similar to that shown below can be followed. If LDL is close to target on simvastatin, then either simvastatin should be titrated to 80mg (often 2 x 40mg tablets) or a switch should be made to atorvastatin 40mg or rosuvastatin 20mg and, if necessary, further titration performed.

If, however, LDL is more than 1mmol/l away from target, it is unlikely to be achieved on dose titration/statin switching. In this case, adding 10mg ezetimibe or in extremis cholestyramine should be sufficient.

Ideally, all cases of familial hyperlipidaemia and ultra-high risk cases, or resistant or consistently intolerant patients should be assessed by a lipid or cardiovascular prevention clinic.


statin matrix cholesterol timeline statin flowchart

Rate this article 

Click to rate

  • 1 star out of 5
  • 2 stars out of 5
  • 3 stars out of 5
  • 4 stars out of 5
  • 5 stars out of 5

0 out of 5 stars

Have your say