Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Cholesterol fairy home to roost

Study
Results
Comment

Bandolier 86 we asked why it was that we have to take our statins (usually) in the evening. We thought it might have been a simple pharmacokinetic explanation (short half-lives meaning that statins were more effective in the evening). Some of you wrote with the half-lives (many thanks), but pointed out that with time this should not matter much, even if cholesterol synthesis was higher at night.

Hard evidence was hard to find. What was required was a large study demonstrating that normal doses of evening statin produced convincingly lower cholesterol levels than normal doses of morning statin. A number of readers sent suggestions about papers we should read, and some of these we had read ourselves.

But even so, no convincing evidence. As best we can understand it, the evidence, such as it is, comes from a single study in Japan done 10 years ago [1]. It is a good study, but it doesn't answer the question.

Study


Patients with hyperlipidaemia (cholesterol at least 5.6 mmol/L) were allocated to one of five groups:


It was not stated whether allocation was randomised, but double-dummy methods maintained blinding. Fasting blood samples were taken at the end of a four-week placebo run in period (baseline) and then after 4, 8 and 12 weeks. A number of parameters were measured together at a single laboratory.

Results


The average total cholesterol was about 7 mmol/L in each of the five groups (29-31 patients per group), with standard deviations of about 1 mmol/L or so. Statins, but not placebo, caused falls in total and LDL-cholesterol (Figure 1) over the 12 weeks. Evening statin, and higher dose, were associated with lower total and LDL-cholesterol.

Figure 1: Changes in LDL-cholesterol over 12 weeks caused by different simvastatin regimens in groups of about 30 patients with initial total cholesterol above 5.6 mmol/L (average about 7.2 mmol/L)



Comment


Are these differences credible, and is any credible difference worthwhile?

Credibility is stretched, somewhat. For instance, examine Figure 1, and remember that a standard deviation of over 1 mmol/L has to be superimposed on each of the numbers, meaning that plus and minus two standard deviations is 4 mmol/L when the mean is 4. Hmm! The statistics might tell us there are differences, but we have to strain to see them, so we are not greatly impressed.

Then there's the issue of dose. The 4S study ( Bandolier 15 [2]) had a target range for total cholesterol of 3.0 to 5.2 mmol/L for patients with similar starting cholesterol as here. That study used daily doses of 20 mg or 40 mg simvastatin (and 10 mg in two of 4000 patients). Prescribing Cost Analysis in primary care in England shows prescribed simvastatin as predominantly 10 mg and 20 mg tablets (with some 40 mg). So doses of 2.5 mg or 5.0 mg a day are unrepresentative of doses taken by our patients.

So the evidence we have is this: in a relatively small number of patients, low doses of simvastatin taken in the evening produce slightly more reduction in cholesterol than when taken in the morning. What we do not have is a large study demonstrating that normal doses of evening statin produced convincingly lower cholesterol levels than normal doses of morning statin.

Given that most patients are likely to forget to take their statin in the evening some of the time, one wonders whether messing up dosing regimens for some theoretical benefit is worthwhile. Of course, there may be more convincing evidence, but we can't find it.

References:

  1. Y Saito et al. Comparison between morning and evening doses of simvastatin in hyperlipidaemic subjects. A double-blind comparative study. Arteriosclerosis and Thrombosis 1991 11: 816-826.
  2. Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994 344: 1383-9.

previous or next story in this issue