This week, the BMJ has published a study that adds further to the debate regarding the safety of statins. It is officially recognised that one potential adverse-effect of statins is the development of diabetes. In this study, Canadian researchers pooled the results of several studies to see if the risk of diabetes was higher in individuals (with known cardiovascular disease) taking higher potency statins rather than lower ones. It was: compared to lower potency statins, the taking higher potency ones were associated with a 15 per cent increased risk of diabetes.
Now, I should make clear that evidence assessed in this study was epidemiological in nature. This means it shows that while the review found a link between higher dose statins and increase diabetes risk, this is just an association. However, we do know from other clinical studies known as randomised controlled trials (RCTs) that statins do have a genuine capacity to cause diabetes.
Perhaps in the ideal world we would be able to rely more heavily on RCTs to judge the safety of statins. The problem is, as I wrote this week, there are many reasons why RCT-derived data is utterly unreliable in this respect. The fundamental problem is that RCTs are often designed or reported in a way that can massively downplay any hazards or harms.
A case in point, is potential hazard of diabetes. Earlier this year, saw the publication of a study which claimed that side-effects from statins were no more common than those seen in people taking placebo . Actually, as I have pointed out before, the authors of this study acknowledge several of the deficiencies in the data.
Within the discussion of this paper we find the following text:
…commercial sponsors of clinical trials may not be motivated to search exhaustively for potential side effects. One pointer towards this is that…new diagnosis of diabetes was only documented in three of the 29 trials.
Yes, you read the correctly: only in about 10 per cent of the trials assessed in this review was the development of diabetes even documented. You don’t find if you don’t look, and of course what this means is that we simply can’t rely on the RCTs to assess the diabetes risk that statins bring.
While the potential diabetes risk from statins is now established, the conventional response from those generally in favour of statin therapy has been that the benefits ‘massively outweigh’ the risks. Actually, the evidence on adverse effects is so corrupted and incomplete, this claim does not necessarily stand up at all.
The authors of the recent review question the conventional line here too, when they write that this it is based on data from “..trials specifically designed to measure cardiovascular events but where diabetes events were not recorded in a consistent or required way.”
But the authors have other questions to ask too, when they point out that while statins reduce risk of overall mortality in those with known cardiovascular disease (secondary prevention), higher potency statins do not work any better than lower one here. Bearing this in mind, the authors suggest that doctors should consider the potential diabetes risk of prescribing high potency statins to patients in the secondary prevention setting.
1. Dormuth CR, et al. Higher potency statins and the risk of new diabetes: multicentre, observational study of administrative databases. BMJ 2014;348:g3244 (Published 29 May 2014)
2. Finegold JA, et al. What proportion of symptomatic side effects in patients taking statins are genuinely caused by the drug? Systematic review of randomized placebo-controlled trials to aid individual patient choice. European Journal of Preventive Cardiology 2014;21(4):464-74
And in fact if you look at the numbers needed to treat, then statistically only a relatively small number of people will benefit from statins anyway, while there have been several studies showing that, while they may reduce the incidence of cardiovascular events, statins have no effect on overall mortality rates in women.
When I stopped taking the statin (40mg Simvastatin) over a year ago my HbA1c dropped quite dramatically from 6.7 to 5.6. It seems very odd to me that as a T2 diabetic with no history of CV disease I should have been prescribed a drug which would drive my BG levels up. Now my cholesterol level has gone up from 4 point something to 7.3 but my triglycerides are very low and the ratios are spot on and I feel fine. I’m a bit on the old side and understand that elevated blood lipids are good for me and infinitely better than elevated BGs.
Thank you, Dr. Briffa, for all your words of wisdom. I look forward to your blog every week. Thank you.
I just read Stephanie Seneffs essay on sulphur deficiency that sits on the weston price foundations website. Seneff is well up on sulphur and its involvement on glucose metabolism. Consequences of a low fat diet and statin therapy might result in insufficient cholesterol / cholesterol sulphate being supplied to Gi track and this may have knock on issues for glucose metabolism.
The results of the lipid profile test do not inform much, so why fret over them, but improved BG control is the way to go to avoid complications that can follow poor BG control.
There is no need to fear fat in the diet and fat won’t necessarily lead to weight gain, the odds favour the reverse. Stand your ground if you feel happy with your lipids, and there is nothing to fear from saying no to a statin. Cholesterol is good. Oxidation of cholesterol is ‘bad’. Steer clear of veg oil and marg, eat fresh veggies, consider a multivitamin.
Thank you, Chris, for your reply. Happy to say I do all the right things. I wish I could persuade my friends that fat doesn’t make you fat. I’m a living example but I think they’re all waiting for me to drop down dead. They are horrified that I eat eggs, butter, fatty meat, cheese and loads of nuts in addition to lots of vegetables and not very much fruit. Why, I have even learned to love the Brussels Sprout – (with garlic and coconut oil)
CORRECTION to published table
Cholesterol categories No dementia %age AD + VaD %age TOTAL %age
NO TREAT (239 mg/dl) 2932 50.31 215 3.69 3147 54.00
Totals 5477 93.98 351 6.02 5828 100
Why it did not come out like this – Don’t know
What is truly mind-boggling is that part of the “protocol” for treating diabetes is to prescribe statins!
Study says that statins may lead some patients to pig out (JAMA Internal Medicine April 2014). May be an independant effect on the diabetes risk in statins users? Statins as an excuse to overeat, mainly bad fried fatty foods…
Pig Out???? Not I, my friend. And not all diabetics are overweight and/or given to ‘pigging out” whether on statins or not.
The Great Cholesterol Myth by Jonny Bowden and Stephen Sinatra is a very enlightening book which may safe your life:
• The benefits of statin drugs have been widely exaggerated, and
any benefit of these drugs has nothing to do with their ability to
• Statin drugs deplete coenzyme Q10, one of the most important
nutrients for the heart. Depletion of CoQ10 can cause muscle
pain, weakness, and fatigue.
• The brain depends on cholesterol to function optimally. Cholesterol
helps stimulate thinking and memory.
• Statin drugs lead to a reduction in sex hormones, as shown by
several studies. Sexual dysfunction is a common (but underreported)
side effect of statin drugs.
• Statins interfere with serotonin receptors in the brain.
• There are troubling indicators that statin drugs may be associated
with a higher risk for cancer and diabetes.
• A comprehensive study by a University of California, San Diego,
School of Medicine researcher showed that a majority of doctors
dismiss complaints of side effects from statins and do not report
them to MedWatch, the FDA’s system for reporting any undesirable
experiences associated with the use of medical products or
drugs (experiences collectively known as “adverse events”). In
other words, side effects are grossly underreported.
Your tendency to express yourself using exaggerated language does your articles no favours at all. How is someone reading your article supposed to reconcile “we do know from RCTs that statins do have a genuine capacity to cause diabetes” with “RCT-derived data is utterly unreliable”?
On a related issue, I think you’ve missed a key finding of this study. The authors compared their results with the RCT meta-analysis data. They stated in their discussion:
“The overall rate ratio for current treatment that we observed in our analysis of statins and diabetes (rate ratio 1.15) was similar to the relative risks reported in the meta-analyses by Rajpathak et al (relative risk 1.13), by Sattar et al (odds ratio 1.09), and by Preiss et al in a meta-analysis of trials comparing higher potency and lower potency statins (odds ratio 1.12).”
(It’s also only fair to point out here that the RCT meta-analyses are based on many more than the “3 of 29 trials” you mention – see their introduction for the numbers.)
Put simply, the results from this study are similar to the those of the RCT meta-analyses, which you think were based on “corrupted” and “utterly unreliable” data. Maybe if you’d thought more critically about your own (very strongly held) opinions, you may have come to the conclusion that perhaps, just perhaps, the RCT data isn’t as corrupt and unreliable as you thought it was?
“Your tendency to express yourself using exaggerated language does your articles no favours at all. How is someone reading your article supposed to reconcile “we do know from RCTs that statins do have a genuine capacity to cause diabetes” with “RCT-derived data is utterly unreliable”?”
Well, how about this, Mark: The RCTs have traditionally been designed and reported in a way that means many adverse effects will not be picked up or logged. There’s myriad reasons why RCTs may not accurately reflect reality. So, when RCTs pick up an increased risk of diabetes, we can be pretty sure it’s a real effect. The problem is, with the evidence as it stands, we cannot be at all certain on the extent of this effect as well as the burden of other adverse effects. Make sense?
“(It’s also only fair to point out here that the RCT meta-analyses are based on many more than the “3 of 29 trials” you mention – see their introduction for the numbers.)”
Yes, Mark, but at the risk of patronising you, the number of studies included in a meta-analysis is not the relevant number: it’s the number of studies that have reported the adverse effect we’re looking at, right?
“Maybe if you’d thought more critically about your own (very strongly held) opinions, you may have come to the conclusion that perhaps, just perhaps, the RCT data isn’t as corrupt and unreliable as you thought it was?”
Well, Mark, I would tend to conclude that if you are simply failing to understand the reason why the data are corrupt and unreliable. I can’t make you understand it or make you want to understand it, but that does not detract from the fact that the issues are real.
I say we turn a blind eye to them at our patients’ peril.
“The RCTs have traditionally been designed and reported in a way that means many adverse effects will not be picked up or logged”
“the number of studies included in a meta-analysis is not the relevant number: it’s the number of studies that have reported the adverse effect we’re looking at, right?”
With respect to statins and diabetes – the focus of your article – you’re exaggerating again (and making incorrect assumptions about the recording of adverse effects). Let’s look at the largest meta-analysis mentioned in the paper. 16 trials met the selection criteria: 13 out of 16 trials definitely recorded incidence of diabetes. Six of them had previously published data on incident diabetes, and seven had not, but they provided it when it was requested for this collaborative meta-analysis. Three studies did not provide the data (I don’t think the reasons were given in the paper, so we don’t know whether or not they recorded incidence of diabetes). Bear in mind that some of these 16 trials are quite old, and many were performed before the risk of getting diabetes from statins was known; they nevertheless still recorded incidence of diabetes.
Remember that there’s an important difference between the recording, and the reporting, of adverse effects. I’m sure you understand that published papers have word and space limits. I’m sure you also understand basic medical statistics: that if you analyse data for 100 different adverse effects, on average 5 of them will be statistically significantly different across treatment groups by chance alone. So, do you report results for all 100 in the paper (impractical)? do you cherry pick the 5 and risk wrongly suggesting that the (chance) results are due to treatment? Or do you focus on the known areas of concern from earlier studies? This is where the meta-analyses really help. (They mention this in their discussion: “Our meta-analysis incorporated almost all completed large trials, providing good statistical power.”)
“I say we turn a blind eye to them at our patients’ peril.”
I think you’re turning a blind eye to a key point I made: “put simply, the results from this study are similar to the those of the RCT meta-analyses, which you think were based on “corrupted” and “utterly unreliable” data.” You correctly point out that RCTs may not reflect the reality of clinical practice. So a huge, well conducted study was done on patients in real clinical practice to see whether or not this may be the case for statins and diabetes. The result was the same as the RCT data.
Something awful! If you want to reduce the diabetes risk associated with statins, no problem… Take another drug with your statin, glyburide (diabetes drug)! But when we look at the adverse effects of glyburide …
But in what world do the researchers live?