Statin Wars: Why Has The Latest JAMA Study Been Ignored?

Statistical games, the effect of the pharma-media complex on reporting and why asking questions is now 'statin denial'

After three years during which the pharmaceutical malpractice has been repeatedly exposed, statins are set to become another battle-ground. The global market is now worth $20bn a year. That is to say, enough to determine the survival of many powerful multinationals. The American College of Cardiology and the American Heart Association provide guidelines so aggressively in favour of statins that almost every single elderly or diabetic individual would be subject to treatment. The fact that the majority of both of these panels have financial links to statin manufacturers is presumably something we should ignore.

So it is within this climate that the results of a meta-study on statin effectiveness was published in the Journal of the American Medication Association was published several weeks ago. Perhaps unsurprisingly, its findings have been universally ignored in the mainstream press, despite them fawning over two meta-studies of 2022 which contained similar data (more on these studies shortly).

Two things separated this paper from those that were heralded by the media just months ago: the format used to report the results and the inclusion of additional analysis on the presumed mechanism of statins.

What exactly did they do and what did they find? The researchers assessed the outcomes of 21 different studies on statins, encompassing a total of 143,532 participants and tracked outcomes over an average of 4.4 years of medication use. The international team looked into the effects of the medication on ‘hard’ outcomes (death, heart attacks and strokes) and did so using only verifiable, published data.

Their findings on efficacy largely mirrored previous trials and showed that the risk of heart attacks was reduced by 1.3% (from 4.5% to 3.2%) and risk was reduced by 2.2% in sub-groups who had already had already had a heart attack.

However, the paper is noteworthy because it published the figures as actual risk, rather than the more common practice of sharing only the relative risk for the efficacy and then switching back to actual figures when reporting the rate of side-effects. This practice is done for a very simple reason; it makes the drugs look better. Imagine a drug that reduces your risk of death of a disease from 1% to 0.5%, but causes serious side effects in 5% of cases. That’s 10 serious side-effects for every success, a risk/benefit profile that would cause understandable alarm to the public. Enter the statistical sleight of hand, aka relative risk when reporting the benefit. Suddenly the same agent becomes a ‘miracle’ drug that ‘cuts risk of death by a whopping 50%, with side effects only seen in 5%’. Misleading, but technically true.

If these researchers followed the well-trodden course, they would instead have reported that statins reduced cardiovascular events by an impressive-sounding 29%. However, they provided the figures in the same format normally saved only for the side-effects. Not only that, they then went to share an important finding that runs contrary to the accepted narrative on statins and heart disease: there was no consistent relationship between lowering LDL-C and death, heart attack or stroke, following statin therapy (that is to say, there was a relationship in some studies but not others).

This accepted narrative will likely be familiar to most readers. It states that heart disease is driven by atherosclerotic plaque, plaque formation is a consequence of higher cholesterol levels and that lowering these cholesterol levels will thus prevent heart disease.

Pushback

This simplistic narrative has failed to win over many critics. Opposition often centres on how this inductive reasoning is subject to flawed assumptions, pointing that multiple studies that ask questions of it are ignored. Critics point to studies that show older adults with low cholesterol levels experience atherosclerosis at the same rate and how, in a 30-year follow-up of the famous Framingham Heart Study, it was shown that ‘for each 1 mg/dl drop in TC per year, there was an eleven percent increase in coronary and total mortality’. While we will leave the nuances of this discussion for another day, its fair to say that the body of literature provides good evidence that high serum cholesterol clearly can drive atherosclerosis formation – see an assessment of lifetime exposure to LDL cholesterol and so-called ‘Mendelian randomisation’ studies (that look at those with genetic tendencies towards lowered LDL levels) - but that this process is subject to regulation from multiple metabolic factors.

Which brings us to the crux of the more pragmatic disagreement on statins, one that concerns itself less with the mechanisms in play on more so on the outcomes. That is to say: do the benefits of statins outweigh the side-effects?

The ability to simply take statins is a problem for many. A 2019 study used a medical database to track statin compliance over five years and found that only 47% were able to adhere to their prescription for an entire year, and this dropped to a mere 19% after five years.

The biggest concern for these drugs is their effect on Co-Q10 levels, a key nutrient to maintain efficient energy production at the mitochondria. Because both cholesterol and Co-Q10 are manufactured from the same metabolic pathway and statins inhibit this pathway, an inevitable downstream effect is lowered Co-Q10 levels. Studies show this drop to be around 40%.  

The main sign of this mitochondrial burden is muscular weakness/pain. The most commonly-cited study found that 40% of the women developed fatigue from taking statins for six months. Sinzinger et al found that a quarter of statin users that are physically active develop muscular weakness/pain. It may be that the more active that the individual is, the more likely they are to suffer from such issues; a study on 22 athletes taking the drugs for genetic cholesterol issues found that 17 had to stop taking them.

Concerns extend beyond simply wellbeing and physical performance. Statins have never been able to shake off the cancer connection. The CARE trial found a small reduction (1.83%) in cardiovascular events in their study group, 4159 cardiovascular patients tracked over five years, but saw a 3.9% increase in breast cancer risk in the women (which was shown to be highly statistically significant, p = 0.002). Their discussion of these worrying findings? That they should be ignored. To quote, “these findings could be an anomaly”.

A similar story played out in the PROSPER trial, which showed that there were 28 fewer fatal heart attacks from in those taking statins (3.3% instead of 4.2%, a 0.9% reduction). However, the statin group had 24 more cancer deaths. Only one of these findings made it into the conclusion (and was, believe it or not, reported as a massive 24% reduction in risk of heart attacks). Their comments on the spike in cancer? “most likely… was a chance finding”.

The SEAS trial tracked 1.873 cardiovascular patients over 4.3 years and found no significant benefit in the main cardiovascular endpoints, but saw rates of cancer rise from 7.5% to 11.1% (p < 0.01). It may come as no surprise to hear that the authored declared that “the difference in cancer rates in the study may have been the result of chance” and made no mention of this whatsoever in their abstract. There has only ever been one long-term study of cancer risk from statin use; it found that the risk of breast cancers doubled. Could that also be chance?

While the cancer concerns are hypothesized to relate to lowered fat-soluble vitamin status that inevitably accompanies lowered cholesterol levels, there are also concerns for brain health (as cholesterol used by the brain to manufacture both cell membranes and neurosteroids like pregnenolone, DHEA and cortisol. Multiple studies have correlated low cholesterol with increased violence (Asellus et al), increased suicides and other violent deaths (Muldoon et al). While association-based studies should be treated with extreme caution (as they cannot ever separate causation from correlation), the Muldoon group went on to undertake primate studies and saw that the administration of statins actually caused an increase in violent behaviour (the lower the cholesterol levels, the more violent the behaviour). Further impacts on brain health were revealed in a study of 143 patients with cognitive decline who were taking statins; it found that 90% of them improved, often within days, upon removing the medication. A further study replicated these findings. Other studies that examined the link between the drug and peripheral neuropathy found that statin users showed a 16-fold increase in risk. Prescribing statins in elderly populations is associated with an 9% increase in the relative risk of developing diabetes (0.4% in actual risk). Other studies have found associations between statin use and cataracts, kidney disease and liver dysfunction.

Is this simply a price we have to pay for saving lives?

If a drug has a biological effect, we can expect side-effects. Therefore, drugs that cause side-effects are not unusual and should not be thrown out on this basis. But it calls us to ask an important question: do the side-effect justify the benefits?

Statin proponents point reflexively to the tremendous benefits shown in the keystone trials. The first ever trial to actually demonstrate any benefit from statins was the BHF Heart Protection Study in 2002. The project, which coincidentally was jointly funded by Merck and Roche , studied 20,000 individuals over five years, with half taking a statin and half taking a placebo. They recorded 781 deaths (7.6%) in the statin group, compared to 937 (9.1%) in the placebo group; a drop of 1.5% or 13% (depending if you are using actual or relative risk). Put another way, this study produced an NNT (Number Needed to Treat) of 67. To clarify, this means that 67 people needed to take the drug for five years to prevent one death. Worth clarifying, because these trials never tend to report NNT. Sceptics would argue that this is because providing this number would inevitably lead to easy comparison against NNH (Numbers Needs to Harm). We can only guess what this actually was, because researchers chose not to comment further on a key facet of their experiment; that 26% of those taking statins pulled out within the first month of doing so. That’s around 2,600 people.

This brief analysis clearly demonstrates the split in perspectives. The pharmaceutical companies can state with accuracy that the drugs are proven to reduce deaths, while sceptics point out that statins in general rate among the least effective drugs ever put into production, with the results only appearing impressive through statistical manipulation and the rate of side-effects hidden through deliberate sleight of hand.

Such positions became even more entrenched following the ‘landmark’ JUPITER trial. This was the monumental study that was ‘so successful that it became unethical to continue it’. Researchers reported with glee that, after pairing two equally-sized groups (of 8901 patients apiece), those receiving treatment recorded a spectacular reduction in ‘cardiovascular endpoints’ (87 versus 156 in the placebo group) and an “unequivocal reduction in cardiovascular mortality”. They did not point out that the endpoints were made up of a hodgepodge of highly-specific ‘soft’ outcomes that depend on a subjective judgement call from doctors, such as revascularization and hospitalization for unstable angina.

And the ‘unequivocal’ reduction in cardiovascular mortality? A bald-faced lie. There were 12 cardiovascular deaths in the treatment group and 12 in the placebo. The claim lies on the fact that those taking statins recorded only 3 deaths from stroke (compared to 6 in the placebo group); such a drop (33% relative reduction) may have carried weight if it were not for the fact that there were 9 heart attacks in the statin group and 6 in the placebo group. Hence, 12 cardiovascular deaths apiece. In a woeful farce, the authors decided that a death from heart attack, the archetypal cardiovascular death, was not a cardiovascular death. They also omitted cardiovascular deaths from their publication (leaving readers to calculate this only from subtracting ‘non-fatal strokes’ from ‘total strokes’, and doing similar for heart attacks).

But why did they stop the trial early? Many pointed to the fact that the treatment group took an ‘early lead’ in the death counts, that the outcomes-over-time graph showed that the placebo group had since shown a strong trend towards better figures the curves were converging at the time the trial. This led to speculation that the researchers had shut down the trial to avoid the impending possibility that the results would show negative efficacy. Such speculation was further fuelled when the researchers put out a follow-up piece to defend their incoherent calls, in doing so removing the graph that had initiated such discussion.

Beyond the spectacularly unspectacular benefits shown in the trial, just as worrying was the rate of side-effects. In the rosuvastatin group, there were 270 new cases of diabetes compared to 216 in the control group (3.0% vs 2.4%, p = 0.01). True to form, the researchers did not report this figure using RRR/Relative Risk Ratio (25% increase in diabetes risk) as they did for the benefits, instead slipping this into the small text in ARR/Actual Risk Ratio, which amounted to 0.6%.

Similar datacrimes can be found in the much-feted ASCOT trial. This was launched with the intention to track 10,000 participants over 5 years but they stopped early, at 3.3 years (sound familiar?). They did so after they determined that, if they constructed a custom endpoint (there it is again) of cardiovascular deaths plus non-fatal heart attacks, this was seen in only 1.9% of the atorvastatin group, compared to 3% of the control group. So a 1.1% drop in this one measure, with no significant benefits found in cardiovascular mortality or overall mortality. That’s right, there was no difference in the amount of people died of heart disease in both arms of the trial; the sub-type of cardiovascular death varied but the numbers did not. There were no benefits shown in those with type II diabetes or metabolic syndrome (the population most likely to receive the drugs), and neither was there any benefit in the under-60s or females. Weirdly, none of this ever seemed to be mentioned in the press. They simply parroted the conclusions of the paper itself, that ‘the trial needed to be stopped due to a massive 36% drop in risk’. Because a drop of 3.0% to 1.9% is a relative drop of 36%. No attention was paid to the cold hard fact that, when it came to cardiovascular deaths, this life-saving intervention wasn’t shown to save lives.

Is there any benefit shown in the literature on statins? Yes, most definitely. The argument is not whether statins are effective in reducing cholesterol – they do this very successfully – or even if there is a benefit in reducing cardiovascular risk – they do – but on whether this reduction is substantial enough to warrant an intervention that so heavily impairs lipid metabolism in the body. A 2018 review notes that there are 33 studies undertaken on the effectiveness of statins, with 29 reporting cardiovascular mortality, but only 11 (38%) showed an actual risk reduction in cardiovascular mortality of more than 1% when used for several years.

Interestingly, all 29 of these trials showed substantially benefits in total mortality than cardiovascular mortality, indicating either that statins are having a powerful biological effect that is independent of cardiovascular health, or b) the groups were never properly balanced or c) a bit of both. Clearly a discussion for another day.

What is clear from the literature is that there is a disconnect between the huge body of evidence and what patients are told at the point of prescribing. Would we see a change in statin uptake if the findings from the literature was fully conveyed to patients, eg. “if you take this drug, your likelihood of avoiding a fatal heart attack in the next year would increase from 98% to 99%, but you have a 40% chance of experiencing major side-effects”? What if patients were made aware of the benefits of regular cardiovascular exercise, itself associated with 9x the benefit on reducing total mortality (reducing actual risk by 9%), without any such side-effects? Or similar discussions were undertaken on the expected benefit of supporting metabolic health, be that through removing mitochondrial blockages, improving sleep or regulating inflammation, instead of taking a pill?

Can statins survive?

As explained above, the widespread use of statins stands out as just another medical oddity that we future generations are destined to look back on with disbelief. Yet statins remain one of the most profitable markets in the industry. And, perhaps predictably given the size of the commercial interests, the pharma-media complex seems fully committed to parroting two messages:

1.       That the reductions in cardiovascular risk, when reported in relative terms, makes statins a miracle drug

2.       That the side-effects, naturally referred to in actual terms, are over-stated and should not stand in the way of widespread use of statins

Such positions are evident in a 2021 paper that tracked just 60 individuals. The paper, unsurprisingly, came away with positive conclusions for statins; they tracked the participants for six months, providing them with a statin, a placebo or no pills for a month at a time throughout the study period. They found that there was a substantial increase in symptoms when participants were taking statins, but that this was only slightly higher than those reported when they took the placebo. The researchers concluded that, should any patient report problems while taking a statin, that ‘clinicians should not interpret these as being caused by statins’.

They made this conclusion without any reference to some important facts buried in the Results section; here, they admitted that eleven of the participants had to quit the study. Eight of these coming while taking statins, with only one occurring from placebo. Five had to stop due to ‘severe’ side-effects (which are normally defined as life-threatening complications that require medical attention, although the researchers did not provide any further details). In any sane world, the headline of such a trial, one where 1 in 12 quickly stopped due to severe side-effects, writes itself. Not here. They instead chose to ignore this data - some of the most concerning results we could expect to encounter in a modern trial - and declared that any side effects are simply all in the patient’s head. Brazen.

As one may imagine, such deliberate blindness to what the data was actually saying did not hold much weight with sceptics, who pointed out the obvious: that you cannot produce an accurate estimate of side-effects if your process for doing so deliberately does not count those who were not able to tolerate statins due to side-effects. In other words, by calculating side-effect rates only from those that finished each study, that you are calculating such rates in a sub-group that excludes those with the most serious challenges.

The response? February 2022 saw the release of the much-hyped article in the European Heart Journal. The paper billed itself as an authoritative analysis to settle the ongoing debate on statin and the side-effects that plague so many people who take them. To settle the debate – one that rested on the masses of side-effects that were missed through systematically ignoring the outcomes in trial participants who were unable to complete the study – they chose to exclude all trials that “only investigating statin discontinuation without specifying intolerance”. In other words, using a methodology that excludes almost all patient who experienced the most severe side-effects. Believe it or not, such methods lead to the conclusion that statin intolerance was only 9.1%. Cue rapturous headlines proclaiming that the side-effects of statins were officially overblown, ie. that such effects were all in the heads of those taking them.

We were treated to more of the same a few months later, when a meta-analysis in the Lancet also cleared statins of suspicion. How did they reach a conclusion on the issue that stems on systemic under-reporting of side-effects? Easy. They did the same as the February article, and note in their methodology that their assessment was based on “participant data on all recorded adverse events”. In other words, as long as you ignore all the individuals that needed to drop out due to side-effects and thus went unrecorded, the rate of side-effects is only 3%! The press ate it up.

As the PR cycles continue to spin, most clinical trials now adopt a ‘wash-in’ period, which sees participants take the drug for a month before the trial begins. I am not aware of anyone attempting to justify such a practice that, from an outside perspective, appears a laughable way of reducing drop-outs (you can’t drop out of a trial if you never made it to the start!)

Meanwhile, while sceptics patiently wait for a new drug class to come online and for the same consultants to suddenly ‘discover’ the downsides to statins, the mainstream position remains exemplified by editorials that are dually published in both medical journals and pharma-friendly newspapers. Such contributions make it clear that requests to assess the data on the risk/benefit profile of statins is ‘statin denial’ and that ‘statins have developed a bad reputation with the public, a phenomenon driven largely by proliferation on the Internet of bizarre and unscientific but seemingly persuasive criticism of these drugs.” Bizarre indeed.