A rapid response was published in the BMJ here.
The Cholesterol Treatment Trialists’ (CTT) Collaboration has been quiet for a while, so a major publication in The Lancet (the favoured journal for CTT publications) was perhaps overdue. One duly appeared on 1st February 2019. It was entitled “Efficacy and safety of statin therapy in older people: a meta-analysis of individual participant data from 28 randomised controlled trials” (Ref 1).
It attracted much media attention on both sides of the pond. CNN reported “After years of uncertainty, study finds statins can benefit all ages, including those over 75” (Ref 2). The UK Times reported “Giving statins to all older people could save 8,000 lives every year” (Ref 3). The story took up the front page of the UK Express newspaper “Statins really do save lives” (Ref 4).
As I shared in this post (Ref 5), “the CTSU is the Oxford Clinical Trial Service Unit. Where this ends and the CTT (Cholesterol Treatment Trialists) Collaboration begins or where CTT ends and where Sir Professor Rory Collins and Colin Baigent and co. begin or end, I know not. One gets the impression that the web between the parties is not intended to be clear.”
It is important to read that post for context to this note. That post went through the attacks made by Collins on authors of two articles published in the BMJ and attacks on the BMJ itself. Before this happened (in the summer of 2014), I had already identified approximately £115 million of funding from drug companies to the CTSU. In the disclosures that were required as part of the full BMJ independent investigation, it emerged that the full extent of the funding from pharma to the CTSU was beyond £268 million. That was back in 2014; it can only have increased since then.
The CTSU is also the group that, despite requests from journalists and the BMJ, has refused to share information from their statin trials, so that it can be independently examined. The CTSU has also refused to share data about serious adverse events, so we have the outrageous situation that data exist that could inform prescription practices and the holders of these data refuse to share them.
The lives saved claim
The claim about lives saved was remarkably consistent across the media reports:
“But scientists said up to 8,000 lives could be saved annually in the UK alone if everyone over the age of 75 received statin therapy.” (The Express)
“Everyone over the age of 75 should be considered for cholesterol-lowering statins, experts have urged, after an analysis found up to 8,000 lives a year could be saved.” (The Times)
“Researchers said up to 8,000 deaths a year could be prevented if GPs simply prescribed drugs costing pennies a day.” (The Telegraph) (Ref 6).
“Up to 8,000 pensioners a year are needlessly dying of heart disease because they are not being given statins, leading experts warn.” (The Daily Mail) (Ref 7).
The 8,000 lives saved claim came from a press conference, which was held on January 30th to launch the paper. The press conference was reported in a BMJ article, which quoted Colin Baigent as saying “Only a third of the 5.5 million over 75s in the UK take statins and up to 8000 deaths per year could be prevented if all took them” (Ref 8).
The problem with the 8,000 lives saved/deaths prevented claim is that it cannot be supported from evidence in the paper.
The study published on 1st February was a meta-analysis of trials for which the CTSU holds data. The objective of the study was set out in the introduction: “We aimed to do a meta-analysis of data from all large statin trials to compare the effects of statin therapy at different ages and explore the effects of statin therapy among older individuals.” The specific age of interest in the abstract of the paper and throughout the paper and in media coverage was the over 75 age group.
The paper reported that, from the available data on 186,854 people, 14,483 people were over the age of 75. Of these 14,483 people, 55% had a history of heart disease of some kind (so 45% didn’t). The average total cholesterol for the over 75s was 5.1 mmol/l and the average LDL cholesterol for this group was 3.2 mmol/l. The abstract (summary) of the paper reported that “Overall, statin therapy or a more intensive statin regimen produced a 21% (RR 0·79, 95% CI 0·77–0·81) proportional reduction in major vascular events per 1·0 mmol/L reduction in LDL cholesterol” (Ref 9).
There are five Figures in the paper. The only one that looked at deaths was Figure 5. It reported that the rate ratio (RR) for statins (achieving a 1 mmol/l reduction in LDL) vs. the control group in the over 75s was 0.95 (95% CI 0.83–1.07). That includes the line of no effect (1.0) and thus could have happened by chance and thus was not statistically significant. Interestingly the result for the age group 70-75 was also not statistically significant. Neither of these can be reported as a finding therefore.
The CTT personnel even re-calculated the RRs after leaving out four trials where “statin therapy had not been shown to be effective.” Yes – really – ponder on that for a second. Even leaving out four trials that didn’t show support for statins, they still could not achieve a statistically significant result in the over 75 age group.
All the reports in the media that “up to 8,000 lives a year could be saved”, with reference to over 75s (or over 70s without the fudge) are false.
The evidence for cholesterol lowering in the elderly
I knew to doubt this claim as soon as I saw it because I worked on a paper published in 2016 with Dr Uffe Ravnskov as the lead author (Ref 10). The paper was called “Lack of an association or an inverse association between low-density-lipoprotein cholesterol and mortality in the elderly: a systematic review.“ I spent several weeks doing the data extraction with Uffe and cross checking each other’s tables. I withdrew my name from the paper before publication because the correct (PRISMA) systematic review/meta-analysis methodology was not followed. This did not impact the results and, having seen the data first hand, the conclusion is robust. The conclusion was: “High LDL-C is inversely associated with mortality in most people over 60years.” i.e. high LDL-cholesterol is associated with lower deaths in most people over 60.
I also knew to doubt the “lives saved” claim because statin patient leaflets caution against statin use in the over 70s. Here’s an extract from the Lipitor leaflet (the most prescribed statin) (Ref 11):
What about events?
Even if the death claims are lies, what about the claimed reduction in events? The newspapers reported that, “for every 10,000 people aged 78, who take statins, but have no history of cardiovascular problems, 80 heart attacks or strokes would be prevented every year.”
This came from the press release (Ref 12) and the Lancet paper itself. Both identically reported “In the primary prevention setting…” [i.e. in individuals with no known history of vascular disease], “Reducing those risks by a fifth with a 1.0 mmol/L LDL cholesterol reduction would prevent first major vascular events from occurring each year in 50 individuals aged 63 years and 80 individuals aged 78 years per 10,000 people treated.”
This is also false.
– Figure 1 reported effects on major vascular events by age. This is the Figure that gave the top level claim in the abstract of the paper as shared in “The Study” section above: “Overall, statin therapy or a more intensive statin regimen produced a 21% (RR 0·79, 95% CI 0·77–0·81) proportional reduction in major vascular events per 1·0 mmol/L reduction in LDL cholesterol.”
– Figure 2 reported major vascular events by age and by type of trial (heart failure trials vs. dialysis trials vs. other trials).
– Figure 3 reported major vascular events by age and by type of event (coronary event, stroke etc).
– Figure 4 reported major vascular events by age and by previous vascular disease. There was no statistically significant difference between the statin group and the control group in people over 75 (or in the 70-75 age group). Benefit in those who have “no history of cardiovascular problems” cannot be claimed, therefore.
Serious Adverse Effects
Notwithstanding that no benefit can be claimed for those without previous vascular disease – and hence there will not be 80 fewer events per 10,000 78-year olds given statins – there will be serious adverse events in those people if they take statins.
The Number Needed to Treat (NNT) is the important measure to examine. The web site for this is www.thennt.com Frustratingly, these numbers seem to change every time I look at the page (I suspect that there is intense pressure from drug companies for these figures to be ‘re-visited’). Currently, the NNT numbers advise that 1 in 10 people who take statins without known heart disease will be harmed by muscle damage and 1 in 50 will be harmed by developing (type 2) diabetes (Ref 13). So, for every 10,000 people given statins, 1,000 are likely to develop muscle damage and 200 are likely to develop (type 2) diabetes. For no benefit in heart disease.
The Daily Mail article reported Colin Baigent as having said that “a number of misleading studies – which he branded ‘fake news’ – had created confusion over the effectiveness and side-effects of statins among doctors and patients.”
The only fake news that I can find is coming from Baigent himself.
– A study was published in The Lancet, which was reported worldwide as “giving statins to people over the age of 75 could save thousands of lives.”
– This claim came from a press conference to launch the Lancet Paper, where a lead author, Colin Baigent, was quoted as saying: “Only a third of the 5.5 million over 75s in the UK take statins and up to 8000 deaths per year could be prevented if all took them.”
– This is false. It relies upon evidence in the over 75s for both deaths and primary prevention (people who do not already have heart disease) and neither was found.
– Figure 5 in The Lancet paper confirmed that the Rate Ratio (RR) for vascular deaths for over 75s was not statistically significant. Nor was it for those aged 70-75 for that matter. Even with the attempt to achieve a significant result, by excluding trials that failed to show benefit of statins, the RR for vascular deaths for over 75s was not statistically significant.
– Figure 4 in The Lancet paper confirmed that the RR for major vascular events for over 75s without vascular disease was not statistically significant. Nor was it for those aged 70-75 for that matter.
– This article was reported worldwide as ‘statins can save thousands of lives in the over 75s’. No statistical significance was achieved for deaths or primary prevention in this age group. The falsehoods need to be corrected.
Ref 1: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(18)31942-1/fulltext
Ref 2: https://edition.cnn.com/2019/01/31/health/statins-elderly-cholesterol-study/index.html
Ref 3: https://www.thetimes.co.uk/article/giving-statins-to-all-older-people-could-save-8-000-lives-every-year-6vtt2kzll
Ref 4: https://www.express.co.uk/life-style/health/1081016/statins-save-lives-study-cholesterol-elderly
Ref 5: http://www.zoeharcombe.com/2014/08/ctsu-funding-from-drug-companies/
Ref 6: https://www.telegraph.co.uk/news/2019/01/31/75s-should-offered-statins-ageism-failing-patients/
Ref 7: https://www.dailymail.co.uk/news/article-6655239/Thousands-pensioners-dying-ageism-denies-statins-aged-75.html
Ref 8: https://www.bmj.com/content/364/bmj.l522.full
Ref 9: An email group of academics/medics, of which I’m a member, can’t work out what this means. Did they only include people who achieved a 1.0 mmol/l reduction in LDL (which is large, from a starting point of 3.2 mmol/l) in the comparisons with the controls? Did they ‘adjust’ actual benefit to what they estimated it would had there been a reduction of 1.0 mmol/l? Who knows?
Ref 10: https://bmjopen.bmj.com/content/6/6/e010401
Ref 11: https://www.medicines.org.uk/emc/product/1059/pil
Ref 12: https://www.eurekalert.org/pub_releases/2019-01/tl-tls013019.php
Ref 13: http://www.thennt.com/nnt/statins-for-heart-disease-prevention-without-prior-heart-disease-2/