Statins: numbers needed to treat and personal decision making

I see now why Richard Watson took personal offence at my phrase "his stealth provides a Trojan Horse for drug company profiteering." I apologise for any misunderstanding. Having not understood why you accused the Professor of shouting, whilst you were calm and reasonable, I described it as stealth. You introduced JBS3 into the debate, which I feel IS a vehicle by which the JBS extols the virtues of statins, whilst telling us nothing of their harms, trouble, or costs - hence "Trojan Horse". This is how Pharmaceutical companies promote their products, and yes the JBS consultants do acknowledge their financial interests.

NICE also declares a finacial interest - the cost-effectivene use of NHS resources.

I had not understood, for my part, that capitalisation can be construed by some as as 'shouting'. Unfortunately capitalisation is the only way in a textual rapid response to lay emphasis.

I again apologise to Richard for any unintended implication that he was personally deceitful, and hope this clears the air between us.

With best wishes,

In the current heated debate over the benefits and harms of statins, many voices have requested the calculation and publication of the NNTs (number-needed-to-treat) and NNHs (number-needed-to-harm)1 in order to facilitate decision making with respect to individual patient treatment with statins in primary prevention.

However, the validity of extrapolating data from meta-analyses that incorporated trials of relatively short duration (median follow-ups from 1.8 year to 4.9 years), with the aim to calculate a 5-year (or 10-year) NNTs and NNHs may be questionable.

Truncated trials overestimate the reported benefits of any treatment. This overestimation is independent of prespecified rules and greater in truncated trials having fewer than 500 events.3 Therefore a 5-year (or a10-year) NNT based on extrapolation from data obtained by meta-analyses that included truncated trials with a relatively short duration, like ASCOT-LLA,4 or CARDS,5 JUPITER, 6 whose median follow-ups were 3.3 years, 3.6 years, and 1.8 years respectively, may overestimate the benefits and cannot be considered a reliable indicator for decision-making when considering starting a "lifelong" treatment for a specific patient.

Moreover the definitions of the so-called major adverse cardiovascular events (MACE), to which the calculated NNTs apply, should also be taken into consideration. These MACEs are most often composite end-points (made of “hard” end-points like deaths or cardiovascular deaths, mixed with “soft” end-points like diagnosis of unstable angina, non-fatal stroke, decisions for hospitalisations or for catheterisation and/or for stenting). The respective weight and the subjectivity to biases may differ for each individual MACE, especially for the “soft” end-points, which are quite prone to biases as they are not end-points per-se, but medical decisions, and therefore quite easily influenced by judgement or perception, particularly in non-blinded or poorly blinded trials.6

Much has been written about the proven and associated harms of statins, the evidence for which is accumulating. Meta-analyses do suggest that statins indeed increase the risk for new-onset diabetes.8 However, the 5-year (or 10-year) NNHs, if extrapolated from the data of these truncated trials, may still be underestimating the real harm for 3 main reasons. Firstly, randomised patients usually have been screened carefully and pre-exposed to the active drug through a run-in period thus excluding those intolerant to the study drug.9 Secondly, adverse events may be minimized10 or considered subjectively by the investigators as non-related to the study drug such as cases of rhabdomyolysis.11 Thirdly, the most serious adverse events may take variable time-periods to develop, such as occurrence of acute kidney injuries, as demonstrated by an elegant cohort study12 or a more recent observational study.13

Therefore, we caution that the 5-year (or 10-year) NNTs and NNHs calculated from meta-analysis that incorporated truncated trials may be inaccurate and misleading. There is a clear risk of underestimating the NNTs and overestimating the NNHs.

REFERENCES

1. Tresidder A. NICE should publish numbers needed to treat and harm for statins. BMJ 2014; 348: g3458.

2. Cholesterol Treatment Trialists (CTT) Collaborators. The effect of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: Meta-analysis of individual data from 27 randomised trials. Lancet 2012: 380: 581-90.

3. Bassler D, Briel M, Mailovi VM, et al. STOP-IT-2 Study group. Stopping randomized trials early for benefits and estimation of treatments effects. Systemic reviews and metaregression analysis. JAMA 2010; 303 : 1180-7.

4. Sever PS, Dahlöf B, Poulter NR et al. Prevention of coronary and stroke events with atorvastatin in hypertensive patients who have average or lower than average cholesterol concentrations in the Anglo-Scandinavian cardiac Outcomes Trial-Lipid Lowering Arm (ASCOT-LLA): a multicentre randomised trial. Lancet 2003; 361(9374): 1149-58

5. Colhoun HM, Betteridge DJ, Durrington PN, et al. Collaborative atorvastatin diabetes study (CARDS). Lancet 2004; 364(9435): 685-96.

6. Ridker PM, Danielson E, Fonseca FAH, et al. Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein. N Engl J Med 2008; 359: 2195-207.

7. Nguyen PV. Electronic health records may threaten blinding in trials of statins. BMJ 2014; 349:g5239.

8. Sattar N, Preiss D, Murray HM, et al. Statins and risk of incident diabetes: a collaborative meta-analysis of randomised statin trials. Lancet. 2010; 375:735-742.

9. Anonymous. MRC/BHF Heart Protection Study of cholesterol-lowering therapy and antioxidant vitamin supplementation in a wide rage of patients at increase risk of coronary heart disease death: early safety and efficacy experience. Eur Heart J 1999; 20: 725-741.

10. Fernandez G, Spatz ES, Jablecki C, Phillips PS. Statin myopathy : A common dilemma not reflected in clinical trials. Clev Clin J Med 2011; 78: 393-403.

11. Ravnskov U, Rosch PJ, Sutler MC. Letter. N Eng J Med 2005, 353 (1): 94.

12. Hippisley-Cox H, Coupland C. Unintended effects of statins in men and women in England and Wales : population based cohort study using the QResearch database. BMJ 2010; 340: c2197.

13. Dormuth CR, Hemmelgam BR, Paterson JM, et al. Use of high potency statins and rates of admission for acute kidney injury : multicentre, observational study of administrative databases. BMJ 2013:346: f880.

Authors:

Paul v Nguyen MD, Internal Medicine Specialist, Centre Hospitalier Universitaire de Montréal, Montréal, Canada
Pierre Biron, MD, Pharmacology Professor (Retired), Université de Montréal, Montréal, Canada

Dear Dr Richard Watson,

False charges and ad hominem attacks are in my view worse failings than the ‘shouting’ which you accuse this discussion of thus far.

“Dr Lewis says that I am working covertly for the drug companies who make statins and antihypertensives and am failing to declare it in the compulsory declaration of interests.”

I said no such thing, and therefore have no need to retract. But I do believe that to advocate treatment of a healthy 55-year-old man with a normal BP and cholesterol is a poor use of limited funds. You are such a normal male, with a low CVD event risk, a normal cholesterol, a rather favourably high HDL, and a marginally high BP, for which NICE advises lifestyle change and ambulatory/home BP monitoring.. JBS3 shows that such a person has an unsurprisingly ’55-year-old heart’. Add the most relevant factor – FH of a CV event - and your risk goes up by 1.5 x normal if it were an MI in first-degree relative.. There are many such persons in my practice , some more keen than Dr Raj Bhopal to start preventive treatment. Personal decision-making is subject to cost-effectiveness NICE guidance in the NHS, as well as carrying harms and marginal benefits for the hundreds of individuals requiring to be treated for just one to gain.

“By implication all the eminent authors and bodies behind JBS3 (1) are also acting solely out of desire for personal profit. I simply do not believe that so many doctors are so unethical.” Nor do I believe that such eminent authors would act ‘solely out of desire for personal profit’, if at all. The JBS3 Declaration of Interests is to be found at http://www.jbs3risk.com/pages/DOI%20for%20JBS3.pdf .

I do believe that ‘lifetime risk’ engenders a serious deviation from cost-effectiveness, and much prefer the NICE expert guidance focussed on ten-year risk to the misguided JBS3 ‘lifetime risk’ calculator, for the reasons which I stated in my first posting. I do believe that scarce NHS funds must be rationed cost-effectively, and am happy to be labelled by you as a ‘leftist’ in that regard. May I ask that you read the responses to your reference (2) - http://www.bmj.com/content/348/bmj.g3047/rapid-responses to see the rational critique expounded in detail. The trouble with JBS3 is that it over-emphasises benefits for everyone, and pays no attention whatever to harms or costs per event avoided. In your case JBS3 claims you have put off your heart attack by 4 years after just 6 weeks of Amlodipine, Ramipril, and Atorvastatin ! Look carefully and JBS3 also tells you you are likely to live into you 80s and not suffer a heart attack.

Returning to the individual personal decision regarding “a healthy 55-year-old man with a normal BP and cholesterol”, the fact remains that your Ramipril will need to be swallowed for 800 years ( or by 80 men for 10 years ) to be confident of avoiding one serious CV event, whilst simvastatin would need to be swallowed for 400 years.to avoid one serious CV event. Even at a conservative estimated cost of £2.50 per month this implies somewhere between £15000 and £24000 per event. If you pays your money you takes your choice, and take the consequences !

Dr L S Lewis

Richard Watson’s response sounds very measured and calm, but I fear his stealth provides a Trojan Horse for drug company profiteering.
As a fellow GP I would agree he has every right to delay his first cardiovascular event by four years, by starting ramipril, amlodipine and atorvastatin, if it were his free choice and at his expense. But Richard calculated his own ASSIGN risk to be very low at 7.5% in 10 years. The chances are that he will not have a heart attack in his lifetime, and delaying the one after his 100th birthday and long dead is not a good use of NHS funds (ref 1). NICE calculates quite correctly that below 15% ten-year risk is not a cost-effective use of Statin. Cash could be better spent on cancer drugs, for example.
Richard has been beguiled by JBS3’s ‘lifetime risk’ calculator, which http://www.jbs3risk.com/ says ‘represents an opportunity for investment in future cardiovascular health’. It certainly represents a major opportunity for drug manufacturers to profit from NHS expenditure. The deceit of ‘Lifetime risk’ has been ably criticised elsewhere (ref 2), and should not traduce the NICE cost-effectiveness tools. Drug treatment of people with 10-year CVD risk below 15% yields very little benefit for individuals, with a large number of people exposed to side-effects, at a high net cost.
Richard is correct to say that ‘the arguments against the use of statins for primary prevention surely also apply to treatment of high blood pressure’. NICE has always argued that hypertension treatment should also be restricted to those at significant CVD risk. Mild hypertension in low-risk men (MRC trial last century, for example) required 800 man-years of treatment to prevent one stroke.
References
1. Pills in the sky
L Sam Lewis
BMJ 2008;336:174 (Published 24 Jan 2008)

2. Is estimating lifetime cardiovascular risk useful?
Rod Jackson, Andrew Kerr, Sue Wells
BMJ 2010;341:c7379 (Published 31 Dec 2010)