In light of the recent decision by the UK health authorities to approve inclisiran for use in patients, I thought it would be interesting to delve in to how good the class of drugs it belongs to, the PCSK9 inhibitors, really are. PCSK9 inhibitors can be thought of as “next generations statins”. Like statins, they work under the assumption that the lipid transport molecule LDL (often incorrectly referred to as “bad cholesterol”) is the main driver of cardiovascular disease, and their goal is thus to lower LDL levels in the bloodstream.
Unlike statins, however, they work in a much more direct fashion, which should resut in significantly less side effects. Statins inhibit the ability of cells to produce cholesterol (a molecule that is critical for the correct functioning of cell membranes and for the production of many hormones), and through a sequence of downstream effects, this results in the liver hoovering up more LDL from the blood stream. PCSK9 inhibitors instead work in a much less roundabout way to increase the number of receptors on the surface of liver cells that pluck LDL from the blood stream.
In theory, as mentioned, this should result in significantly less side effects, since the side effects seen with statins, such as muscle pain and weakness, likely for the most part have to do with impaired cholesterol synthesis, and PCSK9 inhibitors don’t impair cholesterol synthesis. This more targeted and side-effect free mechanism should then allow for higher dosing, which should cause LDL to drop lower than is possible with statins. This should then result in less heart attacks and strokes.
PCSK9 inhibitors are new drugs and thereby still on patent and hugely expensive. Additionally, the PCSK9 inhibitors that have so far been developed are all broken down in the intestine if ingested, which means they can’t be given as a pill – they need to be injected. That means they’re not going to replace statins as the main LDL lowering drug any time soon. But if they can be shown to actually lower cardiovascular events and prolong life, and especially if new PCSK9 inhibitors are developed that can be taken as a pill, then it’s quite likely that they will eventually come to replace statins as the main LDL lowering drug.
Ok, let’s get to the evidence. An observational study was published in the New England Journal of Medicine in 2006 that provided support for the idea that PCSK9 inhibiting drugs might be effective at preventing heart disease. 12,973 people in the US were followed for 15 years to see if they developed cardiovascular disease. The participants were tested to see if they had a normal PCSK9 gene (which results in normal recycling of the LDL receptor on liver cells, and thus normal levels of the receptor) or if they had a defective gene (resulting in defective recycling and thus higher levels of the LDL receptor on liver cells). 97% had the normal gene, while 3% had an abnormal version of the gene.
As would be expected, those with the defective genes had lower levels of LDL in the blood stream, 21% lower to be precise. More importantly, they also experienced less heart disease. A lot less. While 11% of those with the normal gene developed heart disease over the fifteen year follow-up period, only 5% of those with an abnormal gene developed heart disease.
When it comes to stroke there was no difference, however, with 4% experiencing a stroke in each group. This is surprising at first glance, but could perhaps be explained by the fact that strokes often aren’t caused by LDL induced damage to the vascular wall, but rather by other things, such as atrial fibrillation, artery dissections, and intracranial bleeds.
In terms of overall mortality, there was a small benefit seen, with 12.5% dying in the group with the normal PCSK9 gene over 15 years, as compared with 9.6% in the group with the defective gene. The difference in mortality didn’t reach statistical significance however (which of course doesn’t mean it wasn’t real, just that there weren’t enough deaths overall to be able to tell if the difference was real or not).
This was an observational study, so it’s impossible to draw firm conclusions about cause and effect. The study was therefore suggestive of a health benefit associated with inhibiting PCSK9, but it didn’t prove anything. Nevertheless, the massive reduction in heart disease seen in the study led to a frenzy among drug companies to develop PCSK9 inhibiting drugs. There are now three such drugs authorized for use in patients: evolocumab, alirocumab, and inclisiran.
So, how well do they work?
A systematic review and meta-analysis was published by the Cochrane Collaboration in October 2020 that sought to answer this question. It looked specifically at evolocumab and alirocumab (inclisiran wasn’t yet approved for use in patients). 24 randomized controlled trials, with a total of 60,997 participants, were included in the review. It’s worth keeping in mind that all but one of the trials was industry sponsored, which means that we can be pretty sure that whatever the results turn out to be, they are a best case scenario. Industry sponsored trials usually overestimate benefit (and underestimate harm).
The trials varied in length from six months to three years, which is a bit short if you want to detect a meaningful difference in heart attacks, and especially if you want to detect a meaningful difference in mortality. The observational study described above ran for fifteen years, and although there was a clear difference in heart disease, the difference in overall mortality wasn’t statistically significant. That’s another thing that’s worth keeping in mind as we get to the results – if the studies fail to find a mortality benefit, the reason might simply be that the follow-up period was too short, not that they don’t work. I do think, however, that a good general principle is to assume that drugs don’t work until it’s been proven otherwise.
Let’s get to the results.
Nine trials, with 23,352 participants, looked at the ability of alirocumab to prevent heart attacks. 9.8% of participants in the alirocumab group had a heart attack, compared with 12.5% in the placebo group. That is a 22% relative risk reduction, and it is statistically significant. Considering the short follow-up period, that’s actually not too bad. It would mean that you would need to treat 37 people for somewhere between six months and three years to prevent one heart attack.
Three studies, with 29,432 participants, looked at the ability of evolocumab to prevent heart attacks. 3.2% of the participants in the evolocumab group had a heart attack, compared with 4.5% in the placebo group. That’s a 29% reduction in relative risk. Again, the difference was statistically significant.
Let’s look at at overall mortality, which is after all what matters most. When it comes to alirocumab we see that 2.6% died in the alirocumab group compared with 3.6% in the placebo group, and when it comes to evolocumab we see that 3.0% died in the evolocumab group compared with 3.0% in the placebo group.
Hmm… So, basically alirocumab appears to result in a massive 28% reduction in risk of death, while evolocumab doesn’t appear to impact risk of dying at all. That is odd. Considering how large a number of participants the trials have gathered, and that the drugs basically function in exactly the same way, and that the effect on heart attacks is similar, you would expect the effect on overall mortality to be similar. Yet one results in a massive drop in overall mortality and the other doesn’t affect mortality at all. Which result is to be believed?
To be honest, I think the difference might simply be due to differences in luck and above all patient selection. 12.5% of patients in the placebo group in the alirocumab trials had a heart attack, compared with only 4.5% in the evolocumab trials – clearly the participants in the alirocumab trials were sicker to begin with, which would have increased the odds of finding a difference in mortality.
My guess is that both drugs probably do result in a reduction in overall mortality. If you compare the confidence intervals for the two drugs you find that there is some overlap, at between a 9% relative risk reduction at best for evolocumab and a 4% relative risk reduction at worst for alirocumab. The true benefit is probably somewhere in that overlap. If we assume the relative risk reduction is halfway between those two numbers – 6.5% – that would mean someone with a 3% risk of dying over the next few years could decrease the risk to 2.8%.
Impressive? Or not? Well, it would mean that you would need to treat 500 people for a few years to prevent one death. Which isn’t very impressive at all. It’s possible that the benefit increases over time, but it’s also possible that it stays at the same level or even diminishes over time. Without any long term follow-up it’s impossible to know. And the studies were virtually all industry funded, so the real world benefit may well be smaller.
What about the third PCSK9 inhibitor, inclisiran, that’s just been approved in the UK?
Two trials, with a total of 3,178 participants, were published in the New England Journal of Medicine in April 2020. Participants received inclisiran or placebo and were followed for 18 months. So, what did the researchers find?
1.8% of people had a heart attack in the inclisiran group, compared with 2.5% of people in the placebo group. Not bad, especially when you consider that the trials only ran for a year and a half. The relative risk reduction is 28%, which is similar to what the studies of alirocumab and evolocumab found.
If we look instead at overall mortality, we see that 1.6% of participants died in the inclisiran group, compared with 1.9% of people in the placebo group. Again, not bad for such a short follow-up if real, since it represents a 16% reduction in the relative risk of dying, but the small number of participants in the study means the result isn’t statistically significant. It is thus impossible to say whether it’s real or not.
Ok, what can we conclude from all this?
People who take PCSK9 inhibitors appear to lower their relative risk of a heart attack by somewhere between 20% and 30%. Whether that is a meaningful difference or not in real terms will depend on your starting risk. For someone at high risk of a heart attack, it might be useful to take a PCSK9 inhibitor. For someone at low risk, the benefit will be negligible.
There is still too little evidence available on PCSK9 inhibitors to be able to determine whether they can actually prolong life, although there is a suggestion of a small benefit. As mentioned earlier, the results presented here should be viewed as best case scenarios since virtually all the evidence comes from pharmaceutical companies. Maybe in another twenty years there will be some large independent studies done and we’ll know for sure how effective these drugs are.
Blaming cholesterol for heart disease is akin to accusing emergency services like the fire brigade of causing fires.
Why the enthusiasm for relative risk reduction (RRR), and no mention of absolute risk reduction (ARR)? Because ARR 0,07% and 0,03% would not look “impressive”? This kind of spin has become typical of the pharma industry sales pitches but from you Sebastian I expected something different. I am disappointed.
I present all the real numbers, so people have access to all the data they need to come to their own conclusions. In this case I think it’s helpful to mention relative risk reduction because the studies are so short, and relative risk reduction is more likely to stay the same as the studies are prolonged while absolute risk reduction increases. Short term studies underestimate benefit if they only look at absolute risk reduction. Additionally, relative risk reduction is possible to compare across studies with varied populations, while absolute risk reduction isn’t.
My issue is the WAY you present the real numbers, using RRR rather than ARR, and it seems to be an issue for many of us.
You mention – for me, at least – benefits of presenting results as RRR.
Can you give me the sources of the statements you just made
– relative risk reduction is more likely to stay the same as the studies are prolonged while absolute risk reduction increases
– short term studies underestimate benefit if they only look at absolute risk reduction.
– relative risk reduction is possible to compare across studies with varied populations, while absolute risk reduction isn’t
Thank you.
It doesn’t require sources, just logical thinking. The longer you follow people for, the more heart attacks there are going to be overall, and the bigger the absolute difference is likely to be between the treatment group and the placebo group. The same is true if you compare a study of low risk people with a study of high risk people. The study of high risk people will have more total events, and will therefore likely show a bigger absolute difference between groups.
We are told that observational studies in medicine has no value. Relative risk reduction of 20% to 30% may be statistically significant but it is not real life significant.
Well, that all depends on what the absolute risk is. For someone with a high risk, a 20% to 30% risk reduction is good, while for someone with a low risk it’s meaningless.
Sebastian, if someone has a high risk for a disease, say 50%, and there is a high RRR, let’s say 30%, then the absolute risk reduction will also be “high” just like the RRR.
There is still no legitimate reason to present the RRR. The RRR is never relevant in real life. It is just a tool to mislead.
The actual or absolute risk reduction will not change on a percentage basis no matter how long a study is drawn out. Accordingly, your reason to use RRR as an extrapolation tool for short-term studies just doesn’t hold water.
I don’t understand your last paragraph. Say you run a study for five years and the reduction is from 10% to 5 %, then you do one for ten years, and the reduction is from 20% to 10%. The absolute risk reduction is 5% in the first study but 10% in the second study. It changed.
Sebastian, I am saying that the absolute risk reduction of death from all causes (the only thing we should pay attention to as you yourself have pointed out), expressed as a percentage, should remain stable as time goes on if the study was done properly.
It MAY change but there is no good reason to assume that.
Moreover, if the absolute risk reduction does change, it could certainly change in the unwanted direction as toxicity from the intervention could pile up over time.
I still see ZERO reason to ever use RRR when talking about the benefits of any proposed medical intervention.
The absolute reductions are not so great.
Statins have no effect on prolonging life in people who have not undergone a cardiovascular event. Beware the use of relative and absolute risk in statistical analysis. Around 40% of people who have a heart attack have cholesterol and blood lipids within normal criteria. The model of cholesterol and heart disease is very dated and problematical
Interestingly, dr. Kendrick is not enthusiastic about inclisiran:
https://drmalcolmkendrick.org/2021/09/23/inclisiran-sneaks-through-under-cover-of-covid19/
And he is right to be so, especially considering how expensive it is at present. Like I write above, it hasn’t been proven to prolong life, so, as he says in his article, with currently available information it is impossible to estimate what the cost per QALY is.
https://drmalcolmkendrick.org/2020/08/05/cholesterol-lowering-has-no-impact/
I don’t think we should ever use relative risk reductions in considering whether to use a drug or any other medical intervention to prevent disease — because they don’t matter in real life. What we care about is the actual reduction in risk. Please consider this Sebastian. Relative risk reduction percentages only serve to mislead people which is the opposite of what you are trying to do.
As far as the studies, you did not mention morbidity of the drugs. You only mentioned the drug companies covering up morbidity. This is a huge issue.
You also failed to mention alternatives to the drugs. If good alternatives exist, such as better management of diabetes for example, then this should be weighed against the benefits/risks of the intervention.
Pharma companies are also notorious for cherry-picking study participants. I don’t trust the studies for this reason. They find a small benefit (on an absolute basis) in a select population, advertise a large relative risk reduction, and then they obtain approval and the drugs are given to classes of people who were never studied.
Finally, one has to look at the results of these studies and consider if those results are congruent with what we know of human biology/physiology. There is no reason to believe that a cause of heart disease is the over-production of LDL molecules. Barring a genetic disease or an external factor affecting normal physiology (like another drug or toxin), the production of LDL molecules is something the body does normally, and normal physiologic processes do not cause disease!!!
Moreover, LDL levels are very weak markers of heart disease, and if correlation is very weak, causation must be accordingly weak if present at all. In summary, there is no biologically plausible mechanism I know of whereby LDL molecules can cause atherosclerosis itself or the rupture of the plaques and subsequent intra-coronary thrombosis that causes MIs and some ischemic CVAs.
For all these factors, I think you should consider revising the conclusions of your post. I appreciate your content.
Thanks , I agree with all your points.
I had not read your post before I made my own comment regarding Sebastian’s use of RRR instead of ARR.
Completely misleading and should never be done.
So, I think the absolute vs relative risk discussion is interesting. My thinking on the matter has evolved over the course of the last year, and it’s not because I’m in the pocket of big pharma.
I have to call you out about relative risk being forbidden in evidence based medicine. That is clearly incorrect. Meta-analyses, such as those done by the Cochrane Collaboration, always present relative risk, never absolute risk. The reason is that relative risk can be compared across studies, while absolute risk can’t.
The most important numbers to show are the real numbers, for example that risk decreased from 10% to 7%, so I alway present those if I can find them. After that we have absolute risk reduction and relative risk reduction, which are both flawed metrics, and I think both have their place.
As an example of how ARR is flawed, it definitely matters if a drug decreases risk of death over the next year from 6% to 3% or from 56% to 53%, yet absolute risk reduction makes no distinction.
Relative risk is the more useful metric when studies are short and it’s reasonable to assume that the absolute difference will be larger with time. A good example is the covid vaccine studies. Since they presented results after just two months, it was completely right to focus on relative risk reduction. Focusing on absolute risk reduction would have been absurd, considering that only a tiny proportion of the study population had been exposed to the virus.
I assume that my readers are mathematically literate enough that they understand what I’m saying if I say ”the risk reduction is from 10% to 7%, which is a 30% relative risk reduction”, and thus won’t be manipulated by the numbers.
For what it is worth from a non-medical doctor’s perspective, my understanding was that both relative and absolute risk factors were supposed to be presented with results in a publication. I would prefer to see absolute differences.
Presenting relative risks may make it easier to compare across studies, but usually I’m interested in exactly what I can expect from the results of a study. I have no problem with meta-analyses, I’m not sure there is any other way to assess a field of study objectively, but typically that is a heterogenous mess they are studying.
“Controlled” trials are all well and good, but we all know that at best we only control what we think are relevant factors – and usually not all of them can be well controlled. In a dietary study, for example, you could try to control for the proportion of a diet that is vegetables/red meat, but probably not what kind of vegetables or red meat (well, maybe in a prison). Now multiply that across studies.
Sebastian brings up the early vaccine trials as an example of where relative risk factors were informative. I found them misleading for efficacy – they generated very high expectations (while at the same time showing that protection wasn’t absolute).
Absolute risks, though, were very informative for how very few people in the study were actually coming down with covid and how unclear the results really were.
Recently a nurse told me “your cholesterol level is good”. I replied that it had been good ever since it was first measured many years ago. What I didn’t say was that there had been a spell when my GP strongly urged me to take statins. “But my cholesterol is fine” said I. “It can never be too low” he replied.
There’s the thing that needs explaining: how can an intelligent chap with medical training say “it can never be too low”? Maybe someone should develop a pill to stop GPs talking rubbish.
big smile…….
Prolong life? Why would that be a medical concern, when so many doctors are stating they are unable to adhere to the oath they originally took. From almost everything I read the world’s goal is to reduce the planet’s over-population by 50%, openly stated now on powers-that-be websites.
You fall to mention that all PCSK9 trials are performed in patients who are already treated with an optimal dose of a statin. We will never be told what these drugs will do by themselves in terms of reducing mortality because it has been postulated that not giving a statin is unethical. Which is all rubbish of course. Pcsk9
Inhibitors are a waste of money.
Interesting that you say that the LDL is hoovered up by the liver, I had thought that the ‘hooks’ were present on all cells that needed to ‘catch’ LDLs.
Even if I am wrong, the question is really, is enough LDL entering cells in those ‘unlucky enough’ to not have defective a PCSK9 reabsorption process?
Maybe the benefit is actually nothing to do with circulating LDL but everything to do with a fully requited LDL need of processes inside cells.
Absolutely, which is why I don’t personally make any claims about whether the LDL hypothesis is right or not here, I just report the study findings. I agree with Malcolm Kendrick that the primary cause of heart disease is that vascular wall repair happens more slowly than vascular wall damage. That doesn’t mean LDL isn’t involved in some way. The fact that people with defective PCSK9 genes have a much lower rate of heart attacks suggests that LDL does contribute in some way to the process of atherosclerosis, even though it clearly is neither necessary nor sufficient.
It is, unfortunately, the case that the genetic ‘defect’ has additional effects on cell metabolism as well as different effects on different circulating lipoprotein subclasses. https://lipidworld.biomedcentral.com/track/pdf/10.1186/s12944-020-01280-0.pdf
We are probably still in the World of hubris and unintended consequences – think monkeys mending Swiss watches with hammers or, at worst, playing with hand-grenades.
Or am I being too cynical?
Interesting findings. What was the actual cause of the heart attacks? Was it atherosclerotic plaque blocking coronary arteries or other heart pathology? Were all subjects that died in both the drug group and the placebo groups autopsied to confirm cause of heart attack and death?
It would be good to provide CTA of the hearts on all subjects in both groups prior to the trials to see if PCSK9 inhibitors had any effect on calcification and atherosclerotic plaque build up in coronary arteries.
What is the cost of these drugs?
NICE in the UK has approved these PCSK9 inhibitors but no detailed costings.
Dr. Rushworth;
I am one of a growing group of folks who have become familiar with Dr. Chris Knobbe’s work and I’ve changed my diet as a result. I’m new to your blog but in evaluating some recent tests you didn’t comment on those aspects of test that completely failed to account for the toxic effects, if present, due to the use of seed oils. What is your opinion of Knobbe’s work and the test data that he uses when making his case. Personally, i’ve found it convincing.
Dr Rushworth
Then what harm does the cholestetol do?
thank you for the review.
However I am really disappointed that you keep on calculating Relative Risk Reduction instead of Absolute Risk Reduction. You should know, as does anyone conversant in EBM, how misleading that is.
For example , you say “So, basically alirocumab appears to result in a massive 28% reduction in risk of death”, which sounds so impressive I would expect the drug company to have said that.
What you should have said is that ” alirocumab results in a 1% Absolute Risk Reduction of death”.
Which is not so impressive, and means that 100 people have to be treated for one death reduction.
Sebastian, thanks for your reply on this debate. You are correct that giving the figures as well as the RRR means that “as long as your audience is mathematically literate” they can work it out. But of course, you know it is not as simple as that in real life.
I suppose my point was that giving RRR figures alone is misleading, and there can be no argument on that point.
Being able to compare different studies is not that useful if the figures are misleading.
In the example you gave : ” As an example of how ARR is flawed, it definitely matters if a drug decreases risk of death over the next year from 6% to 3% or from 56% to 53%, yet absolute risk reduction makes no distinction.”, the flaw and potential to mislead would be much greater if you report RRR.
Had you given RRR, there would have been a “massive” RRR of 50% in the first case, but only 5% RRR in the second.
Yet both have ARR of 3%.
Does that not mean that NNT is 33 in both examples?
In that case should we not be talking about NNT ( and hence ARR) and forget RRR?
NNT is the same in both, as you say. I think NNT, like RRR and ARR, is flawed. Like the other two it provides a partial picture. The best thing from my perspective is to have access to all these different metrics, that gives the most complete picture. In the example above, I think I’d take the drug that dropped risk from 6% to 3% but I’m not sure I’d take the drug that dropped risk from 56% to 53%, even though ARR and NNT are the same in both.
On average, a drug that decreases risk of dying per year from 56% to 53% is only prolonging life by a week or two, while a drug that decreases risk from 6% to 3% is prolonging life by a decade or two. So although the two drugs have the same ARR and NNT, the real world difference is enormous.
ARR and NNT is basically the same metrics, where the one is the inverse of the other.
Sebastian, this debate on ARR vs RRR is fascinating to me, because it highlights how these abstract mathematical constructs we use daily in EBM are not intuitive to the human brain when we try to apply them to real patients in the real world, so thank you for engaging in this debate.
E.g. in the example you gave I don’t think I would take either drug.
If the ARR was reduced from 6% to 3%, then that means there is a 94% chance that I don’t need the drug at all, so 94% of the population is risking side effects for no benefit.
So once again, the figure of 3% ARR is more important to me than to know the RRR of 50%, which sounds so good, but really is not in the real world.
So I can’t agree that “on average it would extend life by decades” , I don’t see how that conclusion can be drawn.
Thanks again for the interesting discussion.
The thing is that in this example the 3% probability of benefit is after one year. If 6% are dying each year in the untreated group, then after ten years, 1-(0,94^10) = 46% are still alive. In the treated group 1-(0,97^10) = 74% are still alive. So the absolute reduction in risk after ten years would be 74 – 46 = 28%. While more than half are dead in the untreated group, only a quarter are dead in the treated group. While the one year NNT is 33, the ten year NNT is 4.
This of course assumes that the risk of death and benefit of treatment stays constant, just for the purpose of this hypothetical example. In reality both are likely to vary over time, but it illustrates that shorter studies that only look at ARR and NNT can grossly underestimate benefit.
Thanks Sebastian,
you make a good point , I had not thought of calculating it year by year.
But as you say, this assumes the death rate stays constant in each group which probably is unlikely, but still, I appreciate the calculation.
Won’t short term studies underestimate risk as well as benefit? Isn’t this an argument against rolling out vaccines without conducting long term studies?
Absolutely, the shorter a study is the lower the absolute number of events, so the lower the risk will appear to be.
I’m on a Ketogenic diet, and have been for three years. I’ve learnt that the problem is vegetable oils, or to give them their proper name, seed oils.
Seed oils oxidise during the manufacturing process, in storage, while cooking, and once digested.
It is these oxidised particles that damage the LDL, which breakdown and release the cholesterol. It is the broken and damaged LDL, and the released triglycerides, that cause arteries to become congested.
LDL and HDL carry Cholesterol to every cell in the body, so why inhibit it’s production?
I’ve been taking Vitamins K2 & D3, which will have cleared most of the deposits in circulation system. As a male aged 68yrs, my bp is now around 125/70.
Cooking with olive oil, avocado oil, butter, and animal fats, is how we used to cook.
Artery problems have only increased since the early 1900’s when seed oils were introduced.
Heart disease and diabetes have increased since sugar became cheap and plentiful, and especially since Mr Kellog started to add sugar to his new cereals to increase flavour and sales.
There is no need to interfere with the bodies production of lipids, but there is a need to stop the damage of lipids from seed oils.
I can furnish you with relevant literature if you require.
Best regards
Steve
What dose of K2 do you take?
since we all die eventually, surely the measure of success should not be mortality after treatment with drug or placebo but life expectancy – do the drugs significantly lengthen the time that the drug taker is on the planet.
How do I contact Sebastian privately with a question please? I am a patron.
Send me a direct message through Patreon.
There is no mention of side effects of PSK9 inhibitors. Since most are already on statins & other lipid lowering drugs, it is probable that the side effects are due to further decreases in LDL. They do seem to be not great for quality of life, and argue that we need LDL. The following is out of 38 patients.
” Twenty three percent reported mild-to-moderate side effects; most common fatigue (12%), and abdominal symptoms, skin reactions, nasopharyngitis, flu-like symptoms (all 5%). Two patients (5%) experienced serious side effects i.e. a major depression and an attempted suicide. Three patients discontinued treatment due to side effects. Five patients (11 %) dropped one or more doses of PCSK9-inhibition. ”
https://www.jacc.org/doi/full/10.1016/S0735-1097%2818%2932295-2
A recent phone conversation brought it to light for myself that I’m not the only one who is fed up with assessments based entirely on number of years with zero consideration of quality.
For each of us in that conversation the option: ‘happy and short’ far exceeded ‘long and miserable’. The latter is, for us, totally unacceptable – an absolute ‘no thank you’
Jane,
Thanks. You said it better than I. Welcome to the machine.
Dr. Rushworth,
Would you please analyze the following study–at least to point out any glaring holes that you see?
“Hydroxychloroquine / azithromycin in COVID-19: The association between time to treatment and case fatality rate”
https://www.sciencedirect.com/science/article/pii/S1477893921002040
Oh, I found the limitations statement about RECOVERY’s trial of HCQ:
“These findings indicate that hydroxychloroquine is not an effective treatment for hospitalized patients with Covid-19 but do not address its use as prophylaxis or in patients with less severe SARS-CoV-2 infection managed in the community.”
So RECOVERY says nothing about outpatient HCQ treatment effectiveness.