I want to thank our friend ‘OLD BOB’ for alerting me to Patrick Holford’s comment on a recent trial of vitamin C for COVID-19. Here are three short quotes from Holford:
… Overall, 5 out 26 people (19%) died in the vitamin C group while 10 out of 28 (36%) receiving the placebo died. That means that vitamin C almost halved the number of deaths. Those on vitamin C were 60% more likely to survive.
… Of those most critically ill, 4 people (18%) in the vitamin C group died, compared to 10 (50%) in the placebo group. That’s two-thirds less deaths. Statistically this meant that of those most critically ill who were given vitamin C, they were 80% less likely to die…
… now there is another proven treatment – vitamin C…
And here is the abstract of the actual trial Holford refers to:
Background: No specific medication has been proven effective for the treatment of patients with severe coronavirus disease 2019 (COVID-19). Here, we tested whether high-dose vitamin C infusion was effective for severe COVID-19.
Methods: This randomized, controlled, clinical trial was performed at 3 hospitals in Hubei, China. Patients with confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in the ICU were randomly assigned in as 1:1 ratio to either the high-dose intravenous vitamin C (HDIVC) or the placebo. HDIVC group received 12 g of vitamin C/50 ml every 12 hours for 7 days at a rate of 12 ml/hour, and the placebo group received bacteriostatic water for injection in the same way. The primary outcome was invasive mechanical ventilation-free days in 28 days(IMVFD28). Secondary outcomes were 28-day mortality, organ failure, and inflammation progression.
Results: Only fifty-six critical COVID-19 patients were ultimately recruited due to the early control of the outbreak. There was no difference in IMVFD28 between two groups. During the 7-day treatment period, patients in the HDIVC group had a steady rise in the PaO2/FiO2 (day 7: 229 vs. 151 mmHg, 95% CI 33 to 122, P=0.01). Patients with SOFA scores ≥3 in the HDIVC group exhibited a trend of reduction in 28-day mortality (P=0.06) in univariate survival analysis. IL-6 in the HDIVC) group was lower than that in the placebo group (19.42 vs. 158.00; 95% CI -301.72 to -29.79; P=0.04) on day 7.
Conclusion: This pilot trial showed that HDIVC might show a potential signal of benefit for critically ill patients with COVID-19, improving oxygenation even though it failed to improve IMVFD28.
The following points are, I think, worth mentioning:
- This was, according to its authors, a PILOT study.
- It was far too small (n=56) to provide reliable results on mortality.
- The trial authors know that and interpret their findings with sufficient caution.
- The primary endpoint, the IMVFD28, showed NO significant difference between the groups.
- The secondary endpoint: HDIVC infusion exhibited a non-significant trend of reduction in 28-day mortality (P=0.06).
- In more severe patients (SOFA score ≥3), univariate survival analysis and Cox regression showed a similar results (P=0.07, HR, 0.32 [95% CI 0.10-1.06]).
And what does all of this mean? It means that, in this pilot study, vitamin C failed to produce a significant result. Only in a subgroup analysis related to a secondary endpoint was there a slight advantage of vitamin C. This effect is, of course, interesting and needs further investigation (I am sure that is happening as we speak). It could have some clinical significance but, just as likely, it could just be due to chance. There is not way of knowing which is which.
In other words, to hype the findings and to even make statements such as ‘now there is another proven treatment, vitamin C’ is not just exaggerated, it is irresponsible.
This begs the question: why does Mr Holford do it? In case you don’t already know about this man, go on the Internet, and you will quickly find possible answers. Here is an excerpt from his Wiki page which might give you a clue:
Patrick Holford is a British author and entrepreneur who endorses a range of controversial vitamin tablets. As an advocate of alternative nutrition and diet methods, he appears regularly on television and radio in the UK and abroad. He has 36 books in print in 29 languages. His business career promotes a wide variety of alternative medical approaches such as orthomolecular medicine, many of which are considered pseudoscientific by mainstream science and medicine.
Holford’s claims about HIV and autism are not in line with modern medical thought, and have been criticised for putting people in danger and damaging public health.
In 2006 Holford was discovered to be using his PR advisor to delete critical content from his Wikipedia page…
Holford has been the subject of criticism for his promotion of medically dubious techniques and products including hair analysis, his support of the now struck off doctor Andrew Wakefield, and advocating the use of “non-drug alternatives for mental health” for which he has been given an award by the Church of Scientology-backed Citizens Commission on Human Rights.
SAY NO MORE!
28 Responses to Vitamin C, Patrick Holford, and dangerously misleading COVID-nonsense
Leave a Reply to Richard Rawlins Cancel reply
This site uses Akismet to reduce spam. Learn how your comment data is processed.
Haven’t heard of Holford in a long time. I’m hardly surprised he is trying to make money out of a pandemic.
Why indeed? A complete mystery.
I can’t follow his arithmetic here. Two-thirds as a percentage is 66%, not 80%. Though it is bad practice to use percentages when referring to such small numbers.
The authors of the paper have already performed their own statistical analysis and come to a different conclusion, which is that their results may well be due to chance but are worth investigating further.
I see that patients were recruited in March 2020 when the main risk factors for a poor outcome with Covid-19 had not yet been identified. Although the two groups were well-matched on most of these, roughly 50% more of the placebo group had hypertension and roughly double had coronary heart disease. This might have had some influence on the outcome observed. though in truth the numbers are so small that chance is a bigger factor here.
Perhaps in the original sense of the word “proven” meaning “tested”.
Are you seriously suggesting that Holford is trying to make money? Really? Wow, just take a look at the toxic drugs being used to treat patients all over the world and this radically fast-tracked experimental vaccine, which will undoubtedly fetch billions in profits for the pharma cabal. Mainstream medicine blatantly ignores the role of vitamin D deficiency and other nutrient deficiencies with regard C19 and general illness despite the mound of evidence, but no, Holford is trying to profit off vitamin C! What rubbish.
While this study was too small to offer significant results, there is still a mountain of evidence for vitamin C as an effective treatment to C19/viral infections.
Also, it is well known that Wikipedia is heavily biased in favour of big pharma and mainstream medicine. It is basically a propaganda tool at this point.
” there is still a mountain of evidence for vitamin C as an effective treatment to C19/viral infections”
I’d be interested to see one or two studies which you think are the most compelling.
Despite lack of media coverage (unsurprising), there’s been a lot of success reported from the use of IV Vitamin C. One example: https://nypost.com/2020/03/24/new-york-hospitals-treating-coronavirus-patients-with-vitamin-c/
Vitamin C was also recommended as part of the C19 treatment plan in an expert consensus from Shanghai (note: it’s all in Chinese): http://rs.yiigle.com/yufabiao/1183266.htm
Vitamin C significantly lowered the mortality rate of patients with ARDS:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6777268/
Shortens ICU stays:
https://pubmed.ncbi.nlm.nih.gov/30934660/
@ ryan matters
in the first study you quote they were also using hydroxychloroquine and azithromycin as well as “biologics” – they were clearly throwing the kitchen sink at these poor= people. I don’t think we should take their advice any too seriously.
Studies from China (in Chinese) need to be taken with large doses of salt unless they can be replicated elsewhere. There are serious problems with research in China which even the Chinese admit.
And your study on ARDS actually states that the net effect was “zero.” It might help to check what the result of a study is before you post it – a non-committal study is neither here nor there.
So you are left with the ICU study which appears to show some mild benefit from giving Vit C. Well blow me down – but that doesn’t appear to be a very rousing support for Vit C in Covid-19 does it?
And in what way does that give any credence to the nutty ramblings of Patrick Holford and his wildly inappropriate interpretation of the study in question?
He has totally extrapolated from a tiny number and claimed effects that are not supported by the evidence.
No-one would be objecting if he were trying to make money from an effective treatment – but he is making wild claims not supported by the evidence he is hawking.
In addition are you prepared to back up the charges he is making for his “treatments”?
Vitamins and supplements are by and large cheap to produce – but look at the prices he charges!
“£56.95 Vitamins For Life” printing money is more like it.
£67.93 Turmeric and Honey – can it really cost that much?
Profiteering might be a more apt phrase
As for big pharma doing it – who do you think makes his vitamin pills and supplements? The tooth fairy?
He is part of the exact same industry – but you just give him a pass because “vitamins and supplements” are supposedly natural! Illogical!
At least there’s a reasonable of a targeted drug treating a disease – whereas unnecessary vitamins or supplements taken instead of a proper treatment may well lead to injury or death – and SCAM artists often give bad advice or discourage people from seeking appropriate care.
Excess Vit C can cause renal stones and other ills. Excess Vit D is even more dangerous. Just because something is natural doesn’t mean it is good for you in excess.
Arsenic and rattlesnake venom are natural too.
“Studies from China (in Chinese) need to be taken with large doses of salt unless they can be replicated elsewhere. There are serious problems with research in China which even the Chinese admit.”
I agree: https://edzardernst.com/2016/10/data-fabrication-in-china-is-an-open-secret/
Someone’s been misleading you Ryan.
Glad you follow this blog – there’s much to learn.
The Origin of the 42-Year Stonewall of Vitamin C
https://www.seanet.com/~alexs/ascorbate/199x/landwehr-r-j_orthomol_med-1991-v6-n2-p99.htm
“Today there are areas of the world where polio vaccine is still not used and where the incidence of polio is increasing.”
oh really?
From here:
https://en.wikipedia.org/wiki/Confidence_interval
this:
“Misunderstandings
See also: § Counter-examples, and Misunderstandings of p-values
Confidence intervals and levels are frequently misunderstood, and published studies have shown that even professional scientists often misinterpret them.[8][9][10][11][12]”
From here:
https://en.wikipedia.org/wiki/P-value
“Since the precise meaning of p-value is hard to grasp, misuse is widespread and has been a major topic in metascience.[4][5]”
At the head of the CI above (Confidence Interval) it says:
“This article needs attention from an expert in statistics. The specific problem is: Many reverts and fixes indicate the language of the article needs to be checked carefully. WikiProject Statistics may be able to help recruit an expert. (November 2018)
This article may require cleanup to meet Wikipedia’s quality standards. The specific problem is: Prose is confusing, cluttered, and I’m not sure about the accuracy of some things Please help improve this article if you can. (September 2020) (Learn how and when to remove this template message)”
From P-value link it says this:
“Misconceptions
Main article: Misuse of p-values
According to the ASA, there is widespread agreement that p-values are often misused and misinterpreted.[3] One practice that has been particularly criticized is accepting the alternative hypothesis for any p-value nominally less than .05 without other supporting evidence. Although p-values are helpful in assessing how incompatible the data are with a specified statistical model, contextual factors must also be considered, such as “the design of a study, the quality of the measurements, the external evidence for the phenomenon under study, and the validity of assumptions that underlie the data analysis”.[3] Another concern is that the p-value is often misunderstood as being the probability that the null hypothesis is true.[3][12] Some statisticians have proposed replacing p-values with alternative measures of evidence,[3] such as confidence intervals,[13][14]”
In other words, there is so much confusion about P-values that “some statisticians have proposed [replacing them with] confidence intervals” – i.e. the one of the “Prose is confusing, cluttered,” type above.
The whole thing is a sublime weapon of confusion, if only Wellington had had this, he could have defeated Napoleon without firing a shot.
what is your point?
Old Bob,
It looks as though you are saying that we shouldn’t use confidence intervals because a Wikipedia article about them is badly written.
I hope Wikipedia isn’t where medical statisticians learn their trade. I would even go as far as to suggest that if you bought and read a statistics textbook you might be better informed before posting.
If you would prefer something non-technical aimed at the lay reader, I can recommend “The Art of Statistics: Learning from Data” by David Spiegelhalter. This is a fascinating and very readable overview of some of the essential areas of statistics and probability with lots of examples of how people and institutions have got into trouble by not understanding numbers.
I’ve heard about Patrick Holford and his pseudo-theory about vitamin C for a long time. And finally someone dared to give an answer to this phenomenon. You’re great. But do you know what surprises me the most? And how most people today blindly believe everything they see or hear. They are educated people, most of them graduated from prestigious universities, so why are we so blind? It’s scary. Let’s get smarter and check the information we have.
Ahhh, Vitamin C and other bogus ‘medicines’. Readers of this blog may be interested to peruse this recent published study suggesting that a common fruit can influence cardiovascular health for the better -could it really be true? I appreciate this ‘evidence’ is typical of SCAM practitioners and even a published clinical trial may be fabricated in some way, so keep taking the medicine gentlemen!
https://www.tandfonline.com/doi/full/10.1080/10641963.2020.1764015
an RCT with 14 patients. BRAVO!
Edzard on Friday 16 October 2020 at 20:48 said:
“an RCT with 14 patients. BRAVO!”
What should the minimum number be?
the question confirms your cluelessness – even if I wanted to, I could not teach you stats and trial design.
you can start here: https://www.wikihow.com/Calculate-Sample-Size
Edzard on Saturday 17 October 2020 at 09:15 said:
“the question confirms your cluelessness – even if I wanted to, I could not teach you stats and trial design.
you can start here: https://www.wikihow.com/Calculate-Sample-Size”
Thanks, but how do you know, without trying?
So, using the dummy example with its default settings and a population of 320 million:
CI == 95%
Therefore z == 1.96 (from lookup table, supplied)
n == population size == 320 million
p == standard of deviation == 0.5
e == margin of error == 0.05
Sample Size == ((z^2 * p(1-p)) / e^2) / (1 + ((z^2 * p(1-p)) / e^2 * n))
== ((1.96^2 * 0.5(1-0.5)) / 0.05^2) / (1 + ((1.96^2 * 0.5(1-0.5)) / 0.05^2 * 320,000,000))
== 3.84 * 0.25 / 0.0025 / (1 + ((3.84 * 0.25) / 800000))
== 384 / 1 + 0.0000012
== 384
it only requires a sample size of 384 for the USA.
Old Bob,
That is not the correct method for calculating the sample size for a clinical trial.
I think what you are doing is a calculation related to the standard error, which is an estimate of how confident you can be that a sample taken from a population is representative of the population itself (so in this case, say you wanted to know the average height of a person in the US you might want to know how many people you should measure in order to be confident of your answer to a specific degree). Though I have no idea where you got a standard deviation of 0.5 from. Standard deviation of what measurement?
In any case if what you are measuring is categorical data (i.e. alive or dead) then it doesn’t have a standard deviation.
Dr Julian Money-Kyrle on Saturday 17 October 2020 at 20:26 said:
“That is not the correct method for calculating the sample size for a clinical trial.”
It comes from EE’s suggested link above:
https://www.wikihow.com/Calculate-Sample-Size
Where is says:
“Since this value is difficult to determine you give the actual survey, most researchers set this value at 0.5 (50%). This is the worst case scenario percentage, so sticking with this value will guarantee that your calculated sample size is large enough to accurately represent the overall population within your confidence interval and confidence level.”
Old Bob,
It seems that Edzard gave the wrong link, as this site gives the method for calculating the sample size for surveys, not the number of subjects required for clinical trials (which aren’t usually referred to as samples).
The site itself makes this quite clear:
Indeed, it takes only a moment’s thought to realise that establishing the effectiveness of a treatment for Covid-19 sufferers in China (or anywhere else, US included) should not depend on the size of the population of the USA.
In a survey, you want to know whether your sample is representative of the population as a whole. In a clinical trial you want to know that your results are unlikely to be due to chance. They are not the same question and they are not answered in the same way.
Thanks JMK, I don’t mind it being the wrong type for now; it’s headed in the right direction 🙂 – wrong planet, right solar system.
Old Bob,
That is actually a very sensible question, and one of the things that you need to think about at the beginning of designing a trial. The answer depends primarily on what size of effect you are looking for (e.g. if it is a reduction in mortality, are you expecting to find 0.5%, 5% or 50%?), how confident do you want to be of your findings (e.g. a significance level of 0.05, known as ALPHA meaning that if there is no effect then you have only a 1 in 20 chance of finding a spurious one, won’t require as many subjects as a level of 0.01), and what probability do you want of the study finding or missing the effect if it is real (e.g. a 9 out of 10 chance won’t require as many subjects as a 19 out of 20 chance; this is called the POWER of the study).
There are standard statistical tools for making this calculation, including on-line calculators such as this one:
https://clincalc.com/stats/samplesize.aspx
So if we are looking at something that we expect might reduce mortality from 30% to 20%, say, at a significance level of 0.05, and we will accept a power of 80% (i.e. that the stucy has a 4 out of 5 chance of finding the effect) then this will require a sample size of 586 patients.
Note that this is a lot more than 14 in order to be reasonably sure that your results aren’t entirely due to chance.
Thanks for the worked example and nice short explanation!
This is the source of the trial above:
https://www.researchsquare.com/article/rs-52778/v2
This is the second comment (at the bottom, below the tweets):
[quote]
“david austin
commented on 14 October, 2020
Mortality is a “categorical variable”, in fact it is the worst case categorical variable because it is boolean (patient is either dead or alive) so requires extremely high sample size to be statistically significant. STOP USING IT. Stop even discussing it with such small populations! (dumb tweets). It’s shouldn’t even be a secondary measure if the population is too small, which these are.
However, continuous variables on the other hand do NOT require large sample sizes, so they are ideal, additionally if they are absolutely predictive of patient recovery in the pathology of the disease then they should be the primary outcome measures. Thankfully low SOFA scores and low inflammatory markers like IL-6 are both continuous AND essential to the pathology of a recovering covid19 patient, and in this study they are both profoundly improved for Vit C group and profoundly significant from a statistical standpoint. More so than ANY OTHER TREATMENT TO DATE….
[end of quote]
In other words, according to this poster, the problem is not with vitamin C, it is with human misunderstanding of statistics (as reflected in Wikipedia above).
You have made a fundamental error in only reading the abstract. If you read the outcomes (eg all results) on page 6 table 2 you’ll see that:
Hospital mortality of those more critically ill (defined as SOFA >3) showed 80% reduction, and statistically significant p.04 –
ICU mortality of those more critically ill (defined as SOFA >3) showed 80% reduction p.04
The mortality at 28 days (includes those who died not in hospital or ICU) shows 70% reduction, close to significant p.07 (p.05 is the definition of significant so the first two results are significant and the last is close to significant – a clear trend in an underpowered study.
(HR means Hazard Ratio, so a hazard Ratio of 0.2 means 80% less risk of mortality)
Also, the primary outcomes, as in the abstract:
During the 7-day treatment period, patients in the HDIVC grouphad a steady rise in the PaO2 / FiO2 (day 7: 229 vs. 151 mmHg, 95% CI 33 to 122, P = 0.01), which was not observed in
the control group. IL-6 in the HDIVC group was lower than that in the control group (19.42 vs. 158.00; 95% CI -301.72 to -29.79; P = 0.04) on day 7.
This means that lung function got better and better in the vitamin C group, and worse in the placebo group.
The measure of inflammation (IL-6) got much better in the vitamin C group (19.42) versus the placebo group (158)
This study was meant to include 140 patients for all results to be able to be significant, if there was an effect.
It is underpowered due to Wuhan running out of critical covid-19 patients towards the end of April 2020.
For this reason it is called a ‘pilot’ study. The fact that there was a statistically significant 80% reduction in hospital/ICU mortality in the most critically ill, despite the small numbers, illustrates that this is a big effect and not one that should be ignored. A larger trial is called for.
According to Emeritus Pharmacology Professor David Smith, from the University of Oxford “When you consider that the steroid dexamethasone reduced mortality in ventilated patients by 34% but in the Wuhan study, we are seeing a significant 80% reduction in mortality in the most critically ill in hospital after vitamin C, we have to take this seriously. We urgently need a larger trial to see if this finding can be replicated.”
You might also want to read our review on ‘Review: Vitamin C – an Adjunctive Therapy for
Respiratory Infection, Sepsis and COVID-19’ – https://www.mdpi.com/2072-6643/12/12/3760/pdf -before jumping to unfounded conclusions.