Covid-19 and vitamin D: still "not enough evidence" unless more people die
 |
| So many deaths. How many could have been avoided? |
But what is really going on? Why are so many deaths allowed to continue when we are ignoring the evidence supporting vitamin D?
The proposal is that:
“Vitamin D deficiency is common, especially in the elderly, the obese, in UK citizens with an ethnic melanin-rich skin, and in the winter;
vitamin D deficiency creates a serious biological disadvantage, at present in respect of Covid-19 infection;
vitamin D is essential for the escalation of defensive immunity at the time of serious infection;
correction of vitamin D deficiency is a medical duty;
correction of vitamin D deficiency by an oral supplement corrects the biological disadvantage.”
This would seem to be a non-controversial proposal in that it is based on many observations and much knowledge.
During the past eleven months there have many substantiated claims of benefit from vitamin D, but also many denials. The reason why vitamin D has turned out to be so controversial is very difficult to understand, but the reality is that correction of vitamin D deficiency has been neglected by our supposedly scientifically-directed government and by public health policy, at very great human cost. It has been suggested by the research group in Heidelberg that the UK might have experienced 90,000 fewer deaths from Covid-19 if the public health bodies had corrected widespread vitamin D deficiency, especially in the population groups know to be at risk of both serious vitamin D deficiency and death from Covid-19.
Desperate times demand desperate measures
However, correction of vitamin D deficiency appears to be a step too far. Why the hesitation? Widespread vitamin D deficiency is well-known in medial science, if not by the general public. Vitamin D supplementation, ideally preceded by checking blood level, should be a natural almost knee-jerk response to what can only be described as an overwhelming pandemic of a novel virus that has caused more that 4 million cases and more than 120,000 deaths in the UK alone. Optimisation of defensive immunity (knowing of the importance of vitamin D in this process) would be a sensible and obvious action.
What on earth has gone wrong? I have been overwhelmed by evidence supporting the above proposal and yet NICE (the UK National Institute for Clinical and Care Excellence with considerable international influence) repeatedly claims in a parrot-like fashion that there is not enough of it. How much is “enough”? Is evidence to be measured by weight of paper, or in the current era by accumulated word-count? The knee-jerk reaction of NICE is vitamin D denial.
Pascal’s wager
The practical objective is to activate a decision. Are we as a nation going to correct vitamin D deficiency or are we not? It is an application of Pascal’s wager.
If vitamin D deficiency is an important and effective factor in serious and fatal Covid-19:
What can be gained by promoting vitamin D?
Answer: many deaths prevented.
What is to be lost if it turns out that vitamin D deficiency is
not important and vitamin D is not effective?
Answer: just wasted effort, as vitamin D is cheap and safe.
If vitamin D deficiency is an important factor in serious and fatal Covid-19:
What would be gained if we refuse to correct wide-spread vitamin D deficiency, as NICE proposes?
Answer: nothing.
What is to be lost
Answer: many avoidable deaths.
If vitamin D is not of value, there is no loss from either promoting or not promoting it.
It is glaringly obvious that the choice with the greatest utility, potential benefit, is to promote vitamin D, with the possibility of the prevention of many deaths (perhaps many thousands) and with no disadvantage.
The choice with the greatest loss is to refuse to promote vitamin D (which turns out to be effective as all evidence indicates) with the result of many avoidable deaths, and with no gain. This is the position of NICE. There is much evidence, but allegedly "not enough". Utility is not a consideration of NICE.
Why is NICE involved in decision-making?
NICE is an advisory body. It has no executive function and it does not make decisions. Unlike clinical doctors it has no direct clinical responsibility to individual named people. NICE has a main function of evaluating pharmaceutical agents and clinical procedures, and advising government and hospital trusts on best practice.
We are dealing with widespread vitamin D deficiency, which is responsible for many deaths from Covid-19. This is beyond doubt but it is carefully avoided by NICE. It would seem to be a straightforward clinical responsibility to detect and correct vitamin D deficiency, like hypothyroidism (underactive thyroid) or diabetes (especially Type 1).
It is very strange that NICE has become involved in this artificial evaluation of the need to correct deficiency of vitamin D, which is converted into a hormone within the body. What can be controversial about correcting a vitamin or hormone deficiency? If NICE did not exist would there even have been a controversy? It is interesting to note that NICE has been completely bypassed by the government in the vaccine roll-out programme, even though vaccines are pharmaceutical products and not natural substances.
NICE avoids any mention of vitamin D deficiency and treats vitamin D as a pharmaceutical product. That is its modus operandi. NICE has stated that there is no proof that Vitamin D is of value in Covid-19, but introduces the word “proof” without defining it. Proof is the fulfilment of predetermined criteria, but like so many others, NICE has not stated its criteria of proof. It demanded a randomised controlled trial (RCT), but when one appeared (from Córdoba) it demanded another. Now that it has appeared (from the University of Barcelona) we await with interest the official response of NICE. So far, just silence.
I learned about "proof" in Euclidian geometry when I was at school. It was followed by quod erat demonstrandum (QED), "what was to be shown", or it has been demonstrated, perhaps quite easily done! "Demonstration" is an important concept: I can demonstrate gravity but I cannot prove it. The extent of vitamin D deficiency can be demonstrated. The disadvantages of vitamin D deficiency, including with Covid-19, can be demonstrated. The benefit of vitamin D supplementation can be demonstrated. How does demonstration differ from proof?
RCT from Barcelona
Since the appearance a week ago of the results of the Barcelona RCT we have had no official response from NICE. However we have had a response from Professor Adrian Martineau, one of its members. Professor Martineau, together with Professor Sattar from Glasgow, set the scene in a dismissal of the Barcelona study, critical of aspects of the methodology. Guess what? They demand more RCTs, but they must be “robust” (good management-speak).
One criticism was an apparent "absence of registration" of the trial before it commenced. This is an administrative issue that would have no bearing on outcome. The study had permission from the hospital ethics committee. Without this there would have been a valid criticism, but lack of prior registration can surely not invalidate the findings of the study.
Another criticism is a lack of clarity of the process of randomisation. The 930 patients were randomly allocated to one of eight dedicated Covid-19 wards. On three of them the patients received high quality standard care. On the other five wards, the patients received in addition calcifediol, a natural substance, part-activated vitamin D with a very rapid action.
Randomisation
The purpose of randomisation is to produce two groups as near identical as possible, except for the the treatment under investigation, in this case calcifediol. Randomisation could be made in a number of ways: the spin of a coin, the random selection of an envelope containing an allocation code, a random number generated by a computer, year of birth odd or even number, etc. The Córdoba study involved electronic randomisation. The Barcelona study just tells us that the patients were "randomized" to the wards, with no indication of the method. At a time of very busy hospital activity, randomisation could simply have been the on the basis of the next available bed on any of the eight wards, as there would be very few empty beds to choose from. The absence of method of randomisation in the paper could be corrected easily, but would it have any impact on the result?
The success of randomisation can be determined by comparing characteristics of the two groups of patients. Ideally they should show no significant differences, but the variations of humankind mean that randomisation is seldom perfect, even with a sample size of 930 as in the Barcelona study. Experiments in physics, chemistry, or plant and animal studies can be controlled rigorously, but not so with humankind and civil liberties. Pragmatic randomisation must be used, with as large numbers as possible to minimise the effect of natural variation.
The randomisation profile of the Barcelona study is shown in the table.
There are inevitable differences between the two columns, but the majority are of no significance. We can see that randomisation was generally successful.
Baseline vitamin D (calcifediol) testing
The median average blood level of the 752 of the 930 patients tested at baseline was 14ng/ml, very low at less than 20ng/ml. This indicates a high risk of serious or fatal Covid-19, above 30mg/ml being safe.
Of 495 patients with blood vitamin D less than 20ng/ml, 332 (67% of 495, 44% of 752) were found to be calcifediol treated patients, a higher proportion than 163 (33% of 495, 22% of 752) in the controls. This would give a disadvantage to the calcifediol treated group but it was not brought out in paper.
The study reported the only significant difference of randomisation to be the baseline measurements of vitamin D, blood levels of 25(OH)D which is calcifediol. The median average is 12ng/ml in the controls and 15ng/ml in the group randomised to receive calcifediol.
The difference between 12 and 15 does to appear to be very great but the report states that it is statistically significant. The sample sizes are good, but we are not given the standard deviations and so we are not able to calculate the standard errors, from which, with the sample sizes, we would be able to calculate for ourselves the statistical significance of the differences.
There is a conflict between the two methods of assessing differences of baseline blood calcifediol levels between the two groups. We cannot be clear whether or not the calcifediol treated group had an initial advantage as suggested in the randomisation table. Does this invalidate the trial? No. We must accept the result but ask clarification from the authors.
The primary endpoint
The most serious criticism by Professor Martineau is a departure from the protocol during the Barcelona RCT. This was for a reason that I regard as for strongly valid ethical reasons.
Conducting a trial of 930 critical ill patients is not easy. We are not dealing with laboratory animals, and the over-riding objective of the attending physicians in their clinical practice must have been to minimise possible deaths of their patients.
The primary endpoint was reached, with 21.1% of the control patients requiring admission to the ICU compared to 5.4% of the calcifediol-treated patients, a highly significant result, 74% efficacy (21.1-5.4=15.7; 15.7x100÷21.1=74%). This is a large difference in outcome that cannot be ignored, and it appears that it was not ignored during the trail.
The second endpoint: is human sacrifice essential?
The next endpoint was to be death following transfer to ICU. Calcifediol was of demonstrated great benefit in the first stage. The trial could have been stopped at this point, with great help if the protocol were to be used in other hospitals.
For the second phase in the ICU something had to be sacrificed. Was it to be rigour of the protocol or was it to be human life? – a conflict between experimental purity and the welfare of the very ill patients. The latter option was chosen as the way forward, the welfare of the sick.
It had to be decided whether the control patients who had been randomised not to receive calcifediol were to continue without it when on the ICU. The pragmatic decision was made to allow the clinical judgement of the attending physicians. Of the 80 control patients admitted to the ICU, 50 were started on calcifediol. Human sacrifice was thereby minimised, to the displeasure of Professor Martineau as implied in his criticism of departure from the protocol.
Deaths
The paper from Barcelona first presented analysis of deaths on the basis of "intention to treat". The result was 56% efficacy. But this method is not really justified as the majority of the control patients received calcifediol when on the ICU.
The analysis was repeated for the sum of the 551 randomised calcifediol patients and the 50 control patients given calcifediol when on the ICU, a total of 601. Of these 49 (8.15%) died. Of the 329 patients who did not receive calcifediol, 44 (13.4%) died. This is a 39% efficacy of the reduction of deaths by calcifediol (13.4-8.15=5.25; 5.25% is the gain from taking calcifediol; 5.25x100÷13.4=39% efficacy).
Had calcifediol been withheld from all control patients on the ICU, we would anticipate more deaths among the control patients, thereby increasing statistical efficacy, but this cannot be proved without further human sacrifice.
1,25(OH) vitamin D is consumed
A further "criticism" was the usual comment by Professor Sattar that a low blood level of vitamin D as calcifediol is the result of the disease. There is some truth in this statement in that it is well-known that a molecule of intracellular 1,25(OH)D can only be used once, and after use it is inactivated to 24,25(OH)D. This means that at a time of escalation of defensive immunity it is essential for the body to have a good supply of vitamin D as calcifediol circulating in the blood. Without this there is a high chance of critical disease and death.
Professor Sattar dismisses the RCT from Barcelona as “….not a useful study….”, and follows “We must await robust randomised trials to form appropriate conclusions”.
How do we define "conclusions"?
Professor Martineau does not discuss the ethical issues of the trial. He states: “Overall, more methodological detail is needed before the claims of treatment benefit can be substantiated.” Professors Sattar and Martineau do not define what they mean by "substantiation" and “conclusions”.
If we await a conclusion we will wait indefinitely while counting the dead. In science a conclusion is never reached, just a revolving wheel of research, often changing direction. There is no absolute truth. The paradigm, the clinical action, the acceptance of Pascal’s Wager, is pragmatic proof based on the best evidence we have at present, warts and all.
The meaning of Proof
The RCT is only just a part of proof, and a part that is not essential. This was understood by Sir Austin Bradford Hill, whose wisdom should be central to the current so-called controversy concerning the correction of vitamin D deficiency at the time of the Covid-19 pandemic and the 120,000 associated deaths in the UK.
Using Bradford Hill’s criteria, the evidence to support the use of vitamin D in the prevention and treatment of Covid-19 is far stronger than the evidence that cigarette smoking causes lung cancer, or that driving a car under the influence of alcohol, or without a seat-belt, increases risk of death. The governments of the day took action, with good results, and without RCTs.
The government of today should take action and advise the use of vitamin D in a dose to correct deficiency so as to optimise immunity rather than simply to avoid rickets. This should not be controversial to those with understanding of the issues. Had it been done, there might have been 90,000 fewer Covid-19 deaths.
The disappearance of the Barcelona study
The power of Professor Martineau and NICE must not be underestimated.
The result from Barcelona was published on a pre-publication website under the control of The Lancet. This is useful as publication can be a slow process. The Barcelona study was undertaken in March, April and May 2020. The result has taken a long time to come to attention, at a time when we have required "desperate measures" to reduce the pressure on ICU beds and the number of Covid-19 deaths.
Five days after its appearance and after a further two days following the comments by Professors Martineau and Sattar, the paper was taken down from its web-site. It can no longer be viewed, but no doubt many people (such as me) will now have a PDF. This is censorship, the burial of important information. Perhaps it will reappear with clarification of the process of randomisation. Time will tell.
The RCT from Córdoba was rubbished in a few sentences by Professor Neil Gittoes, who was acting on behalf of NICE. As a result of his cursory comments, NICE stated:
"The clinical management of patients with COVID-19 should not be changed based on the results of this study."
A team from the Massachusetts Institute of Technology undertook a very thorough review of the Córdoba study and came to the opposite conclusion from Professor Gittoes, reporting that process was acceptable and that the probability of a chance finding was less than one in a million. NICE has not commented on the MIT analysis.
And so we have three individuals, Professors Gittoes, Sattar, and Martineau who have been allowed to use extreme power to prevent the nation-wide treatment of patients with Covid-19 pneumonia using vitamin D as calcifediol, despite great evidence of the dangers of vitamin D deficiency and the benefit of correction. Their analyses have been very superficial. There views have not been challenged in the medical or national press.
Accountability
The Córdoba RCT result became available on September 3rd 2020. Since then there have been more than 56,000 Covid-19 deaths in the UK. Had treatment with calcifediol (price €10 per patient, no side-effects) been instituted, several thousand of these patients would not have died. Are these deaths of no importance?
Who is to be held accountable?
This Blog post contains a lot of numbers as I felt that my comments should be justified. If any reader spots a fault in the calculations. please inform me immediately.
davidgrimes1@mac.com
