Posted 8 May 2020
I shall leave the destruction of Dr Judy Mikovits to rhetoric rather than statistical science – it seems attempts to be critically curious of her story are too easily misinterpreted.
Instead let me return to something a little less conspiratorial: Vitamin D.
The thread so far is that there have been studies and indeed there continue to be new studies showing that poor COVID19 outcomes are correlated with patients showing the lowest serum levels of Vitamin D. A new study from India published on the 5 May reports the same correlations.
But there is still a disjoint. If there are low levels of Vitamin D in those patients does that mean supplementing with Vitamin D would have made a difference? Alas, unless Vitamin D supplementation is a specifically tested protocol in a study and outcomes of the COVID19 illness compared, we cannot say for sure.
You could be pragmatic of course. As long as you do not take supplemental Vitamin D at levels that could be harmful, all you have to lose is the cost of the supplement. And you might gain some benefit against the severity of COVID19.
Today I came across a paper that tries to fill that gap in our knowledge. I will quickly dash hopes and just say that the paper is not reporting on a trial where Vitamin D has been administered to COVID19 patients. Instead it is performing what you might call ‘statistical detective work’.
It looks at the problem from three perspectives.
A: Are there clinical trials where Vitamin D supplementation has been used to improve outcomes from other viral infections?
B: Are there clinical trials where Vitamin D supplementation has been used to improve outcomes from clinical presentations associated with COVID19?
C: Are there any clinical trials where Vitamin D supplementation has been use to improve any of the co-orbidities associated with poor COVID19 outcomes?
Perspective A is asking if Vitamin D supplementation has been successful in treating other infections caused by viruses and then ponders if there is any reason why such findings are likely or unlikely to be applicable to infections from SARS CoV2 virus.
Perspective B then looks at the specific characteristics associated with COVID19 (pneumonia, respiratory distress, sepsis etc) and examines if any trials have used Vitamin D supplementation in the critical care of those conditions (but from other pathogenic causes).
Perspective C asks if any Vitamin D trials have looked at improving type 2 diabetes, obesity, hypertension and so on, all of which are correlated to poorer COVID19 outcomes.
And seemingly there are plenty of historic studies to reference in each of these perspectives.
The paper comes to a number of conclusions.
1. Safety first. No study has shown any harm from daily doses of Vitamin D as high as 10,000IU per day for three years. Many studies show dosages much higher than this for one month or less and have no measurable side effects. In two trials with patients on ventilators, single dose administrations of 500,000IU reduced time on ventilation by around a third.
2. Vitamin D daily doses of around 3-4000IU are correlated with better outcomes with some respiratory tract infections (RTIs).
3. Vitamin D daily doses of 2000-5000IU reduced inflammatory cytokines and reactive proteins in critically ill patients.
4. Vitamin D daily doses of 4000IU reduced likelihood and severity of infection with some viral infections. None of those were coronavirus infections.
5. Studies looking at Vitamin D supplementation to reduce HbA1c scores in type 2 diabetics were inconclusive. Both studies showed dietary change was more strongly correlated to improvement.
What does that mean?
Well, it can be interpreted that regular daily doses of Vitamin D up to 10,000IU are not harmful. And doses of around 4000-5000IU per day have some benefit in reducing RTIs and inflammation, both conditions associated with COVID19 infection. Those doses may also improve your glucose control if you are type 2 diabetic, a known co-morbidity in COVID19 infection, but dietary management is more effective.
A little closer to an answer, maybe?