A few months back I wrote an article about the only randomized controlled trial that had at that point been done on vitamin D as a treatment for covid. That trial, which was carried out in Spain, showed very impressive results. There was an incredible 96% reduction in the relative risk of requiring ICU treatment among those treated with 25-hydroxyvitamin D (an activated form of vitamin D).
Unfortunately, that study had some significant flaws. It was small, with only 76 patients in total, and it was open-label, meaning that both the doctors treating the patients and the patients themselves knew who was in which group. These two flaws mean there are lots of different ways in which the trial could have ended up showing an impressive result that isn’t real.
Since then, two more randomized controlled trials of vitamin D for covid have posted their results. The first study was published in the Post-Graduate Medical Journal in November 2020. This was a double-blind randomized controlled trial carried out at a hospital in India. The authors did not receive any specific funding for their work, and reported no conflicts of interest.
Healthy patients who were admitted to the hospital with asymptomatic or mildly symptomatic covid were eligible to be included in the study.
Why were they admitting people who were only mildly symptomatic?
Why were they admitting people who were completely asymptomatic?
Are there so many hospital beds available in India that they admit people who are completely healthy?
I don’t know, it seems very strange to me.
Apparently the only requirement to be admitted to the hospital was a positive covid PCR test, which for unclear reasons was taken from healthy people who weren’t showing any symptoms. This is odd, and it limits the validity of the study, because the probability increases that a large proportion of those being admitted to hospital will be false positives, and not really covid patients at all. Obviously, this will make it harder to find any meaningful clinical effect of a treatment.
Apart from being healthy and having a positive covid PCR test, potential participants also had to have vitamin D deficiency, which they defined as a level of circulating 25-hydroxyvitamin D in the blood of less than 20 ng/ml. This part kind of makes sense. A previous study I’ve written about found a huge reduction in respiratory infections in people with low vitamin D levels who were treated with vitamin D supplements, while there was a much smaller effect in people who had decent levels in the blood to begin with.
It’s reasonable to think that people with a deficiency would be the ones to benefit most from treatment with vitamin D, so from that perspective it makes sense to check vitamin D levels before treating. On the other hand, if vitamin D is shown to be effective against covid, then in the real world doctors probably aren’t going to check the vitamin D levels of patients before they start treating them – they’re just going to give them the vitamin D. So adding this requirement makes the study less applicable to the real world.
After inclusion in the study, patients were randomized to receive either 60,000 IU of vitamin D per day (1,500 mcg), or a matching placebo. In total, 40 patients were recruited in to the study. Of these, 16 received vitamin D and 24 received the placebo. Yes, I agree, 40 is a very small number. Yes, the odds that this study will be able to detect a meaningful difference, even if one does exist, are small at best.
Unfortunately, the researchers topped things off by choosing a completely meaningless endpoint for the study. They chose to look at the amount of time it takes for patients to have a negative PCR test. Patients don’t care about whether they are still positive on a PCR swab or not (at least, I imagine they don’t). They care about whether they feel sick or not, and about how much time they need to spend in hospital, and about whether they live or die, and whether they have complications. Unfortunately the researchers didn’t gather any data on symptoms or anything else that actually matters to patients, so that data isn’t even available for us to analyze.
Let’s get to the results. On average, patients in the vitamin D group became PCR negative at 18 days after inclusion in the study. Patients in the control group also became PCR negative at 18 days. So, there was no difference between the groups.
Would I have expected a difference? No. A study that only has 40 participants, and that chooses to only recruit people with disease ranging from mild to non-existent, is almost certain to fail to find any meaningful effect. This was a very low quality study.
Let’s move on to study number two. This one hasn’t yet been peer-reviewed, and is currently only available as a pre-print at MedRxiv. It was a double-blind randomized controlled trial involving 240 patients. So, so far, it ticks all the important boxes for quality. Unlike the Spanish study, this one was double-blind. And unlike both the Spanish and the Indian study, this one was also quite large. In fact it was more than twice as large as both the other two studies put together.
So, just from this limited information, we know that this is the highest quality study yet that looks at vitamin D as a treatment for covid. The study was carried out in Brazil, and was funded by the Brazilian government and by the state of Sao Paolo. The authors reported no conflicts of interest.
In order to be included in to the study, participants had to be over the age of 18 and have covid, as diagnosed either by PCR or a CT-scan of the chest. They also had to have symptoms that were severe enough to require hospitalization, or risk factors for more serious disease, such as diabetes, heart disease, cancer, or obesity. So far so good – they were looking primarily at people who were quite sick, or at least at increased risk of becoming quite sick. Those are the people we want to find an effective treatment for. The researchers excluded people who had previously supplemented with vitamin D, presumably to increase the odds of getting a significant result.
Participants were split evenly between the intervention group and the placebo group. Those in the intervention group received a single bolus containing 200,000 IU of vitamin D (5,000 mcg), dissolved in 10 ml of peanut oil, on the day of admission. The placebo group just received 10 ml of peanut oil, without the added vitamin D.
The average age of the participants was 53 years and the average BMI was 32. In other words, the average participant was obese. Apart from that, chronic health conditions were common among the participants. 53% had hypertension, and 35% had diabetes.
This is a good thing, because the sicker the patients are before getting covid, the greater the probability that the study will show a meaningful impact on things like mortality and ICU-admission (assuming the treatment actually works). If all the patients are healthy to begin with, as in the previous study, then mostly likely no-one is going to die, and no-one is going to end up in the ICU, regardless of what you do.
The mean time between symptom onset and receiving the bolus of vitamin D was 10 days. This is quite late in the disease course. In general, the people who get really sick from covid start to get worse at around day 7 after the onset of symptoms, and are at their worst around day 10-14, after which they usually start to get better. So this actually creates a problem. If the study fails to show benefit, it could simply be due to the fact that the vitamin D was given too late to have an effect.
This is especially true considering that the participants were given regular vitamin D, which needs to be activated in the body, first to 25-hydroxyvitamin D, and then further to 1,25-hydroxyvitamin D, before it can start exerting its effects. Remember that in the Spanish study that showed a huge reduction in ICU-admissions, the patients weren’t getting regular vitamin D, they were getting 25-hydroxyvitamin D, so the first step had already been carried out, which would have meant that it started exerting its effects on the body more quickly.
On average, both groups had serum levels of 25-hydroxyvitamin D of 21 ng/ml, which is pretty low, and in line with the large amount of observational evidence showing that people with low vitamin D levels are more likely to become seriously ill if they get covid.
So, what were the results?
Average length of stay in the hospital was seven days in both groups. 7,0% of the participants in the vitamin D group died, compared with 5,1% in the placebo group. 15,8% in the vitamin D group ended up receiving intensive care, compared with 21,2% in the placebo group. None of the differences was even close to being statistically significant.
That’s disappointing. Giving a high dose bolus of vitamin D at ten days after symptom onset had no effect on the disease course. This is the biggest, highest quality study so far, and on the face of it, this would suggest that vitamin D is ineffective as a treatment for covid-19.
However, as mentioned, there are two reasons to think that there still might be a place for vitamin D as a treatment for covid. The first is that it was given late in the course of the disease, and the second is that the patients were given regular vitamin D, not 25-hydroxyvitamin D. So, it’s still possible that high dose vitamin D given early and/or in the form of 25-hydroxyvitamin D is an effective treatment. That is certainly what the Spanish study suggests. Hopefully more studies will be coming out in the near future that can answer this question.
Luckily there is now mounting evidence that ivermectin is a highly effective treatment for covid, so we don’t have to get too despondent if vitamin D doesn’t pan out. And even if vitamin D does turn out to be ineffective as a treatment for covid, that doesn’t mean it’s ineffective as a prophylactic.
As I’ve written about previously, there is strong evidence that a daily vitamin D supplement decreases the relative frequency of viral respiratory infections by almost 50% in people who are vitamin D deficient. And, as mentioned already, there is a lot of observational evidence showing that people who are vitamin D deficient do worse when they get covid. So I think it’s a good idea to take a daily vitamin D supplement. 4,000 IU (100 mcg) is probably a sufficient daily dose to get the full benefit, and is within the bounds of what the FDA considers a completely safe dose to take on a daily basis.