Search strategy

The evidence review was developed using NICE interim process and methods for guidelines developed in response to health and social care emergencies.

The original NICE recommendations on the use of vitamin D supplementation for the treatment of COVID-19 were published in December 2020 (NG187). The following recommendation was made regarding the efficacy of vitamin D for the treatment of COVID-19:

• Do not offer a vitamin D supplement to people solely to treat COVID-19, except as part of a clinical trial.

Continuous surveillance of the literature using a broad COVID-19-wide search was undertaken weekly. The results were processed on a weekly basis and all records with relevance to COVID-19 and vitamin D, and which were added since the original review search was conducted, were brought together into a single group. Automated pattern matching was applied to the records in this group in order to further focus the results and identify the most relevant records for screening. The pattern matching code was created in Python using keywords relevant for the topic to enable automated study categorisation through natural language processing (see appendix B for full details).

All studies included by the pattern matching code were considered for inclusion in this review.

As quality assurance, the pattern matching code was run on all studies which were identified in the continuous surveillance, but had been excluded as not relevant to NICE’s current reviews. No additional studies were identified through these checks.

Review protocol

A review protocol was developed by NICE for the effectiveness of vitamin D supplementation for treating COVID-19 (NG187) (published in December 2020). As part of the current update to this guideline, the existing review protocol was updated to include the following subgroups:

(i)

baseline 25(OH)D levels

(ii)

dosage >800 IU/day versus ≤800IU/day

(iii)

single versus multiple doses

The updated protocol is presented in appendix A.

Review question: What is the effectiveness d safety of vitamin D supplementation for the treatment of COVID-19 in adults, young people and children?

There remains a high degree of uncertainty over whether vitamin D supplementation is more effective than placebo plus standard care or standard care for treating COVID-19. But, there is currently no clear evidence of benefit.

What is the evidence informing this conclusion?

This is an update to the December 2020 review. During this update, we have added 6 extra studies (Cannata-Andia 2022, Maghbooli 2021, Mariani 2022, Murai 2021, Rastogi 2020, Sanchez-Zuno 2021) and retained the existing study (Entrenas Castillo 2020).

Evidence comes from 7 randomised controlled trials that compared vitamin D supplementation with standard care (Cannata-Andia 2022, Entrenas-Castillo 2020, Maghbooli 2021, Rastogi 2020) or standard care plus placebo (Murai 2021, Mariani 2022 and Sanchez-Zuno 2021) in 1,262 people with COVID-19. Studies were conducted in Argentina, Brazil, Chile, Guatemala, India, Iran Mexico and Switzerland. All the studies were conducted in a hospital setting, regardless of disease severity.

All the trials included vitamin D doses higher than 800 IU/day. Three studies used a single-dose intervention of vitamin D (Cannata-Andia 2022, Mariani 2022 and Murai 2021). The remaining studies used a multiple-dose regimen (Entrenas-Castillo 2020, Maghbooli 2021, Rastogi 2020 and Sanchez-Zuno 2021). Two studies included only participants with vitamin D deficiency at baseline, defined as <75nmol/L (Maghbooli 20210) and <50nmol/L (Rastogi 2020) in the studies. It should be noted that these levels are not usually considered deficient in the UK.

Due to variability in dosage, disease severity and baseline vitamin D status, subgroup analyses were carried out to measure the effects of vitamin D supplementation.

Publication status

All studies are peer-reviewed manuscripts.

Study characteristics

The mean or median age in the studies ranged between 43 and 58 years and the proportion of men ranged between 50% and 65%. The severity of COVID-19 in 3 of the studies was reported as moderate-severe (Cannata-Andia 2022, Maghbooli 2021 and Murai 2021), mild to moderate in Mariani (2022) and asymptomatic-mildly symptomatic in Rastogi (2020) and Sanchez-Zuno (2021). Disease severity in one study was not clear (Entrenas-Castillo 2020).

The dose and duration of vitamin D supplementation varied across the trials. Three studies included a single dose of vitamin D of 100,000 IU (Cannata-Andia 2022), 200,000 IU (Murai 2021) or 500,000 IU (Mariani 2022). The remaining studies used multiple doses between 1000 IU/day and 60,000 IU/day for a duration ranging from 7 to 60 days.

Due to the variability in dosage, disease severity and baseline vitamin D concentration, subgroup analyses were conducted where the data allowed.

Children under 18 were excluded from the trials.

What are the main results?

Vitamin D supplementation does not result in statistically significant differences in mortality, ICU admission, requirement for oxygen therapy, adverse events, symptoms at day 7 or 14, PCR test results at day 7 or 14, or duration of: ICU stay, hospitalisation or mechanical ventilation.

Data on adverse events (for example cardiovascular or gastrointestinal serious adverse events) was only reported in one study (Mariani 2022).

The evidence suggests that, compared with control groups, there was a non-statistically significant reduction in mechanical ventilation and an increase in negative COVID-19 test results within 3 weeks in the vitamin D group.

Our confidence in the results

Overall, the studies are heterogeneous with both clinical and methodological diversity. Most studies were assessed to have some concerns with risk of bias due to insufficient reporting around randomisation and allocation and a lack of blinding in studies which did not use placebo.

Other reasons for downgrading evidence included inconsistency (for example, when point estimates varied widely between studies) and imprecision (with outcomes rated as having serious imprecision when the confidence interval crossed the line of no effect and outcomes further downgraded as having very serious imprecision when fewer than 300 people contributed to the outcome). The variance in the duration of symptoms prior to randomisation across the studies may impact the certainty of outcomes as well as the effect of standard care regimens. The certainty of the evidence was moderate to very low for most outcomes.

Benefits and harms

All included trials compared the effects of vitamin D for treating COVID-19 with placebo or standard care. The trials included diverse populations with different COVID-19 severity, and used varying dosages of vitamin D. All trials compared the effects of vitamin D3 and not D2 for treating COVID-19.

The evidence shows no statistically significant difference in mortality, admissions to intensive care, hospitalisation or oxygen therapy in people having vitamin D compared with standard care or placebo. The panel noted that the certainty of the evidence for mortality and oxygen therapy was moderate, although certainty was lower for the other outcomes. The panel noted a non-statistically significant reduction in progression to mechanical ventilation (relative risk 0.55, 95% confidence interval 0.31 to 1.00) (Mariani 2022, Murai 2021 and Maghbooli 2021). While this result was at low risk for bias, the panel did not consider the result to be certain enough to justify recommending vitamin D.

The panel noted no statistically significant difference in presence of symptoms or positive polymerase chain reaction tests at 7 and 14 days. But in 1 study with 40 people there was increased SARS-CoV-2 negativity within 3 weeks (Rastogi 2020). This outcome was downgraded because of very serious risk of bias, and the panel did not think it represented a meaningful benefit, or that it could be attributed to treatment with vitamin D.

The panel were also presented with evidence on subgroup effects of vitamin D. These analyses show no statistically significant effect on admission to intensive care or mechanical ventilation of different numbers of doses (single or multiple), baseline vitamin D levels or COVID-19 severity. A meta-analysis of 5 studies (Cannata-Andia 2022, Mariani 2022, Murai 2021, Entrenas Castillo 2020, Maghbooli 2021) suggested an association between multiple doses of vitamin D and reduced mortality compared with single doses. But the difference was not statistically significant.

The panel had concerns about adverse events with the high doses of vitamin D used in the studies. Only 1 study reported adverse events as an outcome but showed no statistically significant difference between vitamin D and placebo. The panel acknowledged that there are some well-known adverse effects of vitamin D overdose, including raised concentrations of calcium and phosphate in plasma and urine, and nausea and vomiting (for more details, please see the BNF). They agreed that treating COVID-19 with vitamin D at the dosages used in the included studies could have potential harms, and that more research is needed in this area.

The panel noted that the study populations did not include pregnant women or older populations who may be at more risk of severe COVID-19 outcomes. They also did not include children and young people under 18 years. So, they agreed that more research is also needed in the area.

Certainty of the evidence

Evidence comes from 7 randomised controlled trials (Cannata-Andia 2022, Entrenas Castillo 2020, Maghbooli 2021, Mariani 2022, Murai 2021, Rastogi 2020 and Sanchez-Zuno 2021).

All of the outcomes from the trials were rated as moderate to very low certainty, but the panel agreed that there was no clear evidence of benefit.

Outcomes from Maghbooli 2021, Mariani 2022 and Murai 2021 were assessed to be at low risk of bias. Outcomes from Cannata-Andia 2022 and Entrenas Castillo 2020 were downgraded because of insufficient detail on randomisation or the allocation process. Outcomes from Rastogi 2020 and Sanchez-Zuno 2021 were rated as at high risk of bias because neither trial reported details of study design and analysis plans.

The panel decided not to downgrade for indirectness despite reported vitamin D deficiency in the studies. This was because the deficiency threshold for vitamin D reported in the studies differed from the UK threshold. Vitamin D is a negative acute-phase reactant. This means its serum concentration falls during a systemic inflammatory response, which may occur with severe COVID-19. But, the panel noted that the potential mechanism of action of vitamin D in the context of COVID-19 is unknown. The effect of deficiency on COVID-19 outcomes is also unclear.

Overall, the panel noted that the evidence comes from diverse populations with varying care regimens and vitamin D doses. As such, evidence from the trials may not be generalisable to the UK.

The panel discussed the populations in different ongoing trials and noted that more research into specific subgroups of interest (for example, older people, children and pregnant women) could help determine the effects of vitamin D.

Values and preferences

The panel discussed that vitamin D3 (colecalciferol) supplements can be derived from an animal source. They noted that care providers need to consider people's concerns about using animal products because of a religious or ethical belief when they are discussing vitamin D products and their provision.

Resources

Vitamin D supplements are widely available in the NHS and public stores.

The panel discussed any resource implications that vitamin D provision for treating COVID-19 might have. They recognised that the doses from the trials are substantially higher than those used for daily supplementation in the UK. They agreed that, because vitamin D is not being recommended for this, there would be no resource implications.

Equity

The panel recognised the existing inequalities in vitamin D status, including those relating to location, health and family background. They noted that people who have dark skin and people who have low or no sunlight exposure, including people who spend more time indoors because of frailty or disability, are more likely to have vitamin D deficiency. They also acknowledged that people with dark skin are at greater risk of infection by SARS-CoV-2.

The panel also discussed that there is a lack of evidence on the effectiveness of vitamin D for COVID-19 in children, older people and when pregnant or breastfeeding, so the effect of treatment in these groups cannot be determined.

Acceptability

The acceptability of vitamin D for treating COVID-19 has not formed part of the evidence review. The panel believed that the lack of effect against COVID-19 shown in the studies and the limited information about adverse events may reduce the acceptability of vitamin D for treating COVID-19.

Feasibility

Vitamin D does not have marketing authorisation for the treatment of COVID-19.

PICO table

Review question: What is the effectiveness and safety of vitamin D supplementation for the treatment of COVID-19 in adults, young people and children? (PDF, 135K)