Does D Stand for Decrease Deterioration + Death? – The Jury is Out for Vitamin D in COVID

Update: This article was removed from the Lancet server on February 19th, 2021 (Link)

Background: Vitamin D deficiency is common amongst critically ill patients and there has been ample speculation about a possible role for supplementation (with calcifediol) in the treatment of patients with COVID19. Calcifediol is cheap, readily available and, has minimal side effects making it an ideal therapeutic – if it works. To date, there is limited high-quality evidence on efficacy.

Paper: Nogues X et al. Calcifediol treatment and COVID-19-related outcomes (Barna-COVIDIOL). Lancet Preprints (22, January 2021). [Link is HERE]

ADDENDUM 02/20/2021:

• This paper has been removed from the Lancet preprint server.  As in our analysis, the authors call this a randomized clinical trial, however it is more an observational cohort study.  For more info checkout Retraction Watch [Link is HERE]

Clinical Question: Does treating patients admitted to general hospital wards with calcifediol reduce the risk of decompensation and/or death?

Population: Patients > 18 years of age who tested PCR positive for SARS-CoV-2 and were admitted to a single hospital in Barcelona, Spain.

Intervention: Calcifediol (266 mcg/capsule):

• Day 0: 532 ug (2 capsules); 
• Day 3, 7, 15, & 30: 266 ug (1 capsule) along with standard treatment (HCQ, azithromycin + ceftriaxone if bacterial superinfection was suspected)

Control: Standard treatment only (no placebo)

Outcome (Primary): Admission to the ICU or death (no time point specified in manuscript)

Outcomes (Secondary): None specified in this manuscript

Design: Prospective, observational, open-label trial. Eight wards were randomized to treating patients with calcifediol or not (individual patients were not randomized)

Exclusion: Not detailed in this manuscript. Presumed trial registration specifies already being treated with calcifediol or cholecalciferol, intolerance/allergy to calcifediol and pregnancy

Results:

Primary Results:

• Study enrolled 930 patients
  • Calcifediol treated: n = 551
  • No calcifediol: n = 379
• Crossed over: 50 patients in the non-treated arm crossed over to treatment
• > 85% of patients were white
• >55% of patients were men

Critical Results

*Intention-to-treat analysis

• ICU Admission (with logistic regression)
  • RR 0.18 (95% CI 0.11 – 0.29)
• Deaths (with logistic regression)
  • RR 0.65 (95% CI 0.34 – 1.18)

Strengths:

• Study asks a clinically important question
• 1st study to evaluate the effects of calcifediol supplementation on COVID-19 related mortality
• Largest study to evaluate vitamin D supplementation on ICU admission
• Patients in two groups were similar in terms of measured characteristics except for baseline vitamin D level

Limitations:

• Non-consecutive patient enrollment (convenience sample of patients)
• Wards were randomized not patients. Patients were not randomized to which ward they went to. Essentially, this is a prospective, observational study
• Clinicians and patients were unblinded to treatment
• Patients given calcifediol had statistically higher baseline vitamin D levels
• No information on why patients were sent from the ward to the ICU
• >85% of patients enrolled were white limiting generalizability to other groups
• Single center study limiting generalizability
• No information on escalation of therapy or criteria for admission to ICU
• The investigators do not report any information about side effects

Discussion:

• Prospective, observational study
  • The authors state this is a randomized trial but it is not
  • Wards were randomized to calcifediol treatment or non-treatment. Individual patients were not randomized to treatment or to which ward they were placed in
  • This introduces significant selection bias to the study as it’s unclear what criteria were used (if any) to decide which patients went to which ward
  • Clinicians could have cherry picked patients for the treatment arm (ie less sick patients sent to calcifediol wards)
• Open-label
  • Introduces significant bias particularly when one of the primary endpoints is subjective
    • While admission to ICU is a yes/no question, decision to upgrade to the ICU is subjective and no information is provided as to why that decision was made
  • Hawthorne effect likely to have affected outcomes in the calcifediol wards
• Patients sent to wards where calcifediol was not part of treatment had lower vitamin D levels at baseline. 
  • Authors state patients with higher baseline 25(OH)D were significantly more likely to have decreased mortality RR 0.40; 95% CI 0.24 to 0.67
  • This is likely a signal that these patients were sicker on admission to the hospital
• The manuscript appears to be missing critical information including secondary outcomes, exclusion criteria and complete primary outcome information (ie how many days were patients followed for outcome). This clinicaltrials.gov number appears to be for this study but, we are not clear on this point (https://clinicaltrials.gov/ct2/show/NCT04366908)

Author Conclusion: “In patients hospitalized with COVID-19, calcifediol treatment at the time of hospitalization significantly reduced ICU admission and mortality.”

Our Conclusions: This poorly done study adds little in the way of meaningful evidence of benefit. There is an association between calcifediol treatment and reduced ICU admission and mortality. This weak-methodological study cannot show causality and even the association seen is heavily influenced by significant biases.

Bottom Line: While calcifediol is an inexpensive and safe medication, there is no high-quality evidence that it improves outcomes in COVID-19.

References:

1. Nogues X et al. Calcifediol treatment and COVID-19-related outcomes (Barna-COVIDIOL). Lancet Preprints (22, January 2021). [Link is HERE]