key: cord-0857259-7w52ya3k
authors: Gupta, D.; Menon, S.; Criqui, M. H.; Sun, B. K.
title: Temporal association of reduced serum vitamin D with COVID-19 infection: A single-institution case-control and historical cohort study
date: 2021-06-06
journal: nan
DOI: 10.1101/2021.06.03.21258330
sha: c3f24587f57fc91021cc76be1d0ea73eabaea33c
doc_id: 857259
cord_uid: 7w52ya3k

Objectives: Vitamin D supplementation has been proposed for the prevention and treatment of COVID-19, but the relationship between the two is inconclusive: Reduced serum vitamin D may predispose to COVID-19, but it may also be a secondary consequence of infection. The objective of this study was to assess the temporal association between serum vitamin D levels and COVID-19. Design: A single-institution case-control study and a historical cohort study Setting: University of California San Diego (UCSD) Health System in San Diego, California Participants: Patients testing positive for COVID-19 from January 1, 2020 to September 30, 2020 with serum 25-hydroxy-vitamin D (25(OH)D) measured within 180 days of diagnosis. Patients were separated based on whether 25(OH)D was measured before (n=107; "pre-diagnosis") or after (n=203; "post-diagnosis") COVID-19 diagnosis. Subjects with 25(OH)D assessments prior to COVID-19 diagnosis were evaluated using a case-control study design, while subjects with 25(OH)D measured after COVID-19 diagnosis were analyzed with a historical cohort study design. Primary and Secondary Outcome Measures: In the pre-diagnosis study, a conditional logistic regression was performed using COVID-19 infection status as the binary dependent variable. In the post-diagnosis study, an ordinary least squares regression was performed using serum 25(OH)D levels as the continuous dependent variable. Results: Serum 25(OH)D levels were not associated with the odds of subsequently testing positive for COVID-19 (OR 1.00, 95% CI: 0.98 to 1.02, p = 0.982). However, COVID-19 positive individuals had serum 25(OH)D measurements that were lower by 2.70 ng/mL (95% CI: -5.19 to -0.20, p = 0.034) compared to controls. Conclusions: In our study population, serum 25(OH)D levels were not associated with risk of testing positive for COVID-19 but were reduced in subjects after being diagnosed with COVID-19 infection. These results raise the possibility that reduced serum 25(OH)D may be a consequence and not a cause of COVID-19 infection.

SARS-CoV-2, the coronavirus that causes COVID-19, has claimed over 3.5 million lives globally. Although there is optimism that vaccination efforts will eventually bring the pandemic under control, there has been widespread interest in complementary measures to mitigate the risk of viral infection and mortality. Vitamin D, available as an inexpensive supplement, gained extensive attention for its potential role in the prevention and treatment of COVID-19. However, the relationship between vitamin D and COVID-19 is unclear.

Mechanistically, vitamin D may enhance the immune response to SARS-CoV-2 in several ways. Vitamin D boosts innate immunity by augmenting the expression of human cathelicidin antimicrobial peptide (CAMP) 1 in lung and skin epithelial cells 2 , which has been shown to attenuate the infectivity and viability of viruses 3 . In addition, vitamin D prevents excessive adaptive immune responses 4 . Because systemic inflammatory responses have been associated with respiratory distress and mortality in patients with severe COVID-19 5 , it has been proposed that vitamin D supplementation could mitigate these harmful inflammatory reactions.

for COVID-19 6 , using serum 25-hydroxy-vitamin D (25(OH)D), a standard laboratory serum marker of vitamin D stores, as a primary biomarker for vitamin sufficiency. Recommendations for supplementation have particularly been endorsed for populations with the highest COVID-19 risk who also have higher rates of vitamin D deficiency, including the elderly, individuals with chronic diseases, darker-complected individuals such as African Americans, and those who live at high latitudes 7 .

The studies that have evaluated the correlation between serum vitamin D and COVID- 19 have shown mixed results. Some reports have concluded that there is an association between . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint vitamin D deficiency and increased susceptibility to COVID-19 infection 8, 9 , but others have not 10, 11 . Other studies, focusing on different endpoints, have found that 25(OH)D deficiency is correlated with a higher risk of intensive care unit admission, ventilation dependency, and survival rate [12] [13] [14] [15] [16] . Beneficial outcomes from vitamin D supplementation trials have been reported 17 , although these therapeutic studies have not always been done in randomized groups 18 , and have had modest sample sizes. It has been postulated that any beneficial effect of vitamin D on severe COVID-19 could be masked by the effect of other adjunctive treatments such as dexamethasone 19 . Highly powered, randomized controlled trials will be needed to definitively test for causality. 8, 9 , others relied on measurements taken after diagnosis 12, 13 , and others included both 26 .

Because severe illness itself can cause rapid reduction of serum 25(OH)D 27 , the timing of laboratory measurement is critical: Vitamin D deficiency may predispose to COVID-19, but it is also possible that COVID-19 infection can reduce 25(OH)D as well.

In this report, we sought to address these methodological differences in examining the temporal correlation between serum 25(OH)D and COVID-19 infection. We performed two complementary single-center studies examining patients testing positive for COVID-19 between . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint January 1, 2020 to September 30, 2020 in the University of California San Diego (UCSD) Health system who had a serum 25(OH)D assessment within 180 days of diagnosis. These dates capture the vast majority of COVID-19 cases in the UCSD Health system from the onset of the COVID-19 pandemic until the initiation of vaccinations.

We evaluated the study populations with complementary approaches: In the first study, we used a case-control design to compare 25(OH)D levels drawn prior to COVID-19 positive diagnosis against COVID-19 negative controls matched by age, sex, body mass index (BMI), diabetes, hypertension, time from vitamin D draw, and season that the 25(OH)D test was performed. In the second study, we applied a cohort study design to assess serum 25(OH)D levels drawn after COVID-19 diagnosis, applying the same matching criteria. Finally, we performed a subgroup analysis of the second study to specifically examine COVID-19 patients whose disease was severe enough to require hospitalization, comparing them against a matched hospitalized cohort that was COVID-19 negative.

Our overarching approach was based on the reasoning that if vitamin D deficiency . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint

Baseline characteristics of the study cohorts are shown in Table 1 . Covariates consisted of established risk factors for COVID-19 susceptibility 28 that included age, sex, obesity, and medical comorbidities. In addition, due to seasonal variation in serum vitamin D levels, control subjects were matched by meteorological season that serum vitamin D was measured, as well as to -0.20, p = 0.03) ( Table 2 ). These data indicated that COVID-19 positive subjects showed a significant reduction in 25(OH)D compared to matched COVID-19 negative subjects.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint

To assess for potential correlation between serum 25(OH)D and severe COVID-19 infection, a subgroup analysis was performed to compare COVID-19 subjects requiring hospitalization against a COVID-negative hospitalized control group. To promote matching of patients with comparable disease severity, control patients were additionally matched by length of hospital stay, an indicator of disease severity 29 . Baseline characteristics of the matched analysis for the hospitalized cohort are shown in Table 3 . No significant differences were observed for matching characteristics between cases and controls. Table 4 ). These data indicated that COVID-19 positive hospitalized subjects showed a significant reduction in 25(OH)D compared to COVID-19 negative hospitalized subjects.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. infection 30 . To examine this possibility, we also stratified our pre-diagnosis subjects into these categories. We found no statistically significant association with increased odds of COVID-19 with either vitamin D insufficiency or deficiency compared to the sufficient (>30 ng/mL) reference group (Supplemental Table 1 ). Viewed together, our results are consistent with other reports that identified lower 25(OH)D in association with COVID-19 8, 9, [12] [13] [14] 16 , but raise the possibility that lower 25(OH)D levels may be an outcome of COVID-19 infection rather than a cause of it.

These findings are consistent with prior observations that have examined the relationship between vitamin D and other respiratory illnesses. While vitamin D deficiency has been associated with increased risk of respiratory infections 31 , acute illness itself can also reduce serum 25(OH)D through fluid shifts, depletion of serum binding proteins, and renal wasting 27 .

Acute inflammation following surgery has been associated with a reduction of serum 25(OH)D within 48 hours 32 . Inflammatory mediators increase the activity of CYP24A1 and CYP27B1, . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint enzymes that metabolize vitamin D pathway compounds 33 . Vitamin D metabolism is dysregulated in patients with asthma and chronic obstructive pulmonary diseases, leading some investigators to suggest that the relationship between airway inflammation and vitamin D deficiency is bidirectional 34 . These findings support the biological plausibility that elevated rates of vitamin D deficiency observed in subjects infected by COVID-19 could be, at least in part, secondary to the infection itself 35 .

Our study has several strengths. First, our data from a single institution directly assessed 25(OH)D and COVID-19 laboratory results using uniform, standardized assays, minimizing the significant variations that can occur between different testing methodologies 36 Our study also has several limitations. The retrospective design does not allow for determination of causality, and our modest sample size is not powered to detect smaller but potentially significant correlations. Additionally, although we matched baseline characteristics between cases and controls, other unaccounted confounders could have affected the results. In particular, data on the racial identity of our study subjects was incomplete, which did not allow us to include this factor in cohort matching. Racial disparities in COVID-19 illness have been clearly established, with one retrospective study finding that a positive COVID-19 test was significantly associated with lower vitamin D levels in Black individuals, but not with lower . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint vitamin D levels in White individuals 38 . Our health system's general catchment area (San Diego county) has a lower Black population (~5.5%) than the U.S. national average (~13.4%). Therefore, the results from our analysis may differ from other study populations, and do not argue against the potential utility of Vitamin D supplementation for specific populations. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint

The UCSD Institutional Review Board approved this study and approved a waiver for . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint Analyses were performed using RStudio software, version 1.3.959. Continuous covariates are reported as mean ± standard deviation (SD) and compared using unpaired t-tests; categorical variables are reported as numbers and percentages and compared using chi-square tests. The significance level for all analyses was set to a two-sided p-value <0.05.

Subjects or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this study.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint

This study was reviewed and approved by the UC San Diego Institutional Review Board (protocol #200768).

The datasets for this study are not publicly published due to presence of potential patientidentifiable information but will be made available from the corresponding author on reasonable request.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint Table 1 . Characteristics of COVID-19 study population and controls. Continuous covariates are reported as mean ± standard deviation (SD) and compared using unpaired t-tests. Categorical variables are reported as numbers (no.) and percentages (%) and compared using Chi-square tests. A p-value < 0.05 was considered a significant difference for covariates. Season dates were defined by meteorological seasons within the study interval.

Characteristic In the pre-diagnosis study, a conditional logistic regression was performed in which the independent variable is continuous (i.e., serum 25(OH)D level) and the dependent variable is binary (i.e., COVID-19 infection status). In the post-diagnosis study, an ordinary least squares regression was performed, in which the independent variable is binary (i.e., COVID-19 infection status) and the dependent variable is continuous (i.e., serum 25(OH)D level).

Abbreviations: CI = confidence interval; 25(OH)D = 25-hydroxy vitamin D *significant at p < 0.05 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint Table 3 . Characteristics of hospitalized COVID-19 study population and controls. Continuous covariates are reported as mean ± standard deviation (SD) and compared using unpaired t-tests. Categorical variables are reported as numbers (no.) and percentages (%) and compared using chi-square tests. A p-value less than 0.05 was considered a significant difference for covariates. Season dates were defined by meteorological seasons within the study interval. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 6, 2021. Table 4 . Association of serum 25-hydroxy vitamin D with severe COVID-19 infection. Serum 25(OH)D for cases and controls are reported as mean levels in ng/mL ± standard deviation (SD). An ordinary least squares regression was performed in which the independent variable is binary (i.e., hospitalization due to COVID-19 or hospitalization due to another cause) and the dependent variable is continuous (i.e., serum 25(OH)D level).

Abbreviations: CI = confidence interval; 25(OH)D = 25-hydroxy vitamin D *significant at p < 0.05 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint

Mechanisms Underlying the Regulation of Innate and Adaptive Immunity by Vitamin D

Vitamin D, respiratory infections, and asthma

Antiviral Activity and Increased Host Defense against Influenza Infection Elicited by the Human Cathelicidin LL-37

Vitamin D and the Immune System

Is a "Cytokine Storm" Relevant to COVID-19?

Evidence that Vitamin D Supplementation Could Reduce Risk of Influenza and COVID-19 Infections and Deaths

Geographic location and vitamin D synthesis

Association of Vitamin D Status and Other Clinical Characteristics With COVID-19 Test Results

Low plasma 25(OH) vitamin D level is associated with increased risk of COVID-19 infection: an Israeli population-based study

Vitamin D concentrations and COVID-19 infection in UK Biobank

Vitamin D Deficiency and Outcome of COVID-19 Patients

Vitamin D sufficiency, a serum 25-hydroxyvitamin D at least 30 ng/mL reduced risk for adverse clinical outcomes in patients with COVID-19 infection

Interaction between age and vitamin D deficiency in severe COVID-19 infection

Does Serum Vitamin D Level Affect COVID-19 Infection and Its Severity?-A Case-Control Study

Low serum 25-hydroxyvitamin D (25[OH]D) levels in patients hospitalized with COVID-19 are associated with greater disease severity

Effect of calcifediol treatment and best available therapy versus best available therapy on intensive care unit admission and mortality among patients hospitalized for COVID-19: A pilot randomized clinical study

Vitamin D and survival in COVID-19 patients: A quasi-experimental study

Vitamin D for COVID-19: a case to answer?

The role of vitamin D in the prevention of coronavirus disease 2019 infection and mortality

COVID-19 mortality increases with northerly latitude after adjustment for age suggesting a link with ultraviolet and vitamin D

Vitamin D (25OHD) Serum Seasonality in the United States

Intraindividual Variation in Plasma 25-Hydroxyvitamin D Measures 5

Years Apart among Postmenopausal Women

Seasonal variations in serum 25-hydroxy vitamin D levels in a Swedish cohort

25-Hydroxyvitamin D Concentrations Are Lower in Patients with Positive PCR for SARS-CoV-2

Vitamin D in acute stress and critical illness: Current Opinion in

Covid-19: risk factors for severe disease and death

The relationship between a severity of illness indicator and mortality and length-of-stay

Sorting Out Whether Vitamin D Deficiency Raises COVID-19 Risk

International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted

Vitamin D: a negative acute phase reactant

Allergen specific immunotherapy enhanced defense against bacteria via TGF-&#x03B2;1-induced CYP27B1 in asthma

Vitamin D Metabolism Is Dysregulated in Asthma and Chronic Obstructive Pulmonary Disease

Letter to the Editor: Vitamin D deficiency in COVID-19: Mixing up cause and consequence

& Ausimmune Investigator Group. Variability in vitamin D assays impairs clinical assessment of vitamin D status: Variability in vitamin D assays

Vitamin D status as a predictor of Covid-19 risk in Black, Asian and other ethnic minority groups in the UK

Association of Vitamin D Levels, Race/Ethnicity, and Clinical Characteristics With COVID-19 Test Results

Effect of a Single High Dose of Vitamin D 3 on Hospital Length of Stay in Patients With Moderate to Severe COVID-19: A Randomized Clinical Trial

International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 6, 2021. ; https://doi.org/10.1101/2021.06.03.21258330 doi: medRxiv preprint