Exploring the link between vitamin D and clinical outcomes in COVID-19.

The immunomodulating role of vitamin D might play a role in COVID-19 disease. We studied the association between vitamin D and clinical outcomes in COVID-19 patients. This is a retrospective cohort study on COVID-19 patients with documented vitamin D levels within the last year. Vitamin D levels were grouped as ≥ 20 ng/mL or < 20 ng/mL. Main outcomes were mortality, need for mechanical ventilation, new DVT or pulmonary embolism, and ICU admission. A total of 270 patients (mean ± SD) age, 63.81 (14.69) years); 117 (43.3%) males; 216 (80%) Blacks; 139 (51.5%) in 65 and older age group were included. Vitamin D levels were less than 20 ng/mL in 95 (35.2%) patients. During admission, 72 patients (26.7%) died, 59 (21.9%) needed mechanical ventilation, and 87 (32.2%) required ICU. Vitamin D levels showed no significant association with mortality (OR = 0.69; 95% CI, 0.39-1.24; P = 0.21), need for mechanical ventilation (OR = 1.23; 95% CI, 0.68-2.24; P = 0.49), new DVT or PE(OR= 0.92; 95% CI, 0.16-5.11; P = 1.00) or ICU admission (OR = 1.38; 95% CI, 0.81-2.34; P = 0.23). We did not find any significant association of vitamin D levels with mortality, the need for mechanical ventilation, ICU admission and the development of thromboembolism in COVID-19 patients.NEW & NOTEWORTHY Low vitamin D has been associated with increased frequency and severity of respiratory tract infections in the past. Current literature linking clinical outcomes in COVID-19 with low vitamin D is debatable. This study evaluated the role of vitamin D in severe disease outcomes among COVID-19 patients and found no association of vitamin D levels with mortality, the need for mechanical ventilation, ICU admission, and thromboembolism in COVID-19.

INTRODUCTION

Coronavirus disease (COVID-19) originated in Wuhan, China, and has now become a pandemic resulting in 926,000 deaths worldwide as of mid-September 2020. The lack of evidence-based information and highly variable clinical presentation of individuals infected with this novel virus has perplexed clinicians worldwide. Elderly patients, especially those with underlying comorbidities, are at a higher risk for severe infection and worse clinical outcomes (1, 2). Prior studies point that 25-hydroxyvitamin D plays a role in immune regulation and induction of antimicrobial peptides to both viral and bacterial infections (3–6). Review of literature points towards the association of low levels of vitamin D and increased frequency and susceptibility to acute respiratory tract infections including COPD exacerbation (7–10). Many reports also suggest that vitamin D supplementation reduces the risk of respiratory tract infections, decreases symptom duration, and length of stay in hospitalized patients (11–14).

Vitamin D has immunomodulating properties (3) and acts at various levels, i.e., maintains cellular junctions (15), enhances innate immunity (16), induces antimicrobial peptides (cathelicidins and defensins) (5, 17), which lowers viral replication, decreases proinflammatory Th1 cytokines (18–20), increases anti-inflammatory cytokines (21), and modulates adaptive immunity (22). Cytokine release syndrome has been reported in some of the critically ill COVID-19 patients (23, 24), and given its immunomodulating properties, one can hypothesize that vitamin D levels might have a role in this syndrome. Vitamin D has been documented to have antioxidant (25) and antifibrotic properties (26), and modulate renin-angiotensin-aldosterone-system (RAAS) and angiotensin converting enzyme-2 (ACE2) expression (27). A leading cause of mortality in COVID-19 patients has been reported to be acute respiratory distress syndrome (ARDS) (28). Vitamin D has been shown to reduce lung permeability in ARDS and regenerate lung lining (29). This led us to explore a possible correlation between vitamin D levels and clinical outcomes in COVID-19 patients. Interestingly, vitamin D deficiency is very prevalent in the United States and ∼41% of the adult US population has inadequate vitamin D levels (30).

A high incidence of thrombotic complications has been reported in COVID-19 patients who need intensive care (31). There have been conflicting results reported by studies on the role of vitamin D and venous thromboembolism. Some studies have documented that 25-hydroxyvitamin D plays a critical role in the pathogenesis of deep vein thrombosis (DVT) (32) and reported an association between decreased vitamin D levels and increased risk of venous thromboembolism (33). However, other studies have refuted any such correlation (34).

The immunomodulating role of vitamin D may play a role in COVID-19 disease progression and there is a paucity of literature on the role of vitamin D in COVID-19. The main objective of this study is to understand the association between vitamin D levels and mortality among COVID-19 patients. Our study also explores if vitamin D levels have any association with other clinical outcomes such as the need for mechanical ventilation, development of new DVT or pulmonary embolism (PE), and intensive care unit (ICU) requirement in COVID-19 patients.

METHODS

Study Design

We conducted a retrospective cohort study on 2001 adult patients with a confirmed COVID-19 diagnosis. The study was exempt by the Detroit Medical Center (DMC) and Wayne State University Institutional Review Board (IRB application #20-06-2422). No external funding was received for conducting the study.

Study Site and Patient Population

Adult patients (≥18 years of age) with a confirmed COVID-19 diagnosis (either via nasopharyngeal or oropharyngeal swab) were included. Testing for COVID-19 was done at DMC, one of the largest academic medical centers and healthcare providers in southeast Michigan. DMC comprises four distinct hospitals in Michigan and all four hospital locations were included in the study. These hospitals primarily serve the Detroit metropolitan area.

Data Collection

A list of 2001 patients who visited DMC between March 10, 2020, and June 30, 2020, with a laboratory-confirmed COVID-19 PCR diagnosis was collected in collaboration with institutional information technology services. Patients under the age of 18, any readmission during the time frame, and pregnant patients were excluded from the study. A total of 67 patients were excluded initially as they met the above criteria. We reviewed 1,944 electronic medical records to screen if these patients had a previously documented vitamin D, 25-OH level within the past 12 months. After the initial screen, there were a total of 277 patients with a documented vitamin D level within the past 12 months. However, 7 of these patients were excluded from the study as they presented only for ambulatory surgery leaving a total of 270 patients, who were included in the study (Fig. 1). We then classified the patients based on their vitamin D levels into 2 groups as ≥ 20 ng/mL (patients with normal vitamin D levels) and < 20 ng/mL (patients with low vitamin D levels). Data points were manually collected and coded for each patient. Data regarding the prescription of vitamin D supplements (weekly/daily) were also collected. For additional analysis based upon stratified vitamin D levels, patients with normal vitamin D levels were further divided into two subgroups, patients with vitamin D level 20–30 ng/mL and patients with vitamin D level > 30 ng/mL.

Figure 1.

Flowchart depicting patient inclusion criteria. Adult patients (≥ 18 years of age) with a confirmed COVID-19 diagnosis and a documented vitamin D level in the past 12 months were included. Patients under the age of 18, any readmission during the time frame, ambulatory surgery and pregnant patients were excluded from the study.

Outcomes

The main outcomes for this study were mortality, the need for mechanical ventilation, new DVT or PE during hospitalization, and ICU admission among COVID-19 patients. All of the patients included in the study had a documented outcome (mortality/discharged status) at the time of data collection. Additionally, the number of prior comorbidities, BMI, disposition upon emergency department (ED) visit (discharge home, inpatient admission, and direct ICU admission), and maximum oxygen requirement during admission were collected. Charts were screened to determine if the patient required transfer to ICU from inpatient floors. Demographic data collected included age, sex, and race.

Statistical Analysis

Categorical variables have been described as frequency and percentages. We categorized age into two groups (18–64 years, and 65 and older). A crude relative association measure (odds ratio, OR) was calculated for each correlation using the Pearson chi-square and Fisher test. An adjusted odds ratio was calculated using binary logistic regression. We adjusted for age, sex, BMI, and presence of comorbidities. Age and BMI were taken as continuous variables and the presence of comorbidities as a categorical variable for the adjusted model. Subgroup analyses were done based on sex and age groups as defined earlier. Subgroup analysis based on race was limited to Blacks and Whites due to the limited sample size of other races. The 95% confidence intervals (CI) were estimated using a binomial distribution. A P value of less than 0.05 was determined to be significant. Bonferroni correction was used to protect from the inflated type 1 error while performing multiple analyses. Additional analyses were performed on the stratified vitamin D levels (< 20 ng/mL, 20–30 ng/mL and > 30 ng/mL). Among the patients with low vitamin D levels, further comparison was made between the patients who were prescribed vitamin D supplements and those who were not prescribed any vitamin D supplements. Statistical analyses were completed using IBM SPSS Statistics software (version 26).

RESULTS

Baseline Characteristics

There were 2001 patient records with positive COVID-19 test at the 4 DMC hospitals with a nasopharyngeal/oropharyngeal PCR swab between March 10, 2020, and June 30, 2020. Based on the exclusion criteria, only 270 patients were included in the study. In the cohort analysis, there were 117 males (43.3%) and 153 females (56.7%). The mean age of patients was 63.81 years (mean ± SD, 14.69). More than half of the patients (n = 139, 51.5%) were in the 65 and older age group, with Blacks being the predominant race (n = 216, 80%). Distribution of vitamin D levels showed that more than one-third of the patients had levels less than 20 ng/mL (n = 95, 35.2%). Among the patients with low vitamin D levels, only 27.4% (n = 26) were prescribed vitamin D supplements. About 70% of patients had three or more comorbid diseases (n = 187, 69.3%). The mean BMI of patients was 32.09 (mean ± SD, 9.12), and more than 50% of patients (n = 139) were in the obese category as per the World Health Organization criteria. The baseline characteristics of the population included are detailed in Table 1.

Table 1.

Baseline characteristic of patients

Characteristic	Cohort (n = 270)
Age group, n (%)
18-30 years	5 (1.9)
31-45 years	26 (9.6)
46-64 years	100 (37)
65+ years	139 (51.5)
Sex, n (%)
Male	117 (43.3)
Female	153 (56.7)
Race/ethnicity, n (%)
Blacks	216 (80)
Whites	48 (17.8)
Asian	3 (1.1)
Middle Eastern	3 (1.1)
Number of comorbidities, n (%)
0	14 (5.2)
1	30 (11.1)
2	39 (14.4)
3 or 3+	187 (69.3)
Vitamin D levels, n (%)
≥20 ng/mL	175 (64.8)
<20 ng/mL	95 (35.2)
BMI categories, n (%)
Underweight (BMI < 18.5)	5 (1.9)
Normal (18.5 to < 25)	52 (19.3)
Overweight (25 to < 30)	74 (27.4)
Obese (> 30)	139 (51.5)
Vitamin D supplementation, n (%)
≥20 ng/mL	58 (33.1)
<20 ng/mL	26 (27.4)

n = Number of patients. BMI, body mass index.

Clinical Course

The total mortality in this cohort was 26.7% (n = 72). About 14.8% (n = 40) of the patients were admitted straight to ICU from the ED. An additional 47 patients were later transferred to ICU from the inpatient service. Approximately one in every three patients in this study (n = 87, 32.2%) who came to ED ended up requiring ICU. Around 3.7% of the total patients were sent home from ED (n = 10), whereas 81.5% (n = 220) were admitted to the inpatient service. Close to 81% (n = 219) of patients required supplemental oxygen during their admission stay and 21.9% (n = 59) required mechanical ventilation. About 2.2% of the patients (n = 6) developed new DVT or PE during their hospitalization. The clinical course of the patient population is further detailed in Table 2.

Table 2.

Admission characteristics of patients, n (%)

Mortality	72 (26.7)
Mechanical ventilation	59 (21.9)
ICU admission	87 (32.2)
Admission disposition
ER visit only (discharged from ER)	10 (3.7)
Inpatient admission	220 (81.5)
Direct ER to ICU admission	40 (14.8)
Maximum supplemental oxygen during admission
Room air only	51 (18.9)
Nasal canula	100 (37)
Venti-mask	15 (5.6)
Non-rebreather	37 (13.7)
High flow oxygen	7 (2.6)
BPAP/CPAP	1 (0.4)
Mechanical ventilation	59 (21.9)
New DVT or PE	6 (2.2)

n = Number of patients. BPAP, bilevel positive airway pressure; CPAP, continuous positive airway pressure.

Vitamin D and Mortality

In the cohort analysis vitamin D levels showed no significant association with mortality (OR = 0.69; 95% CI, 0.39–1.24; P = 0.21). No correlation between mortality and vitamin D levels was seen in either males (OR = 1.10; 95% CI, 0.46–2.63; P = 0.83) or females (OR = 0.49; 95% CI, 0.22–1.09; P = 0.08). With subgroup analysis based on age groups and race, no significant association was found between vitamin D levels and mortality in patients less than 65 years old (OR = 0.90; 95% CI, 0.37–2.18; P = 0.81), or in patients 65 years and older (OR = 0.83; 95% CI, 0.35–1.92; P = 0.66), Blacks (OR = 0.78; 95% CI, 0.41–1.48; P = 0.44) or Whites (OR = 0.51; 95% CI, 0.12–2.19; P = 0.36). Similarly, no correlation between vitamin D and mortality was noted in the total cohort when adjustment was made for age, sex, BMI, and presence of comorbidities (adjusted OR = 1.04; 95% CI, 0.55–1.97; P = 0.90).

Vitamin D and Mechanical Ventilation/ICU Admission

We found no significant association between vitamin D levels and need for mechanical ventilation (OR = 1.23; 95% CI, 0.68–2.24; P = 0.49) or ICU admission (OR = 1.38; 95% CI, 0.81–2.34; P = 0.23). No correlation between the need for mechanical ventilation and vitamin D levels was seen in either males (OR = 1.24; 95% CI, 0.52–2.91; P = 0.63), females (OR = 1.20; 95% CI, 0.52–2.76; P = 0.67), Blacks (OR = 1.36; 95% CI, 0.71–2.62; P = 0.34) or Whites (OR= 0.77; 95% CI, 0.13–4.49; P = 0.77). The need for ICU admission was higher among males with low vitamin D levels compared to the males with normal vitamin D levels (OR = 2.32; 95% CI, 1.07–5.03; P = 0.03) in the unadjusted models, and in the models adjusted for age, BMI, and comorbidities (adjusted OR = 2.60; 95% CI, 1.07–6.28; P = 0.03). However, after Bonferroni correction was made to protect from the inflated type 1 error due to multiple comparisons, these results were noted to be statistically non-significant. No association between vitamin D levels and need for ICU admission was seen in females (OR = 0.82; 95% CI, 0.38–1.76; P = 0.61). Similarly, no significant association was noted between vitamin D levels and the need for ICU admission or mechanical ventilation among patients less than 65 years, or 65 years and older for unadjusted models, as well as when the models were fully adjusted for age, sex, BMI, and presence of comorbidities. Further details on the results, unadjusted and after adjusting for age, sex, BMI, and comorbidities, are summarized in Tables 3 and 4.

Table 3.

Association between vitamin D levels and mortality, mechanical ventilation and ICU admission, unadjusted odds ratio

	Mortality		ICU Admission		Mechanical Ventilation
Characteristic	Odds Ratio (95% CI)	P Value	Odds Ratio (95% CI)	P Value	Odds Ratio (95% CI)	P Value
Total cohort	0.69 (0.39–1.24)	0.21	1.38 (0.81–2.34)	0.23	1.23 (0.68–2.24)	0.49
Males	1.10 (0.46–2.63)	0.83	2.32 (1.07–5.03)	0.03	1.24 (0.52–2.91)	0.63
Females	0.49 (0.22–1.09)	0.08	0.82 (0.38–1.76)	0.61	1.20 (0.52–2.76)	0.67
Less than 65 years	0.90 (0.37–2.18)	0.81	1.39 (0.66–2.94)	0.39	1.81 (0.75–4.40)	0.19
65+ years	0.83 (0.35–1.92)	0.66	1.67 (0.74–3.75)	0.21	1.04 (0.42–2.59)	0.94
Blacks	0.78 (0.41–1.48)	0.44	1.33 (0.73–2.41)	0.35	1.36 (0.71–2.62)	0.34
Whites	0.51 (0.12–2.19)	0.36	2.56 (0.74–8.89)	0.14	0.77 (0.13–4.49)	0.77

Table 4.

Association between vitamin D levels and mortality, mechanical ventilation, and ICU admission, adjusted model

	Mortality		ICU Admission		Mechanical Ventilation
Characteristic	Odds Ratio (95% CI)	P Value	Odds Ratio (95% CI)	P Value	Odds Ratio (95% CI)	P Value
Total cohort adjusted*	1.04 (0.55–1.97)	0.9	1.51 (0.85–2.69)	0.16	1.36 (0.71–2.60)	0.35
Males**	1.94 (0.72–5.25)	0.19	2.60 (1.07–6.28)	0.03	1.32 (0.51–3.43)	0.56
Females**	0.62 (0.26–1.49)	0.28	0.91 (0.40–2.05)	0.82	1.29 (0.52–3.20)	0.58
Less than 65 years*	1.14 (0.43–2.99)	0.79	1.31 (0.58–2.99)	0.51	2.00 (0.75–5.34)	0.16
65+ years*	0.96 (0.40–2.31)	0.93	1.66 (0.73–3.79)	0.23	0.99 (0.39–2.53)	0.99
Blacks*	1.17 (0.57–2.39)	0.66	1.36 (0.71–2.60)	0.36	1.48 (0.72–3.02)	0.29
Whites*	0.86 (0.16–4.53)	0.86	3.71 (0.81–16.91)	0.09	1.08 (0.15–7.66)	0.94

*Adjusted for age, sex, body mass index (BMI), and comorbidities; **adjusted for age, BMI, and comorbidities.

Vitamin D and New DVT/PE

Vitamin D levels showed no significant association with development of new DVT or PE among COVID-19 patients (OR= 0.92; 95% CI, 0.16–5.11; P = 1.00). Further subgroup analysis was not done due to a limited number of patients developing thromboembolic episodes during the course of their admission.

Stratified Vitamin D Levels and Clinical Outcomes

The analyses performed for stratified vitamin D levels (< 20 ng/mL, 20-30 ng/mL, and > 30 ng/mL) showed no statistically significant association of these vitamin D levels with mortality, the need for mechanical ventilation and ICU admission in our cohort. Additionally, among patients with low vitamin D levels (< 20 ng/mL), no significant association was noted between vitamin D supplementation and clinical outcomes in COVID-19. More details on these results have been summarized in Table 5.

Table 5.

Association of stratified vitamin D levels/vitamin D supplementation with mortality, mechanical ventilation, and ICU admission, adjusted odds ratio for age, sex, BMI and comorbidities

	Mortality		ICU Admission		Mechanical Ventilation
Characteristic	Odds Ratio (95% CI)	P Value	Odds Ratio (95% CI)	P Value	Odds Ratio (95% CI)	P Value
Vitamin D supplements (among patients with low vitamin D levels)
Supplements (Yes vs no)	0.86 (0.26–2.80)	0.8	0.96 (0.35–2.59)	0.93	0.68 (0.22–2.13)	0.51
Stratified vitamin D levels
<20 ng/ml	1 (Ref)		1 (Ref)		1 (Ref)
20-30 ng/mL	1.20 (0.57–2.54)	0.63	0.70 (0.35–1.40)	0.31	0.69 (0.31–1.53)	0.36
>30 ng/mL	0.81 (0.39–1.66)	0.56	0.63 (0.33–1.22)	0.17	0.77 (0.37–1.60)	0.48

DISCUSSION

This retrospective cohort study demonstrated no significant association between vitamin D levels and mortality among COVID-19 patients. A review of literature points toward equivocal evidence linking clinical outcomes in COVID-19 patients with low vitamin D levels. Risk factors for low vitamin D levels include elderly, obesity, and males (35), and higher mortality in COVID-19 patients is also noticed among these patient populations (36–38). Another study suggests higher mortality seen in Italy and Spain could be attributed to a higher prevalence of low vitamin D seen in these countries (39). The study by llie et al. (40) also reports an association between vitamin D levels and COVID-19 mortality in European countries. However, these studies only provide inferred evidence based on the high prevalence of vitamin D deficiency in these countries and do not account for potential confounders like other underlying comorbidities and BMI. Also, the study by Laird et al. (39) has relied on literature and data as old as 20 years ago to derive the mean vitamin D levels in the patient population and correlated it with the current data of mortality among COVID-19 patients. Recent studies have also tried to elucidate if low vitamin D levels are associated with increased risk of testing positive for COVID-19, but the results have been conflicting (41–43).

Some of the literature suggests increased morbidity and mortality among Blacks with COVID-19 (44–46) and a plausible explanation is low levels of vitamin D seen in Blacks (47). However, our study demonstrated no such correlation between low vitamin D levels and mortality among COVID-19 patients who were Blacks. Hence further studies are needed before any such link between mortality and low vitamin D levels can be established.

High incidence of vitamin D deficiency has been reported among critically ill patients admitted to intensive care, resulting in increased length of stay and mortality (48–54). Literature review also suggests that low vitamin D levels may also be associated with worse disease outcomes especially in pneumonia (55, 56) and with the development of ARDS and acute lung injury (29, 57). A recent metanalysis by Munshi et al. (58) reported poor outcomes in COVID-19 patients with low vitamin D levels, however, in their study, poor outcomes were clubbed together as the development of ARDS, mortality, need for ICU admission, and mechanical ventilation. Our study looked at each of these severe disease clinical outcomes separately to identify individual correlations. We found that the need for ICU admission was higher among males with low vitamin D levels, however, after Bonferroni correction was applied, these results failed to reach the level of statistical significance. A study conducted by Carpagnano et al. (59) noticed a high prevalence of low vitamin D among patients admitted to ICU, and the majority of patients in the vitamin D deficiency group of their study were males.

We did not find any significant correlation between low vitamin D levels and the development of new DVT or PE in COVID-19 patients. This could be, in part, due to the low number of patients in this study who developed new DVT or PE. The role of vitamin D and thromboembolism is debatable in the literature. In the past, some large population studies have shown no correlation between vitamin D levels or vitamin D supplementation, on the risk of development of thromboembolism (60–62).

Some of the literature on vitamin D levels recommends maintaining vitamin D levels at a minimum of at least 30 ng/mL to achieve optimal health benefits (63). In our study, we did not see any significant association between vitamin D levels < 20 ng/mL, 20-30 ng/mL, and > 30 ng/mL with mortality, the need for ICU admission, and the need for mechanical ventilation in COVID-19 patients.

We acknowledge that our study has several limitations that need to be addressed. Although we had a large database of over 2000 patients, a large number of patients did not have recorded vitamin D levels within in the last year. This significantly reduced the number of patients who could be included in this study. Also, we relied on electronic medical records and clinical notes to gather data including the presence of comorbidities and documentation of vitamin D levels. Hence there is a possibility of both selection and information bias. The data for this study were collected from 4 hospitals in southeast Michigan, predominantly serving the underserved population having multiple comorbidities. Our sample size consisted of very few patients with other races besides Blacks and Whites, thereby limiting analysis in this population group. Also, very few patients in our cohort developed new DVT or PE during their hospital stay, hence more studies with a large sample size are needed before any conclusive inference can be made in this regard. Although the use of vitamin D levels before the patient developed illness helped avoid the negative acute phase impact of the illness on vitamin D levels, it would have been ideal if we had the measurements immediately preceding the infection from COVID-19. But it was not a possibility given the nature and design of the current study. We believe that further community-based studies will provide a better understanding of the possible role of vitamin D in the disease progression and severity of symptoms in COVID 19 patients.

Conclusions

This study did not find any significant association of vitamin D levels with mortality, the need for mechanical ventilation, ICU admission, and the development of thromboembolism in patients with COVID-19. Further studies are warranted before any conclusive association can be made between vitamin D levels and the clinical course of COVID 19 patients.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

P.L., P.N., and N.P. conceived and designed research; P.L. and S.K. analyzed data; P.L. and S.K. interpreted results of experiments; P.L., P.N., N.P., and S.K. prepared figures; P.L., P.N., and S.K. drafted manuscript; P.L., P.N., N.P., and S.K. edited and revised manuscript; P.L., P.N., N.P., and S.K. approved final version of manuscript.