Effect of Single High Dose Vitamin D Substitution in Hospitalized COVID-19 Patients with Vitamin D Deficiency on Length of Hospital Stay

Vitamin D and its role in the coronavirus-19 disease (COVID-19) pandemic has been controversially discussed, with inconclusive evidence about vitamin D3 (cholecalciferol) supplementation in COVID-19 patients. Vitamin D metabolites play an important role in the initiation of the immune response and can be an easily modifiable risk factor in 25-hydroxyvitamin D3 (25(OH)D3)-deficient patients. This is a multicenter, randomized, placebo-controlled double-blind trial to compare the effect of a single high dose of vitamin D3 followed by treatment as usual (TAU) of daily vitamin D3 daily until discharge versus placebo plus TAU in hospitalized patients with COVID-19 and 25(OH)D3-deficiency on length hospital stay. We included 40 patients per group and did not observe a significant difference in the median length of hospital stay (6 days in both groups, p = 0.920). We adjusted the length of stay for COVID-19 risk factors (β = 0.44; 95% CI: −2.17–2.22), and center (β = 0.74; 95% CI: −1.25–2.73). The subgroup analysis in patients with severe 25(OH)D3-deficiency (<25 nmol/L) showed a non-significant reduction in the median length of hospital stay in the intervention group (5.5 vs. 9 days, p = 0.299). The competing risk model with death did not reveal significant differences between the group in the length of stay (HR = 0.96, 95% CI 0.62–1.48, p = 0.850). Serum 25(OH)D3 level increased significantly in the intervention group (mean change in nmol/L; intervention: +26.35 vs. control: –2.73, p < 0.001). The intervention with 140,000 IU vitamin D3 + TAU did not significantly shorten the length of hospital stay but was effective and safe for the elevation of serum 25(OH)D3 levels.


Introduction
For more than two years, the world has been facing a global pandemic, with a then new coronavirus (CoV). The underlying pathogen of this new disease was later identified as Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) [1]. Within a short time period, the virus spread quickly around the world, and led the World Health Organization (WHO) to reclassify the epidemic as a global pandemic in March 2020 [2]. on length hospital stay. All patients that were hospitalized due to COVID-19 in one of the participating centers (Cantonal hospital Baselland and Cantonal hospital St. Gallen) were considered as potentially eligible for participation in this trial and therefore checked for the inclusion and exclusion criteria ( Table 1). The patients were recruited between December 2020 and August 2021. Participants were randomized in a 1:1 ratio to either group and randomization was stratified by center using block randomization (block size = 4). Table 1. Inclusion and exclusion criteria.

Inclusion Criteria
Informed consent as documented by signature >18 years old Ongoing COVID-19 infection (PCR confirmed) Hospitalized in one of the participating study centers Laboratory-confirmed serum 25(OH)D 3 < 50 nmol/L(<20 ng/mL)

Exclusion criteria
Known hypersensitivity against one of the used vitamin D 3 products Active malignancy Hypercalcemia defined as serum calcium >2.2 mmol/L Granulomatous disease (e.g., sarcoidosis) History of renal stones in the past 12 months Pregnancy or breastfeeding Admission to ICU before inclusion or at day of inclusion To elevate the serum 25(OH)D 3 level quickly and sufficiently, a single dose of 140,000 IU (3500 µg) vitamin D 3 was administered orally. This dose was found to be safe and effective in a study of high-dose vitamin D 3 in patients with tuberculosis [29,30]. The dosage recommended for the standard treatment of a vitamin D 3 deficiency varies from 400 to 1000 IU per day [31]. In consultation with a pharmacovigilance expert (University Hospital of Basel), the study team considered that the dose of 800 IU was adequate to elevate the serum 25(OH)D 3 level in the included patients.

Outcomes
The primary outcome was defined as the length of the hospital stay (LoS), defined as the time from randomization to discharge. Further, the time from symptom onset to discharge and the time from hospital admission to discharge were assessed and analyzed.
As secondary outcomes, we defined COVID-19-related complications, namely ICU admission, mechanical ventilation, death, and other complications linked to COVID-19. Additionally, we looked at the serum 25(OH)D 3 levels, the health-related quality of life on days 28 and 90, self-reported symptoms, and vital signs. For safety outcomes, the occurrence of adverse events (AEs) and serious adverse events (SAEs) were assessed [32].

Data Collection and Management
Each participant underwent the same procedure regarding the study-related visits from inclusion until discharge and for the follow-up period (Supplementary Table S1). The patients had daily visits until discharge by treating physicians and nurses. The Short Form 12 item health survey (SF-12) to assess the health=related quality of life (hrQoL) on day 28 and day 90 days after randomization was assessed by a telephone visit. The data were collected in an electronic data capture system (SecuTrial ® ), which is accessible via a standard browser and personalized and password protected.

Statistical Analysis
To perform the statistical analysis, either R Statistical software (Version 4.0.3) or SPSS Statistics, (Version 24 (IBM)) was used. We performed a sample size calculation based on the assumption of a true difference in the mean length of stay of two days (SD = 2.96) [33].
According to the calculation, accounting for 10% dropouts and the chosen block size randomization, a total number of 80 patients (40 per group) were included.
For the primary outcome, we used the Mann-Whitney-U, due to non-normal distribution, unless our assumptions for the sample size calculations were based on a two-sided t-test. For continuous data, independent two-sample Student t-test or Mann-Whitney-U test was used for secondary analyses of group differences in the variables of interest. To compare the proportion of patients between the two groups regarding the secondary outcomes, a Chi 2 test was performed and if values had expected frequencies <5, the Fisher exact test was used with simulated p-value (based on 2000 replicates). To evaluate the change in the SF-12 questionnaire over three time points, linear mixed effects models were used to determine time-by-treatment interactions using the Satterthwaite approximation to estimate p-values. Sensitivity analysis of the primary outcome was performed with a robust linear regression model adjusted for known COVID-19 risk factors. To evaluate the safety of short-term high dose vitamin D 3 administration in the above-described population, linear/logistic mixed-effect models were used to estimate treatment effects on adverse and serious adverse events across time.
A competing risk analysis that accounts for death as competing risk (besides the event of interest which was time to discharge) using the Fine and Gray model and a subgroup analysis for patients who had severe vitamin D deficiency at randomization, defined as 25(OH)D 3 level <25 nmol/L, for the primary endpoint with the same method used in the full dataset was performed post hoc. We only performed the binary analysis for the subgroup based on the small number of patients in the subgroup. The detailed methodology of the trial has been published previously [34].

Results
A total of 80 patients were randomized, 40 patients in each group. Two patients were excluded after randomization because the serum 25(OH)D 3 level at the day of inclusion was >50 nmol/L. The data of 78 patients were used for the analysis. A total of 39 patients were in the intervention group and 39 patients in the placebo group. Median time from hospital admission to randomization was two days for both groups (SD ± 1 day).

Primary Outcome
There was no significant difference between intervention group and control group in terms of length of hospital stay (Table 3, Figure 1). Median time from symptom onset to discharge was 16 days for the intervention group and 15 days for the control group (p = 0.193). Median time from hospital admission to discharge was 8 days for both groups (p = 0.924) and the median time from randomization to discharge was 6 days for both groups (p = 0.920).

Post Hoc Analysis
The post-hoc analysis of patients with severe 25(OH)D3 deficiency showed a median time from symptom onset to discharge of 15 days in both groups (p = 0.764). Median time from hospital admission to discharge was 5.5 days for the intervention group and 9 days for the control group (p = 0.299) and the median time from randomization to discharge was 4.5 days for the intervention group and 7 days for the control group (p = 0.444) ( Table  4).

Post Hoc Analysis
The post-hoc analysis of patients with severe 25(OH)D 3 deficiency showed a median time from symptom onset to discharge of 15 days in both groups (p = 0.764). Median time from hospital admission to discharge was 5.5 days for the intervention group and 9 days for the control group (p = 0.299) and the median time from randomization to discharge was 4.5 days for the intervention group and 7 days for the control group (p = 0.444) ( Table 4). In the competing risks model with the full dataset and death as the competing risk, groups did not differ in (a) time from symptom onset to discharge (HR = 0.88, 95% CI 0.57-1.36, p = 0.570), (b) time from hospital admission to discharge (HR = 0.99, 95% CI 0.64-1.52, p = 0.950), or (c) time from randomization to discharge (HR = 0.96, 95% CI 0.62-1.48, p = 0.850) Figure 2.
On day five after randomization, it was observed that significantly more patients reported headache within the intervention group compared to the control group (intervention: 56.3% vs. control: 15%, p = 0.014) (Supplementary Table S7). There was no other significant difference in self-reported symptoms observable at day 5 after randomization. The groups did not differ in terms of vital signs (Supplementary Table On day five after randomization, it was observed that significantly more patients reported headache within the intervention group compared to the control group (intervention: 56.3% vs. control: 15%, p = 0.014) (Supplementary Table S7). There was no other significant difference in self-reported symptoms observable at day 5 after randomization. The groups did not differ in terms of vital signs (Supplementary Table S8). Ten days after randomization, there were no more significant differences in terms of self-reported symptoms observed between the groups. In terms of vital signs, the intervention group had a lower systolic blood pressure (p = 0.029), but no difference was observed in other vital signs (Supplementary Table S9, Supplementary Table S10).
Non-significantly fewer adverse events were reported in the intervention group compared to the control group (intervention: 10.3% vs. control: 25.6%, p = 0.138). For serious adverse events, the intervention group had non-significantly more SAEs reported compared to the control group (intervention: 10.3% vs. control: 5.1%, p = 0.675) (Supplementary Table S11). We did not observe a significantly higher mean calcium level in the intervention group compared to the control group (intervention: 2.30 mmol/L (SD ± 0.99) vs. control: 2.22 mmol/L (SD ± 0.063); p= 0.101).
We observed a significant higher mean PTH level in the control group compared to the intervention group (intervention: 2.98 pmol/L (SD ± 1.30) vs. control: 1.60 pmol/L (SD ± 2.16); p = 0.007) and a significant lower mean level of phosphate in the control group compared to the intervention group (intervention: 1.01 mmol/lL(SD ± 0.289) vs. control: 0.86 mmol/L (SD ± 2.66); p = 0.003) (Supplementary Table S12).

Discussion
We could not observe a difference in the primary endpoint length of hospital stay. The comparison of the length of stay variables between the two groups did not reveal a significant difference that would indicate a beneficial effect of a single high dose vitamin D 3 in addition to standard treatment. The findings of our study are in line with previous results [17,18,[22][23][24]35]. The median length of stay in our population was similar to the findings presented by Murai et al. (2021) [22].
In the post-hoc subgroup analysis, which only included patients with a 25(OH)D 3 deficiency (<25 nmol/L), we observed a non-significant shorter length of stay for the intervention group. There was no difference in the median time from symptom onset to discharge, but in the other length of stay variables ("hospital admission to discharge" and "randomization to discharge"). This trend was already observed by other studies, therefore it might be worth having a closer look at patients with severe 25(OH)D 3 deficiency in bigger studies, as our sub-group was underpowered to detect a clinically significant difference [18,36].
In terms of secondary outcomes, we could not find significant differences in the need of ICU treatment, intubation, death, or other COVID-19 related complications. Those findings indicate that high dose vitamin D 3 had no impact on the course of the disease regarding severity. We observed a significantly longer mean duration in the ICU within the intervention group. In the overall population, there were fewer admissions to the ICU compared to other studies, which was possibly influenced by the fact that our population had fewer co-morbidities and therefore the risk of ICU admission was smaller [22,37]. With our study, we could not support previous findings that high dose vitamin D 3 supplementation reduces ICU admissions [37]. The fact that more patients in the intervention group were intubated compared to the control was not significant and we therefore assume that individual predisposing factors or pre-existing conditions have had a greater influence than the intervention with high doses of vitamin D 3 in patients with 25(OH)D 3 deficiency [5]. This might also contribute to the significantly longer length of ICU treatment in the intervention group. Our study could not confirm the results that fewer patients needed ICU admission or less mechanical ventilation when treated with high dose vitamin D 3 in addition to the best available other therapies [37][38][39]. However, the number of patients with COVID-19-related complications was very small and therefore the results need to be interpreted with the utmost caution.
We observed a significant increase in the mean serum 25(OH)D 3 level in the intervention group compared to the control group, which resulted in a significant higher mean level at discharge in the intervention group and significantly more patients had serum 25(OH)D 3 levels ≥50 nmol/L in the intervention group. At the same time, there was no clinically relevant increase in calcium or phosphate in the intervention group, which is in line with previous results [24,[40][41][42]. This indicates that a single bolus of high dose vitamin D 3 in addition to daily low dose maintenance therapy is effective to elevate serum 25(OH)D 3 levels quickly. This finding is in line with previous observations that investigated the effect of a single high dose vitamin D 3 with or without daily maintenance dose in various conditions [29,[43][44][45]. Another factor that might have influenced the outcome of our study is the individual response to vitamin D 3 supplementation, as only 23.1% of the participants in the intervention group reached sufficient 25(OH)D 3 serum levels [46].
At time of inclusion, we observed significantly lower SF-12 mental scores in the intervention group than in the control group. This might have led to a bias in the primary outcome variable, since higher hrQoL scores are associated with a reduction of LOHS [47]. The difference between the groups in the SF-12 mental sub-score was not observable anymore 28 days and 90 days after randomization. The increasing number of missing data over time might mediate this observation. Another aspect that should be considered in the discussion of the observation is the range of the SF-12 mental score in our population. In the intervention group, both the range and the variance were larger than in the control group. However, the sample with a completed hrQoL from baseline to day 90 was rather small and therefore possible effects of the two groups did not differ significantly in terms of self-reported symptoms or recorded vital signs. At day five after randomization, we observed that a significantly higher number of patients in the intervention group reported headache compared to the control group, but no differences in other symptoms or vital signs. Ten days after randomization, no more differences were observed in self-reported symptoms. When looking at the vital signs, we observed a significantly lower systolic blood pressure in the intervention group compared to control group, but still the values were in the normal range. Orthostatic hypotension can be caused by longer bedrest or infections [48,49]. Further, only 10 and 11 patients respectively were still hospitalized at day 10 after randomization and therefore the results must be interpreted with caution. Our study could not confirm the findings which reported faster symptom recovery [17,18].
Our results did not show significant differences in the occurrence of adverse and serious adverse events between the intervention group and the control group. This indicates that the intervention with high dose vitamin D 3 is safe and not associated with more adverse or serious adverse events.
Our study population showed similar characteristics at the time of randomization such as age, median time from symptom onset to hospitalization, laboratory parameters, or 25(OH)D 3 level compared to other trials, except the sex distribution, as our study had fewer female participants [17,18,[21][22][23][24]35,40,41,[50][51][52][53]. In the early stage of the pandemic in Switzerland, it was observed that the majority of hospitalized patients were male (female: 37% vs. male: 63%), which might explain the lower number of females in our study [54]. In terms of hematological and blood chemistry parameters, we observed several values that were above or beyond the normal value range, which can be explained by viral infections, including SARS-CoV-2, medication, systemic disorders, or malnutrition [55][56][57][58][59][60][61][62]. The intervention group showed significantly higher baseline CRP values than the control group. This difference might imply that patients in the intervention group were suffering from a more severe disease compared to the control group, as a higher CRP could be an early sign for more severe COVID-19 [63,64]. However, since there were no significant differences in baseline respiratory status suggesting a more severe disease in the intervention group (respiratory rate, SpO2, dyspnea, O2-supplementation), the authors cannot conclude if this imbalance had an effect on the overall results of this study.
Our population had a lower frequency of comorbidities reported compared to other studies, namely arterial hypertension and other CVDs [22,24,35,53,65,66], which might influenced the overall outcome.
The time between symptom onset and randomization was rather long with more than 7 days in both groups, which might have had a influence on the primary outcome, as already reported by Murai and colleagues [22].

Strengths and Limitations
Among the strengths of our study are the study design as a randomized, controlled double blind multicenter trial, that the groups did not differ significantly across demographic and clinical variables at time of randomization, and the fact that all patients completed the trial. Limitations for this trial were the small sample size of 80 patients included and 78 analyzed, and missing values in some laboratory parameters at randomization and discharge and the health-related quality of life in the follow-up period. Further, we did not assess COVID-19 severity with a validated scale. In addition, we did not include outpatients in this trial and the period between symptom onset and randomization was rather long.

Conclusions
In this trial, we could not confirm previous findings that a bolus of high dose vitamin D 3 had a beneficial effect on the investigated outcomes patients with COVID-19 and 25(OH)D 3 deficiency, except for the elevation of the serum 25(OH)D 3 level. In the subgroup analysis, we observed a non-significant difference in median LOS, which might indicate that patients with severe 25(OH)D 3 deficiency would benefit more from the high dose intervention compared to the overall 25(OH)D 3 deficient population in our study [36,67,68]. Whether this finding was due to the beneficial effect of vitamin D 3 or due to chance remains unclear, since the subgroup analysis was underpowered, and the assumption needs further investigation for verification. In further studies focusing on patients with severe 25(OH)D 3 deficiency, non-hospitalized patients or patients in nursing homes should be included and the outcome of fewer hospital admissions needed should be added. We could show that a single high dose vitamin D 3 in addition to standard vitamin D 3 is not associated with adverse or serious adverse events, and therefore a safe and effective way to elevate the serum 25(OH)D 3 level quickly, whereas daily small dosages of vitamin D 3 supplementation did not elevate the serum 25(OH)D 3 level sufficiently.

Participant Privacy and Confidentiality
All individual medical information that was obtained during this trial is considered confidential and disclosure to third parties is prohibited. Only ethics committees or regulatory authorities are allowed to have access to source data within the framework of an audit of the study. The confidentiality is further ensured by utilizing subject identification numbers for data record in the database.
Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/biomedicines11051277/s1, Table S1: Visit procedure and collected information for each patient at different time points during the study. Table S2: Co-Morbidities at time of randomization. Table S3: Concomitant medication during hospitalization. Table S4. Robust linear regression results presenting the effect of group. Intervention vs. Control on time to discharge (from randomization) (univariable and adjusted for COVID-19 risk factors, prognostic imbalances, and center effects. Table S5: Mean SF-12 mental and SF-12 physical scores at randomization, 28 days and 90 days after randomization. Table S6: Mixed-effects linear regression results on SF-12 mental and physical outcome with a time and group interaction term. Table S7: Frequency of self-reported symptoms at day 5 after randomization. Table S8: Mean vital signs at day 5 after randomization. Table S9: Frequency of self-reported symptoms at day 10 after randomization. Table S10: Mean vital signs at day 10 after randomization. Table S11: Frequency of adverse and serious adverse events occurred during the trial per study group. Table S12: Levels of calcium, PTH and phosphorus at time of discharge. Figure S1: Group differences in changes in SF-12 mental (top) and SF-12 physical (bottom) across time (randomization = 0m, 28 days = 1 m, 3 months = 3 m). The mean and 95% confidence interval for the observed mean values are displayed.  Institutional Review Board Statement: The study was approved by the local ethical committees of northwestern and eastern Switzerland on 23 September 2020 and 19 October 2020 (Project ID: 2020-01401) and Swissmedic on the 19 November 2020 (Project ID: 2020DR33170). The study was registered on an international and a national clinical trials registry platform with the registration number NCT04525820.
Informed Consent Statement: Written informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy policy and ethical statement.

Conflicts of Interest:
The authors declare no conflict of interest, and the funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.