Vitamin D Status and COVID-19: Some Implications

Background Vitamin D is essential for the maintenance of good health and its status is dened by the level of serum 25-hydroxyvitamin (25(OH)D). Negative correlations between mean levels of 25(OH)D per country and the number of COVID-19 conrmed cases per one million population, and COVID-19 mortality per one million population, were recently observed. The aim of this study was to identify levels of 25(OH)D below which, rates of COVID-19 conrmed cases, mortality and lethality, increase signicantly. Methods A data table found in the literature, containing a list of twenty countries and their corresponding mean level of 25(OH)D was updated with COVID-19 latest numbers of conrmed cases and mortality rates. Cut points of 25(OH)D below which rates were signicantly higher were found according to various statistical criteria: absolute difference of means, t-test p-values, between class variance, entropy. Thresholds of 25(OH)D below which there can be a signicant rise of COVID-19 conrmed cases, mortality and lethality, were found performing a Gaussian kernel regression.


Page 3/18
The recent outbreak and rapid spreading of severe acute respiratory syndrome COVID-19 are a global threat and primary concern worldwide, with a still uncertain outcome. Previous observational studies report independent associations between low serum concentration of 25-hydroxyvitamin D and susceptibility to acute respiratory tract infections [4]. Some authors found that vitamin D might play a protective role for COVID-19 with an association between the mean levels of vitamin D in various countries and the morbidity, and the mortality, caused by COVID-19 [5].
Our purpose was to estimate 25(OH)D levels below which rates of COVID-19 con rmed cases per one million population (/1 M), rates of COVID-19 mortality/1 M, and lethality, were higher.

Methods
We used the aggregated data table built by Ilie et al. in [5] with the same sample of n = 20 countries and their corresponding mean level of 25(OH)D, but the number of COVID-19 con rmed cases/1 M and the number of COVID-19 deaths/1 M population were updated according to the recent records of Coronavirus Disease 2019  data in the WHO records [6]. Lethality variable, ratio of the number of deaths/1 M to the number of con rmed cases/1 M, was computed. It was an estimation of the probability of dying knowing that the patient is infected.
K-means algorithm was applied with K = 2 to Cases variable in order to get a partition of two clusters Δ and ∇ composed of countries having a higher number of cases/1 M and a lower number of cases/1 M, respectively.
In order to nd a 25(OH)D cut point below which the behavior of Cases variable statistically differs signi cantly from the behavior of Cases variable above that cut point, the table was rst ordered in increasing order of 25(OH)D (Table 1)  BCV indicates how Class 1 and Class 2 differ from the whole sample and, when it is maximal, the Within Class Variance is minimal. BCV is a standard evaluation of the classi cation quality.
Con dence Intervals of 95% were adjusted bootstrap percentile intervals obtained using the 'boot.ci' function, option type = 'bca', in the 'boot' package. They were presented in the form original [le, re], where original denoted the estimation from the original sample Table 1, le and re denoting the con dence interval (CI) left and right endpoints, respectively. The number of bootstrapped samples of size n = 20, was R = 1,000. Actually, the value original was not the midpoint of the CI [le, re].
Linear regressions were applied to our sample, con dence intervals and p-values were derived from Dupont-Plummer formulas [7]. The rise of Cases variable was investigated by performing a Gaussian kernel regression of Cases variable on 25(OH)D concentrations using the 'npreg' function in the 'np' (nonparametric) package.
The preceding methods were applied similarly to Mortality and Lethality variables, respectively.

Results
The results presented below in Table 1, were obtained using the COVID-19 data recorded on 2020 July 27 th [6]. No signi cant differences were observed in the results when using COVID-19 data recorded two weeks before or two weeks after that date because infection and mortality rates per one million population did not vary signi cantly. The results for the number of cases, for mortality and for lethality were summarized in Table 2, Table 3 and Table 4, respectively.
K-means algorithm applied with k = 2 for example to Mortality variable, using the original Table 1 sample, yielded two clusters of countries, one cluster of 8 countries (Spain, United Kingdom, Belgium, Italy, Ireland, The Netherlands, France, Sweden) having an elevated mortality rate and a second cluster of 12 countries (Portugal, Switzerland, Denmark, Germany, Estonia, Turkey, Island, Hungary, Czechia, Norway, Finland, Slovakia) having a moderated or low mortality rate, see Table 3.   For the Cases variable, the cut point c was found around 50 nmol/L for any of the four criteria from Table  1 sample, see Fig. 1 and Fig. 2. The boot.ci function performed the procedure on each of the R = 1,000 bootstrapped samples to get a 95% CI displayed in Table 2.
Linear regressions applied to our data were not signi cant, con dence intervals were not acute enough and p-values derived from Dupont-Plummer formulas [7] were bad, even using bootstrap methods. The nonparametric Gaussian Kernel regression curve of the morbidity variable on 25(OH)D, showed a rising starting at 61 nmol/L when using Table 1 data, see Fig. 3. That of mortality variable showed a rising at 62 nmol/L when using Table 1 data, see Fig. 4.
In our Table 1 sample, we noticed that there was only one country, namely Sweden, over the 6 countries having 25(OH)D mean strictly greater than 60nmol/L, which had an elevated mortality rate, while 7 countries over the 14 having 25(OH)D less than 60 nmol/L, had an elevated mortality rate, see Fig. 4.
However, 25(OH)D is certainly not the only factor that could predict mortality clusters and : we observed that the AUC (Area Under Curve) of ROC curves were lower than 0.65 when performing the prediction function in the ROCR package of R software.

Discussion
The general metabolism and actions of vitamin D are well-known, in particular as a natural immune modulator [8]. Vitamin D enhances cellular innate immunity partly through the induction of antimicrobial peptides [9,10,11]. It also reduces the cytokine storm induced by the innate immune system in response to viral and bacterial infections, as it was observed in COVID-19 patients [12].
In the literature, we found several thresholds of 25(OH)D concentrations which de ned vitamin D de ciency. The institute of Medicine de ned vitamin D de ciency threshold at 25(OH)D level of 30 nmol/L [13]. Vitamin D de ciency is de ned at 100 nmol/L for the vitamin D Council [14], and at 50 nmol/L for the Endocrine society [15]. Our ndings, which seems to be consistent with the Endocrine Society estimate, suggested that a 25(OH)D levels below 50 ± 10 nmol/L may be linked to severe clinical outcomes of COVID-19 infection. Thus, maintaining a serum 25(OH)D levels above 50 nmol/L could be recommended in COVID-19 disease. Indeed, Casey et al. [16] concluded that daily supplementation of vitamin D potentially offer additional protection against COVID-19 and Ebadi et al. [17] noticed that a high-dose vitamin D intervention could have a potential bene t in decreasing risk of COVID-19 severity and mortality.
Many investigators have shown that there is a threshold for serum 25(OH)D below which secondary hyperparathyroidism may occur [18] and this threshold has been estimated in the region of 40-50 nmol/L [19,20,21]. As we found that COVID-19 mortality rates below vitamin D level of 50 ± 10 nmol/L are statistically higher than the rates above 50 nmol/L, it could be assumed that elevation of PTH concentrations are involved in COVID outcomes on most of patients who have a 25(OH)D level below 50 nmol/L. Indeed, some studies have shown that elevated PTH levels are associated with increased cardiovascular risk in the general population [22,23]. Mitnick et al. [24] have found that the liver is an important source of the circulating interleukin-6 generated in response to PTH while Casey et al. [16] have indicated that the most severe cases in COVID involving a pro-in ammatory state can lead to harmful outcomes mediated by a deregulated immune response involving interleukin-6 and other in ammatory signaling molecules. Moreover, Cheng et al. [25] have found an association between higher serum concentrations of PTH and several in ammatory markers.
It is known that 1,25(OH)2D concentrations depend mainly of 25(OH)D levels. According to White et al. [26], the in uence of 1,25(OH)2D on the immune system is one of the most important factors to consider.
Notably Concerning the role of calcium concentration in COVID-19 mortality, Lippi et al. [28] found that COVID-19 pandemic severity is in part associated with lower serum concentrations of calcium. It is well-known that the ionic concentration of calcium, which is highly correlated to 25(OH)D and to 1,25(OH)2D, appears as the major factor in uencing the secretion rate of PTH [29]. PTH plays a critical role in calcium homeostasis, defending against hypocalcemia by acting on target organs such bone and kidney in order to stimulate bone resorption, to promote renal conservation of calcium, and to induce production of 1,25(OH)2D, which in turn enhances intestinal calcium absorption [30]. Moreover, Goodman et al. [31] noticed that a slight decrease in ionized calcium triggers the release of PTH from the parathyroid glands, suggesting that patients with severe COVID-19 and low calcium concentrations may have high levels of PTH that could worsening COVID-19 issues. More studies are needed to investigate this suggestion.
In the past, vitamin D de ciency was identi ed by the presence of bone diseases, essentially either rickets or osteomalacia. More recently, the term vitamin D insu ciency has been used to describe suboptimal levels of serum 25(OH)D that might be associated with other disease outcomes [32]. According to Holick [33], in the normal range [34]. This was con rmed in the results presented in Emilion et al. study [20]: the regression curve of PTH concentrations on 25(OH)D concentrations showed a small rise of PTH levels at 59 nmol/L, a plateau in the range 49-59 nmol/L and a main rise at 49 nmol/L. Moreover, when patient becomes vitamin D insu cient and de cient, the increase in PTH levels result in normal or elevated levels of 1,25(OH)2D [34]. Then, 1,25(OH)2D assay could be useless as a measure of vitamin D status [33].
In this study, we were not informed of PTH levels, 1,25(OH)2D levels, and calcium concentrations of the patients with COVID-19 from countries. Though the vitamin D cut-points of 50 ± 10 nmol/L should not be interpreted as an optimal vitamin D status, it could be used in the monitoring of vitamin D supplementations in COVID-19 syndrome. Future works need at least to take in account, for patients with COVID-19 from different stages of clinical evolution, measurements of concentrations of 25(OH)D, 1,25(OH)2D, calcium and PTH in order to elaborate reliable protocols of vitamin D-calcium supplementations and to realize a more e cient clinical monitoring of COVID-19 infection.

Conclusion
We identi ed 25(OH)D levels, of 60 ± 6 nmol/L, in the range of vitamin D insu ciency, below which rates of COVID-19 con rmed cases per one million population, and rates of COVID-19 mortality per one million population increased. We found that below 25(OH)D levels, of 50 ± 10 nmol/L, mainly in the range of vitamin D de ciency, rates of COVID-19 mortality were the highest. Therefore, we suggest that 25(OH)D concentrations should be above 60 ± 6 nmol/L to reduce morbidity and mortality during the COVID-19 pandemic. Future works could investigate other potential roles of vitamin D status in the evolution of the COVID-19 syndrome. Availability of data and materials The dataset used and analyzed during the current study is available from the corresponding author on reasonable request.