Discussion

The current umbrella meta-analysis summarises 13 meta-analyses, 57 observational studies and 23 RCT. According to results, vitamin D supplementation was efficient in reducing mortality, and vitamin D deficiency significantly increased mortality, severity of COVID-19, and risk of infection among patients. In addition, lower serum levels of vitamin D were significantly associated with poor prognosis. However, there was no significant relationship between serum vitamin D and positive cases of COVID-19, and the results regarding ICU admission and vitamin D were contradictory. Furthermore, results did not show any significant positive effect of vitamin D serum status or vitamin D supplementation on reducing risk of invasive or/and non-invasive mechanical ventilation. Due to limited number of studies for each variable, sub-group analyses were not possible.

In this umbrella meta-analysis, we discussed the multiple aspects of vitamin D deficiency and risk of mortality and COVID-19 health status outcomes. Vitamin D is a fat-soluble vitamin with anti-inflammatory, antioxidant and antiviral features^{(Reference Nikniaz, Akbarzadeh and Hosseinifard21)}. The regulatory role of vitamin D on acquired and innate immunity explains its possible role in infectious diseases such as COVID-19^{(Reference Teshome, Adane and Girma18)}. Based on the findings of clinical trials, vitamin D supplementation is efficient in reducing mortality. The beneficial effects of vitamin D in treating COVID-19 are by preventing ‘cytokine storm’ and subsequent acute distress syndrome, known as the main cause of mortality^{(Reference Maghbooli, Ebrahimi and Shirvani22)}. After activation of the angiotensin-converting enzyme 2 (ACE2) receptor by the coronavirus, vitamin D provides its protective role via activating the renin–angiotensin–aldosterone system, modulating the cytokine storm and neutrophil activity, maintaining the pulmonary epithelial barrier, stimulating epithelial repair and reducing the damage caused by pro-inflammatory cytokines. Moreover, vitamin D augments the activity of the ACE2/Ang (1–7) axis, which has anti-inflammatory and antioxidant functions and also suppresses renin and the ACE/Ang II/AT1R axis, thereby enhancing the expression and concentration of ACE2, MasR and Ang-(1–7)^{(Reference Rawat, Roy and Maitra5,Reference Bassatne, Basbous and Chakhtoura15,Reference Oristrell, Oliva and Subirana23)} .

Vitamin D increases cathelicidin (LL-37)/defensin expression and displays antimicrobial and antiviral activities. Cathelicidin and defensin, furthermore, stimulate the expression of antiviral cytokines and chemokines involved in the recruitment of monocytes/macrophages, natural killer cells, neutrophils, and T cells and eventually enhance the host defence. The vitamin D receptor and CYP27B1 dignify the expression and cellular production of cathelicidin and defensin, which is effected by the interactions of pathogens and membrane pattern recognition receptors, including toll-like receptor and toll-like receptor 2^{(Reference Kaya, Pamukçu and Yakar4)}. Additionally, vitamin D indorses the up-regulation of IL-10 (anti-inflammatory cytokine) and down-regulation of IL-1, IL-6 (pro-inflammatory cytokines) and TNF-alpha^{(Reference Pal, Banerjee and Bhadada12)}. Vitamin D also increases the expression of genes involved in the antioxidant system, such as the glutathione reductase gene^{(Reference Kazemi, Mohammadi and Aghababaee17)}.

Although the majority of studies confirmed the efficiency of vitamin D supplementation in declining mortality, accurate evidence-based recommendations on circumstances of vitamin D administration in clinical practice can be confirmed by well-designed RCT on health outcomes of COVID-19^{(Reference Bassatne, Basbous and Chakhtoura15)}. In this regard, different aspects of vitamin D supplementation in COVID-19 in RCT must be discussed thoroughly. For example, some studies were accomplished on aged individuals which already have several comorbidities, are less exposed to sunlight, display lower 7-dehyrocholesterol values in the skin, have enhanced markers of cytokine release syndrome and are at high risk of respiratory failure^{(Reference Szarpak, Rafique and Gasecka2,Reference Pereira, Dantas Damascena and Galvão Azevedo9,Reference Nikniaz, Akbarzadeh and Hosseinifard21)} . Also, study population was not stratified based on serum vitamin D status at baseline, since vitamin D-deficient patients benefit more from supplementation. Differences in the dose of supplementation, frequency of supplementation, route of prescription and duration are other limiting factors^{(Reference Nikniaz, Akbarzadeh and Hosseinifard21,Reference Oristrell, Oliva and Subirana23)} . Heterogeneity in the study design, population characteristics, methodology, baseline characteristics and small sample size of the population enrolled have also been mentioned in a number of studies^{(Reference Rawat, Roy and Maitra5,Reference Shah, Saxena and Mavalankar14,Reference Nikniaz, Akbarzadeh and Hosseinifard21)} . Differences in the type and timing of vitamin D supplementation are another confounding factor. In regard to source of vitamin D, it has been mentioned that cholecalciferol supplementation may lead to faster recovery from COVID-19^{(Reference Bassatne, Basbous and Chakhtoura15)}. Most studies administered 1,25-hydroxy cholecalciferol, as the active form of vitamin D and few studies used calcifediol^{(Reference Rawat, Roy and Maitra5)}. Moreover, one study indicated that patients supplemented with vitamin D after COVID-19 diagnosis benefited more than those supplemented with the drug prior to the diagnosis^{(Reference Pal, Banerjee and Bhadada12)}.

According to observational studies, there was an inverse relationship between vitamin D deficiency and mortality. Vitamin D deficiency is related to reduced innate cellular immunity and cytokine storm stimulation^{(Reference Liu, Sun and Wang11)}. The mechanism of action of vitamin D and ACE has been discussed earlier. High levels of ACE have been observed in patients with severe COVID-19 with low vitamin D level^{(Reference Oristrell, Oliva and Subirana23)}. Vitamin D receptors are present on the nuclei membrane and are responsible for regulating different defensive proteins and receptors. Receptors recognise pathogens and their interaction affect the expression of pathogenic genes. Vitamin D inhibits T helper type 1 proliferation and shifts towards T helper type 2 proliferation, leading to decline in oxidative compounds synthesised via T helper type 1, affecting T-cell maturation, and producing anti-inflammatory subtypes^{(Reference Nikniaz, Akbarzadeh and Hosseinifard21)}. McGregor et al. claimed that CD4 + T cells present in the bronchoalveolar lavage fluid of patients diagnosed with COVID-19 are Th1-skewed and the genes induced by severe acute respiratory syndrome coronavirus-2 are regulated by vitamin D receptor (VDR)^{(Reference McGregor, Chauss and Freiwald24)}. Furthermore, vitamin D induces transcription factors including STAT3 (signal transducer and activator of transcription 3), c-JUN and BACH2 (BTB Domain And CNC Homolog 2) that cooperatively suppress Th1 and Th17, and eventually induce IL-10 via IL-6-STAT3 signalling^{(Reference Oristrell, Oliva and Subirana23)}. Jain et al. reported that inflammatory markers such as IL-6, TNF-α and serum ferritin levels were shown low in COVID-19 patients with serum vitamin D level below 50 nmol/l^{(Reference Jain, Chaurasia and Sengar25)}. Additionally, high concentrations of transforming growth factor β have been observed in the acute phases of COVID-19 and are relatively suppressed by VDR^{(Reference Kazemi, Mohammadi and Aghababaee17)}. Mechanistic pathways are comprehensively and schematically demonstrated in Fig. 5.

Fig. 5 Mechanistic pathways demonstrating how vitamin D is affective on COVID-19 patients

The association between vitamin D deficiency and COVID-19 mortality must be discussed from other perspectives as well; for example, it is not clear whether low vitamin D is the cause or consequence of COVID-19. Multiple factors may affect the reduced vitamin D level in patients diagnosed with COVID-19, including age, sex, region, season, sun exposure, BMI, comorbidities and race. In favour of age, in the majority of studies, patients were over 50 years old with basic low vitamin D level^{(Reference Liu, Sun and Wang11,Reference Bassatne, Basbous and Chakhtoura15,Reference Nikniaz, Akbarzadeh and Hosseinifard21)} . Obesity alone is an independent risk factor for severe sequences of the disease^{(Reference Szarpak, Rafique and Gasecka2)}. COVID-19 broke out in winter when in the northern hemisphere, sunlight was low and individuals in that region had low 25-hydroxyvitamin D level^{(Reference Kaya, Pamukçu and Yakar4,Reference Liu, Sun and Wang11)} . Moreover, patients were enforced to be isolated or hospitalised, which prevented them from obtaining sunlight and a balanced diet^{(Reference Liu, Sun and Wang11)}. Ecological studies have revealed that people living in higher latitude with decreased vitamin D level are prone to infection, related complications and mortality^{(Reference Nikniaz, Akbarzadeh and Hosseinifard21)}.

In regard to studies, Liu et al. claimed inconsistency in the number and sample size of included studies, significant heterogeneity, publication bias and variations in ES estimates as reasons for the inconsistent results observed^{(Reference Liu, Sun and Wang11)}. Bassatne et al. reported low quality and inevitability of evidence, as well as variation in the definition of vitamin D deficiency, serum 25(OH)D cut-offs and the timing of blood sampling and COVID-19 diagnosis and related outcomes in the included studies as the reason for the observed controversies among included studies. Also, decline in the synthesis of vitamin D binding protein and increase in 25(OH)D renal excretion which significantly regulate vitamin D level in critical illnesses could be considered as an additional explanation for discrepancies^{(Reference Bassatne, Basbous and Chakhtoura15)}.

Vitamin D deficiency also significantly enhanced the risk of COVID-19 infection and severity of COVID-19. According to the D-CIMA meta-analysis, patients with serum 25(OH)D < 20 ng/ml or 50 nmol/l were 1·64 times more likely to be infected with COVID-19 and also individuals with serum 25(OH)D < 20 ng/ml or 50 nmol/l were 2·42 times more likely to have severe COVID-19^{(Reference Kaya, Pamukçu and Yakar4)}. One study claimed that vitamin D supplementation declined the frequency of infection and was beneficial in patients receiving daily or weekly doses of 25(OH)D, protective effects were stronger in patients with baseline 25(OH)D less than 25nmol/l, and that this relationship was NS in those receiving bolus doses^{(Reference Oscanoa, Amado and Vidal19)}. The mechanism of action is related to the disruption of the parathyroid-vitamin D-axis^{(Reference Pizzini, Aichner and Sahanic26)}. Moreover, vitamin D acts by stimulating the chemotaxis of T-lymphocytes and eliminating respiratory pathogens by inducing apoptosis and autophagy in the infected epithelium^{(Reference Kaya, Pamukçu and Yakar4)}. Hence, vitamin D declines the risk of microbial infection by modulating the innate and adaptive immunity, inhibiting cytokine storm, and declining pro-inflammatory cytokine production, due to its antiviral and anti-inflammatory properties^{(Reference Kazemi, Mohammadi and Aghababaee17,Reference Teshome, Adane and Girma18)} . Several aspects of this association must be further discussed. It is not clear whether the low concentrations of 25(OH)D in patients with severe COVID-19 infection are a cause or consequence of severe COVID-19 infection. Three perspectives have been mentioned: First, absence of baseline 25(OH)D measurement prior to infection; second, the concentration of C-reactive protein was not measured for patients with severe COVID-19 infection; third, 25(OH) D concentration decrease, as a consequence of inflammation, is considered solely as a negative acute phase reactant. Furthermore, a majority of studies did not report whether 25(OH)D concentrations were measured before or during COVID-19 infection^{(Reference Oscanoa, Amado and Vidal19)}.

Patients with poor prognosis had significantly lower serum levels of vitamin D compared to those with good prognosis. One study claimed 25(OH)D concentration may be considered as a negative acute phase reactant and a poor prognosis in COVID-19 infection^{(Reference Oscanoa, Amado and Vidal19)}. In Sun et al.’s study, 74 % of patients with severe COVID-19 had low calcium and 25(OH)D level and hypoproteinaemia. They reported hypocalcaemia as a biomarker of clinical severity and prognosis^{(Reference Sun, Zhang and Zou27)}. As mentioned earlier, calcitriol as the active form of vitamin D is the regulator of renin–angiotensin system and this overactivation is related to poor prognosis^{(Reference Szarpak, Rafique and Gasecka2)}.

According to the results of the present study, there was no significant relationship between serum vitamin D and positive cases of COVID-19. Bassatne et al. reported uncertain evidence regarding the association between positive cases of COVID-19 and serum 25(OH)D levels <20 ng/ml; however, increasing the cut-off of low 25(OH)D levels to 30 ng/ml showed significant results^{(Reference Bassatne, Basbous and Chakhtoura15)}. Other studies showed that COVID-19-positive cases had lower vitamin D level compared to negative cases. However, significant heterogeneity and publication bias was reported in these studies^{(Reference Szarpak, Rafique and Gasecka2,Reference Liu, Sun and Wang11)} .

The results regarding ICU admission and vitamin D were contradictory. Bassatne et al. claimed an increased risk of ICU admission in COVID-19 patients with 25(OH)D levels < 20 ng/ml and also indicated that calcifediol supplementation may have a protective effect on COVID-19-related ICU admissions^{(Reference Bassatne, Basbous and Chakhtoura15)}. Similarly, a pilot trial showed that only 1 out of 50 patients receiving calcifediol needed ICU admission, while 50 % of patients not receiving vitamin D were admitted to ICU (OR = 0·03). However, the reported OR was unreliable mainly due to indeterminate allocation concealment and patient blinding^{(Reference Castillo, Costa and Barrios28)}. One study^{(Reference Nikniaz, Akbarzadeh and Hosseinifard21)} observed decline in ICU admission rate after vitamin D administration. However, this study did not include a RCT that had major influence on the findings of other studies which showed no association between ICU admission and vitamin D supplementation^{(Reference Rawat, Roy and Maitra5)}. The main reason for the contradictory findings observed was the limited number of studies assessing the relationship between ICU admission and vitamin D.

The current study also did not show any significant positive effect of vitamin D serum status or vitamin D supplementation on reducing risk of invasive and non-invasive mechanical ventilation. One study showed that COVID-19 patients who required mechanical ventilation had at least one nutrient deficiency^{(Reference Szarpak, Rafique and Gasecka2)}. Hence, a clear association between vitamin D serum status and mechanical ventilation cannot be obtained. The main reason for the inconsistent results observed is the small number of studies assessing this association. The majority of studies did not observe any significant results, and the few ones lacked important methodological qualifications^{(Reference Szarpak, Rafique and Gasecka2,Reference Rawat, Roy and Maitra5,Reference Pereira, Dantas Damascena and Galvão Azevedo9,Reference Kazemi, Mohammadi and Aghababaee17)} .