In the current study, we performed an MR analysis to investigate the causal association of 25OHD concentration with the risk of COVID-19. Our results indicated that there might be no linear causal relationship of 25OHD concentration with COVID-19 susceptibility and severity.
An observational study based on the UK Biobank data claimed that no link between vitamin D concentrations and risk of COVID-19 infection either overall or separated ethnic groups (16). However, one study used causal inference analysis, supported the hypothesis that vitamin D plays a causal role in COVID-19 outcomes via modification of host responses to SARS-CoV-2 (24). In addition, there were also systematic reviews and meta-analyses to explore the association between vitamin D and COVID-19 (9, 10, 25). These studies provided a biological hypothesis and evolving epidemiological data supporting a role for vitamin D in COVID-19. But these results only based on the observational study design, which may be confounder bias. In our study, the SNPs associated with vitamin D as IVs were used to estimate the overall causal association of 25OHD concentration on COVID-19 susceptibility and severity, based on the MR design. The MR study could potentially avoid many biases and confounding issues existing in conventional observational studies and thus help to identify causally related risk factors. Using MR design, we found no evidence supporting that genetically predicted 25OHD concentration was significantly associated with COVID-19 susceptibility and severity.
There are some possible explanations for these negative findings. First, these null findings suggest that the associations of 25OHD concentration with COVID-19 susceptibility and severity could attribute from the reverse causation bias and confounder bias. Vitamin D from environment with across to adequate sunshine or diet was metabolized in the liver to 25OHD, which was used to determine a patient’s vitamin D status (23, 26). Vitamin D deficiency may be common in COVID-19 patients, as a consequence of quarantined and reduced outdoor behavior. The observed association between 25OHD concentration with increased risk of COVID-19 could be confounded by outdoor behavior which may be corrected with the genetic liability to COVID-19 (23). Our findings suggested that COVID-19 susceptibility and severity are expected to decrease the prevalence of vitamin D deficiency, which are needed to be proved by more bi-directional MR studies. Second, it is also important to note that MR study considers the lifelong effect of genetic modification of COVID-19. However, the association between the vitamin D level and the risk of COVID-19 may be not fixed for a lifetime, but perform time-varying (27). The cross-sectional observational nature of all current MR studies limits the evaluation. The future MR studies incorporating follow-up data should be considered the effect of vitamin D level on COVID-19 and how genetic variants effects change with time may impact the interpretability and validity of their results. Third, as shown by previous studies, vitamin D supplementation only shows treatment effects among individuals with baseline 25OHD concentration of no more than 30 nmol/L, indicating that the relationship between the 25OHD concentration and the risk of diseases may be nonlinear (28–30). However, we noted that there is a linearity assumption in our Mendelian randomization analyses (23), then non-linear relationship could not be tested and might equate to the null hypothesis of no effect of the exposure on the outcome. Therefore, our results indicated that there might be no linear causal relationship of 25OHD concentration with COVID-19 susceptibility and severity.
The evidence of findings from MR studies sit at the interface between observational studies and RCTs (31). RCTs provide interventions for disease, while MR studies could therefore not be directly extrapolated for this purpose, but could rather be used to provide evidence of a causal relationship. It should be also focused on the effect of vitamin D supplementation on COVID-19. Making a trade-off between vitamin D deficiency and treatment aims to create an opportunity for prevention and intervention of COVID-19. The previous findings suggest that vitamin D deficiency and treatment has a long-term effect on preventing overall mortality (11, 32). In addition, future research should pay attention to not only the impact of vitamin D deficiency and treatment on the incidence of COVID-19, but also the impact of vitamin D deficiency and treatment on the COVID-19 mortality and lost life in COVID-19.
Some limitations should be noticed. It is important to note that the results of the MR analyses are based on numerous assumptions. First, we selected genetic variants as IVs based on the recent large-scale GWAS (23), which showed a strong association with 25OHD concentration; therefore, the bias of weak instrument might be less likely. Second, the genetic variants are not associated with measured and unmeasured confounders that influence both vitamin D and COVID-19. However, the unmeasured confounders or alternative causal pathways may be still affected our results because of the limitation of the method. Third, the existence of horizontal pleiotropy may distort MR results. In our study, there was limited evidence of heterogeneity and horizontal pleiotropy. In addition, the GWAS of the severe COVID-19 cases included small sample size, which might lead to small effect for the MR estimate and limit the IVs for COVID-19 for reverse MR analysis. The findings were based on European population, which made it difficult to represent the universal conclusions for other ethnic groups. Therefore, the future studies with larger sample size and more ethnic groups are needed to verify and explore the observed associations.