Association of 25 hydroxyvitamin D concentration with risk of COVID-19: a Mendelian randomization study

Introduction

Coronavirus disease 2019 (COVID-19) has become a global pandemic ¹. Aimed at delaying disease onset by modulating modifiable risk factors, primary prevention has been proposed as a potentially effective and feasible tool to address the global challenge posed by COVID-19.

Vitamin D is a hormone, produced in the skin during exposure to sunlight, helped regulate the amount of calcium and phosphate in the body, which are needed to keep bones, teeth, and muscles healthy, and played a critical role in the immune system ^2-4. Vitamin D deficiency is a candidate risk factor for a range of adverse health outcomes, such as cancer, cardiovascular diseases, metabolic disorders, infectious diseases, as well as autoimmune diseases ^5,6. There were well recognized evidences that vitamin D deficiency contributes to the seasonal increase of virus infections of the respiratory tract, from the common cold to influenza, and now possibly also COVID-19 ^7,8. Some systematic reviews indicated that vitamin D deficiency may increase infection risk of COVID-19 by discussing the biological mechanism of virus infection^4,9,10. In addition, numerous population-based studies have evaluated the Vitamin D deficiency in COVID-19 patients relative to controls; however, the findings have been inconsistent. Some studies showed that COVID-19 patients had a lower 25 hydroxyvitamin D (25OHD) concentration compared with healthy controls ^7,11-13, while others found no association ^14-16. The inconsistent findings from these epidemiological studies may be due to differences in the study design, study population, assays used for measuring Vitamin D deficiency. It is important to note that as the nature of the above studies is observational, such traditional epidemiological studies are particularly vulnerable to reverse causality and residual confounding.

A promising approach known as Mendelian randomization (MR), which uses inherited genetic variants as instrumental variables, provides stronger evidence for the causal effect of exposure on the diseases largely overcoming the traditional limitations due to confounding and reverse causality ^17,18. Summary based MR is an excellent strategy to evaluate the causality using summary statistics from Genome Wide Association Study (GWAS) data^19,20. Therefore, we conducted MR analysis to assess whether 25OHD concentration was causally associated with risk of COVID-19.

Methods

Results

As shown in Table 1, the MR analysis showed no significant association of genetically instrumented 25OHD concentration with COVID-19 using 89-SNP (OR, 1.168; 95% CI, 0.956-1.427; P=0.128). The association of 25OHD concentration with COVID-19 was robust in the weighted median and penalized weighted median methods, except in the MR-Egger regression (OR, 1.302; 95% CI, 1.011-1.676; P=0.044). Pleiotropy bias (P for MR-Egger intercept=0.174) and heterogeneity (P=0.656) were also not observed. In terms of various estimates for different measures, we reported the results of the IVW method. In addition, the “leave one out” results showed that by omitting the included 89 SNPs one at a time, no individual genetic variants seem to have any significant effect on the overall results (Figure 1). The association of 25OHD concentration with COVID-19 remained robust using 57-SNP instrument (Table 1, Figure S1).

• Download figure
• Open in new tab

Figure 1. MR leave-one-out sensitivity analysis for ‘25OHD concentration’ on ‘COVID-19’

Leave-one-out analysis: each row represents a MR analysis of 25OHD concentration on COVID-19 using all instruments expect for the SNP listed on the y-axis. The point represents the odds ratio with that SNP removed and the line represents 95% confidence interval.

COVID-19: Coronavirus disease 2019; MR: mendelian randomization; SNP, single-nucleotide polymorphism; 25OHD: 25-hydroxyitamin D

The MR analysis showed no significant association of genetically instrumented 25OHD concentration with severe COVID-19 using 82-SNP (OR, 0.889; 95% CI, 0.549-1.439; P=0.632) (Table 2). The result in the weighted median, penalized weighted median, and MR-Egger regression methods were robust. There was limited evidence of heterogeneity and horizontal pleiotropy based on the Q test (P=0.793) and MR-Egger intercept test (P=0.760). Sensitivity analyses using different instruments yielded similar findings, suggesting the robustness of the causal association (Table 2, Figure 2, Figure S2).

• Download figure
• Open in new tab

Figure 2. MR leave-one-out sensitivity analysis for ‘25OHD concentration’ on ‘severe COVID-19’

Leave-one-out analysis: each row represents a MR analysis of 25OHD concentration on severe COVID-19 using all instruments expect for the SNP listed on the y-axis. The point represents the odds ratio with that SNP removed and the line represents 95% confidence interval.

COVID-19: Coronavirus disease 2019; MR: mendelian randomization; SNP, single-nucleotide polymorphism; 25OHD: 25-hydroxyitamin D

Discussion

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-2²³. In addition, there were also systematic reviews and meta-analyses to explore the association between vitamin D and COVID-19 ^9,10,24 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 ^22,25. 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 ²² 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, the association between the vitamin D level and the risk of COVID-19 may be not fixed for a lifetime, but perform time-varying ²⁶. However, 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 ^27-29. However, we noted that there is a linearity assumption in our Mendelian randomization analyses ²², 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 ³⁰. Of the three study designs, RCTs provide interventions for disease, while observational studies and MR studies demonstrate the risk factors that increase the risk of disease. It should be focused on the effect of vitamin D supplementation on COVID-19. The previous findings suggest that vitamin D deficiency and treatment has a long-term effect on preventing overall mortality ^12,31. 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 ²², 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 COVID-19 host genetics initiative from UK Biobank individuals 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.

Conclusion

Using 25OHD concentration-related SNPs as IVs from GWAS data, the MR analysis results indicated that there might be no linear causal relationship of 25OHD concentration with COVID-19 susceptibility and severity. In future, the bi-directional MR and non-linear MR study was needed to further prove these results. In addition, we should pay more attention to the randomization control trials about association between vitamin D treatment and the improvement of the COVID-19 in the long-term benefits.

Availability of data and materials

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Funding/Support

The study was supported by grants from the National Natural Science Foundation of China (81872682 and 81773527), and the China-Australian Collaborative Grant (NSFC 81561128020-NHMRC APP1112767).

Role of the Funder/Sponsor

The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.