Ang Zhou and Elina Hyppönen
Key Points
-
Genetic research from the University of South Australia shows a direct link between low levels of vitamin D and high levels of inflammation, providing an important biomarker to identify people at higher risk of or severity of chronic illnesses with an inflammatory component.
-
Findings suggest that boosting vitamin D in people with a deficiency may reduce chronic inflammation.
Introduction
Systemic low-grade inflammation, characterized by prolonged release of inflammatory mediators and activation of harmful signal-transduction pathways, is associated with various complex somatic and neuropsychiatric diseases and disorders.
It is considered that nutritional factors can influence many aspects of inflammation. Vitamin D is a pro-hormone and an essential micronutrient, and although its classical roles are related to the regulation of calcium homeostasis, various types of immune cells express both the vitamin D receptor and metabolizing enzymes, suggesting that hormonal vitamin D could also play a role in modulating inflammatory responses. This is supported by an inverse association between serum 25-hydroxyvitamin D [25(OH)D] concentrations and C-reactive protein (CRP), which is frequently reported in observational studies.
Serum 25(OH)D is the best indicator for vitamin D status whereas CRP is one of the most widely used inflammatory biomarkers in clinical practice. However, there is an ongoing debate about the causal nature of the association between 25(OH)D and CRP, and the observational association has not been supported by randomized trials. Indeed, it has been suggested that the association between serum 25(OH)D and CRP simply reflects reverse causality or confounding, where a low 25(OH)D concentration is either a consequence of chronic inflammation or results from behaviours such as less time outdoors in people who are unwell.
Sitting at the interface between observational studies and randomized–controlled trials (RCTs), Mendelian randomization (MR) has been increasingly used to strengthen causal evidence in observational studies. It uses genetic variants associated with the exposure of interest to approximate the exposure and, conditional on the key method assumptions being met, MR has the benefit of reducing bias due to confounding and reverse causation.
The association of 25(OH)D and CRP has previously been investigated using the MR approach, with no evidence to support a causal effect. However, all previous studies have only used the standard linear MR method, which cannot rule out the possibility of a threshold effect restricted to vitamin D deficiency. Indeed, it is logical to expect that improving vitamin D status would be relevant only in the presence of vitamin D deficiency, whereas any further additions may be redundant and, in the high extreme of supplementation, might become toxic. These types of non-linear dose–response relationships can be tested by the non-linear MR approach, which allows us to interrogate the shape of the association.
This method has been recently used to provide evidence for the adverse effect of vitamin D deficiency on cardiovascular disease (CVD) risk and mortality, which is not visible using the standard linear MR approach. In this study we set out to examine evidence for the direction and causality of the association between serum 25(OH)D and CRP, also allowing for possible threshold effects. We performed these analyses using data from 294?970 participants in UK Biobank, representing the largest cohort to date with measured serum 25(OH)D concentrations.
Discussion
In this large-scale genetic analysis, we observed evidence for a causal effect of vitamin D status on CRP with no support for CRP as a determinant of 25(OH)D concentrations. The association between 25(OH)D and CRP was largely restricted to the deficiency range, where only individuals with low serum 25(OH)D concentrations have elevated serum CRP. The shape of the observed association supports the previously proposed threshold effect, suggesting that correction of vitamin D deficiency in the affected individuals is likely to reduce systemic low-grade inflammation and potentially mitigate the risk or severity of chronic illnesses with inflammatory components.
The earlier RCTs or MR studies have failed to provide evidence for an effect of vitamin D on CRP, which seemingly contradicts our finding. However, if the causal effect of vitamin D is truly L-shaped and restricted to concentrations within the deficiency range as seen in our study, it would have been overlooked both by the existing supplementation trials and linear MR studies. Severe deficiency is relatively rare and as it is unethical to subject participants to undue harm, supplementation trials often can only include individuals who are already vitamin D replete, rendering the health effect of supplementation in the deficiency range largely unassessed.
Previous linear MR studies would have had limited statistical power to capture the threshold effect restricted to the deficiency range, which is indeed what was reflected in our linear MR analysis. We also found no evidence for an effect by genetically instrumented CRP on serum 25(OH)D concentration. This provides evidence against the notion that both molecules serve as acute phase reactants and that their association arises merely from confounding by inflammation. Indeed, if inflammation truly drives low serum 25(OH)D concentrations, we would have expected genetically instrumented CRP values to be associated with serum 25(OH)D concentration.
Overall, our bidirectional MR analyses suggest that rather than vitamin D acting as a bystander [where 25(OH)D-CRP association arises merely from confounding by inflammation], increasing 25(OH)D concentrations to alleviate a state of severe deficiency may help to mitigate the severity of inflammation. This said, it is important to note that these findings provide no support for a need to use high-dose vitamin D supplementation, as the observed benefits appeared to become largely saturated by the time 25(OH)D concentrations reach 50 nmol/L. It should also be noted that a higher vitamin D status may benefit some subpopulations or with respect to other disease outcomes, which is an area warranting further investigation.
Vitamin D is a pro-hormone. Its anti-inflammatory property captured by our analysis could be mediated through its hormonal effect on vitamin D receptor-expressing immune cells, such as monocytes, B cells, T cells and antigen-presenting cells. Indeed, cell experiments have shown that active vitamin D can inhibit the production of pro-inflammatory cytokines, including TNF-αα?, IL-1ββ?, IL-6, IL-8 and IL-12, and promote the production of IL-10, an anti-inflammatory cytokine. Further, the anti-inflammatory effect also raises the possibility that having adequate vitamin D concentrations may mitigate complications arising from obesity and reduce the risk or severity of chronic illnesses with an inflammatory component, such as CVDs, diabetes, autoimmune diseases and neurodegenerative conditions, among others.
If the related effects are indeed true, given the high prevalence of serum 25(OH)D levels of <50 nmol/L across the world (≤40% in some European countries) population-wide correction of low vitamin D status (e.g. by food fortification) could potentially be a cost-effective measure to reduce the burden of chronic disease. In fact, in linear MR analyses higher 25(OH)D concentrations have been associated with a lower risk of type 2 diabetes and multiple sclerosis (a chronic inflammatory disease of the central nervous system), with recent non-linear MR analyses providing evidence that correction of vitamin D deficiency can decrease the risk for CVDs and all-cause mortality.
To the best of our knowledge, this is the first non-linear MR study to explore the bidirectional association between serum 25(OH)D and CRP concentrations. We used several strategies to ensure that our MR analyses are not affected by horizontal pleiotropy, where variants may influence the outcome through pathways other than through the exposure of interest.9 First, we restricted our vitaminD-GS to 35 variants with robust replicated evidence for an association with serum 25(OH)D concentrations.
Second, we confirmed that the L-shaped association of serum 25(OH)D and CRP was consistent between the non-linear and stratified MR analyses, with the inverse association of serum 25(OH)D and CRP in the deficiency range consistently observed across pleiotropy-robust methods with largely independent assumptions on the pattern of pleiotropy.
Third, the L-shaped association of serum 25(OH)D and CRP was confirmed to be robust using a spectrum of alternative instruments, including when using a broader set of variants, when excluding variants related to lipids and other metabolic traits and when restricting the set of variants to those directly related to vitamin D metabolism.
Despite these strengths, our study also has some limitations. Although CRP is a widely used inflammatory biomarker, it certainly cannot capture the full complexity of the immune system and hence investigation of more specific biomarkers (such as TNF-αα and IL-6) is required to provide a more detailed understanding on the anti-inflammatory effects of hormonal vitamin D. We restricted our analysis to participants of White-British descent to minimize bias due to population stratification; however, this may limit the transferability of our findings to other ethnic groups.
As with all MR studies, genetic instruments approximate the average effects over the life course and the true biological association between serum 25(OH)D and CRP may be more complex than that indexed in our study. With only a 5% response rate at the recruitment, UK Biobank is not representative of the general public in the UK despite its large sample size. It is uncertain to what extent this selection could affect the non-linear MR analysis. However, given that risk factor–disease associations show close agreement between UK Biobank and nationally representative studies and that an earlier publication from UK Biobank using the same non-linear MR approach has replicated the expected J-shaped association between BMI and mortality, this lack of representativeness may not be affecting our findings.
Conclusion
In conclusion, using a large population-based cohort, we provide genetic evidence for an L-shaped association of serum 25(OH)D with CRP, suggesting that the benefit of increasing 25(OH)D is restricted to individuals with low vitamin D status. Our finding suggests that improving vitamin D status in the deficiency range could reduce systemic low-grade inflammation and potentially mitigate the risk or severity of chronic illnesses with an inflammatory component.
Original source: https://academic.oup.com/ije/advance-article/doi/10.1093/ije/dyac087/6586699?login=false
