High serum concentration of vitamin D may protect against multiple sclerosis

Background High 25-hydroxyvitamin D concentrations have been associated with a reduced risk of multiple sclerosis, with indications of a stronger effect among young individuals. Objective Investigate the 25-hydroxyvitamin D association with multiple sclerosis and test if this association is age dependent. Methods Prospectively drawn blood samples from individuals later developing relapsing–remitting multiple sclerosis and controls matched for biobank, sex, age and date of sampling, were analysed with liquid chromatography tandem mass spectrometry. Results High levels of 25-hydroxyvitamin D (top quintile) were associated with a reduced multiple sclerosis risk (odds ratio 0.68, 95% confidence interval 0.50–0.93). Conclusion These findings further support a role for vitamin D in MS aetiology.


Introduction
Higher serum concentrations of 25-hydroxyvitamin D (25(OH)D) have repeatedly been associated with a decreased risk of multiple sclerosis (MS) development in nested case-control studies [1][2][3] with one study showing a larger effect before 20 years of age. 1 Additional support for a causal role of vitamin D in MS aetiopathogenesis comes from Mendelian randomisation studies 4,5 but this subject remains controversial. 6 this study we aimed to test the hypothesis that high 25(OH)D concentrations reduce the risk of developing MS, with a more pronounced effect among young individuals, by comparing blood samples from healthy controls to samples from individuals who later developed relapsing-remitting multiple sclerosis (RRMS).To achieve these goals, we accessed six Swedish biobanks specifically chosen because they include plasma or serum drawn at a young age.

Case ascertainment
In this nested case-control study we accessed five Swedish microbiological biobanks associated with university hospitals in Umea ˚, € Orebro, G€ oteborg, Ska ˚ne and Link€ oping and one biobank from the Public Health Agency of Sweden (PHAS), to obtain serum or plasma from a total of 670 individuals who later developed RRMS and 670 controls matched for biobank and sex, and with decreasing priority for date of sampling and age.These biobanks contain remainders from serological analysis in routine clinical practice.Cases were identified either through crosslinking with the Swedish MS registry (www.neuroreg.se)or a local MS/possible MS database, and a total of 665 complete sets of cases and controls were included in the final analysis (see Supplementary Figure 1).All samples from MS patients were collected 8 years prior to symptom onset (in median) and all participants were below 40 years of age at the time of sampling.The absolute mean difference between cases and controls was 6 days for the date of sampling and 152 days for the age at sampling.

Laboratory analysis
The concentration of 25(OH)D 3 was analysed using liquid chromatography tandem mass spectrometry (LC-MS/MS) as described previously. 7amples for matched cases and controls were analysed immediately after each other and in random order, and technicians were blinded to case-control status.

Statistical analysis
We modelled 25(OH)D 3 levels as quintiles, derived from the distribution among controls, separately for each biobank as the levels differed significantly between them (Kruskal-Wallis test P 0.001).These quintile assignments were used in a pooled analysis that included all individuals.The Mann-Whitney U-test was used to test differences between groups, odds ratios (ORs) and P for trend over quintiles were calculated using conditional logistic regression.Age was stratified into three groups based on the age at sample draw, less than 20, 20-29 and 30-39 years of age.If a case-control

Ethical considerations
This study was approved by a local regional ethical review board in Umea ˚(2011-198-31M).No written informed consent was required for participation.

Results
Median 25(OH)D 3 did not differ between cases and controls (Table 1).Being in the top 25(OH)D 3 quintile was significantly associated with a decreased risk of MS in the total cohort (OR 0.68, 95% confidence interval (CI) 0.50-0.93)(Table 2).A sensitivity analysis excluding the PHAS biobank, which had higher levels compared to the others, yielded an OR of 0.69 (95% CI 0.49-0.97)when using the median cut-off for the remaining biobanks (72 nmol/L).Subgroup analyses in different age strata were not significant and we found no trend over 25(OH)D 3 quintiles.

Discussion
Although cases and controls did not significantly differ in median levels of 25(OH)D and there was no significant trend over quintiles, we did find a decreased MS risk among individuals with concentrations in the top quintile.These findings suggest In the total cohort, P for trend over quintiles was 0.24.
that there may exist a threshold located within the higher range of 25(OH)D levels (cut-off 70-82 nmol/L in the six biobanks) above which the effect of 25(OH)D modulates MS risk.This is in line with the findings of one earlier study using 75 nmol/L as a cut-off. 2 Data from the currently largest pre-symptomatic study, performed in a Finnish maternity cohort, 3 seem to indicate that seasonally corrected levels above 50 nmol/L are protective when compared to less than 30 nmol/L.In that study, 6% of cases and 7.5% of controls were above 50 nmol/L, compared to 54.6% and 56.1% in our study.Although we found higher vitamin D levels, they are in line with previously published population-based studies in our region. 8ifferences in methodology, including 25(OH)D assay and the use of seasonal correction of multiple samples from each individual in the Finnish study, may explain some of the differences in absolute 25 (OH)D and a direct comparison between the studies may therefore be inappropriate. 9strength of our study is the relatively large number of individuals below 20 years of age, enabling comparisons of different age strata.These analyses did not yield any significant findings, however, but the effect sizes converge with earlier studies. 1,2Furthermore, this is to our knowledge the first study applying the gold standard method LC-MS/MS.
The main limitation in our study is that the samples came from six unique biobanks, with different preanalytical procedures and geographically distinct catchment areas, both of which may influence the results as well as provide a geographical explanation of why serum concentrations of vitamin D differed between biobanks.To minimise this, we matched cases and controls from the same biobank and defined quintiles separately for each biobank.Pooling of site-specific quintiles has been used previously 10 and enabled analysis of the total cohort by applying a similar relative cut-off (i.e.top quintile), despite the biobanks representing a heterogenous material.Also, we did not have access to data on race/ethnicity and the results may therefore not be generalisable to other populations.In addition, the retrospective compilation of data may have implicated other biases affecting the results that we have not considered.
In conclusion, our results further support the hypothesis that relatively higher 25(OH)D concentrations may protect against the development of MS but not that the effect is stronger among young individuals.

Table 1 .
Characteristics of cases and controls.different sides of an age cut-off they were assigned to either the youngest or oldest group containing either a case or control, in order to increase power in the smaller groups.IBM SPSS statistics version 23 (IBM Corporation, New York, NY, USA) was used for statistical analysis.

Table 2 .
Associations of vitamin D 3 concentration and MS stratified by biobank and age.
MS: multiple sclerosis; OR: odds ratio; CI: confidence interval; PHAS: Public Health Agency of Sweden.a