Vitamin D supplementation in multiple sclerosis: an expert opinion based on the review of current evidence

ABSTRACT

Introduction

Vitamin D has long been known for its immune-modulating effects, next to its function in calcium metabolism. As a consequence, poor vitamin D status has been associated with many diseases including multiple sclerosis (MS). Epidemiological studies suggest an association between a poor vitamin D status and development of MS and a poor vitamin D status is associated with more relapses and faster progression after patients are diagnosed with MS.

Area’s covered

The aim of the authors was to review the role of vitamin D supplementation in the treatment of MS. Pubmed was used to review literature with a focus of vitamin D supplementation trials and meta-analyses in MS.

Expert opinion

There is no solid evidence to support the application of vitamin D therapy, based on current available supplementation trials, although there are some promising results in the clinically isolated syndrome (CIS) patients and young MS patients early after initial diagnosis. The authors recommend further larger clinical trials with selected patient groups, preferable CIS patients and young patients at the time of diagnosis, using vitamin D₃ supplements to reach a 100 nmol/l level, to further investigate the effects of vitamin D supplementation in MS.

KEYWORDS:

• Vitamin D
• Multiple Sclerosis
• 25-hydroxyvitamin D
• vitamin D supplementation

1. Introduction

Multiple sclerosis (MS) is the most common chronic inflammatory disease of the central nervous system. The overall number of people affected by the disease is estimated at more than 2 million people worldwide [1]. The exact cause of MS remains unknown, but it has been proposed to arise from interplay between environmental factors and genetic factors. Most patients (80%) start with episodes of (partially) reversible neurologic deficits called relapses, eventually developing in a gradually progressive course, on average 10–20 years after onset. A small proportion of patients experience progressive disability from disease onset onwards without relapses and remissions, called primary progressive MS (PPMS) [2].

Medications used in MS are immune modulating, which improved the perspective in MS patients; however, MS remains a chronic and incurable disease.

Vitamin D has long been known for its principal function in calcium metabolism, where it stimulates intestinal calcium absorption [3]. Vitamin D is also recognized as an important immune modulating agent and as a consequence a poor vitamin D status is associated with a large number of diseases, including MS [4–6]. In addition, multiple vitamin D related genes have been investigated regarding their role in the development of MS (reviewed by Scazzone et al. [7]).

Most Vitamin D₃ is synthesized in the skin through sunlight (ultra violet B) and a smaller part is obtained via nutrition [4]. Vitamin D₃ is biologically inactive and is therefore first metabolized by members of the cytochrome P450 family (Cyp) enzymes into 25-hydroxyvitamin D (25(OH)D) and is subsequently converted by Cyp27B1 to the biologically active form of Vitamin D₃ 1,25dihydroxyvitamin D (1,25(OH)₂D), also known as calcitriol [reviewed by Peelen et al [4]. In the circulation, 25(OH)D is the predominant metabolite. Excessive amounts of 1,25(OH)₂D are inactivated by Cyp24A1 which is predominantly performed in the kidneys. Interestingly, multiple cell types which are of importance in MS, including T cells, B cells, monocytes, macrophages, dendritic cells, microglia, astrocytes and neurons, can also convert 25(OH)D into 1,25(OH)₂D [8–12]. Furthermore, the expression of vitamin D receptor (VDR), Cyp27B1 and Cyp24A1 in these cell types suggests that the availability and effectivity of 1,25(OH)₂D can be strictly regulated by cells of the immune system at sites of inflammation. Indeed, an up-regulation of the associated enzymes and VDR has been shown in active MS lesions [13]. Furthermore, in vitro, 1,25(OH)₂D attributed to a less inflammatory profile of peripheral blood mononuclear cells (PBMC) [4,14].

The associations found previously between vitamin D and MS and the effects of vitamin D on immune cells and central nervous system (CNS) cells raises the question whether vitamin D supplementation offers a new therapeutic strategy in MS. Here we will review some, selected clinical associations found between vitamin D and MS as well as the latest available, randomized controlled trials (RCT) and meta-analyses on vitamin D supplementation studies in MS followed by our expert opinion on how vitamin D supplementation should be practiced in MS.

2. What is the epidemiological evidence for vitamin D in MS?

Epidemiological evidence for vitamin D in MS can be divided into two categories: research focussing on the role of vitamin D in developing MS and research focussing on the role of vitamin D on disease activity in MS. We will focus on the role of vitamin D on disease activity after patients are diagnosed with MS, however, we will shortly discus both roles here.

2.1. Role of vitamin D in developing MS

Multiple epidemiologic studies suggest an association between a poor sun exposure and/or poor vitamin D status and MS partly based on prevalence rate of MS patients in different latitude areas: MS has a high prevalence rate (> 30 per 100.000) in high latitude areas like north west Europe, southern Canada and northern USA, medium prevalence rate (5–29 per 100.000) in southern Europe, southern USA and Australia, and a relative low prevalence rate (<5 per 100.000) for the rest of the surveyed world [15–21].

Especially very early in life there seems to be a protective effect of exposure to sun and/or vitamin D. In a Norwegian case-control study, cod liver oil use between the ages 13–18 was associated with a reduced risk of MS [22]. In children aged 6–15 years, more hours of sun a day, especially during winter months, was associated with a decreased risk of developing MS [21]. In neonates, an inverse correlation between the neonatal serum 25(OH)D levels and the risk of developing MS later in life has been found [18]. Even a maternal vitamin D deficiency during pregnancy increases the risk of developing MS in the new-born [23].

The epidemiological evidence of a relation between MS and vitamin D is further supported by other studies in adults. Among participants of the Nurses Health Initiatives studies, women that had a high vitamin D intake at baseline, dietary or supplemental, had a lower relative risk on developing MS, compared to women with low vitamin D intake [17]. Other prospective data in military personnel showed an inverse correlation between serum vitamin D levels and the risk of developing MS [24].

2.2. Vitamin D from onset MS diagnosis

Vitamin D levels are described in different phases of MS disease. After patients are diagnosed with MS, lower serum vitamin D levels have been reported in summer and with higher expanded disability severity scale scores (EDSS) compared to healthy controls [20,25–28]. Higher vitamin D levels were related to a lower risk of a relapse and during a relapse serum vitamin D levels in MS patients are lower compared to MS patients in remission [20,25–27]. However, in the largest RRMS onset cohort studied, 25(OH)D levels did not predict long-term relapse rate in patients treated with interferon beta-1b [29]. In addition, an effect of vitamin D status on EDSS progression was also less certain in advanced RRMS or progressive MS patients [29,30].

The associations in longitudinal observational studies should, however, be interpreted with caution, because reverse causality cannot entirely be excluded. It is conceivable that MS patients with a higher EDSS score go less outside and therefore have less sun exposure and thus lower vitamin D status [28]. In addition, sunlight itself is a confounding factor. Since sunlight is the main source of vitamin D synthesis, it should be excluded that previously observed clinical associations with vitamin D are not an epiphenomenon of sun exposure. Indeed, in experimental autoimmune encephalitis (EAE), an animal model of neuroinflammation, sunlight itself has an immune-modulating effect [31,32].

Vitamin D supplementation studies in MS are therefore needed to find out whether these clinical associations are indeed affected by vitamin D itself.

3. How are the effects of vitamin D supplementation investigated?

To this date, there have been multiple trials on vitamin D supplementation in MS with variable endpoints to examine the effects of vitamin D supplementation [33–43](Table 1).

Table 1. Randomized clinical trials of vitamin D supplementation in MS patients with clinical outcome measures

Study	Year	Patients^a	Follow up (weeks)	Choice of supplement and dose	Primary outcomes *	Secondary outcomes*
Burton et al. [33]	2010	49/47	52	D3 40,000 IU vs 4000 IU/day	Mean change in serum calcium	Change in ARR Proportion with relapses
Stein et al. [34]	2011	23/20	26	D2 6,000 IU vs 1,000 IU/day	Number of new Gd+^b T1 lesions Change total volume T2 lesions	Change in EDSS**, number of relapses
Mosayebi et al. [45]	2011	62/59	26	D3 300.000 IU/month vs placebo	Not defined	Change in EDSS Level of serum 25(OH)D, IFN-γ, IL-10, TGF-β^c
Soilu-Hänninen et al. [35]	2012	66/62	52	D3 20,000 IU/week vs placebo	T2 burden of disease on MRI	ARR, changes in EDSS, times 25-foot walk test, timed 10 foot tandem walk test
Shaygannejad et al. [36]	2012	50/50	52	calcitriol 50 mcg/day vs placebo	EDSS progression	change in ARR Relapse free patients
Kampman et al. [37]	2012	71/68	96	D3 20,000 IU/week vs placebo	Bone density	Relapse free patients Change in ARR, EDSS, MSFC components^d, grip strength, FSS^e
Golan et al. [38]	2013	45/30	52	D3 4,370 IU vs 800 IU/day	Serum levels vitamin D, calcium and PTH	EDSS, relapses, quality of life
Achiron et al. [39]	2015	158/143	26	alfacalcidol 1 mcg/day vs placebo	Fatigue impact scale score	quality of life, EDSS progression, relapses, relapse free patients
Sotirchos et al. [40]	2016	40/35	26	D3 10,400 IU vs 800 IU/day	Adverse effects	CD4 T cells, relapses
Hupperts et al. (SOLAR) [41]	2019	229/183	48	D3 14,000 IU/day vs placebo	NEDA-3^b, Proportion of relapse free patients, change total volume T2 lesions	ARR, time to first relapse, confirmed EDSS progression
Camu et al. (CHOLINE) [42]	2019	129/90	96	D3 100,000 IU every other week vs placebo	Change in ARR	Time to first relapse, EDSS progression, change in cognitive abilities, change in quality of life
Dörr et al. (EVIDIMS) [43]	2020	53/38	78	D3 20,400 IU vs 400 IU every other day	Number of new T2-w lesions	Relapse, EDSS, MSFC^d

^arandomized/completed

^bNEDA-3: no relapses, EDSS progression, or combined unique active (CUA) lesions (new gadolinium-enhancing or new/enlarging T2 lesions. Primary outcome was changed during the trial

^cIFN-γ: Interferon-gamma, IL-10: interleukin-10, TGF-β: transforming growth factor-beta

^dMSFC: multiple sclerosis functional composite, composed of timed 25-foot walk test, Nine hole peg test and Paced Auditory Serial Addition Test

^eFSS: Fatigue severity scale

* outcomes that reached statistically significant difference in bold

** After adjusting for entry EDSS, EDSS change was higher in the high dose group (p = 0,05)

Endpoints can be allocated in clinical parameters representing inflammation, i.e. annual relapse rate (ARR), or progression of disability. Next to this, radiological measures of inflammation (gadolinium enhancing T1 lesions) or progression, for example atrophy, are used.

However, patients are usually enrolled in a trial at the time of MS diagnosis when the probability of relapses is high. As time progresses this likelihood of relapses decreases due to the regression to the mean phenomenon. Moreover, relapses may be separated by several years making it a very time consuming endpoint [44]. Time to first relapse seems therefore an appealing alternative to annualized relapse rate. The SOLAR [37] and CHOLINE trial [38] have included this among their secondary outcome measures. In addition, it is suggested that the endpoint time to first relapse requires a lower number of subjects compared to trials using ARR as primary outcome measure [44]. Other possible clinical outcomes for inflammation are portion of patients that remained relapse free, used by Shaygannejad et al. [36], Kampman et al. [37], the SOLAR [41] and CHOLINE [42] trial, or difference in percentage of patients experiencing relapse, used by Burton et al. [33].

Disability or disease progression is often used as primary endpoint in these vitamin D studies (Table 1) [36]. Most used test is EDSS [34–39,41–43,45]. The EDSS score is based on neurological testing and examination of 7 areas of the CNS; pyramidal, cerebellar, brainstem, sensory, bowel and bladder functions, visual and mental function. It is a ordinal scale in half-point increments ranging from 0 (normal neurological examination) to 10 (death due to MS) [46]. Since change in EDSS score in time can take several years, if applicable, variations are used, for example proportion of patients completing a trial with increased EDSS [33,37,41,42]. Other measures for progression used in trials include the functional systems scores (FSS), multiple sclerosis functional composite (MSFC) (composed of the timed 25 Foot Walk test and 9-Hole Peg Test) [35,37,43].

MRI has been widely accepted as a more sensitive biomarker for inflammatory disease activity in MS, compared to relapses. Radiological outcomes of inflammation are most often new or enlarged T2 lesions [35,41–43] or gadolinium enhancing T1 lesions [34,35,41–43,45], both endpoints that are suitable for monitoring MRI activity in MS clinical trials [47]. Other radiological outcomes used in trials are differences in the change of total volume of T2 lesions [34,35,41,43] or new T1 lesions rate ratio [42]. T2 lesion volume is positively correlated with disability measured by EDSS at 2 and 10 years of follow up [48,49] and number of relapses after 2 years of follow-up [48,50].

MRI parameters for progression are sparsely used and include total brain volume, brain gray matter and white matter volume and brain atrophy [42]. Brain volume and atrophy measurements have been shown to correlate with measures of disability [49,51].

Neurofilaments (NFl) are one of the most promising new biomarkers for MS to monitor ongoing axonal injury and neurodegeneration. NFl levels in cerebrospinal fluid (CSF) are suggested to reflect the degree of axonal damage. Cortese et al. demonstrated in 278 clinically isolated syndrome (CIS) patients that each 50 nmol/l increase in mean 25(OH)D levels in the first 2 years after diagnosis was associated with 20% lower NFl-levels and lower odds of poor cognitive performance as measured by paced auditory serial addition test (PASAT) after 11 years [52]. However, no association was found between the natural variation in serum 25(OH)D values and serum NFl-levels [53]. Therefore, we will not further discuss the NFl levels as outcome parameter in this review.

4. Can vitamin D supplementation in MS reduce disease activity?

4.1. Relapses

Observational studies show inconsistent results between serum 25(OH)D levels and subsequent relapse rate. Patients diagnosed with (CIS) with lower serum 25(OH)D levels tend to convert more rapidly to clinically definite MS (CDMS) [54,55]. Simpson et al. and Runia et al. reported an association between lower serum 25(OH)D levels and a lower relapse rate [26,56], although the largest observational study performed by Fitzgerald et al. did not find this association [29]. Muris et al. reported this association only in young patients (<37,3 years) [30], while Mowry et al. reported this in patients with specific genetic risk–allele constitution [57]. Conversely, higher serum 25 (OH)D levels have been reported with a lower relapse rate. Ostkamp et al. found that among patients with higher baseline serum 25(OH)D levels, more patients stayed relapse free compared to patients with low baseline serum vitamin D levels [58]. Ascherio et al. reported that a 50 nmol/l increment in average serum 25(OH)D levels predicted a 57% lower relapse rate [59]. These findings have led to various prospective randomized controlled trials regarding the hypothesis that vitamin D supplementation could lower the relapse rate.

Comparing prospective placebo controlled trials that investigate the effects of vitamin D supplementation in MS is challenging. Most trials are small and on top of that different dosages of vitamin D and supplementation regimes are used. In some trials, for example, patients in the control group are allowed to a daily (lower) vitamin D intake, due to ethical reasons. In the trial by Kampman et al., this led to a considerable group of control patients with 25(OH)D serum levels above 75 nmol/l (24%), even reaching up to 106 nmol/l, likely because control patients were allowed to continue their daily use of supplements [37]. Furthermore, there is a wide spread of duration of the trials (from 26 to 96 weeks).

Considering the outcome measure ARR, multiple trials failed to find a difference in ARR [35,37,41,42]. Shaygannejad et al. did find a change in ARR when patients were given a daily intake of 0.5 µg 1,25(OH)₂D for 12 months [36]. Burton et al. found a trend toward a greater change in ARR in patients receiving high vitamin D supplementation. However, this trial was small (n = 49), unblinded and not powered to study the clinical effects adequately [33].

Results on number of MS patients that remain relapse free after vitamin D supplementation are conflicting. Camu et al. found a larger proportion of patients remaining relapse free (OR 3.24, p = 0.03) [42], as did Achiron et al. (89,5% vs. 67,1%, p = 0.007) [39], while the SOLAR trial found no significant difference in the number of patients that remained relapse-free between the vitamin D supplementation group or placebo group [41].

A meta-analysis by James et al. [60], including 5 trials [33–37], showed that there was no association between treatment with vitamin D supplements and the subsequent relative relapse risk in MS patients. A subanalysis, excluding trials in which control groups were allowed to take a lower dose of vitamin D, also failed to demonstrate an association between vitamin D supplementation and relapse risk [60].

In a more recent meta-analysis by McLaughlin et al. [61], including 12 trials [34–39,41,42,62–64], no statistically significant difference was found in ARR, however there was a trend favoring vitamin D, especially when only placebo-controlled studies were included. However, the authors did pose a warning about prescribing supra-physiological dosages, since two studies reported worse ARR in high-dose groups, compared to low-dose groups [34,38]. These were, however, small studies. Indeed, in one trial results were seriously affected by one patient in the high-dose group that had 3 relapses during the trial [38]. In the other trial by Stein et al [34] vitamin D₂ was supplemented instead of vitamin D₃, which possibly could have affected the results.

A meta-analysis by Zheng et al., including 6 trials [33,35–38,45], found a higher ARR among supplemented MS patients, compared to the control group, in their analysis of 5 of the 6 trials [33,35–38]. The trial by Golan et al. [38] was also included in this analysis, but the largest weight came from the trial by Kampman et al. [37]], in which relapse rates pre- and on-study were overall considerably lower in both treatment and control group, compared to other trials.

A Cochrane review by Jagannath et al [65]., published in 2018, included 12 trials [33–41,45,66,67]. They reported no relapse reduction after vitamin D supplementation, although because of the very low certainty of the evidence, based on the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach, it was advised to interpret the results with caution.

To conclude, until now, trials examining clinical outcomes in vitamin D supplementation in MS failed to show unambiguously effects on disease activity as defined by a reduction in the prevalence of relapses.

4.2. Progression of disease, EDSS progression

The predictive value of serum 25(OH)D levels for risk of EDSS progression in (early) MS is uncertain. In patients with CIS higher serum 25(OH)D levels were associated with lower annual EDSS progression [59]. In patients with RRMS or progressive MS this association was not (convincingly) found [29,30,68]. Among vitamin D supplementation trials, two trials showed slower EDSS progression in the group receiving vitamin D supplements [36,42].

The SOLAR trial found no statistically significant difference between vitamin D supplementation and placebo in time to confirmed EDSS progression at week 48 [41]. Four other trials also found no difference in EDSS progression between vitamin D supplementation or placebo in MS [33,37,43,45]

Kampman et al. [37] investigated the effect on change in MSFC components (composed of the 25 foot timed walk, 9-hole peg test of the non-dominant hand and paced auditory serial addition test), as well as the fatigue severity scale. However, test results did not differ between vitamin D supplemented or placebo treated MS patients. A systematic review and meta-analysis by Hempel et al., investigated the prognostic role of modifiable environmental factors in progressive MS and showed a trend for improved EDSS scores after vitamin D supplementation, although the pooled difference was not statistically significant from placebo [69]. The meta-analysis by McLaughlin et al. showed a trend toward slower EDSS progression with vitamin D supplementation. In the included studies in which a high dose group was compared to a low dose group (instead of a placebo), no therapeutic effect on EDSS progression was found [61]. The meta-analysis by Zheng et al. did not find any therapeutic effect of vitamin D supplements on EDSS progression [70]. A meta-analysis by Doosti-Irani et al [71]., published in 2019, included 6 trials published between 2010 and 2013 [33,35–38,45], reported no effect of vitamin D supplements on EDSS progression. Additionally, the Cochrane review by Jagannath et al. reported no therapeutic effect of vitamin D supplements on EDSS progression, although the certainty of the evidence was classified as very low, complicating interpretation of the results [65].

To conclude, until now, trials examining clinical outcomes in vitamin D supplementation in MS failed to show a clear beneficial effect on disease progression.

4.3. MRI activity

In several longitudinal observational studies, a low serum 25(OH)D level predicts a higher chance on new T2 lesions or active lesions (T1 gadolinium enhancing or new/enlarging T2) on MRI [25,54]. The risk of new/enlarging T2 lesions or T1 gadolinium enhancing lesions on MRI decreased 15–50% for each 25 nmol/l increase in serum 25(OH)D in patients CIS or RRMS. This association seems to be more consistent along the various studies compared to relapse risk [29,59]. In patients with CIS, measures of total brain volume also showed a less marked reduction with higher serum 25(OH)D levels [59]. This association was not found in patients with RRMS [29].

Most of the trials mentioned above in the section relapses also compared the effects of vitamin D supplementation on MRI parameters. MRI parameters used in trials differ widely, making it difficult to compare results (Table 2). Moreover, conflicting results are found.

Table 2. Randomized clinical trials of vitamin D supplementation in MS patients with radiological outcome measures

Study	Year	Patients^a	Follow up (weeks)	Dose	Primary outcome*	Secondary outcomes*
Stein et al. [34]	2011	23/20	26	D2 6,000 IU vs 1,000 IU/day	Number of new Gd+^b T1 lesions Change total volume T2 lesions
Mosayebi et al. [45]	2011	62/59	26	D3 300.000 IU/month vs placebo	Not defined	Number of Gd+ T1 lesions
Soilu-Hänninen et al. [35]	2012	66/62	52	D3 20,000 IU/week vs placebo	T2 burden of disease	Number of Gd+ T1 lesions T1 Gd+ lesion volume New/enlarging T2 lesions Number of patients with T1 Gd+ or new/enlarging T2 lesions
Hupperts et al. (SOLAR) [41]	2019	229/183	48	D3 14,000 IU/day vs placebo	NEDA-3 Proportion of relapse free patients, change total volume T2 lesions	CUA lesions^c New T1 hypointense lesions Percent brain volume change
Camu et al. (CHOLINE) [42]	2019	129/90	96	D3 100,000 IU every other week vs placebo	Change in ARR	T1 lesions rate ratio Total volume T1w(eighted) lesions new or enlarged T1- and T2-w lesions T2-w lesion load Number of T1-w Gd+ lesions total brain volume brain gray and white matter volume
Dörr et al. (EVIDIMS) [43]	2020	53/38	78	D3 20,400 IU vs 400 IU every other day	Number of new T2-w lesions	Change total volume T2 lesions Number of new Gd+ T1 lesions Thalamic volume Brain parenchymal fraction

* outcomes that reached statistical significant difference in bold

^arandomized/completed

^bGd+: gadolinium-enhancing

^cCombined Unique Active lesions, defined as new gadolinium enhancing or new/enlarging T2 lesions

For example, a statistically significant difference in mean number of new gadolinium enhancing (T1) lesions was found by Soilu-Hänninen et al. [35], while the SOLAR trial reported a reduction in combined unique active (CUA) lesions, defined as new gadolinium enhancing T1 or new/enlarging T2 lesions (incidence rate ratio 0.68, p = 0.0045) [41]. Conversely, Soilu-Hänninen et al. and Dörr et al. found no difference in the number of new T2 lesions [35,43]. In addition, no difference was found in T2 lesion count or new gadolinium-enhancing lesions after 18 months [43]. Moreover, no change in total volume of T2 lesion count was found [34,35,43], which was to be the most uniform result among the MRI parameters used. Conversely, Hupperts et al. did find a reduced mean percentage change in total volume of T2 lesions at week 48 in the high-dose group, suggesting a protective effect of high-dose vitamin D on T2 lesion volume increase [41]. As described earlier, T2 lesion volume was correlated with disability measured by EDSS at 2- and 10- year follow-up.

The CHOLINE trial showed a lower T1 lesion rate ratio (p = 0.03) and a larger decrease in the volume of hypointense T1-weighted MRI lesions (p = 0.03), in the group receiving vitamin D supplements [42]. Previously, accumulation of hypointense T1 lesions was significantly correlated to disease progression in progressive MS patients [72].

A meta-analysis by McLaughlin et al. [61] showed no statistically significant difference in new gadolinium enhancing T1 lesions between vitamin D treated MS patients and control groups.

Similar to the results on relapses and EDSS progression, The Cochrane review by Jagannath et al.[65] reported no benefit of vitamin D supplements on the number of MRI gadolinium enhancing T1 lesions, although again the certainty of the evidence was deemed very low.

Concluding, while vitamin D supplements could slow the decrease in total brain volume in patients with CIS, the results in RRMS patients (T1 Gd+ enhancing lesions, total volume T2 lesions, number of new T2 lesions) are conflicting. Vitamin D supplements could possibly benefit RRMS patients regarding these MRI parameters.

5. Conclusion

Though there is increasing evidence indicating low vitamin D levels are associated with increased risk of MS and greater clinical and MRI activity in established MS, based on primary outcome measures current vitamin D supplementation trials fail to support these observations. Until firm establishment of a positive effect of vitamin D supplementation on disease activity in RRMS, there seems to be no final evidence that warrants application of vitamin D therapy, except for maintaining optimal calcium metabolism and bone mineralization. However, secondary outcome measures may indicate positive effects of vitamin D on disease activity in MS. Moreover, the conclusions of published placebo-controlled trials may be hampered by small sample size or unlucky choices of primary outcome measures. Therefore, as elaborated upon in the next paragraph, the quest on the role of vitamin D in MS should not end here.

6. Expert opinion

6.1. Should MS patients receive vitamin D supplementation?

The journey regarding the association between vitamin D and conversion-to or disease activity in MS stretches over more than 50 years. From the first observations regarding the geographical distribution of MS leading to the vitamin D hypothesis, to the more recent results from large prospective vitamin D supplement studies in patients with CIS or MS. Observational studies in MS showed hopeful associations between vitamin D status and disease activity and/or disease progression. However, there is a discrepancy between the observational studies, where a 25 nmol/l increase was associated with a reduced risk of relapse of 14–34%, and the vitamin supplementation trials where an even higher increase in 25(OH)D levels did not show a reduction in disease activity. In addition, data on MRI activity after vitamin D supplementation in MS are conflicting. Therefore, it is our opinion that there is little evidence to support active vitamin D supplementation in MS patients.

This discrepancy between observational studies and supplementation trials might be explained in several ways as already extensively elaborated upon by Smolders et al. [73].

First, it might be that there is indeed no clinical benefit of vitamin D supplementation on disease activity in MS. One could imagine that associations previously found between vitamin D status and disease course are not entirely driven by a direct effect of vitamin D. Sunlight itself, as a confounding factor for vitamin D, may also contribute to this association. UVB exposure has indeed been suggested to be an indirect risk factor for early disease progression [74]. Next, as mentioned before, reverse causality may play a role. Inflammation lowers 25(OH)D levels [75], therefore a subclinical disease activity in MS may lower 25(OH) level by consumption.

Second, present trials may lack results due to current study designs. RCTs on vitamin D have been powered on results from observational studies. Therefore, studies available are relative small and may lack sufficient power to detect small effects of vitamin D supplementation. Furthermore, patients with varying disease duration were included and different doses of vitamin D supplementation were given, making it hard to interpret. But, what would be then a perfect study design to examine vitamin D supplementation in MS? We will further hypothesize on this in the next paragraphs in our expert opinion.

6.2. What are the optimal levels of serum vitamin D? Which dose should be supplemented?

In the general population, persons with a serum 25(OH)D concentration of < 30 nmol/l (< 12 ng/ml) are considered to be vitamin D deficient, while levels of ≥ 50 nmol/l (≥ 20 ng/mL) are called sufficient for practically all people. The committee of the institute of Medicine stated that 50 nmol/l is the serum 25(OH)D level that covers the needs of 97.5% of the population [76]. However, these recommendations are based on the optimal levels for bone metabolism. It might be that in MS patients one should aim on relatively high 25(OH)D levels and thus serum levels >50 nmol/l to show effect on disease activity. Indeed, an inverse association between relapse risk and 25(OH)D levels reached up to 110 nmol/l in an uncontrolled study. Interestingly, above 110 nmol/l this association was lost [77]. Although higher may be better, the risk for side effects such as hypocalcemia, nephrolithiasis, and vascular calcifications which are described at serum levels > 125 nmol/l, should be taken into account. Next to side effects, high vitamin D supplementation induces hypercalcemia, which in turn causes a pro-inflammation state in EAE, the animal model of MS [78]. As described by Smolders et al. these dramatic effects of vitamin D supplementation could not be extrapolated to people suffering from MS [79]. Indeed, in trials with high doses of vitamin D supplementation (ranging from 10.000 IU/w to 14.000 IU/d for 48–96 weeks), no significant adverse events attributed to the high vitamin D supplementation dose were seen [35,37,40–42]. Importantly, the largest vitamin D supplementation trial did not show any convincing evidence supporting one should aim for supraphysiological 25(OH)D levels [41]. Based on these observations, inclusion of patients with vitamin D levels <50 nmol/l, or even <30 nmol/l, and targeting levels of 100 nmol/l would be reasonable.

Another challenge in vitamin D supplementation is that different supplements can be used. Vitamin D3 (cholecalciferol), vitamin D2 (ergocalciferol), alfacalcidol (a vitamin D analog), 25(OH)D (calcediol or calcifediol),or 1,25(OH)2D (calcitriol) can be administered. Indeed, several supplements were used in the supplementation trials described above, in different dosages. In vitro and in animal models, 1,25(OH)2D at high doses had the most profound immune modulating effects [11]. Serum levels of 1,25(OH)2D correlate positively with 25(OH)D. However, 1,25(OH)2D levels have not been extensively correlated with outcomes in MS. The advantage of 25(OH)D or 1,25(OH)2D over vitamin D, would be that the metabolizing steps are circumvented. However, the strict regulation of vitamin D conversion to the active component 1,25(OH)2D could be important in terms of the safety of supplementation. Exploratory studies on 1,25(OH)2D treatment in RRMS reported a good safety profile and signals suggesting potential clinical efficacy [39,80]. However, safety and efficacy of 1,25(OH)2D supplementation remains to be further investigated. Altogether, we suggest vitamin D3 for future supplementation trials.

Lastly, besides genetically en racial differences described in serological response to vitamin D supplements [81], seasonal variation should be taken into account. Due to seasonal variation in UV exposure in different geographical areas, variation in serum levels of 25(OH)D levels are higher in northern parts of Europe and North America compared to more sunny areas. This may influence the dose needed in different seasons. Therefore, a fixed supplementation dose may not be optimal. Although very laborious, in clinical trials serum samples for 25(OH)D can be taken (double blind) and subsequent doses can be adjusted, constantly targeting the 100 nmol/l level.

6.3. Which MS patients should be included in future vitamin D supplementation trials?

When the various studies we described above are considered, certain patient groups are potential candidates to receive vitamin D supplements. CIS patients, young RRMS patients and RRMS patients early in their disease course are possible candidates. As mentioned above, in CIS patients a poor vitamin D deficiency was associated with a faster conversion to CDMS. In an observational study, an association was found between vitamin D status and relapses in previous years, but only in RRMS patients with a disease course <5 years [27]. In a small retrospective longitudinal study among 338 RRMS patients, vitamin D status did not predict the 3-year risk of conversion to SPMS [30]. However, there was an association found between a low vitamin D status measured at the start of disease course and early conversion to SPMS. 25(OH)D serum levels were significantly lower in SPMS patients with a relatively short RRMS duration compared to matched RRMS patients with no conversion to SPMS (38 vs 55 nmol/l, p < 0,01) [82]. Lastly, higher vitamin D serum levels predicted a lower risk of relapses in only the youngest group of RRMS patients (≤37,5 years) [30]. Therefore, we propose that future trials should include CIS patients and, in a separate study arm, newly diagnosed RRMS.

In contrast to CIS patients, supplementing vitamin D as monotherapy in RRMS patients with clinical disease activity is not likely to hold, since there is a potential harm in exposing patients to potentially not efficacious therapy. The largest observational studies showing an association between 25(OH)D and disease activity included predominantly interferon-β treated patients. Previous supplementation trials have mainly been performed in MS patients as add on therapy on interferon-β or glatiramer acetate. However, favorable associations in observational studies have also been published for natalizumab and fingolimod [83–85]. Furthermore, new therapies are continuously entering daily clinical practice. Therefore, these treatments should be considered to study add-on therapy of vitamin D supplementation while taking into account appropriate randomization strategies.

6.4. Which outcomes should be measured in vitamin D supplementation trials?

In a study with CIS patients, time to conversion to definite MS is the most straightforward outcome, but clinical progression after a diagnosis MS is also of interest. In RRMS it is a challenge to capture disease progression or disease activity in a reliable and valid way. Despite its limitations, traditional measures as ARR and EDSS score will probably remain the standard in the near future in MS trials. Taken into account the limitations, a vitamin D supplementation trial should last at least 2 years, but preferentially even as long as 5 years, to assess a reliable ARR and change in EDSS score. In addition, time to first relapse is an interesting additional outcome measure. Different MRI outcome measures should be the number or volume of T2 hyperintensive lesions, T1 gadolinium enhancing lesions and brain volume on MRI. Although data on vitamin D supplementation were not encouraging regarding an effect of vitamin D on brain volume, a relative short follow-up in previous studies should be taken into account. This again underlines the necessity of a longer duration of clinical trials in MS. In addition to established clinical and MRI endpoints, measuring biomarkers as NFl and cognitive performance as measured by PASAT would be of great interest.

Besides the clinically relevant outcome measures mentioned above, a large and long-lasting multicentre study would also enable to obtain more insight in the regulation of vitamin D metabolism and the mechanism of action. In a genome-wide association study, polymorphisms of the genes encoding the main vitamin D-activating and inactivating enzymes (CYP27B1 and CYP24A1) were identified as risk-alleles for MS [86]. Furthermore, a composite score of genetic polymorphisms, leading to a lower vitamin D status, was associated with MS in a case control design [87]. The combination of genetic polymorphisms in the genes encoding the metabolizing enzymes and the (changing) levels of vitamin D metabolites could eventually enhance personalized medicine in applying the optimal dose for vitamin D supplementation in general.

Unraveling the mechanism of action is only of interest if clinical outcome measures are positively affected. It is tempting to speculate that the immune system is involved in the mechanism of action, but despite the clear-cut effects of, in particular, 1,25(OH)2D in vitro and in animal models [4], the results of in vivo human studies are largely disappointing [88,89]. Therefore, it needs to be established how immunological outcome measures are to be defined and analyzed [90]. In our opinion a large and long-lasting multicentre study would benefit from extensive bio-banking for future analyses. Bio-banking should include blood samples for circulating biomarkers, for the presence and function of leukocyte subsets, and for DNA and RNA analyses. Although more of a challenge, the availability of cerebrospinal fluid would be of added value in order to unravel possible effects within the CNS. Finally, stool aliquots should be stored in order to study the effects of vitamin D supplementation on the microbiome [91,92]. The advantage of extensive bio-banking is that all analyses can be performed in a strictly harmonized way at the end of the study. This also enables adaptations in light of increasing knowledge which might be important in the current era of omics and systems biology.

6.5. Five-year perspective

As elaborated on in the previous paragraphs, there is a need for a large, long-lasting multicenter study on the clinical efficacy of vitamin D supplementation in MS. This study should at least cover a 2-year supplementation period, preferably in a selected patient group (for example, CIS patients and young patient at the time of diagnosis), using vitamin D3 supplements that target a 100 nmol/l level. Since there is no initiative yet to design such a study, it is not to be expected that a definitive conclusion will be available within 5 years. At best, the stakeholders would form a task-force to agree on the optimal study design, and, importantly, to tackle the logistic and financial hurdles. Such a task-force would be best organized under the umbrella of the international neurological societies involved in MS research. Awaiting the final conclusion, vitamin D supplementation is to be guided by best practice in clinical care. As such, vitamin D supplementation may be restricted to patients with low 25(OH)D serum levels (<50 nmol/l) and could be targeted to levels of 75–100 nmol/l. This will at least support optimal calcium metabolism and bone mineralization, and additionally, may have beneficial effects on MS disease activity without harmful side effects.

List of abbreviations

1,25(OH)2D	=	1,25dihydroxyvitamin D
25(OH)D	=	25-hydroxyvitamin D
ARR	=	Annual relapse rate
CDMS	=	Clinically definite multiple sclerosis
CUA	=	combined unique active
CIS	=	Clinically isolated syndrome
CSF	=	cerebrospinal fluid
Cyp	=	Cytochrome P450
EAE	=	experimental autoimmune encephalitis
EDSS	=	expanded disability severity scale
FSS	=	functional systems scores/fatique severity scale
Gd+	=	Gadolinium-enhancing
GRADE	=	The Grading of Recommendations Assessment, Development and Evaluation
IFN-γ	=	Interferon-gamma
IL-10	=	Interleukin-10
MS	=	Multiple sclerosis
MSFC	=	Multiple sclerosis functional composite
NEDA	=	No evidence of disease activity
NFl	=	Neurofilaments
PBMC	=	peripheral blood mononuclear cells
PPMS	=	Primary progressive multiple sclerosis
RCT	=	randomized controlled trials
RRMS	=	Relapsing remitting multiple sclerosis
SPMS	=	secondary progressive multiple sclerosis
TGF-β	=	Transforming growth factor-beta
VDR	=	Vitamin D receptor

Article Highlights

• In observational studies low serum 25-hydroxyvitamin D levels is associated with more relapses and more lesions on MRI in patients with CIS and early RRMS.

• Contrary to observational studies, prospective vitamin D supplementation studies are negative on their primary outcomes and conflicting on their secondary outcomes

• The discrepancy may be explained due to reverse causality, confounding factors or differences in study design and outcome measurements between the vitamin D supplementation studies.

Declaration of interest

In the last 3 years, S Knippenberg has received honoraria for lectures and advisory boards from Novartis and Merck. R Hupperts has received institutional research grants and fees for lectures and advisory boards from Biogen, Merck, and Genzyme-Sanofi. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or conflict with the subject matter or materials discussed in this manuscript apart from those disclosed.

Reviewer disclosures

A peer reviewer on this manuscript has received honoraria for consultancy and speaking from Merck Serono, Teva, and Biogen. Peer reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.

Additional information

Funding

This paper was not funded.