Can vitamin D reduce inflammation in relapsing-remitting multiple sclerosis?

Epidemiological studies have pointed out vitamin D as one of the main suspects in the etiology of multiple sclerosis (MS), along with Epstein–Barr virus infection and smoking [1]. There is a plausible biological rationale for a clinically relevant anti-inflammatory effect of vitamin D in MS. Vitamin D exerts anti-inflammatory effects in vitro and in vivo, including in the experimental autoimmune encephalitis model that reflects many inflammatory aspects of MS. Experimental evidence also suggest that vitamin D could affect remyelination, independent of its effect on inflammation [2]. From a clinical point of view, the evidence linking vitamin D with disease activity is even more intriguing, as it indicates that relapsing–remitting MS (RRMS) could be treated by a cheap and safe intervention. This evidence primarily comprises several studies showing a quite consistent association between vitamin D status and inflammatory disease activity, as measured by relapses [3] and by magnetic resonance imaging (MRI) [4,5]. Most association studies show approximately 10% reduction of MRI and relapse rate for each 10 nmol/l increment in the serum concentration of 25-hydroxyvitamin D (25(OH)D). Thus, given that these associations reflects causality, raising the concentration of 25(OH)D from 50 nmol/l to 125 nmol/l should have an effect on disease activity in the same range as conventional disease-modifying drugs.

So why is not the role of vitamin D in MS settled – why do we still ask whether vitamin D exert relevant anti-inflammatory effects in MS?

No suspect – not even vitamin D – should be deemed guilty based on association alone. Although experimental studies support an anti-inflammatory effect and possibly also an effect on remyelination, the interpretation of such studies are limited by constraints such as the frequent use of supraphysiological doses of hormonal 1,25(OH)₂D. One illustrative example is that although 1,25(OH)₂D has been shown to inhibit B cell proliferation and isotype switching in vitro, vitamin D3 supplementation in MS patients did not have corresponding effects [6].

The most obvious potential confounding factor in association studies of vitamin D and RRMS is that 25(OH)D is a proxy for UVB radiation from sunshine, which can induce other immunological effects than synthesis of vitamin D [7]. One possibility is UV-mediated induction of cis-uranoic acid, which was recently demonstrated to induce CD4⁺CD25⁺FoxP3⁺ regulatory T cells and to reduce antigen-presenting capacity of dendritic cells [8].

It was earlier reported that both MRI activity in RRMS patients and serum levels of 25(OH)D display opposite sinusoidal seasonal fluctuation [9]. Accordingly, we found that the association between 25(OH)D and inflammatory MRI activity was entirely attributable to seasonal fluctuation [5]. It could thus in theory be explained by any factor with the same seasonal fluctuation as vitamin D. Another example illustrating what could be called an ‘association/causation paradox’ is that whereas an observational study showed a significant association between serum levels of 25(OH)D and the inflammation markers interleukin 1 receptor agonist and secreted frizzle-like protein [10], neither of these was affected by vitamin D3 supplementation in a randomized placebo-controlled trial [11].

Sufficiently powered randomized clinical trials are the obvious answer to these concerns. It is generally accepted that the trials published so far have been too small and methodologically divergent to be conclusive [12]. Thus, one study comprising 68 patients randomized to placebo or 20,000 IU vitamin D3 per week for 2 years were completely negative, although mean serum 25(OH)D more than doubled to 121 nmol/l in the intervention group [13]. Conversely, another study of comparable size showed a significant and pronounced effect of 14,000 IU vitamin D3 per week on new gadolinium-enhancing lesions after 1 year [14]. The first study included only clinical end points, and the pre-study relapse rate was only 0.11 [13]. The study was therefore obviously underpowered to detect any realistic effect on relapses of any intervention, including vitamin D.

How should a future study on vitamin D supplementation in RRMS be designed? One lesson from previous studies, and from the emerging evidence of other immunomodulatory effects of UVB than induction of vitamin D synthesis, is that the effect size of vitamin D might be smaller than indicated in association studies. Moreover, any intervention study of vitamin D in RRMS will have to be an add-on to a conventional immunomodulatory drug with a proven effect on the outcome measures that are relevant for vitamin D. A brief look at observational studies in patients treated with interferons underscores the need for large sample sizes. Thus, one study of 1482 RRMS patients followed for 2 years found lower rates of new MRI activity with increasing 25(OH)D levels, with the lowest rate of new lesions (relative risk = 0.53) for those with 25(OH)D exceeding 100 nmol/l [15]. In another study 25(OH)D was associated with time to relapse or MRI activity in the 97 patients receiving interferon-beta (hazard ratio = 0.58), but not for those 151 receiving glatiramer acetate (hazard ratio = 0.89) [16]. In 88 untreated RRMS patients who all started on interferon-beta at Month 6, new MRI activity was negatively associated with 25(OH)D before, but not after, introduction of interferon-beta [5]. In fact, interferon-beta treatment reduced MRI activity to the same low levels regardless of 25(OH)D status [17].

The optimal end point should be able to capture an effect on inflammatory disease activity within a reasonable time frame. Although time to conversion to secondary progressive MS is obviously clinically relevant [18], it is not well defined in a clinical setting and would require very long follow-up. In sum, an add-on study will need a large number of patients and should incorporate MRI activity as an outcome. We also suggest that future trials should focus on patients early in the disease course, who have more inflammatory disease activity. Notably, a recent study showed that increasing 25(OH)D levels are associated with low disease activity also in patients with clinically isolated syndrome (CIS), the first clinical symptom suggestive of demyelinating disease [18]. A similar trend was detected in a multicenter study, where it was found that CIS patients with the lowest 25(OH)D levels had a greater risk of converting to MS compared with those in the upper three quartiles [19]. We therefore suggest a randomized controlled trial in CIS patients, with time to conversion to definite MS, either through a new clinical attack or a new MRI lesion, as the primary outcome. The study should be restricted to patients with 25(OH)D below 75 nmol/l at baseline, and the vitamin D3 dose should be sufficient to increase mean 25(OH)D levels to about 125 nmol/l.

It is also important to establish whether any effect of vitamin D differs between patients on different immunomodulatory drugs. To address this question, we urge the pharmaceutical companies that have not already done so to open databases and biobanks from the large phase III trials for such studies. This is particularly important for the new oral drugs and monoclonal antibodies, which will likely be used by many newly diagnosed patients, who may potentially benefit from vitamin D supplementation.

What should clinicians do while waiting for the answers from well-designed and adequately powered clinical trials? Patients will often ask for an advice, so neglecting the topic may not be an alternative. In the lack of high-level evidence, we find it reasonable to emphasize that most epidemiological and observational studies conclude that vitamin D levels in the upper physiological range, approximately 100–125 nmol/l, are associated with the lowest MS risk and disease activity. Such levels are not toxic, and are very close to that in hunter-gatherers living in East-Africa [20], where the human population originated. Notably, they are also associated with good bone health. The latter point should not be neglected, as RRMS patients are prone to develop osteoporosis and high fracture risk [21]. Until randomized clinical trials eventually prove otherwise, we would therefore suggest to aim for a serum level of 25(OH)D of 100–125 nmol/l in all MS patients who ask for an advice. It should be underscored that this should not be regarded as a proven fact, but as a pragmatic attempt to be as specific as possible based on the incomplete knowledge available.

Financial & competing interests disclosure

Trygve Holmøy has received speaker honoraria, unrestricted research grants and travel support form Sanofi-Aventis, Biogen-Idec, Merck Serono, Novartis, Bayer and Genzyme. Øivind Torkildsen has served on scientific advisory boards for Biogen Idec, Genzyme and Merck-Serono and received speaker honoraria and travel grants from Genzyme, Merck-Serono, Novartis, and Biogen-Idec. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.