Understanding how existing vitamin D levels and vitamin D exposure affect clinical relapses and MS lesion activity is critically important to this review. As such, the findings from larger studies investigating these effects are summarized below.
Impact of Vitamin D Levels on Disease Activity in RRMS: Observational Studies
In a prospective longitudinal study from the Netherlands, 25(OH)D was measured every 8 weeks for a mean of 1.7 years in 73 patients with RRMS . Fifty-eight patients experienced a total of 139 exacerbations during the study period. Relapse risk was significantly reduced in those with medium [50–100 nmol/L (20–40 ng/mL)] and high [> 100 nmol/L (> 40 ng/mL)] serum vitamin D levels (vs. < 50 mol/L or 20 ng/mL) compared to those with low levels . For each doubling of serum vitamin D concentration from baseline of 10, 20, 30, MS relapse risk decreased by 27%. Although this suggests a beneficial effect of vitamin D on MS, it must be noted that there is also a possibility that conditions associated with MS relapse had an effect on serum vitamin D levels .
Incident rate ratios (RR) for relapse in relation to serum vitamin D levels were measured in a retrospective study of 110 patients with pediatric-onset MS . After adjusting for several factors (age, gender, race, ethnicity, disease duration, and treatment), the authors found that every 10 ng/mL (25 nmol/L) increase in 25(OH)D level was associated with a 34% decrease in relapse risk. Similar findings were seen in a prospective cohort study from Tasmania, Australia, in a group of 145 adults with RRMS, in which 25(OH)D levels were measured twice a year for a period of 3 years . For each 10 nmol/L increase in serum vitamin D level, there was an associated 12% lower risk of MS relapse. Adjustment for potential confounders, such as timing of the blood testing, did not affect the results. Most of the participants in this study (82%) were receiving immunomodulatory therapy. The authors concluded that raising 25(OH)D levels by 50 nmol/L could decrease the hazard of a relapse by up to 50% (Fig. 4) .
The EPIC natural history study was a 5-year cohort study conducted at the University of California, San Francisco which sought to determine the associations between serum or plasma vitamin D levels and MRI activity in a group of 469 white, mostly non-Hispanic patients with MS or clinically isolated syndrome (CIS) . Sixty-four percent received disease-modifying therapy within the previous 12 months. Vitamin D levels increased significantly during the study, especially for those patients using supplements. Only 9% of patients were taking vitamin D supplements at baseline, but 43% were taking them by year 5. Patients who reported using vitamin D supplements had an 8.7 ng/mL (21.75 nmo/L) higher vitamin D level, on average, compared with those who did not. Additionally, lower vitamin D levels were strongly associated with development of new T2 lesions and with contrast-enhancing lesions on brain MRI. Each additional 10 ng/mL (25 nmol/L) increment of 25(OH)D was associated with a 15% lower risk of new T2 lesions and a 32% lower risk of enhancing lesions (Fig. 5) . Higher vitamin D levels were associated with a lower (but not statistically significantly) risk of MS relapses. Findings from this study also showed strong “within-person” effects of vitamin D levels in individual patients with MS. The authors concluded that “individuals with CIS/RRMS with higher vitamin D levels are at much lower risk of the subsequent development of new lesions and of gadolinium (Gd+)-enhancing lesions on brain MRI, even after accounting for potential confounding factors” .
More recently, Mowry and colleagues examined the association of vitamin D levels with brain volume measures and new lesions in patients with CIS (N = 65) . The scientific rationale for these data are based on the concept that brain volume is thought to reflect neurodegeneration better than classical MRI parameters such as T2 lesion load and Gd+-enhancing lesions . Each 25-nmol/L increase in 25(OH)D level was associated with 7.8-mL higher gray matter volume (P = 0.025). Higher levels of 25(OH)D also were associated with the composite endpoint of ≥ 3 new brain T2 lesions or ≥ 1 relapse within 1 year (P = 0.096). Despite the limitations of the small sample size, these findings suggest that higher vitamin D levels in CIS may slow neurodegeneration .
Lower vitamin D levels also correlate with other surrogates of MS disease activity, including lower odds of remaining relapse free in MS , greater disability and disease severity in MS [53, 55,56,57], conversion from CIS to clinically definite MS (CDMS) , and poorer nonverbal long-term memory performance . These data were largely generated by observational studies that restrict the extent to which inverse associations can be attributed specifically to vitamin D. Properly designed and conducted clinical trials are needed to further define the nature of this association.
Impact of Vitamin D Levels on Disease Activity Based on Post-Hoc Analyses from BENEFIT and BEYOND
To our knowledge, no large randomized, double-blind, controlled, prospectively phase 3 trials have been conducted to study the impact of vitamin D levels on MS activity as a primary endpoint. However, in two phase 3 studies, the BENEFIT study , and the BEYOND study  post hoc analyses were conducted to investigate this potential link.
The BENEFIT (Betaseron® in Newly Emerging Multiple Sclerosis for Initial Treatment) study was a randomized trial originally designed to evaluate the impact of early versus delayed IFNB-1b treatment in patients with CIS [62,63,64]. Patients with a first event suggestive of MS and a minimum of two clinically silent lesions on MRI were randomly assigned to receive interferon beta-1b (IFNB-1b) 250 μg (n = 292; early treatment) or placebo (n = 176; delayed treatment) subcutaneously every other day for 2 years or until diagnosis of CDMS, in which case they could switch to IFNB-1b therapy. All patients were then eligible to enter a prospective follow-up phase with open-label IFNB-1b for up to 5 years after randomization. Patients and study personnel remained unaware of initial treatment allocation throughout the study up to year 5. During the observation period, regular study visits were scheduled to collect clinical and MRI data, with visits at baseline and months 3, 6, 9, 12, 18, 24, 36, 48, and 60 . A post hoc analyses aimed to determine whether vitamin D status [serum 25(OH)D levels] would predict disease activity and prognosis up to 5 years after the first attack in early-disease CIS patients . Serum samples were collected at baseline, 6, 12, and 24 months and levels of 25(OH)D were measured (by ELISA). Of the 468 patients included in BENEFIT, 465 patients had at least one 25(OH)D measurement, 417 had two or more, 396 had three or more, and 303 had all four measurements. 25(OH)D levels were seasonally adjusted to obtain an estimate of long-term 25(OH)D status. To minimize the possibility that lower 25(OH)D levels were a consequence, rather than the cause, of MS severity, the cumulative average 25(OH)D levels at 12 months were related to the outcomes between 12 and 60 months or between 24 and 60 months (thereby allowing inserting a 1-year lag between 25[OH]D measurements and the assessment of MS activity or progression) . Three sets of analyses were performed: (1) continuous 50-nmol/L (20-ng/mL) increments to determine the overall linear trend; (2) quintiles to explore the dose response; and (3) categorical analysis using ≥ 50 nmol/L versus < 50 nmol/L (20 ng/mL) .
Findings indicated that patient characteristics affected vitamin D levels. Those with higher (seasonally adjusted) 25(OH)D levels tended to be younger and to have a lower body mass index (BMI), a lower number of T2 lesions, and a higher brain volume at the CIS stage, but otherwise were similar to patients with lower levels of 25(OH)D .
Over the 5-year follow-up period, 81.3% (377 patients) converted to MS according to the McDonald 2001 criteria that include MRI lesions  and 46.6% (216 patients) converted to CDMS based on exacerbations or progression alone. The hazard of conversion decreased with increasing serum 25(OH)D and mean serum 25(OH)D levels at 12 months predicted subsequent conversions to McDonald MS (P = 0.02) and CDMS (P = 0.05) .
An increasing serum 25(OH)D level was associated with a decreasing rate of new active lesions on MRI; this effect was particularly strong in patients with both 6- and 12-month serum 25(OH)D measurements. A 50 nmol/L (20 ng/mL) increment in average serum 25(OH)D levels within the first 12 months predicted a 57% lower rate of new active lesions (RR, 95% CI: 0.43 (0.26–0.70), P < 0.001) and a 57% lower relapse rate (RR (95% CI): 0.43 (0.20–0.92, P = 0.03). In evaluating the potential progression of MS on MRI, higher levels of serum 25(OH)D were associated with less T2 lesion volume accumulation over time. For a 50 nmol/L increase in serum 25(OH)D, the relative decrease in T2 lesion volume was 20% per year (P < 0.001). Restricting results to patients with both 6-month and 12-month serum 25(OH)D measures, tended to strengthen results .
The dichotomous analysis of serum 25(OH)D levels (< 50 vs. ≥ 50 nmol/L) is shown in Fig. 6 . For instance, the percentage loss of brain volume over time was lower in patients with 25(OH)D levels ≥ 50 nmol/L at the 12-month time point compared with those with serum 25(OH)D levels < 50 nmol/L (P = 0.005). Although a 50 nmol/L increase in 25(OH)D levels did not reach significance for a reduction in the average expanded disability status scale (EDSS) score (P = 0.11), patients with serum 25(OH)D levels ≥ 50 nmol/L had a significantly lower annualized change in EDSS score compared with those patients with serum 25(OH)D levels < 50 nmol/L (P = 0.004) while on IFN-b-1b. Across all analyses, associations were generally stronger for MRI than for clinical outcomes. Nevertheless, “the latter were still remarkable considering the overall low rate of relapses (0.2 per year) and small EDSS score change (median change, 0.0) in BENEFIT” .
Strengths of the BENEFIT study included (1) its longitudinal design, (2) the exclusive recruitment of patients at the CIS stage, (3) the use of repeated serum 25(OH)D measurements, (4) the large number of patients, (5) standardized treatment (e.g., early vs. late IFNB-1b), and (6) rigorous clinical and MRI assessment of all patients during a 5-year period. Limitations of the study included (1) the fact that most patients were eventually treated with IFNB-1b and some crossed over during the 2 years of the study, and (2) while a clear dose response was observed for the most sensitive MRI outcomes, the effects did not reach a plateau level, and, therefore, serum 25(OH)D levels greater than the median 69 nmol/L could have had a greater effect. According to the authors, a low 25(OH)D level early in the disease course is a strong risk factor for long-term MS activity and progression in patients with early MS who were treated with IFNB-1b .
The BENEFIT cohort had an early treatment group and a delayed treatment group. The associations of 25(OH)D levels and MS activity were more pronounced for patients in the early treatment group than for those in the delayed treatment group (Table 2  and Fig. 7 ), although a test for interaction between 25(OH)D levels and treatment assignment was significant only for the time to CDMS (P = 0.04) . These results suggest that early treatment with IFNB-1b may have an additive effect along with 25(OH)D to reduce disease severity and progression in both clinical and imaging outcomes.
To explore the mechanistic rationale for the potential additive effects of 25(OH)D levels and early IFNB-1b treatment, Munger and colleagues conducted a global gene expression analysis in which expression profiles were measured at various time points among participants in the BENEFIT clinical trial . The relationship between genes or gene sets expressed in association with 25(OH)D and those associated with MS activity was examined. The numbers of Gd+-enhancing lesions served as a marker of disease activity. A 50 nmol/L increase in serum 25(OH) levels reduced the Gd+ lesion count by 55%. Adjusting for gender, age, treatment, and treatment − 25(OH)D interaction did not alter the significance of the findings. Gene expression in whole blood was studied in 295 individuals, evaluating approximately 19,000 genes. Reduced Gd+ lesion count was significantly associated with increased expression of 25(OH)D-related genes, an effect that was independent of IFNB-1b treatment. This effect was also noticed when looking at single genes that were associated with regulation of 25(OH)D levels. The authors hypothesized that there was an additive effect of 25(OH)D and IFNB-1bin reducing Gd+ lesion counts .
The second data set from randomized, double-blind, phase 3 trials in MS was derived from the BEYOND (Betaseron® Efficacy Yielding Outcomes of a New Dose) study . Compared with the BENEFIT study, the BEYOND study included patients with established MS (vs. patients with CIS) and was shorter in duration (2 vs. 5 years). It also included considerably more patients (1482 vs. 465) and was conducted in different geographical regions (North America, Western and Eastern Europe, Southern Hemisphere vs. Europe and Canada).
BEYOND was a large, phase 3, prospective, multicenter, blinded, randomized clinical trial. Patients were monitored for at least 2 years. Clinical visits were scheduled every 3 months, and an MRI was performed at baseline and annually thereafter. A post hoc analysis assessed 25(OH)D levels and the subsequent MS disease course and disease progression as characterized by MRI and clinical endpoints . Eligible patients for the vitamin D analyses included 1482 participants randomized to receive 250, or 500 μg of IFNB-1b with at least two measurements of 25(OH)D obtained 6 months apart. Serum 25(OH)D measurements were performed at baseline, 6, and 12 months.
In longitudinal analyses, 25(OH)D was inversely correlated with the cumulative number of active lesions between baseline and the last MRI (average follow-up time, 2 years). A 50-nmol/L higher level of serum 25(OH)D was associated with a 31% lower rate of new lesions [relative rate (RR), 0.69; 95% CI, 0.55–0.86; P = 0.001]. This inverse association was also strong and significant in analyses restricted to patients with 25(OH)D levels > 50 nmol/L (RR, 0.62; 95% CI, 0.46–0.84; P = 0.002) and was consistent in each of the four geographic regions (Fig. 8) . The lowest rate of new lesions was observed among patients with 25(OH)D levels > 100 nmol/L (RR, 0.53; 95% CI, 0.37–0.78; P = 0.002). No significant associations were found between 25(OH)D levels and change in brain volume, relapse rates, or EDSS scores . Strengths of this study include the large number of participants, the regionally diverse population with varying baseline characteristics, and the repeated measurements of 25(OH)D, which helped characterize patients’ long-term vitamin D status. The relatively short follow-up is the most important limitation of this study. This limited follow-up may explain the lack of association between serum 25(OH)D levels and measures of brain atrophy or clinical endpoints, both of which were modified by 25(OH)D in the longer BENEFIT study [60, 61]. Regarding targeted vitamin D levels, the authors stated: “Our observation of the lowest level of MS activity among patients with serum 25(OH)D levels above 100.0 nmol/L [40 ng/mL] is consistent with the results of a previous investigation in the US , and suggests that the 25(OH)D levels in most patients with MS who are not receiving supplemental vitamin D may be suboptimal” .
Effects of Disease-Modifying Therapies on Vitamin D Levels in MS Patients
MS disease activity may be additively affected by vitamin D and IFNB-1b . This hypothesis is supported by investigations from the same researchers suggesting that processes regulated and triggered by 25(OH)D may be additively enhanced by IFNB-1b , and independently by observations from Stewart and colleagues from the Menzies Research Institute in Tasmania . In an observational cohort study, conducted in 178 patients with MS, vitamin D levels were measured every 6 months over an average of 2.2 years. Patients who took an interferon had significantly higher 25(OH)D levels than those who did not (P < 0.001). Each 10-nmol/L increase in serum vitamin D was associated with a 10% lower relapse rate. Interestingly, interferon treatment was protective only against relapse among persons with higher vitamin D levels. Among those with insufficient vitamin D, there was an increased risk of relapse despite interferon treatment. The investigators hypothesized that treatment with IFNB may increase serum vitamin D levels through enhanced responsiveness to sun exposure and recommended that persons being treated with IFNB should have vitamin D status monitored and maintained in the sufficiency range . Also, noteworthy from these data, this group did not find similar associations for glatiramer acetate (GA) therapy and vitamin D.
The notion of complementary or even synergistic effects of IFNB and vitamin D is further supported by observations from Rotstein and coworkers based on the CLIMB (Comprehensive Longitudinal Investigation of MS at Brigham and Women’s Hospital) cohort . The CLIMB cohort is a prospective cohort study that began enrolling patients in 2000. The objective of the study was to determine whether 25(OH)D levels predicted new disease activity in MS patients treated with IFN-β (n = 96) or GA (n = 151). Separately, due to different selection criteria, a similar analysis was conducted for patients treated with fingolimod (FTY, n = 77). Serum 25(OH)D concentration was adjusted for season, and patients were divided into subgroups by 25(OH)D tertile. The primary study endpoint was ‘time to first inflammatory event’, defined as a combination of either first relapse or first Gd+ lesion, using a Cox model adjusted for age, sex, and disease duration. The results demonstrated higher 25(OH)D levels associated with a longer time to the combined first event in the IFNB subgroup [hazard ratio (HR)IFNB = 0.58; P
IFNB = 0.012], but not in GA-treated participants (HRGA = 0.89; P
GA = 0.50). For Gd+ lesions alone, there was a significant association observed in GA and IFNB subgroups, although the effect was more pronounced with IFNB (HRGA = 0.57; P
GA = 0.039 vs. HRIFNB = 0.41; P
IFNB = 0.022). No significant associations were found for relapses. There were some sampling difficulties in this cohort and, therefore, the results need to be interpreted with certain caution. For FTY, due to the mandated first-dose observation, samples were available for all patients. Higher 25(OH)D was associated with a longer time to the first event (HRFTY = 0.48; P
FTY = 0.016) and with relapses (HRFTY = 0.50; P
FTY = 0.046), but not with Gd+ lesions . The large, prospective cohort and the prolonged follow-up times were strengths of this study, as well as the availability of two 25(OH)D measurements for the majority of patients. However, more regular 25(OH)D measurements would have been ideal and offered greater insights into study conclusions .
Studies Contradicting the Association of Vitamin D Levels with Disease Activity
Contradictory to the aforementioned information are findings reported by researchers from Norway . In this small prospective cohort study, 88 patients with RRMS were followed with regular MRI and 25(OH)D measurements during 6 months before and up to 18 months after initiation of IFNB. During the pre–IFNB treatment phase, higher levels of 25(OH)D were associated with reduced MRI activity; each 10-nmol/L increase in 25(OH)D was associated with 12.7% (P = 0.037) lower odds for new T1 Gd + lesions, 11.7% (P = 0.044) lower odds for new T2 lesions, and 14.1% (P = 0.024) lower odds for combined unique activity. However, there was no association between 25(OH)D and disease activity after initiation of IFNB. With clinical measures, neither the occurrence of relapses nor EDSS progression was associated with 25(OH)D levels during both study phases. Strengths of the study were the prospective design and the frequent MRI and 25(OH)D assessments during the observation period. Limitations were the relatively short time on IFNB and the small number of participants, as well as the minimal 4 nmol/L increase in serum vitamin D levels following vitamin D supplementation. In the discussion of the study results, the authors expressed their surprise about the lack of an association between 25(OH)D levels and MRI after initiation of IFNB, “as there is no evidence suggesting that the immunomodulatory effects of vitamin D are counteracted by IFNB or vice versa. A reasonable explanation is that IFNB reduced radiologic disease activity, leaving relatively little left to be reduced” by vitamin D .