Table 1.
Association of 25-hydroxyvitamin D levels with relapse risk
| Study | Year | n | Follow-up (years; median) | Proportional decrease in IR/RR/HR for each 25 nmol/L increase in 25(OH)D (%)a | Specification |
|---|---|---|---|---|---|
| Simpson et al. [82] | 2010 | 145 | 0.5 | 23 | RRMS |
| Mowry et al. [86] | 2010 | 110 | 1.7 | 34 | Pediatric RRMS/CIS |
| Mowry et al. [43] | 2012 | 469 | 5 | 14 | Only HLA-DR15*01 single RRMS/CIS, not in whole cohort |
| Runia et al. [87] | 2012 | 73 | 1.7 | 27b | RRMS |
| Ascherio et al. [40] | 2014 | 465 | 5 | 26 | CIS |
| Fitzgerald et al. [44] | 2015 | 1482 | 2 | NR | RRMS, no effect ARR |
| Kuhle et al. [41] | 2015 | 1047 | 4.3 | 24c | CIS, not in multivariate |
| Muris et al. [42] | 2016 | 340 | 3 | 32 | Only RRMS < 37.5 years, not in whole cohort |
25(OH)D 25-hydroxyvitamin D, ARR annualized relapse rate, CIS clinical isolated syndrome, HLA human leukocyte antigen, HR hazard ratio, IR incidence risk, NR not reported, RR relative risk, RRMS relapsing–remitting multiple sclerosis
aIR, RR, and HR for a relapse (yes/no) values were linearly recalculated to correspond with a 25 nmol/L (10 ng/L) increase of 25(OH)D levels, except for Runia et al. [87] and Kuhle et al. [41]
bFor each doubling of 25(OH)D (non-linear relationship)
cOnly quartiles reported (median level 49.3 nmol/L)