Multiple Sclerosis Journal 23(11) Table 1. Patient population and demographic data. Stable RRMS N = 110 % Female:male Median age (years) Median EDSS Median MSSS Median disease duration (years) Median time on/off treatment (days) Treatment with natalizumab Treatment with fingolimod Treatment with glatiramer acetate Treatment with beta-interferon Off treatment 86:24 44 2.8 3.45 11.8 78:22 Range 23-65 0.0-7.5 0.03-9.09 0.6-38 887 31-6971 31 28.2 Figure 2. Total lymphocyte count and NK (CD56) cell count off treatment and on treatment. 24 21.8 15 13.6 NAT: natalizumab; FING: fingolimod; GA: glatiramer acetate; IFN: beta-interferon. ***p ⩽ 0.001 versus no treatment. 22 20 18 16.4 RRMS: relapsing-remitting multiple sclerosis; EDSS: Expanded Disability Status Scale; MSSS: Multiple Sclerosis Severity Scale.18 Table 2. Clinical outcomes including MRI parameters. Outcome parameters stable RRMS n = 110 % NEDA:NEDA criteria not fulfilled No NEDA status as no MRI data MRI progression EDSS progression Relapse 42:64 38.2:58.2 4 3.6 32 19 47 29.1 17.8 42.7 RRMS: relapsing-remitting multiple sclerosis; NEDA: No Evidence of Disease Activity; MRI: magnetic resonance imaging; EDSS: Expanded Disability Status Scale. and only 3.6% (4 patients) had no MRI imaging in the observation period. Analysis of CD56 counts under different treatments Our analysis revealed statistically significant differences within NK population on treatment compared to off treatment (p = 0.001). Post hoc testing showed circulating NK cell numbers were more than twice as high on natalizumab (p < 0.001; Figure 2). Proportions of NK subsets within the raised lymphocyte count did not differ significantly for natalizumab compared to patients off treatment. The 1482 fingolimod group demonstrated a fivefold higher proportion of NK cells within a suppressed lymphocyte count (p < 0.001; Figure 3). The proportion of CD56 bright NKs within the total NK cell count was almost halved (p = 0.011) and the proportion of CD56 dim NKs increased accordingly (p = 0.011; Figure 4). The proportion of CD56 bright was increased on beta-IFN (p = 0.012) with reciprocal decrease in CD56 dim. No significant changes could be demonstrated for glatiramer acetate (Figure 5). Analysis of treatment response Given that we had previously shown CD56 subsets to vary significantly between treatment groups, we included treatment as a fixed factor covariate in each of the GLM ANOVA models in addition to age and sex. Table 3 shows the ANOVA results for all CD56 subsets and the four clinical outcomes. Significant associations for several CD56 subsets with MRI activity were observed (p < 0.05). These persisted even when the two patients with only MRI brain imaging were excluded from the analysis. There were no associations detected for NEDA, EDSS progression or relapses (Table 3). A lower mean of circulating NK cells as well as NK cell proportion appeared to be associated with stable MRI imaging (Table 4). Increased bright NK proportion and decreased dim NK proportion within the NK cell count were also associated with no MRI activity. This could also be shown for decreased total mean of dim NK cell count, but not for an increased mean of the total number of bright NK cells. Discussion In this study, we correlated NK cells and their subsets in MS patients with clinical and radiological journals.sagepub.com/home/msjhttps://journals.sagepub.com/home/msj