Dynamics of pseudo-atrophy in RRMS reveals predominant gray matter compartmentalization

Abstract

Objective: To assess the dynamics of "pseudo-atrophy," the accelerated brain volume loss observed after initiation of anti-inflammatory therapies, in patients with multiple sclerosis (MS).

Methods: Monthly magnetic resonance imaging (MRI) data of patients from the IMPROVE clinical study (NCT00441103) comparing relapsing-remitting MS patients treated with interferon beta-1a (IFNβ-1a) for 40 weeks versus those receiving placebo (16 weeks) and then IFNβ-1a (24 weeks) were used to assess percentage of gray (PGMVC) and white matter (PWMVC) volume changes. Comparisons of PGMVC and PWMVC slopes were performed with a mixed effect linear model. In the IFNβ-1a-treated arm, a quadratic term was included in the model to evaluate the plateauing effect over 40 weeks.

Results: Up to week 16, PGMVC was -0.14% per month in the placebo and -0.27% per month in treated patients (P < 0.001). Over the same period, the decrease in PWMVC was -0.067% per month in the placebo and -0.116% per month in treated patients (P = 0.27). Similar changes were found in the group originally randomized to placebo when starting IFNβ-1a treatment (week 16-40, reliability analysis). In the originally treated group, over 40 weeks, the decrease in PGMVC showed a significant (P < 0.001) quadratic component, indicating a plateauing at week 20.

Interpretation: Findings reported here add new insights into the complex mechanisms of pseudo-atrophy and its relation to the compartmentalized inflammation occurring in the GM of MS patients. Ongoing and forthcoming clinical trials including MRI-derived GM volume loss as an outcome measure need to account for potentially significant GM volume changes as part of the initial treatment effect.

Figure 1

Graph showing the slope of PGMVC and PWMVC during the double‐blind period (week 0–16) in the two treatment arms. (A) Decrease in PGMVC of −0.14 ± 0.04% per month in the placebo group (blue color) and significantly more rapid decrease in the treated group (red color, −0.27 ± 0.03% per month; P < 0.001). (B) Decrease in PWMVC of −0.067 ± 0.03% per month in the placebo group (blue color) and slightly more pronounced decrease in the treated group (red color, −0.116 ± 0.02% per month; P = 0.27). See text for abbreviations. PGMVC, percent gray matter volume change; PWMVC, percent white matter volume change; IFNβ‐1a, interferon beta‐1a.

Figure 2

Graph showing the slope of PGMVC and PWMVC over the week 16–40 in the group originally randomized to placebo and starting treatment with IFNβ‐1a at week 16. (A) Decrease in PGMVC (−0.24 ± 0.03% per month), which was (B) more pronounced than in PWMVC (−0.016 ± 0.024%, per month). See text for abbreviations. PGMVC, percent gray matter volume change; PWMVC, percent white matter volume change; IFNβ‐1a, interferon beta‐1a.

Figure 3

Graph showing the 40‐week slope of PGMVC and PWMVC in the originally IFNβ‐1a‐treated groups. (A) Decreasein PGMVC with a significant (P < 0.001) quadratic component, indicating a plateauing that was estimated to start at week 20. (B) Over the same time period, it is shown the small decrease in PWMVC with only a trend (P = 0.06) for a quadratic slope with a plateau also starting at week 20. See text for abbreviations. PGMVC, percent gray matter volume change; PWMVC, percent white matter volume change; IFNβ‐1a, interferon beta‐1a.