Background: Investigation of atrophy data from a pivotal natalizumab trial has demonstrated an increased rate of volume loss, compared to placebo, after the first year of therapy. It was considered to be probably due to a pseudoatrophy effect. Objective: To assess grey and white matter volume changes and their relation to global brain volume changes and to baseline inflammation, for patients under natalizumab therapy. Methods: We selected 45 patients on natalizumab therapy for at least 24 months, with magnetic resonance imaging (MRI) scans at baseline, 12 and 24 months. We calculated the percentage brain volume change (PBVC) for the first and second year, using SIENA software. Grey and white matter fractions (GMF and WMF, respectively) for the first year were calculated with SPM5, using lesion masks. After quality checks, six patients were excluded. We studied the predictive variables of change in brain volumes. Results: The PBVC decrease was faster during the first year (−1.10% ± 1.43%), as compared to the second (−0.51% ± 0.96%) (p = 0.037). These differences were more marked in patients with baseline gadolinium-enhancing lesions (p = 0.005). Mean GMF and WMF changes during the first year of treatment were +1.15% (n.s.) and −1.72% (p = 0.017), respectively. The presence of active lesions at baseline MRI predicted PBVC (p = 0.022) and WMF change (p = 0.026) during the first year of treatment, after adjusting for age and corticosteroid treatment. No predictors were found for GMF volume changes. Conclusion: Early brain volume loss during natalizumab therapy is mainly due to WMF volume loss and it is related to the inflammatory activity present at the onset of therapy. We found that the pseudoatrophy effect is mostly due to white matter volume changes.