From 1991-2002, we treated 58 patients with multiple sclerosis (MS) using the humanised monoclonal antibody, Campath-1H, which causes prolonged T lymphocyte depletion. Clinical and surrogate markers of inflammation were suppressed. In both the relapsing-remitting (RR) and secondary progressive (SP) stages of the illness, Campath-1H reduced the annual relapse rate (from 2.2 to 0.19 and from 0.7 to 0.001 respectively; both p < 0.001). Remarkably, MRI scans of patients with SP disease, treated with Campath-1H 7 years previously, showed no new lesion formation. However, despite these effects on inflammation, disability was differently affected depending on the phase of the disease. Patients with SPMS showed sustained accumulation of disability due to uncontrolled progression marked by unrelenting cerebral atrophy, attributable to ongoing axonal loss. The rate of cerebral atrophy was greatest in patients with established cerebral atrophy and highest inflammatory lesion burden before treatment (2.3 versus 0.7 ml/year; p = 0.04). In contrast, patients with RR disease showed an impressive reduction in disability at 6 months after Campath-1H (by a mean of 1.2 EDSS points) perhaps owing to a suppression of on-going inflammation in these patients with unusually active disease. In addition, there was a further significant, albeit smaller, mean improvement in disability up to 36 months after treatment. We speculate that this represents the beneficial effects of early rescue of neurons and axons from a toxic inflammatory environment, and that prevention of demyelination will prevent long-term axonal degeneration. These concepts are currently being tested in a controlled trial comparing Campath-1H and IFN-beta in the treatment of drug-naïve patients with early, active RR MS.