After the placebo-controlled extension of the pivotal US trial of glatiramer acetate for the treatment of relapsing multiple sclerosis ended, 208 participants entered an open-label, long-term treatment protocol Magnetic resonance imaging (MRI) was added to the planned evaluations of these subjects to determine the consequences of long-term treatment on MRI-defined pathology and evaluate its clinical correlates. Of the 147 subjects that remained on long-term follow-up, adequate images were obtained on 135 for quantitative MRI analysis. The initial imaging sessions were performed between June 1998 and January 1999 at 2,447 +/- 61 days (mean +/- standard deviation) after the subject's original randomization. Clinical data from a preplanned clinical visit were matched to MRI within 3 +/- 51 days. At imaging, 66 patients originally randomized to placebo (oPBO) in the pivotal trial had received glatiramer acetate for 1,476 +/- 63 days, and 69 randomized to active treatment with glatiramer acetate (oGA) were on drug for 2,433 +/- 59 days. The number of documented relapses in the 2 years prior to entering the open-label extension was higher in the group originally randomized to placebo (oPBO=1.86 +/- 1.78, oGA=1.03 +/- 1.28; P=0.002). The annualized relapse rate observed during the open-label study was similar for both groups (oPBO=0.2 7, +/- 0.45 oGA=0.28 +/- 0.40), but the reduction in rate from the placebo-controlled phase was greater for those beginning therapy with GA (oPBO reduced by 0.66 +/- 0.71, oGA reduced by 0.23 +/- 0.58; P=0.0002). One or more gadolinium enhancing lesions were found in 27.4% of all patients (number of distinct enhancements=1.16 +/- 2.52, total enhanced tissue volume=97 +/- 26 microl). The risk of having an enhancement was higher in those with relapses during the open-label extension (odds ratio 4.65, 95% confidence interval (CI) 2.0 to 10.7; P=0.001). The odds for finding an enhancement was 2.5 times higher for those patients originally randomized to placebo (CI 1.1 to 5.4; P=0.02) compared to those always on glatiramer acetate. MRI-metrics indicative of chronic pathology, particularly measures of global cerebral tissue loss (atrophy), were uniformly worse for those originally on placebo. These observations enrich our long-term follow up of the clinical consequences of treatment with glatiramer acetate to include its apparent effects on MRI-defined pathology. They show that the effect of glatiramer acetate on enhancements is definite, but modest, consistent with the drug's described mechanisms of action, and that a delay in initiating treatment results in progression of MRI-measured pathology that can be prevented.