As low-field MR imaging is becoming a widely used imaging technique, we aimed at a prospective assessment of differences in imaging quality between low- and high-field MR imaging in multiple sclerosis patients possibly interfering with diagnostic or therapeutic decision making. Twenty patients with clinically proven multiple sclerosis were examined with optimized imaging protocols in a 1.5- and a 0.23-T MR scanner within 48 h. Images were assessed independently by two neuroradiologists. No statistically significant interrater discrepancies were observed. A significantly lower number of white matter lesions could be identified in low-field MR imaging both on T1- and on T2-weighted images (T2: high field 700, low field 481; T1: high field 253, low field 177). A total of 114 enhancing lesions were discerned in the high-field MR imaging as opposed to 45 enhancing lesions in low-field MR imaging. Blood-brain barrier disruption was identified in 11 of 20 patients in the high-field MR imaging, but only in 4 of 20 patients in low-field MR imaging. Since a significantly lower lesion load is identified in low-field MR imaging than in high-field MR imaging, and blood-brain barrier disruption is frequently missed, caution must be exercised in interpreting a normal low-field MR imaging scan in a patient with clinical signs of multiple sclerosis and in interpreting a scan without enhancing lesions in a patient with known multiple sclerosis and clinical signs of exacerbation.