Objectives To compare the magnetisation transfer ratio (MTR) and volume of incidental white matter (WM) lesions of patients who have had anti-tumour necrosis factor-α (TNFα) therapy to that of healthy controls, in order to test the hypothesis that anti-TNFα therapy may cause clinically silent white matter changes.

Methods Ten patients (seven rheumatoid arthritis, three ankylosing spondylitis) who had at least 3 months of anti-TNFα therapy and were free of neurological symptoms and 10 matched healthy controls were recruited and scanned on a 3T MRI with the following protocol-T2 FLAIR, T1 MPRAGE and two sets of T1 gradient-recalled sequences with and without a saturation pulse. WM lesions were defined semi-automatically and graded as small/large (ie, less/more than the 95th centile of lesion sizes in the control group). Lesion MTR histograms for both groups were generated.

Results There was no difference in the total lesion load between both groups but the total number of large lesions was higher in patients (p=0.017). Mean lesion MTR values in both groups were not significantly different from each other (0.54 vs 0.52). Lesion MTR histograms of both groups showed no difference in their peak height and position suggesting that incidental white matter lesions seen in our group of patients are unlikely to be pathologically different to those commonly detected in the healthy population, which are likely ischaemic in origin rather than inflammatory demyelinating. Rheumatoid patients may have more confluent/larger lesions due to an inherently higher ischaemic risk. Therefore we have not shown that anti-TNFα therapy causes white matter changes in neurologically asymptomatic patients.