Background Significant delays still occur in the diagnosis of spinal tuberculosis [sTB] and there is little consensus on the follow-up and duration of treatment. Current UK and international guidelines for sTB still recommend standard short course chemotherapy but this does not reflect clinical practise.

Aims and methods We performed a retrospective study at a UK teaching hospital to identify the demographic and clinical features of our sTB population and the value of spinal magnetic resonance studies [MR] in guiding response to and duration of treatment.

Results Data was available for 26 patients [2005–2011] with an average age of 40 years [60% male]. Mean time to diagnosis was 9.3 months. At presentation the commonest symptoms were back pain –26/26, constitutional symptoms [fevers, sweats, weight loss] – 12/26 and neurological symptoms – 12/26. In 13/26 [50%], the diagnosis was made following an emergency admission. Positive microbiology was obtained in 19/26 and 84% of these were fully sensitive organisms. 42% had disease at multiple spinal levels and 73% had associated paraspinal abscesses. At 6 months only 2/26 [7%] had MR evidence of resolution of active disease and 11/26 [42%] had persisting symptoms [back pain and neurological symptoms]. The median number of follow-up scans/patient was 2 and mean duration of treatment was 10.7 months. Of the 24 patients who received >6 months treatment, 5 [20%] still had symptoms after completion [back pain].

Conclusions MR provides a useful means of assessing response to treatment and can help in deciding duration of treatment. However the frequency/intervals at which scans should be done is not clear.