Aims. HIV predisposes patients to opportunistic infections. However, with the establishment of Highly Active Anti-Retroviral Therapy (HAART), patients’ CD4 counts are maintained, as is a near normal life expectancy. This study aimed to establish the impact of HIV on the bacteriology of spondylodiscitis in a region in which tuberculosis (TB) is endemic, and to identify factors that might distinguish between them. Patients and Methods. Between January 2014 and December 2015, 63 consecutive cases of spontaneous spondylodiscitis were identified from a single-centre, prospectively maintained database. Demographics, presenting symptoms, blood results, HIV status, bacteriology, imaging, and procedure undertaken were reviewed and comparisons made of TB, non-TB, and HIV groups. There were 63 patients (22 male, 41 female) with a mean age of 42.0 years (11 to 78; . sd. 15.0). Results. In total, 53 patients had tuberculous, nine bacterial, and one cryptococcal spondylodiscitis. There were 29 HIV-positive patients, 29 HIV-negative patients, and five patients with unknown HIV status. The local incidence of TB spondylodiscitis was 1.54/100 000 and that of non-TB spondylodiscitis was 0.29/100 000 per annum. TB patients were younger with similar clinical presentation and infective markers, but were more likely to have a paraspinal abscess. They were also less likely to have a hyperintense disc on MRI. There was no difference between the two HIV groups. Conclusion. In areas of endemic TB and HIV, TB remains the most common cause of spondylodiscitis at five times the rate of non-tuberculous causes. HIV managed with HAART increases the rate of infection but does not influence the bacteriology unless there is a low CD4. There was little to differentiate the groups. TB /HIV co-infected patients were younger than non-TB and HIV-negative patients. Non-TB patients were more likely to have disc hyperintensity on MRI and TB patients were more likely to have abscess formation, but as both groups exhibited these features, neither was diagnostic. Culture confirmation remains necessary. Cite this article: Bone Joint J 2019;101-B:617–620