Recurrence of tuberculosis in a low-incidence setting without directly observed treatment: Victoria, Australia, 2002-2014.

Setting: Victoria, Australia, is an industrialised setting with low tuberculosis (TB) incidence and universal health care. Individually tailored adherence support for self-administered daily anti-tuberculosis treatment is provided. Directly observed treatment (DOT) is very rarely used.

Objective: To review the rate of recurrent TB in Victoria between 2002 and 2014.

Design: This was a retrospective cohort study. All recurrent episodes of TB were reviewed and 24-locus MIRU-VNTR (mycobacterial interspersed repetitive units-variable number of tandem repeats) molecular typing was used where possible to determine the likelihood of relapse or reinfection.

Results: Of 4766 notifications, 32 (0.7%) were recurrent episodes. Of 20 episodes that occurred in patients who had previously completed treatment, 11 were culture-positive (0.4% of 3012 culture-positive episodes): 9 were likely relapses (distinguishable at no more than one of 24 loci) and two were likely reinfections, giving a TB relapse rate among culture-positive episodes of 52.5/100 000 person-years (mean time to study end per patient of 5.7 years). The median time until relapse was 18 months (interquartile range 12-30).

Conclusions: The low rate of relapse in our setting demonstrates that individually tailored adherence support for self-administered anti-tuberculosis treatment can achieve excellent treatment outcomes.