Tuberculosis patients may have Mycobacterium tuberculosis in their sputum a
t the end of treatment, and may show new drug resistance, due to either ina
dequate treatment of the original episode or reinfection with a new strain
during therapy. In a cohort study of mineworkers with tuberculosis in South
Africa, 57 of 438 patients had positive sputum cultures 6 months after rec
ruitment in 1995. Of the 31 patients who initially had fully sensitive stra
ins, 3 developed multidrug resistance (MDR) and 3 single-drug resistance (S
DR). Of the 6 who started with SDR, 3 became MDR. HIV infection was nor ass
ociated with drug resistance at enrolment or 6 months later. We compared pa
irs of DNA fingerprints from isolates of M. tuberculosis at recruitment and
6 months later in the 48 patients for whom we had both available. In 45, t
he pairs were identical. In 1 patient, although both isolates were fully se
nsitive, the later fingerprint had 1 less band (transposition). In 2 pairs,
the fingerprint patterns were completely different: one seemed to be the r
esult of laboratory error and the other was a true reinfection with an MDR
strain. Despite a high risk of infection, with a moderate proportion of bac
kground drug-resistant strains (11% SDR, 6% MDR), reinfection is not a comm
on cause of treatment failure or drug resistance at 6 months.