Background Traditionally, patients with drug-resistant tuberculosis are cla
ssified as having acquired drug-resistant or primary drug-resistant disease
on the basis of a history of previous tuberculosis treatment. Only cases o
f primary drug resistance are assumed to be due to transmission of drug-res
istant strains.
Methods This descriptive study of 63 patients with drug-resistant tuberculo
sis assessed the relative contribution of transmission of drug-resistant st
rains in a high-incidence community of Cape Town, South Africa, by restrict
ion-fragment length polymorphism (RFLP). The RFLP results were compared wit
h the results obtained by traditional classification methods.
Findings According to RFLP definitions, 52% (33 cases) of drug-resistant tu
berculosis was caused by transmission of a drug-resistant strain. The propo
rtion of cases due to transmission was higher for multidrug-resistant (64%;
29 cases) than for single-drug-resistant (no cases) tuberculosis. By the c
linical classification, only 18 (29%) patients were classified as having pr
imary drug-resistant tuberculosis (implying transmission). The clinical cla
ssification was thus misleading in 25 patients.
Interpretation The term acquired drug resistance includes patients infected
with strains that truly acquired drug resistance during treatment and pati
ents who were initially infected with or reinfected with a drug-resistant s
train. This definition could lead to misinterpretation of surveillance stud
ies, incorrect evaluation of tuberculosis programmes, and delayed diagnosis
and treatment of patients with multidrug-resistant disease. The clinical t
erm acquired drug resistance should be replaced with the term "drug resista
nce in previously treated cases", which includes cases with drug resistance
due to true acquisition as well as that due to transmitted drug-resistant
strains.