Population-based prospective molecular and conventional epidemiological study of tuberculosis in Hong Kong

To study the transmission of tuberculosis using conventional and molecular epidemiology in Hong Kong.

All patients with positive sputum culture for Mycobacterium tuberculosis residing on the Island of Hong Kong were recruited from May 1999 to April 2002. The restriction fragment length polymorphism technique was used to determine DNA patterns of isolates of M. tuberculosis using the IS6110 probe, supplemented by pTBN12 as a secondary probe.

One thousand five hundred and fifty-three of 2337 (66%) of the patients with bacteriologically confirmed tuberculosis had restriction fragment length polymorphism analysis of their M. tuberculosis isolates. Four hundred and fifty-four (29.2%) patients belonging to 143 clusters were identified; the estimated rate of recent transmission was 20-24%. Significant predictors of clustering included young-age groups (<40 years) versus those >60 years of age (adjusted odds ratio (OR) 1.96, 95% confidence interval 1.47-2.62), permanent residency versus new or non-residents (adjusted OR 3.40, 95% 1.84-6.26) and previous default from treatment versus new cases (adjusted OR 6.12, 95% confidence interval 1.82-20.5). Alcohol and drug abuse, history of imprisonment and HIV infection were not significant risk factors for molecular clustering. Of patients belonging to clusters, 5.1% had definite, 5.5% had probable and 24.4% had possible epidemiological link, suggesting casual contact may be responsible for a high proportion of the clustered cases.

One-fifth to one quarter of the new cases of active tuberculosis in Hong Kong are due to recent transmission. In addition to early diagnosis and successful treatment of all active disease, treatment of latent disease should receive more attention in the control of tuberculosis in Hong Kong.