Drug-resistant tuberculosis in Singapore, 1995 to 1996

Singapore's tuberculosis incidence of 49 to 57 per 100,000 population for 1987 to 1996 presents a picture that is intermediate between developed and developing countries. The proportion of drug-resistant isolates has remained stable at 4.3% to 5.5% from 1992 to 1996 despite rising HIV rates. From 1995 to 1996, of the 199 consecutive drug-resistant isolates at the Central Tuberculosis Laboratory, 66% were mono-resistant, 22% dual-resistant and 12% resistant to more than two drugs. Isoniazid resistance was most prevalent, being found in 72% of isolates, followed by streptomycin resistance in 45%. Resistance to isoniazid and streptomycin (21%) was more common than to isoniazid and rifampicin, i.e. multidrug resistance (MDR) (14%). The small numbers indicated by the low overall prevalence of resistance and the predominance of single-drug resistance support the current initial choice of the standard short course with its three-drug initial phase. Of the 170 cases with matching National Tuberculosis Registry data, 72% of drug-resistant cases represented initial and 28% acquired resistance; testifying to the effectiveness of present day treatment regimens in suppressing resistance when compliance is assured. Case-control analysis using 244 drug-sensitive controls randomly selected from notifications in 1995 to 1996 showed an odds ratio for drug-resistance between subjects with a previous history and no previous history of tuberculosis of 2.47 (95% CI 1.40 to 4.37; P = 0.0007). With each increment in the number of episodes of tuberculosis experienced, there was a trend of resistance to progressively more drugs (P = 0.000004). This association remained even when a logistic regression model including all predictor variables was fitted. No associations were found with age, history of contact with tuberculosis, cavitary disease and, most notably, with human immunodeficiency virus infection. This study reaffirms that a history of previous tuberculosis should increase clinicians' index of suspicion for drug resistance, the urgency with which culture and sensitivity results are sought and the vigour with which patients are followed-up and compliance monitored.