In 2007, the notification rate for tuberculosis (TB) was 63.2 per 100,000 population in Taiwan. Among all regions, Eastern Taiwan (consisting of two counties, Hualien and Taitong) had the highest notification rate and mortality from TB, which was105 and 7.9 per 100,000 population, respectively. World Health Organization (WHO) and the International Union Against Tuberculosis and Lung Disease (The Union) launched the Global Project on Anti-tuberculosis Drug Resistance Surveillance in 1994, based on three principles: surveillance should be based on a samples of TB patients representative of all cases in a geographical setting; drug resistance must be clearly distinguished according to the history of TB treatment; and optimal laboratory performance must be ensured through a quality-assurance program [1], [2]. The latest guideline [3] for surveillance of drug resistance in TB indicates that surveillance, according to the International Health Regulations adopted by the 58^th World Health Assembly, is the systematic ongoing collection, collation and analysis of data for public health purposes and the timely dissemination of public health information for assessment and public health response as necessary. A surveillance system based on routine drug susceptibility testing (DST) of all TB cases provides continuous information on patterns and trends in anti-TB drug resistance and provides an opportunity to monitoring outbreaks of drug-resistant TB.

No official surveillance for anti-TB drug resistance has been conducted in Eastern Taiwan before Taiwan Centers for Disease Control (CDC) signed a contract with the mycobacteriology laboratory of Buddhist Tzu Chi General Hospital, which is the only laboratory in Eastern Taiwan that performs DST for Mycobacterium tuberculosis (M. tuberculosis). All isolates of M. tuberculosis in Eastern Taiwan must to be sent to the mycobacteriology laboratory of Buddhist Tzu Chi General Hospital for DST, which forms the basis of the surveillance program in Eastern Taiwan. We analyzed the DST results of all relevant isolates from 2004 to 2008 to determine the proportion and trend of anti-TB drug resistance in Eastern Taiwan.

A total of 3547 tuberculosis patients were notified in 2004–2008 in Eastern Taiwan, in which 2688 were culture positive. The current study enrolled all of the 2688 patients of culture positive tuberculosis, accounted for 76% of all tuberculosis patients and 100% of culture positive cases in 2004–2008 in Eastern Taiwan. Among the 2688 cases, 2176 (81.0%) were new patients and 512 (19.0%) were previously treated patients. Among the 2176 new cases, 97 (4.5%) underwent re-treatment in 2004–2008 because of treatment failure, default, or relapse and were also included to calculate the proportion of anti-TB drug resistance among previously treated cases (Figure 1). A total of 2,785 non-duplicate isolates from 2688 TB patients were included in this study. Susceptibility of all isolates was determined.

Of the 2785 isolates, 2240 (80.4%) were susceptible to all first-line anti-TB drugs, 407 (14.6%) were resistant to INH, 216 (7.8%) were resistant to RMP, 48 (1.7%) were resistant to EMB, 261 (9.4%) were resistant to SM, and 195 (7.0%) were MDR-TB. Of the 2176 isolates from new cases, 1819 (83.6%) were susceptible to all fist-line drugs; 11.6% were resistant to INH, 4.5% were resistant to RMP, 1.1% were resistant to EMB,8.5% were resistant to SM, and 88 (4.0%) were MDR-TB. Of the 609 isolates from previously treated patients, 421 (69.1%) were susceptible to all first-line drugs; 25.5% were resistant to INH, 19.5% were resistant to RMP, 4.1% were resistant to EMB, 12.5% were resistant to SM, and 107 (17.6%) were MDR-TB. The proportion of resistance to any first-line drugs was significantly higher among previously treated cases than that among new cases (p<0.005) (Table 2).

Table 3 shows the trend of combined resistance to the four first-line anti-TB drugs in all cases. The proportion of INH resistance, RMP resistance, EMB resistance, and MDR decreased significantly, but not SM resistance. The proportion of MDR-TB decreased steadily during the study period, from 11.5% in 2004 to 2.4% in 2008. Tables 4 and 5 show the proportion of resistance over the 5 years in new cases and previously treated cases, respectively. The trend in the decreased proportion of resistance was statistically significant for RMP resistance, EMB resistance, and MDR-TB among new cases and for resistance to all first-line drugs except SM among previously treated patients.

The downward trend of resistance among previously treated patients was more prominent than that among new cases. Figure 2 shows that the proportion of MDR-TB among previously treated cases was 25.9% in 2004, which decreased consistently to 5.0% in 2008 with a cumulative reduction of 80.7%. The respective figures for that among new patients were 4.9%, 2.0% and 59.2%. Similar findings were observed for other first-line drugs except SM. Figure 3 shows the trends in the proportions of drug resistance among previously treated patients and the calendar for introducing new TB control interventions in Eastern Taiwan.

Of the 216 RMP-resistant isolates, 195 (90.3%) were MDR-TB, and the positive predictive value of RMP resistance for MDR-TB was 90.7% (88/97) among new cases and 89.9% (107/119) among previously treated cases.

Anti-TB drug resistance in Eastern Taiwan has been reported previously. The proportions of new and previously treated cases with resistance to any first-line anti-TB drugs in 2001–2002 were 16.8% and 64.5%, respectively, and those with MDR-TB were 2.1% and 45.2%, respectively [4].These data may not be representative of Eastern Taiwan, because isolates were mainly collected from patients treated at a medical center. Hospital-based studies on anti-TB drug resistance in Taiwan commonly reported a relative high proportion of drug resistance to any first-line drug, ranging from 30% to 52.4% [6], [7].

The present study was population-based, enrolled all culture-positive tuberculosis cases in 2004–2008 in Eastern Taiwan.The results revealed that the proportion of anti-TB drug resistance among notified TB cases in Eastern Taiwan remained substantial and that the proportion of new patients with MDR-TB was higher than the global average (4.0% vs. 2.9%). The recommended regimen for new TB patients in Taiwan was a 2-month intensive phase of INH, RMP, PZA, and EMB followed by a 4-month continuation phase of INH, RMP, and EMB [8]. SM was commonly used in the past, and the proportion of SM resistance was as high as 15.4% in 1971 [9]. SM has not been recommended for treating new TB patients for more than 20 years, and the proportion of SM resistance among all cases was 9.4% in this study. The proportions of resistance to INH, RMP, and EMB among previously treated cases were significant higher than those among new cases, with odds ratios of 2.6, 5.2, and 4.0, respectively (Table 2). The high proportions of resistance among previously treated patients highlights the importance of addressing the constraints of TB control program, particularly a patient's adherence to treatment.

The proportion of drug resistance in Eastern Taiwan decreased significantly in 2004–2008, particularly among previously treated patients. Similar findings were observed in recent studies conducted at medical centers in southern and central Taiwan [10], [11]; those studies were hospital-based and study population may not be representative, implying limitation of generalizability. Our study revealed 79.1% cumulative reduction in the proportion of MDR-TB among all cases in 5 years and 80.7% reduction among previously treated cases. Although the 4^th report of Global Project on Anti-tuberculosis Drug Resistance Surveillance demonstrated decrease in the proportion of MDR-TB among all cases from 1996 to 2005 in some settings [2], such as in Hong Kong (from 2.6% to 0.9%) and Singapore (from 0.8% to 0.3%), the magnitude of decline in anti-TB drug resistance was not as striking as our finding in Eastern Taiwan.

Trends in anti-tuberculosis drug resistance were regarded as an indicator of program performance [12], [13]. Several initiatives likely contributed positively to the decline in drug-resistant TB in Eastern Taiwan. First, the National Health Insurance (NHI) program was launched in 1995. The NHI program is government-run, compulsory, and inclusive, which improves access to care among people with low socioeconomic status. The proportion of the population insured under the NHI Program was 93% in 1996, and has been >97% since 2002 [14], [15]. Second, the “Regional Reference Laboratory of Mycobacteriology” program was initiated in 2003. DST has been systematically performed for notified tuberculosis cases, and the DST results provided guidance for treating tuberculosis. Third, the tuberculosis case manager (TCM) program was launched by Taiwan CDC in 2004. The TCM program provided financial support for health care facilities to recruit one or more TCM, depending on the number of TB patients served. The main responsibility of the TCM is tuberculosis case management and contacts examination. The proportion of TB patients successfully treated at Buddhist Tzu Chi General Hospital increased from 65.1% in 1997 to 88.4% after a full-time TCM was recruited in 2004 [16]. Fourth, the Department of Health of Taiwan endorsed “The Global Plan to Stop TB 2006–2015”, which aims to decrease the incidence of TB by 50% in Taiwan in 10 years. The Directly Observed Therapy (DOT) program was launched in April 2006. The proportion of smear positive cases managed with DOT was initially 65% in 2006, and increased to 92.6% in 2007 [17]. A preliminary analysis of data in 2006 revealed that treatment outcomes among patients with DOT were significantly better than those without DOT, namely patients treated with DOT had a higher proportion of sputum conversion in 3 months (47% vs. 33%, p<0.001), higher treatment success (75.1% vs. 51.6%, p<0.001), lower mortality rate (13.6% vs. 37.7%, p<0.001), and fewer defaults (3.5% vs. 5.0%, p<0.01) [17]. Further, the number of previously treated tuberculosis cases in present study has decreased from 166 in 2004 to 80 in 2008, likely due to effective case management rendering permanent cure of TB patients .. Fifth, Taiwan CDC launched a MDR-TB program (Taiwan MDR-TB consortium) in May 2007 to manage MDR-TB patients. A total of 225 MDR-TB patients have been enrolled since August 2008. The proportion of patients with a negative culture at month 18 was 87.6% among those who were enrolled, which was 1.6- (95% confidence interval, 1.4–2.0, p<0.0001) fold higher than those who were not enrolled [18].

The decline of drug resistance among new case was less striking than that among previously treated cases. Transmission of drug-resistant strains in the community prior to effective interventions may build up a sub-population infected with drug-resistant strains, who may develop drug-resistant tuberculosis over time. To clarify whether there is ongoing transmission of drug resistant tuberculosis in the community requires further researches.While our data revealed an encouraging decline in anti-tuberculosis drug resistance, continued surveillance on anti-tuberculosis drug resistance is essential to monitor sustainability of the downward trend.

The positive predictive value of RMP resistance for MDR-TB was 90.7% among new cases and 89.9% among previously treated cases, indicating that rifampin resistance could be used as a MDR-TB proxy, and that rapid testing of rifampin resistance may shorten the time for the diagnosis of MDR-TB.

In conclusion, the proportion of TB patients with drug-resistant TB in Eastern Taiwan remains substantial. However, an effective TB control program has successfully driven the proportion of drug resistance among TB patients downward, which is particularly prominent in previously treated cases. An effective TB control program must continue in Eastern Taiwan to further reduce anti-TB drug resistance.