An evaluation of symptom and chest radiographic screening in tuberculosis prevalence surveys.

SETTING: A tuberculosis (TB) prevalence survey was performed in 2002 in two urban communities in Cape Town, South Africa. The population was 36,334 in 2001, and the TB notification rate was 341 per 100,000 population for new smear-positive TB in 2002. OBJECTIVE: To evaluate the relative contributions of symptom and chest radiographic (CXR) screening in the detection of subjects with smear- and/or culture-positive TB in prevalence surveys. DESIGN: Information on symptoms, CXR abnormalities, sputum smear and culture was gathered from a random cluster sample of 1170 adults (aged > or = 15 years). Smear and/or culture-positive TB was used as the gold standard. RESULTS: Of 1170 adults, 29 had bacteriologically positive TB (smear- and/or culture-positive). The presence of any abnormalities on CXR had the highest sensitivity for detecting subjects with bacteriologically positive TB (0.97, 95%CI 0.90-1.00). Specificity for any abnormalities on CXR was 0.67 (95%CI 0.64-0.70). The specificity of any of five TB-related symptoms was 0.68 (95%CI 0.65-0.71). Individual symptoms had low sensitivities, ranging from 0.10 for fever to 0.54 for cough of > or = 2 weeks. CONCLUSION: In this TB prevalence survey, CXR screening, but not symptom screening, was a sensitive alternative to sputum examination of all participants.     

A survey of tuberculosis prevalence in Hanoi, Vietnam.

OBJECTIVE: To assess the prevalence of tuberculosis (TB) in Hanoi, Vietnam, in 2003/2004. METHODS: A random selection was carried out involving 11624 subjects from 20 communes within the city. RESULTS: On chest X-ray examination, 317 subjects (2.73%) showed abnormal lung opacity, of which 17 were sputum smear-positive, two concentrated smear-positive and three culture-positive, all with active TB. The prevalence of sputum smear-positive pulmonary TB was 146 per 100000 in persons aged >or=15 years (95%CI 65-228). CONCLUSION: This is the first large-scale assessment of the prevalence of TB in Hanoi. The prevalence rate was higher than expected, suggesting that a significant number of patients with active TB, particularly females, remain undiagnosed, thus representing a continuing potential source of transmission in the community.     

A baseline survey of the prevalence of tuberculosis in a community in south India at the commencement of a DOTS programme.

OBJECTIVE: To determine the baseline prevalence of culture-positive and smear-positive tuberculosis and the annual risk of tuberculous infection (ARTI) in a community in south India where DOTS is being implemented. METHODS: Using cluster sampling, 50 rural panchayats (villages) and three urban units in Tiruvallur district were selected randomly. All adults aged > or = 15 years underwent symptom and radiographic examination, and those with abnormal shadows and/or chest symptoms had sputum smear and culture examination. In another cluster sample of 73 villages and three urban units, all children aged < 10 years were tuberculin tested. RESULTS: The prevalence of culture-positive and smear-positive tuberculosis was respectively 605 and 323/100,000. Both increased appreciably with age, and were substantially higher in males than in females at all ages; the overall male:female ratio was 5.5 for culture-positive and 6.5 for smear-positive tuberculosis. The ARTI in children aged under 10 years was 1.6%, and was unaffected by sex. Over three decades there was an overall decline of 1.8% per annum in the prevalence of culture-positive and 2.1% for smear-positive tuberculosis. CONCLUSION: Tuberculosis is a major problem in this rural community in south India, with a prevalence of 605/100,000 for culture-positive tuberculosis and 323/100,000 for smear-positive tuberculosis.     

2月末结核分枝杆菌阳性初治肺结核患者菌型鉴定结果分析

目的分析深圳市南山区2月末结核分枝杆菌阳性肺结核患者的分枝杆菌分布情况,为诊断与治疗分枝杆菌感染提供参考依据。方法 2013年1月至2015年12月深圳市南山区痰结核分枝杆菌阳性的初治肺结核患者治疗2月末仍结核分枝杆菌阳性患者进行菌型鉴定并进行结果分析。结果 617例结核分枝杆菌阳性肺结核患者经2个月强化期治疗后,46例患者出现2月末结核菌阳性,其中涂片阳性为33例,涂片阴性而培养阳性为13例;涂片阳性患者中19例培养阴性,10例为结核药物敏感菌,4例为非结核分枝杆菌;涂阴培阳患者中1例为结核药物敏感菌,12例为非结核分枝杆菌;16例非结核分枝杆菌经分型鉴定后,7例脓肿分枝杆菌,9例戈登分枝杆菌。结论肺结核患者可能继发非结核分枝杆菌感染,抗结核治疗2月末肺结核病人应及时留取痰标本做结核菌涂片,培养和菌种鉴定,为进一步诊断和制定个体化疗方案提供依据。     

Undiagnosed tuberculosis in a community with high HIV prevalence: implications for tuberculosis control

BACKGROUND: Although failure of tuberculosis (TB) control in sub-Saharan Africa is attributed to the HIV epidemic, it is unclear why the directly observed therapy short-course (DOTS) strategy is insufficient in this setting. We conducted a cross-sectional survey of pulmonary TB (PTB) and HIV infection in a community of 13,000 with high HIV prevalence and high TB notification rate and a well-functioning DOTS TB control program. METHODS: Active case finding for PTB was performed in 762 adults using sputum microscopy and Mycobacterium tuberculosis culture, testing for HIV, and a symptom and risk factor questionnaire. Survey findings were correlated with notification data extracted from the TB treatment register. RESULTS: Of those surveyed, 174 (23%) tested HIV positive, 11 (7 HIV positive) were receiving TB therapy, 6 (5 HIV positive) had previously undiagnosed smear-positive PTB, and 6 (4 HIV positive) had smear-negative/culture-positive PTB. Symptoms were not a useful screen for PTB. Among HIV-positive and -negative individuals, prevalence of notified smear-positive PTB was 1,563/100,000 and 352/100,000, undiagnosed smear-positive PTB prevalence was 2,837/100,000 and 175/100,000, and case-finding proportions were 37 and 67%, respectively. Estimated duration of infectiousness was similar for HIV-positive and HIV-negative individuals. However, 87% of total person-years of undiagnosed smear-positive TB in the community were among HIV-infected individuals. CONCLUSIONS: PTB was identified in 9% of HIV-infected individuals, with 5% being previously undiagnosed. Lack of symptoms suggestive of PTB may contribute to low case-finding rates. DOTS strategy based on passive case finding should be supplemented by active case finding targeting HIV-infected individuals.     

Recovery rate of NTM from AFB smear-positive sputum specimens at a medical centre in South Korea.

OBJECTIVE: To investigate the recovery rate of non-tuberculous mycobacteria (NTM) from acid-fast bacilli (AFB) smear-positive sputum specimens at a tertiary care medical centre in South Korea with a high pulmonary tuberculosis (PTB) burden. DESIGN: Retrospective analysis of data from AFB smear- and culture-positive sputum specimens collected between January 1998 and December 2001. RESULTS: Over 4 years, 1328 sputum specimens collected from 616 patients were AFB smear- and culture-positive. NTM were recovered from 9.1% (121/1328) of the smear-positive sputum specimens, and from 8.1% (50/616) of patients with smear-positive sputum. NTM were isolated at least twice in 94% (47/50) of the patients from whom NTM was recovered. The most common organism found was Mycobacterium avium complex, followed by M. abscessus. CONCLUSION: These results suggest that a substantial proportion of patients at a tertiary care medical centre in South Korea with AFB smear-positive sputum specimens may have NTM lung disease rather than PTB.     

Rapid decline in prevalence of pulmonary tuberculosis after DOTS implementation in a rural area of South India

SETTING: Tiruvallur District in Tamilnadu, South India, where the World Health Organization-recommended DOTS strategy was implemented as a tuberculosis (TB) control measure in 1999. OBJECTIVE: To assess the epidemiological impact of the DOTS strategy on the prevalence of pulmonary tuberculosis (PTB). DESIGN: Surveys of PTB were undertaken on representative population samples aged > or =15 years (n = 83000-90000), before and at 2.5 and 5 years after the implementation of the DOTS strategy. The prevalence of PTB (smear-positive/culture-positive) was estimated. RESULTS: TB prevalence declined by about 50% in 5 years, from 609 to 311 per 100000 population for culture-positive TB and from 326 to 169/100000 for smear-positive TB. The annual rate of decline was 12.6% (95%CI 11.2-14.0) for culture-positive TB and 12.3% (95%CI 8.6-15.8) for smear-positive TB. The decline was similar at all ages and for both sexes. CONCLUSION: With an efficient case detection programme and the DOTS strategy, it is feasible to bring about a substantial reduction in the burden of TB in the community.     

Outbreak of tuberculosis in a homeless population involving multiple sites of transmission.

SETTING: During 2002-2003, a large outbreak of tuberculosis (TB) occurred among persons using multiple homeless facilities in King County, Washington. OBJECTIVE: To control the transmission of TB in multiple settings. DESIGN: In 2002, contacts exposed to patients in homeless facilities were screened using tuberculin skin tests (TSTs) and symptom review. Based on these screening results, sites of transmission were identified and prioritised, and exposed cohorts at these sites were offered intensive screening tests in 2003 (e.g., symptom review, TST, chest radiograph [CXR], sputum examination and culture). Mycobacterium tuberculosis isolates from patients were genotyped using PCR-based methods to identify outbreak-associated patients quickly. RESULTS: During 2002-2003, 48 (15%) of 313 patients diagnosed in King County were outbreak-associated; 47 culture-positive patients had isolates that matched the outbreak strain by genotyping. Three facilities visited by >12 patients in 2002 had a higher prevalence of TST positive results (approximately 30%) among clients compared with the background rate (7%) in the homeless community. Screening contacts with one sputum culture was as sensitive as CXR in detecting TB disease (77% vs. 62%, respectively). CONCLUSIONS: A comprehensive, resource-intensive approach likely helped to control transmission. This outbreak highlights the vulnerability of homeless populations and the need to maintain robust TB programs in urban settings.     

Prevalence of pulmonary tuberculosis among HIV-infected persons in a home care program in Phnom Penh, Cambodia.

SETTING: Phnom Penh, Cambodia. OBJECTIVE: To determine the burden of active pulmonary tuberculosis among an HIV-infected cohort and the proportion of drug-resistant strains, using active case-finding adapted to a home care setting. DESIGN: Cross-sectional study. Measures include the prevalence of culture-confirmed Mycobacterium tuberculosis, utilizing a single spot sputum specimen; the proportion of pulmonary TB, detected and undetected; proportion of cases resistant to isoniazid, rifampicin, ethambutol, streptomycin; and the diagnostic value of symptoms. RESULTS: Of 441 persons surveyed, 41 (9%) had active pulmonary TB by culture; 29 were smear-positive (71%), and only one case was on treatment. The total burden of pulmonary TB was 12% (54/441), with a ratio of undetected to detected cases of 3:1. Primary isoniazid resistance was detected in six new cases (15%); no MDR-TB was identified. Symptoms were not predictive of active pulmonary disease. Mortality was high among those not surveyed (20%) and those found to have TB (49%). CONCLUSIONS: Tuberculosis is epidemic in this HIV-infected population. Active case-finding yielded three times the number of cases already detected and should be considered where resources allow. However, effective passive case detection and improved coordination of TB and HIV care programs are required to address HIV-associated TB morbidity and mortality.     

Nationwide study of drug resistance among acid-fast bacilli positive pulmonary tuberculosis cases in Lebanon.

OBJECTIVE: To assess the prevalence of drug resistance among smear-positive sputum specimens from pulmonary tuberculosis (TB) cases in Lebanon. DESIGN: Between July 2002 and April 2004, 224 newly diagnosed TB cases and 21 previously treated TB cases were collected nationwide. Mycobacterium tuberculosis isolates were tested against isoniazid (INH), rifampicin (RMP), streptomycin (SM) and ethambutol (EMB) using the BACTEC-TB system. RESULTS: M. tuberculosis and non-tuberculous mycobacteria (NTM) isolates were recovered from 190 and 15 new cases, respectively, and from 16 and 1 previously treated cases, respectively. Overall drug resistance among new TB vs. previously treated TB cases was 19.5% and 75%, and for single drugs it was INH (12% vs. 63%), RMP (3% vs. 56%), SM (12% vs. 44%) and EMB (3% vs. 44%). The overall rate of multidrug resistance (MDR) was 5.8% (1% vs. 62.5%). The male:female ratio was 1.3:1; most were young adults. CONCLUSION: Relatively moderate single drug resistance and very low MDR rates were found among new TB cases, while among previously treated TB cases very high resistance and MDR resistance rates were detected. Such findings underline the need for ongoing stringent control measures to curb the spread of M. tuberculosis and its deleterious effects.     

A rapid survey to determine the prevalence of smear-positive tuberculosis in Addis Ababa.

SETTING: Four kebeles (lowest administrative units) in a typical district of Addis Ababa, Ethiopia. OBJECTIVE: To determine the prevalence of smear-positive pulmonary tuberculosis (TB) in an adult population by using a screening questionnaire and sputum examination. DESIGN: A house-to-house visit was conducted. All those 14 years of age and above were screened for the presence of symptoms. Three sputum samples were collected from persons who had symptoms for acid-fast bacilli (AFB) examination. RESULTS: A total of 12,149 individuals were screened, of whom 173 had symptoms suggestive of TB. Twenty-three were positive for AFB (189/100,000, 95% CI 112-267). No association was observed between smear positivity and socio-demographic factors. Only two of the 23 patients were on anti-tuberculosis treatment (8.7%, 95% CI 1.5-29.5). CONCLUSION: This simple and cost-effective method provides a fairly realistic estimate of the prevalence of TB. Thus with further development of the method it can be used to monitor the progress of TB programmes in resource-poor settings.     

A community-based TB drug susceptibility study in Mimika District, Papua Province, Indonesia.

SETTING: A district level tuberculosis (TB) control programme in Papua Province, Indonesia. OBJECTIVE: To determine the nature and extent of drug-resistant TB in newly diagnosed sputum smear-positive patients. METHODS: Sputum was collected from previously untreated smear-positive pulmonary TB patients diagnosed in the district over a 10-month period. Sputum specimens were processed and inoculated into a BACTEC MGIT960 tube. Isolates were identified by Ziehl-Neelsen staining, hybridisation with nucleic acid probes and biochemical investigations. Susceptibility testing was performed using the radiometric proportion method. Pyrazinamide testing was performed using the Wayne indirect method. RESULTS: One hundred and seven patients had sputum sent to a reference laboratory; 101 (94.4%) were culture-positive for Mycobacterium tuberculosis, with 87 (86.1%) fully sensitive to first-line anti-tuberculosis drugs. Two per cent were multidrug-resistant (MDR-TB) and 12 (11.9%) had other drug resistance. Each of the MDR-TB isolates was susceptible to amikacin, capreomycin, ciprofloxacin and para-aminosalicylic acid (PAS), but were resistant to rifabutin. One isolate was also resistant to ethionamide. CONCLUSIONS: MDR-TB is present in Indonesia but is not a major problem for TB control in this district. Generalisability to other districts in Indonesia, particularly large urban areas, needs to be confirmed by future studies.     

Normal chest radiography in pulmonary tuberculosis: implications for obtaining respiratory specimen cultures.

SETTING: Urban tuberculosis (TB) clinic, Nashville, Tennessee, USA. OBJECTIVE: Chest radiographs (CXRs) help in the diagnosis of pulmonary TB, but may be normal. Mycobacterium tuberculosis in culture is diagnostic of TB, but cultures are not routinely obtained in resource-poor settings. We examined rates and risk factors for pulmonary TB associated with normal CXR. DESIGN: An observational cohort study was performed among all respiratory culture-positive TB cases referred to the Nashville Health Department from October 1992 to July 2003. Clinical factors, demographics and underlying medical conditions were assessed. RESULTS: Of 601 study patients, 53 (9%) had normal CXRs: 31/138 (22%) were human immunodeficiency virus (HIV) infected and 22/463 (5%) were non-HIV-infected/unknown (P<0.001). Among HIV-infected patients, normal CXR was more likely in persons with renal failure (13% vs. 3%, P=0.048). Among non-HIV-infected/unknown patients, normal CXR was more likely in those who were asymptomatic at presentation (32% vs. 13%, P=0.022). In multivariable logistic regression analysis, HIV infection was associated with an increased risk of normal CXR (odds ratio [OR] 6.61, P<0.0001); factors associated with reduced risk were dyspnea (OR 0.24, P=0.026), positive sputum smear (OR 0.45, P=0.028) and cough (OR 0.48, P=0.038). CONCLUSIONS: The rate of normal CXR among persons with culture-confirmed pulmonary TB was high. Respiratory specimen cultures should be obtained in TB suspects with a normal CXR, particularly HIV-infected persons.     

Excess mortality and tuberculosis among individuals with prolonged cough: a population-based study from Vietnam.

SETTING: A demographic surveillance site in north-west Vietnam. OBJECTIVES: To compare notification rates of sputum smear-positive tuberculosis (TB) and mortality rates between the general population and individuals with prolonged cough. DESIGN: A cohort study of 559 prolonged cough cases in a total population of 35,832, observed over 2.5 years. Outcome measures were sputum smear-positive TB and death. RESULTS: The age-standardised smear-positive TB notification rate among cough cases (553/100,000 person-years [py], 95%CI 268-1143) was significantly higher than in the general population (50/100,000 py, 95%CI 43-57), with a relative risk of 11.06 (95%CI 8.28-14.77). Among those who had a chest X-ray suggestive of TB but negative sputum smears prior to the study start, the rate was 6542/100,000 py (95%CI 2906-12,511). The cough cohort had a 2.61 (95%CI 2.38-2.87) times higher risk of dying than the general population. CONCLUSIONS: Individuals with prolonged cough in this study run a significantly increased risk of developing infectious TB and of dying. We recommend further evaluations of improved follow-up systems aimed at identifying and treating sputum smear-negative TB cases before they convert to an infectious stage.     

How to identify tuberculosis cases in a prevalence survey

The identification of pulmonary tuberculosis (TB) cases in a prevalence survey is a challenge, as diagnostic methods are labour-intensive and large numbers of individuals need to be screened because the prevalence rate is low (almost never greater than 1,200 per 100,000 population). Three testing methods are used: questionnaires, chest radiography (CXR) and bacteriological tests, including sputum smear microscopy and culture. These methods can be applied in four strategies to identify cases. The most sensitive strategy is to apply all methods to each eligible individual. The next most sensitive option is to apply the questionnaire, CXR and sputum smear microscopy to each eligible individual and obtain sputum for culture from those individuals with symptoms, abnormalities on the CXR or a positive smear. If laboratory capacity is limited, screening using symptom enquiry and CXR can be used to select those individuals at highest risk of TB. These individuals are then requested to submit sputum for smear microscopy and culture. If neither CXR nor culture is available, sputum samples may be collected from all eligible individuals and examined by an enhanced microscopy method such as fluorescence microscopy. Case definitions are ideally based on the combined results of symptom enquiry, CXR and bacteriology.     

Factors influencing time to smear conversion in patients with smear-positive pulmonary tuberculosis

Background and objective:   Persistent smear-positivity in patients with pulmonary tuberculosis has been shown to predict an unfavourable outcome. This study was conducted to identify the factors influencing time to sputum smear conversion.
Methods:   From July 2003 to June 2007, all patients with smear-positive and culture-confirmed pulmonary tuberculosis, who had attended a medical centre and a local teaching hospital, were identified. Factors that might have influenced time to smear conversion were investigated using time-to-event analysis.
Results:   Altogether 305 patients (mean age: 58.6 years) were studied. Diabetes mellitus was the most common underlying comorbidity. Eight patients (2.6%) had AIDS. After 2 months of treatment, 34 (11.1%) patients remained smear- and culture-positive. Cox proportional hazard regression analysis indicated that the presence of a cavity on CXR, smear grading and the first 2-month treatment regimen were independent factors influencing the time to sputum smear conversion. Among patients who had received isoniazid in the first 2 months of treatment, the time to sputum smear conversion in the 24 patients whose isolate showed isoniazid resistance was not different from that in the 236 patients whose isolate was isoniazid-susceptible (hazard ratio 1.061; 95% CI: 0.697–1.616).
Conclusions:   This analysis revealed that 11.1% of tuberculosis patients remained smear-positive after 2 months of treatment. Patients with cavitation, higher smear grading and those who had not used isoniazid, rifampicin, ethambutol and pyrazinamide continuously in the initial treatment phase had a longer time to sputum smear conversion.     

Detection of pulmonary tuberculosis in patients with a normal chest radiograph

OBJECTIVES: To describe the early symptoms of pulmonary tuberculosis (TB) when the chest radiograph (CXR) is normal. SETTING: Centralized, provincial TB control program. SUBJECTS: Twenty-five patients with culture-positive pulmonary TB and a normal CXR were identified from a review of 518 consecutive patients with culture-positive pulmonary TB in the province of Saskatchewan from January 1, 1988 to March 31, 1997. Patients with abnormal CXRs at the time of diagnosis were excluded from the analysis. RESULTS: Twenty-three of the 25 patients (92%) were symptomatic at the time of diagnosis, with cough/sputum (76%) being reported most commonly. Eleven patients were identified because of contact tracing from cases of infectious pulmonary TB, while the other 14 patients were identified because of an investigation of symptoms. Twenty-four patients (96%) exhibited one or more symptoms of cough for > 1 month, fever for > 1 week, or skin-test conversion after contact with infectious TB. The sputum smear of only one patient was positive. Two patients were pregnant at the time of diagnosis, one patient was HIV-positive, and one patient demonstrated isoniazid-resistant organisms on sensitivity testing. Five patients were diagnosed as having primary TB associated with Mantoux skin-test conversion. The incidence of culture-positive pulmonary TB with a normal chest radiograph was < 1% in the period from 1988 to 1989 and steadily increased to 10% in the period from 1996 to 1997. CONCLUSIONS: Culture-positive pulmonary TB with a normal CXR is not uncommon, and the incidence of this presentation is increasing. Patients with this presentation of TB are typically symptomatic and/or are detected by contact tracing to infectious cases of pulmonary TB. The results suggest that patients presenting with a cough for > 1 month, with a fever for > 1 week, or with documented skin-test conversion < 2 years after known exposure to infectious TB should have sputum submitted for a Mycobacterium tuberculosis smear and culture despite a normal CXR.     

Direct detection and identification of Mycobacterium tuberculosis in smear-positive sputum samples by fluorescence in situ hybridization (FISH) using peptide nucleic acid (PNA) probes.

SETTING: Peptidenucleic acid (PNA) probesdesigned for specific detection of mycobacteria of the Mycobacterium tuberculosis complex (MTC) and other non-tuberculous mycobacterium species (NTM) are shown to be able to penetrate the mycobacterial cell wall and subsequently hybridize in situ to complementary rRNA. OBJECTIVE: To demonstrate the use of fluorescein-labelled PNA probes for detection and identification of M. tuberculosis in smear-positive sputum samples. DESIGN: The sensitivity and specificity of the PNA probes were investigated by fluorescence in situ hybridization (FISH) using cultures of mycobacterium strains representing species of the MTC and NTM, respectively. RESULTS: M. tuberculosis strains were detected by FISH using specific fluorescein-labelled PNA probes directly in smear-positive sputum samples without changing the morphology of the cells. CONCLUSION: PNA probes allow for rapid diagnosis of tuberculosis in smear-positive cases.     

Cohort study of HIV-positive and HIV-negative tuberculosis, Nairobi, Kenya: comparison of bacteriological results.

We have set up a cohort of human immunodeficiency virus (HIV) positive and negative patients with tuberculosis in order to address the problems associated with HIV-related tuberculosis. We present here the results of sputum smear microscopy, culture, mycobacterial identification tests and drug susceptibility assays from specimens taken at presentation. In this selected population of largely pulmonary tuberculosis cases, HIV infection is not associated with significant differences in sputum smear positivity rate, culture positivity rate or initial drug resistance. No atypical mycobacteria were found. Direct sputum smear examination remains specific for the diagnosis of tuberculosis in Kenya in spite of the presence of HIV. HIV infection was not associated with an increase in the proportion of pulmonary cases still culture-positive at 6 months. However a significant increase in the proportion of cases still culture-positive at 6 months was seen in those with initially resistant strains and also in those treated with standard regimen (streptomycin, thiacetazone and isoniazid for 1 month followed by thiacetazone and isoniazid for 11 months, 1STH/11TH) rather than a short-course, rifampicin-containing regimen (rifampicin, pyrazinamide and isoniazid for 2 months, together with streptomycin for the first month and followed by 6 months of thiacetazone and isoniazid, SHRZ/6TH).     

Increasing TB case detection through intensive referral of TB suspects by village doctors to county TB dispensaries.

OBJECTIVE: To explore new approaches to increase the detection of tuberculosis cases (TB). DESIGN: Thirty counties participated in the study. Patients with TB symptoms were surveyed and referred by trained village doctors to county dispensaries, designated township health centres or general hospitals for free sputum examination. TB patients and suspects notified by general hospitals were traced by TB staff if they defaulted during the transfer. RESULTS: A total of 12,091 new smear-positive TB cases were detected. The registration rate of new smear-positives increased from 36.2 per 100,000 population before the project to 49.9/100,000 after the project, and the case detection rate under the DOTS strategy reached 86%. Of 43,464 registered TB suspects, 15,363 (35.3%) were referred by village and hospital doctors. The referral rate increased significantly (P < 0.01). Of the 15,363 referred patients, 3870 were diagnosed as new smear-positive TB cases. Among three different microscopy centres, there was a statistically significant difference in the sputum examination rates of TB suspects and in the smear-positive rates among the suspects examined. The follow-up rate was 70.9%, but the follow-up success rate was only 33.1%. CONCLUSIONS: Intensive referral of patients with TB symptoms by village doctors to TB dispensaries is an effective way of increasing detection. At the same time, incentives are necessary for patients and village doctors.