The Thailand TB Active Surveillance Network is a demonstration project, implemented by the Thailand Ministry of Public Health (MOPH), Bangkok Metropolitan Administration (BMA), U.S. Centers for Disease Control and Prevention (CDC), and Research Institute of Tuberculosis, Japan (RIT), involving all districts in three provinces (Chiang Rai, Phuket, Ubon-ratchathani), selected districts in Bangkok (2 in 2005; 9 in 2006), and one national infectious diseases hospital (Bamrasnaradura Institute). In 2006, the catchment area included 59 public and 26 private healthcare facilities. In 2005, the last year for which published population estimates were available, the estimated number of persons aged 0 – 14 years old in the catchment area was 743,574, including 218,012 aged 0–4 years and 525,562 aged 5–14 years; for this calculation, we excluded the national infectious diseases hospital because this referral facility has no definable catchment area (unpublished data, Thailand National Statistics Department and BMA Office). A detailed description of this project, including methods, has been published elsewhere [14].

Thai national guidelines recommend that childhood pulmonary TB (PTB) be diagnosed using a combination of factors, including: history of contact with a PTB patient known to be sputum smear-positive for acid-fast bacilli (AFB); suspected TB symptoms and signs; an abnormal chest radiograph, including pulmonary infiltrates and/or hilar or mediastinal lymphadenopathy; sputum or gastric aspirates that are smear-positive; and tuberculin skin testing. Because this project relied on data collected as part of routine TB control and most cases occurred in adults, limited information was collected about diagnostic tests other than chest radiography, sputum microscopy, and sputum culture. We did not collect data about tuberculin skin testing. Gastric aspirate specimens were classified as sputum specimens, in accordance with the routine TB program practice of classifying patients with abnormal chest radiographs and AFB-positive gastric aspirates as having smear-positive, pulmonary TB. In Thailand, criteria for EPTB diagnosis are less clearly delineated.

All public hospitals are able to process sputum for AFB smear microscopy. As part of this demonstration project, physicians were encouraged to send sputum specimens for mycobacterial culture at a province-level facility. All specimens were initially cultured on solid media, either Lowenstein-Jensen (LJ) or Ogawa medium, and, during mid-2005, additionally on liquid media with an automated reading instrument (Mycobacterial Growth Indicator Tube [MGIT]; BACTEC 960, Becton-Dickinson Corporation). Isolates underwent drug-susceptibility testing (DST) against first-line drugs (streptomycin, isoniazid, rifampin, pyrazinamide, ethambutol) at the Bangkok municipal or national TB reference laboratories.

A case of childhood TB was defined as a person aged 0 – 14 years old who was diagnosed with TB and treated for TB disease. A "new" case of TB was considered TB disease in a person who reported no or less than 1 month of previous TB treatment.

Thai national guidelines recommend standardized, internationally-accepted treatment regimens for children. Treatment outcomes were defined using standard WHO categories [4].

In this demonstration project, nurses and physicians from public and private healthcare facilities were trained in HIV counseling and testing. They were encouraged to provide HIV counseling and testing to TB patients and referrals to HIV-related care and treatment to HIV-infected TB patients. No financial incentives were provided to patients or healthcare workers for HIV testing. Individual physicians used their own judgment about whether to measure CD4⁺ T-cell lymphocyte counts (CD4), provide opportunistic infection (OI) prophylaxis or anti-retroviral therapy (ART), and manage other clinical conditions. When performed, blood for CD4 counts was usually drawn during the first month of TB treatment.

Data were collected on standardized forms that include questions about demographics, clinical history, laboratory and radiographic testing, TB treatment regimens, TB treatment outcomes, and HIV-related information (infection status, CD4, use of OI prophylaxis or ART). Data were analyzed using the Statistical Package for Social Sciences (SPSS) for Windows, version 12 software (SPSS Inc., Chicago, IL, USA) for analysis. Rates of childhood TB (per 100,000) were determined from surveillance data and population estimates; when calculating rates, we excluded reports from the national infectious diseases hospital, which has no population denominator. We classified age as 0–4 and 5–14 consistent with international recommendations [13]. Differences between groups were compared with X² test for categorical variables within different age groupings; where appropriate, Fisher's exact test was used. Statistical significance was defined as p < 0.05. Factors significant in univariate analysis or factors likely to be associated with outcomes were included in multivariate models using logistic regression, and automated backward, stepwise logistic regression was used to construct final models. We calculated odds ratios (OR), adjusted OR (AOR) and 95% confidence intervals (CI).

The protocol for this demonstration project was reviewed by the Thailand MOPH and CDC and found to be surveillance and public health program implementation not requiring oversight by a human subjects research institutional review board.

From October 2004 to September 2006, 14,487 TB cases were recorded in surveillance. Of these, 279 (2%) cases occurred in children, including: 56 (20%) pulmonary, smear-positive; 103 (37%) pulmonary, smear-negative; 38 (14%) pulmonary, smear not known or done; and 82 (29%) extra-pulmonary. Lymph nodes were the most common extra-pulmonary site (Table 1). Most (78%) were registered as new cases. From October 2004 to September 2005, the incidence per 100,000 for all new TB cases was 14 for children, 12 for age 0–4, and 14 for age 5–14. For pulmonary TB, the incidence per 100,000 was 10 (3 smear-positive) for all children, 8 (1 smear-positive) for age 0–4, and 10 (3 smear-positive) for age 5–14. For extra-pulmonary TB, the incidence per 100,000 was 5 for age 0–4, and 4 for age 5–14.

Of 74 children aged 0–4, 8 (11%) had pulmonary, smear-positive TB, compared with 48 (23%) of 205 children aged 5–14 (p < 0.01). The proportion with EPTB and sites involved did not differ by age group.

The median age was 8 years, and only 9 (3%) were younger than 1 year (Table 1). Half were male. Non-Thais, e.g., migrants, refugees, visitors, or other foreigners, accounted for 13% of cases. Only 2% of cases were diagnosed at a private facility. Previous treatment for TB disease was documented in 17 (6%) and for latent TB infection in 12 (4%). Almost half (43%) had cough lasting more than 2 weeks, and most (69%) had abnormal chest radiographs.

Of 197 pulmonary TB cases, 159 (81%) had at least one sputum smear performed, and 79 (50%) of these had at least one sputum culture performed. Of cases with a culture performed, 41 (52%) were culture-positive for Mycobacterium tuberculosis. Overall, 34 (17%) of the pulmonary TB cases were smear-positive and culture-positive, 22 (11%) were only smear-positive (4 culture-negative; 5 culture not done; 4 culture-positive for non-tuberculous mycobacterium [NTM]; and 9 culture unknown/missing), 134 (68%) were neither smear nor culture positive (24 culture-negative; 38 culture not done; 15 neither smear nor culture done; 57 culture unknown/missing), and 7 (4%) were smear-negative and culture-positive. Of 41 culture-positive cases, 33 (80%) had DST performed. MDR TB was diagnosed in 1, which was 0.5% of all 197 pulmonary TB cases and 3% of the 33 culture-positive cases that had DST performed.

Of 279 TB cases, 68 (24%) had known HIV status before their diagnosis of TB. Of the remaining 211 cases with unknown HIV status, 148 (53%) underwent HIV counseling; 108 (73%) of those counseled agreed to HIV testing. In all, 75 (27%) cases were known to be HIV-infected (55 or 73% of whom knew their HIV status before TB diagnosis), 102 (37%) were known to be HIV-uninfected, and 102 (37%) had unknown HIV status.

Of the 75 HIV-infected TB patients, 21 (28%) were pulmonary, smear-positive; 32 (43%) smear-negative; 7 (9%) smear not known or done and 15 (20%) extra-pulmonary. Most (77%) were new cases. Half were male, and 81% were more than 5 years old. CD4 counts were available for 46 (61%) cases. The median was 55 (range: 2; 2731) and mean 193 cells/mm³. The distribution of CD4 was similar in those 0–4 years old (in whom absolute CD4 is known to be unreliable) compared with those 5–14 years old. Only 6 (16%) of children aged 5–14 years had CD4 greater than 200 cells/mm³. ART was prescribed to 17 (23%) before the diagnosis of TB and to 16 (21%) during TB treatment. ART was not prescribed to 18 (24%) and was not known for 24 (32%). Of 32 HIV-infected children aged 5–14 years who had CD4 less than 200 cells/mm³, 20 (62%) received ART. Among children receiving ART, about half (45%) received a combination regimen of stavudine, lamivudine, and nevirapine. Rates of co-trimoxozole use were higher than ART; 23 (31%) were taking co-trimoxazole before their diagnosis of TB, and 21 (28%) were prescribed co-trimoxazole during TB treatment. Almost all (97%) children with CD4 less than 200 cell/mm³ received co-trimoxazole.

Of all 279 patients, 200 (72%) were cured or completed TB treatment, 17 (6%) died during TB treatment, 3 (1%) failed treatment, 39 (14%) defaulted, and 20 (7%) transferred out (Table 1). We stratified outcomes by age group and HIV status (Table 2). More defaults occurred in children aged 0–4 (16/74, 22%) than in children aged 5–14 (23/205, 11%), (p < 0.02). No deaths occurred in the 74 children aged 0–4, compared with 17 (8%) deaths in the 205 children aged 5–14 (p = 0.03). Of the 17 deaths, 13 were in patients with pulmonary TB (2 culture-positive and smear-negative, 5 smear-positive only, 6 neither smear- nor culture-positive) and 4 were in extra-pulmonary TB (2 lymph node, 2 not known). Thirteen (76%) deaths occurred in HIV-infected children; of the nine with CD4 recorded, all had CD4 < 65 cells/mm³.

We also stratified treatment outcomes by anatomic site and/or microbiologic characteristics (Table 3). In children with extra-pulmonary TB, treatment success rates were highest in patients with lymph node TB (86%). Of patients with meningitis, 67% had unknown outcomes (6 defaulted; 7 transferred out). All 13 patients with unknown outcome were from one province that borders Burma and has a large population of migrant Thais and non-Thais; 8 (62%) of these were non-Thai migrants. In children with pulmonary TB, treatment success rates were not significantly different when we stratified cases by microbiologic characteristics.

In univariate and multivariate analysis comparing patients who died with those successfully treated, factors statistically associated with death were treatment at the national infectious diseases hospital (aOR 7.3; CI, 1.3–41.6, compared with treatment at other sites), and being HIV-infected (aOR 6.9; CI, 1.4–33.0, compared with HIV-uninfected). No children aged less than 5 years died compared with 17% of children aged 5 – 14 years (Table 4). In univariate and multivariate analysis comparing patients who defaulted with those successfully treated, factors statistically associated with default were being non-Thai (aOR 5.8; CI, 2.3–14.8) compared with being Thai (Table 5).