Abstract
Background:
Diagnosing pediatric pulmonary tuberculosis (PTB) remains challenging due to low bacterial load and difficulty obtaining specimens. Current guidelines recommend 3-day acid-fast bacilli (AFB) and culture, with 1 molecular testing. We aimed to compare diagnostic yield across 1-, 2- and 3-day collection protocols.
Methods:
We conducted a retrospective study of children <18 years with PTB at a tertiary hospital (2015–2023) who submitted 3 gastric aspirates (GA) or sputum specimens. Patients without molecular results, treated >14 days, or non-tuberculous mycobacterial infections were excluded. Diagnostic sensitivity was calculated using the World Health Organization-defined Clinical Reference Standard as the reference comparator. Two diagnostic approaches were assessed: (1) AFB + nucleic acid amplification test (NAAT); (2) AFB + NAAT + culture.
Results:
Of 244 children screened, 165 were included (112 GA, 53 sputum). Mean ages were 4.8 and 14.5 years, respectively. In GA, (1) AFB + NAAT showed an absolute increase of 5.4% (95% confidence interval: −5.5 to 16.2) from days 1 to 2 (cumulative yield: 25.0%) and 2.7% (−8.9 to 14.2) on day 3 (27.7%). (2) Adding culture improved yield by 3.6% (−9.2 to 16.4) on day 2 (42.0%) and 2.7% (−10.3 to 15.7) on day 3 (44.6%). In sputum, (1) AFB + NAAT showed a 1.9% (−16.8 to 20.6) increase on day 3 (from 58.5% to 60.4%). (2) With culture, yielded consistently 73.6% across all days (0%, −16.8 to 16.8).
Conclusions:
Two-day specimen collection may be sufficient for PTB diagnosis. At least 1 culture should be included in the workup due to the highest sensitivity and ability to provide drug susceptibility information.
Keywords: pulmonary tuberculosis, diagnostic yield, gastric aspirate, children, Xpert Ultra
Diagnosing pulmonary tuberculosis (PTB) in children poses significant challenges due to the low bacterial load (pauci-bacillary disease) and the difficulty in obtaining adequate respiratory specimens.1–3 The sensitivity of diagnostic tests varies based on the technique used, the type of specimen collected and the timing of sample collection.2,4–6
Diagnostic techniques include microscopy for acid-fast bacilli (AFB), which offers rapid turnaround time but poor sensitivity; mycobacterial culture, which is the most sensitive method but requires long incubation time; and nucleic acid amplification tests (NAATs), which address the limitation of conventional methods by providing rapid results with higher sensitivity. Among NAAT, Xpert Mycobacterium tuberculosis/Rifampicin (Xpert MTB/RIF) test and Xpert MTB/RIF Ultra assay (Xpert Ultra, Cepheid, Sunnyvale, CA) have become widely accessible due to their automation and low complexity.7 Xpert Ultra using gastric aspirate (GA) demonstrated a sensitivity of 87.5% in confirmed TB cases and 44.4% in probable TB cases,8 with similar yield using sputum.9 A recent Cochrane review10 also demonstrated that the diagnostic accuracy of Xpert Ultra is influenced by the specimen types. Due to their high performance, the World Health Organization (WHO) recommends Xpert MTB/RIF or Xpert Ultra as the initial diagnostic test for adult and children suspected of having PTB.5 Integrating Xpert Ultra with culture and AFB smear increased the proportion of pediatric cases with definite TB from 37.2% to 67.4%,8 underscoring the benefit of combining multiple diagnostic methods.
To improve diagnostic sensitivity, current practice guidelines advocate for the collection of multiple specimens. The U.S. Centers for Disease Control and Prevention recommends testing 3 separate specimens to compensate for poor sample quality.11 Similarly, the Thai national guideline recommends collecting 3 specimens for AFB smear and culture, with a single specimen tested using Xpert MTB/RIF or Xpert Ultra.12 Supporting this, studies conducted in Kenya among children under 5 years of age,13 along with a recent meta-analysis14 of nasopharyngeal aspirates, found that collecting 2 samples increased diagnostic yield.
Among available specimen types for pediatric TB diagnosis, GA is commonly used, especially in young children unable to expectorate. GA often has higher diagnostic performance compared to sputum (eg, AFB 10.4% vs. 5.7%; culture 32% vs. 18%).15 Gastric aspiration involves nasogastric (NG) tube insertion after an overnight fast to collect swallowed bacilli before gastric emptying. This procedure necessitates hospitalization, a negative pressure or single room, and the use of protective equipment such as N95 masks.5 Multiple days of hospitalization increase both direct and indirect medical costs. Additionally, repeated NG tube insertions for specimen collection are invasive and may cause discomfort to pediatric patients.
Given these challenges, there is increasing interest in streamlining diagnostic protocols. This study aimed to evaluate whether reducing the number of collected specimens while maintaining high diagnostic sensitivity is feasible. We compared the cumulative and incremental diagnostic yield of 1-, 2- and 3-day specimen collection to identify an optimized approach for pediatric TB diagnosis.
METHODS
Study Design and Population:
This retrospective study included children under 18 years of age diagnosed with PTB between 2015 and 2023 at Siriraj Hospital, a tertiary care center in Bangkok, Thailand. TB cases were identified using ICD-10 codes A15, A16 and A19. Detailed chart reviews were subsequently performed to confirm eligibility based on clinical case definitions aligned with international expert consensus.16 These criteria included clinical symptoms, radiographic findings, TB contact history, immunologic tests, microbiologic evidence and clinical response to anti-TB therapy. PTB was defined as cases involving the lung parenchyma, tracheobronchial tree or intrathoracic lymphadenopathy as evidenced by chest radiographs (CXR). Miliary TB is also classified as PTB. Patients with both pulmonary and extrapulmonary TB were categorized as PTB.17 Patients were excluded if they met any following criteria: (1) specimen collection (GA or sputum) obtained on fewer than 3 days, (2) absence of molecular testing results, (3) receipt of anti-TB treatment for more than 14 days before specimen collection or (4) infection with non-tuberculous mycobacterial or Mycobacterium bovis.
Data Collection
Demographic, clinical, radiologic and microbiologic data were extracted from medical charts. Results from microscopy, culture and NAATs for each GA and sputum specimen were recorded. Additionally, results of interferon-gamma release assays, tuberculin skin tests, CXR, and computed tomography scans taken closest to the first GA collection were included. TB cases were classified into 2 groups: (1) bacteriologically confirmed TB (B-TB), defined as cases in which a biologic specimen tested positive for Mycobacterium tuberculosis by smear microscopy, culture or NAAT; and (2) clinically diagnosed TB (C-TB) referring to cases without bacteriologic confirmation but diagnosed as active TB by a clinician based on radiographic abnormalities, suggestive histology or extrapulmonary involvement requiring a full course of TB treatment.5
Information on immunocompromised status and HIV infection was also collected. HIV testing is not routinely performed in children under 5 years of age if there are no signs of opportunistic infections and the maternal HIV serology is negative. In this study, children under 5 years meeting these criteria were assumed to be HIV-negative, given the low likelihood of acquired HIV infection.18
All data were recorded and managed using the REDCap electronic data capture system hosted at the Research Data Management Unit, Faculty of Medicine Siriraj Hospital, Mahidol University.19
Specimen Collection
A standardized protocol was used for GA collection. The procedure was performed by a trained nurse in the infectious disease ward over 3 consecutive mornings with the patient in a supine position and fasting overnight. Between 5 and 10 mL of gastric fluid was aspirated using an NG tube. Sputum specimens were collected from patients capable of expectoration. At our institution, specimens are typically collected over 3 consecutive days. On day 1, AFB smear, mycobacterial culture and NAAT [Xpert MTB/RIF, Xpert Ultra or real-time polymerase chain reaction (PCR)] are performed. On days 2 and 3, AFB smear and culture are repeated. For both GA and sputum, a total of 3 specimens over 3 days was considered a complete diagnostic workup. Therefore, the number of collection days equates to the number of GA as well as sputum.
Microbiologic Testing
Direct smear microscopy for AFB was performed using fluorochrome staining. Mycobacterial culture was conducted using both Lowenstein-Jensen solid medium and Mycobacteria Growth Indicator Tube liquid medium, with incubation periods of 60 and 42 days, respectively. Mycobacterium species identification was performed using MPT64 antigen detection. NAATs included various molecular assays used throughout the study period, including Anyplex MTB/NTM real-time PCR (Seegene Inc., Songpa-gu, Seoul, Republic of Korea),20,21 Xpert MTB/RIF and Xpert Ultra.9 Xpert Ultra was implemented starting in May 2019.
Statistical Analysis
Data were categorized and analyzed separately by specimen type (GA vs. sputum) to reflect the distinct patient populations, without direct comparisons between these 2 groups. Categorical variables were reported as frequencies and percentages, while continuous variables were presented as means with standard deviations. Diagnostic sensitivity was calculated using the WHO-defined Clinical Reference Standard22 as the reference comparator, which includes both microbiologically confirmed TB and clinically diagnosed cases. Two diagnostic approaches were assessed: (1) without culture: combining 1–3 AFB smears and 1 NAAT to reflect diagnostic tools with rapid turnaround time, and (2) with culture: combining 1–3 AFB smears, 1 NAAT and 1–3 cultures to represent the full microbiologic diagnostic workup. Cumulative and incremental yields were calculated by comparing the proportion of positive test results across each collection day. Proportions were reported with 95% confidence intervals (CIs), estimated using the Clopper-Pearson (binomial exact) method. All analyses were performed using STATA version 14.0. A significant level of 0.05 was used for all statistical tests.
RESULTS
Among the 244 children screened for PTB, 165 met the inclusion criteria, having completed a full 3-day specimen collection (Fig. 1). The GA group included 112 children with a mean age of 4.8 ± 5.1 years, while the sputum group consisted of 53 children with a mean age of 14.5 ± 2.6 years. Overall, 10.3% of patients had immunocompromised conditions, with HIV infection accounting for the majority (11 of 17 cases). Clinical symptoms were present in 74.6% of patients, and 24.2% were asymptomatic but had a history of TB exposure. Extrapulmonary TB was observed in 37.0% of cases. Bacteriologic confirmation was achieved in 60.0% (99/165) of patients. Among these, 23.6% tested positive by all 3 microbiologic methods—AFB, NAAT and culture—while 15.8% were positive by culture only. Molecular testing was performed using Xpert Ultra in 56.4% of cases and real-time PCR in 42.4% (Table 1).
FIGURE 1.
Flow diagram of patient enrollment and specimen classification. B-TB: Bacteriologically confirmed TB, C-TB: Clinically diagnosed TB, NTM: non-Mycobacterium tuberculosis.
TABLE 1.
Demographic Characteristics of Children With Pulmonary Tuberculosis With Abnormal Chest Radiolography
| Characteristics | Total (N = 165) | Patients Submitted Gastric Aspirates for Diagnosis (N = 112) | Patients Submitted Sputums for Diagnosis (N = 53) |
|---|---|---|---|
| Age | |||
| - Mean (SD), year | 7.9 (6.3) | 4.8 (5.1) | 14.7 (2.1) |
| Age group | |||
| - <1 year, n (%) | 30 (18.2) | 30 (26.8) | 0 |
| - 1–<5 years, n (%) | 43 (26.1) | 43 (38.4) | 0 |
| - 5–<10 years, n (%) | 16 (9.7) | 16 (14.3) | 0 |
| - 10–<15 years, n (%) | 45 (27.3) | 15 (13.4) | 30 (56.6) |
| - >15 years, n (%) | 31 (18.8) | 8 (7.1) | 23 (43.4) |
| Sex | |||
| - Male, n (%) | 72 (43.6) | 56 (50.0) | 16 (30.2) |
| - Female, n (%) | 93 (56.4) | 56 (50.0) | 37 (69.8) |
| Patient with immunocompromised condition*, n (%) | 17 (10.3) | 8 (7.1) | 9 (17.0) |
| HIV status†, n (%) | |||
| - Positive | 11 (6.7) | 5 (4.5) | 6 (11.3) |
| - Negative | 137 (83.0) | 100 (89.3) | 37 (69.8) |
| - Unknown | 17 (10.3) | 7 (6.2) | 10 (18.9) |
| HIV—virologic not suppressed, n (%) | 10/11 (90.9) | 4/5 (80.0) | 6/6 (100.0) |
| HIV mean VL—copies/mL (n = 11) | 474,413 | 335,110 | 567,282 |
| Absolute CD4 count (cells/cu·mm)—median (IQR) (n = 11) | 125 (63-204) | 78 (63-164) | 127.5 (76-204) |
| Clinical symptoms, n (%) | |||
| - Symptomatic | 123 (74.6) | 73 (65.2) | 50 (94.3) |
| - Asymptomatic with TB exposure | 40 (24.2) | 39 (34.8) | 1 (1.9) |
| - Asymptomatic without TB exposure | 2 (1.2) | 0 | 2 (3.8) |
| Present of extrapulmonary TB‡, n (%) | 61 (37.0) | 44 (39.3) | 17 (32.1) |
| Positive tuberculin skin test, n (%) | 30/55 (54.6) | 23/44 (52.3) | 7/11 (63.6) |
| Positive IGRA, n (%) | 24/60 (40.0) | 15/48 (31.3) | 9/12 (75.0) |
| Bacteriologically confirmed TB. n (%) | 99 (60.0) | 54 (48.2) | 45 (84.9) |
| Microbiologic tests for diagnosed bacteriologically confirmed TB§ (n = 99) | |||
| - AFB | 7 (7.1) | 5 (9.3) | 2 (4.4) |
| - NAAT | 4 (4.0) | 3 (5.6) | 1 (2.2) |
| - Culture | 26 (26.3) | 19 (35.2) | 7 (15.6) |
| - AFB + NAAT | 0 | 0 | 0 |
| - AFB + culture | 7 (7.1) | 6 (11.1) | 1 (2.2) |
| - AFB + NAAT + culture | 39 (39.4) | 13 (24.1) | 26 (57.8) |
| - NAAT + culture | 6 (6.0) | 4 (7.4) | 2 (4.4) |
| Methods of NAAT, n (%) | |||
| - Xpert Ultra | 93 (56.4) | 58 (51.8) | 35 (66.0) |
| - Xpert | 2 (1.2) | 2 (1.8) | 0 |
| - Real-time PCR | 70 (42.4) | 52 (46.4) | 18 (34.0) |
Ten patients were diagnosed using extrapulmonary specimens.
Immunocompromised conditions: 11 HIV, 5 immunosuppressive agents, 1 aplastic anemia.
Children less than 5 years old with seronegative mother were assumed as negative, if not tested.
Extrapulmonary TB included: 17 Pleura, 18 LN, 8 Abdomen, 7 bone and joint, 10 CNS, 7 miliary, 1 otitis media, 1 aortitis, 1 paraspinal abscess, 1 chest wall abscess, 1 Pericardium, 1 Congenital, 2 Larynx, 1 Mastoiditis, 1 Ocular.
Cumulative microbiologic testing over 3-day specimen collection. Percentages are based on the number of patients with bacteriologically confirmed TB.
Radiologic assessment revealed that the most common CXR finding was interstitial infiltration (44.9%), followed by reticulonodular (35.2%) and patchy infiltrates (18.8%). Additional radiographic findings included hilar adenopathy (17.6%), pleural effusion (10.9%), atelectasis (9.1%), and cavitary lesions (4.2%). Reticulonodular infiltration was significantly more frequent in B-TB (48.5% vs. 15.2%, P < 0.001), whereas interstitial infiltration was more prevalent in C-TB (74.2% vs. 25.3%, P < 0.001). Miliary patterns, bronchiectasis, calcifications, cavities and pulmonary nodules were found exclusively in the B-TB group. Rates of hilar adenopathy, pleural effusion and atelectasis were similar between B-TB and C-TB groups (Table 2). Diagnostic yield varied by radiologic pattern, with higher yields observed in cases with typical radiographic findings. Details are provided in Table, Supplemental Digital Content 1, https://links.lww.com/INF/G444.
TABLE 2.
Radiologic Findings of Patients With Pulmonary Tuberculosis
| Radiologic Findings | Total (N = 165) n (%) |
Bacteriologically Confirmed TB (N = 99) n (%) |
Clinically Diagnosed TB (N = 66) n (%) |
P value |
|---|---|---|---|---|
| Infiltration pattern | ||||
| - Reticulonodular | 58 (35.2) | 48 (48.5) | 10 (15.2) | <0.001 |
| - Interstitial | 74 (44.9) | 25 (25.3) | 49 (74.2) | <0.001 |
| - Patchy | 31 (18.8) | 23 (23.2) | 8 (12.1) | 0.073 |
| - Consolidation | 17 (10.3) | 15 (15.2) | 2 (3.0) | 0.017 |
| - Miliary | 5 (3.0) | 5 (5.1) | 0 | 0.158 |
| Atelectasis | 15 (9.1) | 9 (9.1) | 6 (9.1) | 1.000 |
| Bronchiectasis | 4 (2.4) | 4 (4.0) | 0 | 0.151 |
| Calcification | 2 (1.2) | 2 (2.0) | 0 | 0.517 |
| Hilar adenopathy | 29 (17.6) | 17 (17.2) | 12 (18.2) | 0.867 |
| Pleural effusion | 18 (10.9) | 12 (12.1) | 6 (9.1) | 0.541 |
| Cavity | 7 (4.2) | 7 (7.1) | 0 | 0.042 |
| Mass/nodule(s) | 4 (2.4) | 4 (4.0) | 0 | 0.151 |
Figure 2 illustrates the diagnostic performance of GA specimens. The sensitivity of AFB microscopy increased from 9.8% on day 1 to 17.0% on day 2 and 21.4% on day 3, reflecting incremental gains of 7.2% and 4.4%, respectively. The sensitivity of a single NAAT was 17.9%. Culture demonstrated only modest improvement over time, increasing from 34.8% on day 1 to 37.5% by day 3. When AFB was combined with NAAT, the integrated diagnostic yield rose from 19.6% on day 1 to 25.0% on day 2 and 27.7% on day 3, with incremental gains of 5.4% (95% CI: −5.5 to 16.2, p =0.336) and 2.7% (95% CI: -8.9 to 14.2, p=0.649). Incorporating AFB, NAAT, and culture, yielded 38.4% on day 1, 42.0% on day 2 and 44.6% on day 3, with incremental increases of 3.6% (95% CI: −9.2 to 16.4, P = 0.586) and 2.7% (95% CI: −10.3 to 15.7, P = 0.686), respectively.
FIGURE 2.
Diagnostic yield of gastric aspirates by testing method and number of collection days. AFB: acid-fast bacilli smear microscopy; NAAT: nucleic acid amplification test.
Figure 3 illustrates the diagnostic performance of sputum specimens. AFB sensitivity increased from 45.3% on day 1 to 52.8% on day 2 and 54.7% on day 3, with incremental gains of 7.5% and 1.9%, respectively. The sensitivity of a single NAAT was 54.7%. Culture showed minimal improvement, rising from 66.0% on day 1 to 67.9% on day 2 and remaining unchanged on day 3. When AFB was combined with NAAT, the integrated diagnostic yield was 58.5% on both day 1 and day 2, and 60.4% on day 3, resulting in incremental gains of 0% (95% CI: −18.8 to 18.8, P = 1.000) and 1.9% (95% CI: −16.8 to 20.6, P = 0.843), respectively. When included AFB, NAAT, and culture, it demonstrated a consistent yield of 73.6% across all 3 days, with no additional diagnostic benefit from extended collection (95% CI: −16.8 to 16.8, P = 1.000).
FIGURE 3.
Diagnostic yield of sputum specimens by testing method and number of collection days.
Detailed diagnostic yields for AFB, culture and NAAT by day of collection are presented in Table, Supplemental Digital Content 2, https://links.lww.com/INF/G444, while integrated yields from different test combinations and collection protocols are summarized in Table, Supplemental Digital Content 3, https://links.lww.com/INF/G444. A subgroup analysis of patients tested using Xpert Ultra revealed a higher diagnostic yield compared to the overall molecular group. In GA specimens, when AFB was combined with NAAT, the integrated yield was 31.0% on day 1, 34.5% on day 2 and 36.2% on day 3 (see Table, Supplemental Digital Content 4 and 5, https://links.lww.com/INF/G444).
DISCUSSION
This study demonstrates that culture remains the most sensitive diagnostic method for detecting tuberculosis in children, yielding 34.8% positivity in GA and 66.0% in sputum specimens. NAATs showed moderate sensitivity, 17.9% in GA and 54.7% in sputum, while AFB smear microscopy had the lowest yields at 9.8% and 45.3%, respectively. These results align with prior studies.8,20 A previous study23 conducted in 2013 at our institution reported a bacteriologic confirmation rate of only 25% using culture or PCR. With the addition of Xpert Ultra, this rate has increased to 60% in our current study.
The GA and sputum cohorts in our study represent children with distinct age profiles. Therefore, the primary aim was not to compare them directly but to assess the incremental yield of multiday specimen collection within each group, reflecting real-world diagnostic workflows. This separation acknowledges that diagnostic strategies may need to be tailored based on age and specimen feasibility. All patients in the sputum group were over 10 years of age, which likely contributed to the high diagnostic yields for all diagnostic tests. This pattern is consistent with adolescent TB that resembles adult-type disease,24 characterized by cavitary lesions and higher bacillary burden. In contrast, the GA group primarily comprised younger children, but 23 of 112 patients (20.5%) were over 10 years old. Subgroup analysis demonstrated that these older children had diagnostic yields comparable to the sputum group across all microbiologic methods (e.g., 1 NAAT; 43.5% in GA vs. 54.7% in sputum, 1 culture; 69.6% vs. 66.0%, and combined 1-day AFB+NAAT+culture; 69.6% vs. 73.6%) (see Table, Supplemental Digital Content 6, https://links.lww.com/INF/G444). While direct comparison between GA and sputum was not the primary objective of this study, these findings suggest that in children older than 10 years, sputum collection is preferable due to similar diagnostic performance and less invasive sampling.
In GA specimens, the AFB yield increased slightly over 3 days; however, combining 1 NAAT and 2 AFB smears achieved comparable or better sensitivity than the full 3-day AFB collection (25.0% vs. 21.4%). Similarly, in sputum specimens, 2-day AFB plus 1 NAAT yielded 58.5%, outperforming the 54.7% obtained from 3-day AFB collection. These findings suggest that integrating NAAT with AFB over 2 days offers diagnostic performance equivalent to extended protocols, enabling prompt diagnosis without reducing sensitivity.
Culture continues to demonstrate the highest diagnostic value. In GA, the 3-day protocol integrated with culture increased diagnostic yield from 27.7% to 44.6%, and in sputum from 60.4% to 73.6%, highlighting the importance of always including culture in the diagnostic workup. However, when comparing 2-day versus 3-day combined AFB, NAAT and culture, the differences were minimal: 42.0% versus 44.6% in GA and 73.6% consistently for sputum, suggesting that a 2-day added culture protocol does not compromise diagnostic performance.
Compared with Venturini et al.,25 our findings in GA (42.0% diagnostic yield in 2-day and 44.6% in 3-day collection) were slightly lower than their cumulative yield of 40.4% (2-day GA) and 47.4% (3-day GA), with a 7% incremental benefit in the third-day collection. Their study employed molecular testing on all 3 specimens, which likely contributed to the higher yield. Similarly, a study by Song et al.13 showed that testing a second GA specimen using either Xpert MTB/RIF or MGIT increased diagnostic sensitivity by 3%–17%. However, given the limited resources in routine clinical practice, performing 1 molecular test along with 2 AFB smears and cultures appears to offer a practical and efficient alternative with acceptable diagnostic sensitivity. These results are consistent with a systematic review by Ioos et al.,26 which reported that a second specimen added 6%–33% diagnostic yield across NAAT and culture.
Our results have practical implications. Reducing specimen collection to 2 days may decrease hospital stays, lower healthcare costs and minimize discomfort for children without compromising diagnostic accuracy. This is especially beneficial in resource-limited settings, where prolonged hospitalization and repeated NG insertions are burdensome.
Microbiologic confirmation of pediatric TB remains challenging due to its pauci-bacillary nature. Consequently, initiation of anti-TB therapy is often guided by clinical judgment without requiring bacteriologic confirmation.27 However, with the increasing prevalence of drug-resistant TB, the clinical priority has shifted from merely confirming TB diagnosis to obtaining drug susceptibility information to guide appropriate treatment decisions.
Currently, there are 2 principal methods for assessing drug resistance: molecular testing and culture-based phenotypic drug susceptibility testing. The WHO recommends molecular diagnostics such as Xpert MTB/RIF or Xpert Ultra as first-line tests due to their rapid turnaround and operational simplicity.5,7 Yet, in our study, Xpert Ultra alone provided limited sensitivity—27.6% in GA and 57.1% in sputum. When combined with culture, the diagnostic yield improved to 46.6% and 71.4%, respectively, underscoring the importance of adding culture for both sensitivity and drug resistance data, particularly for drugs beyond rifampicin.
Although newer molecular technologies are emerging, including moderate-complexity automated NAATs for rifampicin and isoniazid resistance, low-complexity assays for second-line agents, and high-complexity reverse hybridization-based NAATs for pyrazinamide resistance, access remains limited in many high-burden, resource-constrained settings. Until these tools become widely available, conventional culture remains essential, both for diagnosis and for treatment tailoring. We emphasize the continued need to support and strengthen mycobacterial culture capacity in limited-resource settings.
Strengths and Limitations
A major strength of this study is that it was conducted in Thailand, a high TB burden country (157 cases per 100,000),28 the findings are generalizable to other high TB burden settings. The study also included a high volume of pediatric TB cases with a broad age distribution, enabling age-specific analysis of diagnostic yield. Another strength is that bacteriologically confirmed TB is observed across all age groups. Among those undergoing GA, B-TB confirmation rates ranged from 20% to 25%, while in the sputum group (mostly adolescents), rates reached 42%–57%, confirming the relevance of diagnostic performance across the pediatric spectrum. However, we acknowledge that age is a key confounder in both specimen selection and diagnostic performance. As such, we included an age-stratified subgroup analysis as shown in Table, Supplemental Digital Content 6, https://links.lww.com/INF/G444.
Several limitations should be noted. First, multiple molecular assays were used over the study period, which may have introduced variability. To address this, we conducted a sensitivity analysis restricted to Xpert Ultra, which showed detection rates consistent with overall trends (GA: 27.6% vs. 17.9%; sputum: 57.1% vs. 54.7%), supporting the robustness of our findings (Tables, Supplemental Digital Content 2 and 4, https://links.lww.com/INF/G444).
Second, rifampicin resistance data from Xpert Ultra were not captured, preventing assessment of drug-resistant TB. As drug resistance becomes increasingly prevalent, future studies should incorporate resistance profiling to guide treatment decisions.
Lastly, we excluded patients with fewer than 3 days of specimen collection, which may introduce selection bias and limit applicability in settings where same-day or 2-day protocols are feasible. However, this approach ensured consistency in evaluating incremental diagnostic yield by day, which was our main objective.
CONCLUSION
Combining molecular testing with AFB yielded approximately 25% sensitivity in GA specimens, with limited gain after 2 days, while in sputum specimens, the yield approached 60% with minimal improvement on the third day. These findings suggest that a 2-day specimen collection protocol may be sufficient for pediatric pulmonary TB diagnosis, particularly in adolescents able to produce sputum. At least 1 culture should always be included in the diagnostic workup due to its highest sensitivity and ability to provide drug susceptibility information.
Acknowledgments
The authors thank Prof. Pope Kosalaraksa, Division of Infectious Diseases, Department of Pediatrics, Faculty of Medicine, Khon Kaen University, for his contribution to the manuscript.
Supplementary Material
Footnotes
The authors have no funding or conflicts of interest to disclose.
N.T.: writing—original draft preparation; K.L., N.V., and K.C.: conceptualization, writing—review and editing; N.T. and W.L.—data collection and interpretation. W.P., O.W., S.R., and N.L.—resources. All the authors have read and agreed to the published version of the manuscript.
Ethics approval with waiver of consent: SI 622/2024.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (www.pidj.com).
Contributor Information
Nattakarn Tantawarak, Email: tantawarak.n@gmail.com;natttan@kku.ac.th.
Nirun Vanprapar, Email: sinvpar@gmail.com.
Orasri Wittawatmongkol, Email: orasriw@hotmail.com.
Wanatpreeya Phongsamart, Email: wanatpreeya@gmail.com.
Supattra Rungmaitree, Email: sp.rt@live.com.
Nichkamol Lertamornkitti, Email: nichkamol.ler@gmail.com.
Wannaporn Limpitikul, Email: poodle_jib@hotmail.com.
Kulkanya Chokephaibulkit, Email: kulkanya.cho@mahidol.ac.th.
References
- 1.Piccini P, Chiappini E, Tortoli E, et al. Clinical peculiarities of tuberculosis. BMC Infect Dis. 2014;14:S4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Swaminathan S, Rekha B. Pediatric tuberculosis: global overview and challenges. Clin Infect Dis. 2010;50(suppl_3):S184–S194. [DOI] [PubMed] [Google Scholar]
- 3.Togun TO, Kampmann B. Diagnostics for childhood tuberculosis: a marathon rather than a sprint. Lancet Infect Dis. 2024;24:110–112. [DOI] [PubMed] [Google Scholar]
- 4.Marais BJ, Pai M. New approaches and emerging technologies in the diagnosis of childhood tuberculosis. Paediatr Respir Rev. 2007;8:124–133. [DOI] [PubMed] [Google Scholar]
- 5.World Health Organization. WHO Operational Handbook on Tuberculosis. Module 5: Management of Tuberculosis in Children and Adolescents. World Health Organization; 2022. [PubMed] [Google Scholar]
- 6.Coulter JBS. Diagnosis of pulmonary tuberculosis in young children. Ann Trop Paediatr. 2008;28:3–12. [DOI] [PubMed] [Google Scholar]
- 7.World Health Organization. WHO Consolidated Guidelines on tuberculosis. Module 3: Diagnosis – Rapid Diagnostics for Tuberculosis Detection. 3rd ed. World Health Organization; 2024. [PubMed] [Google Scholar]
- 8.Sun L, Zhu Y, Fang M, et al. Evaluation of Xpert MTB/RIF ultra assay for diagnosis of childhood tuberculosis: a multicenter accuracy study. J Clin Microbiol. 2020;58:e00702–e00720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chakravorty S, Simmons AM, Rowneki M, et al. The new Xpert MTB/RIF ultra: improving detection of Mycobacterium tuberculosis and resistance to rifampin in an assay suitable for point-of-care testing. Nacy CA, ed. mBio. 2017;8:e00812–e00817. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kay AW, Ness T, Verkuijl SE, et al. Xpert MTB/RIF Ultra assay for tuberculosis disease and rifampicin resistance in children - Kay, AW - 2022 | Cochrane Library. Available from: https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD013359.pub3/full. Accessed September 17, 2025. [DOI] [PMC free article] [PubMed]
- 11.Lewinsohn DM, Leonard MK, LoBue PA, et al. Official American Thoracic Society/Infectious Diseases Society of America/Centers for Disease Control and prevention clinical practice guidelines: diagnosis of tuberculosis in adults and children. Clin Infect Dis. 2017;64:e1–e33. [DOI] [PubMed] [Google Scholar]
- 12.Division of Tuberculosis. National Tuberculosis Control Programme Guideline, Thailand 2021. Division of Tuberculosis; 2021. [Google Scholar]
- 13.Song R, Click ES, McCarthy KD, et al. Sensitive and feasible specimen collection and testing strategies for diagnosing tuberculosis in young children. JAMA Pediatr. 2021;175:e206069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Khambati N, Song R, MacLean EL-H, et al. The diagnostic yield of nasopharyngeal aspirate for pediatric pulmonary tuberculosis: a systematic review and meta-analysis. BMC Glob Public Health. 2023;1:18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mukherjee A, Singh S, Lodha R, et al. ; Delhi Pediatric TB Study Group. Ambulatory gastric lavages provide better yields of Mycobacterium tuberculosis than induced sputum in children with intrathoracic tuberculosis. Pediatr Infect Dis J. 2013;32:1313–1317. [DOI] [PubMed] [Google Scholar]
- 16.Graham SM, Cuevas LE, Jean-Philippe P, et al. Clinical case definitions for classification of intrathoracic tuberculosis in children: an update. Clin Infect Dis. 2015;61Suppl 3(Suppl 3):S179–S187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.TB case and treatment outcome definitions. In: Guidance for National Tuberculosis Programmes on the management of tuberculosis in children. 2nd Edition. World Health Organization; 2014. Available from: https://www.ncbi.nlm.nih.gov/books/NBK214446/. Accessed April 6, 2024. [PubMed] [Google Scholar]
- 18.Thailand | UNAIDS. Published March 24, 2025. Available from: https://www.unaids.org/en/regionscountries/countries/thailand. Accessed September 15, 2025.
- 19.Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sawatpanich A, Petsong S, Tumwasorn S, et al. Diagnostic performance of the Anyplex MTB/NTM real-time PCR in detection of Mycobacterium tuberculosis complex and nontuberculous mycobacteria from pulmonary and extrapulmonary specimens. Heliyon. 2022;8:e11935. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Anyplex™ MTB/NTM Real-time detection. Simultaneous detection and identification of MTB dissociated from NTM. © 2019-2025. Seegene Inc. Available from: https://www.seegene.com/assays/anyplex_mtb_ntm_realtime_detection#. Accessed September 15, 2025.
- 22.Drain PK, Gardiner J, Hannah H, et al. Guidance for studies evaluating the accuracy of biomarker-based nonsputum tests to diagnose tuberculosis. J Infect Dis. 2019;220(Supplement_3):S108–S115. [DOI] [PubMed] [Google Scholar]
- 23.Lapphra K, Sutthipong C, Foongladda S, et al. Drug-resistant tuberculosis in children in Thailand. Int J Tuberc Lung Dis. 2013;17:1279–1284. [DOI] [PubMed] [Google Scholar]
- 24.Margarit A, Simó S, Rozas L, et al. Adolescent tuberculosis: a challenge and opportunity to prevent community transmission. An Pediatr (Barc). 2017;86:110–114. [DOI] [PubMed] [Google Scholar]
- 25.Venturini E, Bortone B, Cini G, et al. Does multiple gastric aspirate collection increase sensitivity of M. tuberculosis detection in children with pulmonary tuberculosis? Eur J Pediatr. 2024;183:425–434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ioos V, Cordel H, Bonnet M. Alternative sputum collection methods for diagnosis of childhood intrathoracic tuberculosis: a systematic literature review. Arch Dis Child. 2019;104:629–635. [DOI] [PubMed] [Google Scholar]
- 27.Hanrahan CF, Dansey H, Mutunga L, et al. Diagnostic strategies for childhood tuberculosis in the context of primary care in a high burden setting: the value of alternative sampling methods. Pediatr Int Child Health. 2019;39:88–94. [DOI] [PubMed] [Google Scholar]
- 28.World Health Organization (WHO). 2024. Global tuberculosis report. Data by country. Tuberculosis incidence Thailand, 2023. Available from: https://www.who.int/publications/i/item/9789240101531. Accessed April 2, 2025.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.