This retrospective study was based on discharge data from the Chengdu Public Health Clinical Center (PHCC), located in the central city of Sichuan Province in southwest China. Patients were recruited from January 2017 to December 2024 at PHCC based on either etiology [cases defined as bacteriologically confirmed when biological specimens were positive by smear microscopy, polymerase chain reaction (PCR), or culture, in accordance with World Health Organization (WHO) guidelines] or clinical confirmation [cases without etiological support but diagnosed through imaging findings, Bacillus Calmette-Guérin (BCG) scar assessment, interferon-gamma release assay (IGRA), or response to effective anti-TB chemotherapy]. After eliminating duplicate cases, a total of 2,208 pediatric TB patients aged ≤14 years were included in the study. All information about these cases was obtained from the hospital medical record management system and hospital information system (HIS), including detailed demography, the geographical distribution of habitual residence, clinical characteristics, outcomes, and other relevant information (Figure 1).

The medical record information of each patient extracted from the HIS included sociodemographic, TB-related and outcome variables. Sociodemographic variables consisted of gender, age, ethnicity, and place of habitual residence. The TB-related variables included TB type, specific diagnosis, complications, and hospital admission time. The outcome variables included treatment outcomes, length of stay in hospital, and TB hospitalization costs. Information about the habitual residence was collected for all cases, then coded according to their residence (using geographic codes), and matched to a 1:100,000 digital map of China using Python 3.7. The cases were displayed using detailed geographical distribution charts of China and Sichuan Province.

Diagnosis of pediatric TB was based on three key standards: the Chinese Pulmonary Tuberculosis Diagnostic Criteria (WS 288–2017), the Chinese TB guidelines for clinical diagnosis and treatment (Chinese Medical Association, 2005), and the updated WHO guidelines (World Health Organization, 2024). The diagnostic workup included smear microscopy, symptom-based clinical evaluation, culture isolation, and molecular testing. Following WHO guidelines, pulmonary tuberculosis (PTB) was defined as TB disease involving the lung parenchyma and/or tracheobronchial tree, including miliary TB cases, with or without extrapulmonary tuberculosis (EPTB), while all other disease sites were classified as EPTB. Concurrent pulmonary and extrapulmonary TB (combined TB) was defined as any TB case involving both lungs and other organs. For mycobacterial culture, we used the BACTEC™ MGIT 960 System. For species identification, MPT64 antigen detection (using colloidal gold immunochromatography), p-nitrobenzoic acid (PNB), and thiophene-2-carboxylic acid hydrazide (TCH) were used to distinguish MTB from Nontuberculous mycobacteria (NTM) (Cao et al., 2021). Polymerase chain reaction (PCR) was then used for further species/complex identification following the manufacturer’s instructions (CapitalBio Corp., Chengdu, China) (Wang et al., 2020).

Drug sensitivity testing (DST) was performed using both phenotypic (conventional) and genotypic (molecular) methods. The rapid molecular detection methods included Xpert MTB/RIF and fluorescence PCR melting curve analysis, both capable of detecting isoniazid and rifampicin resistance within 1 day. The fluorescence PCR melting curve method using the MeltPro® TB integrated detection system (Zeesan Biotechnology Co., Ltd., Xiamen, China) could identify drug-resistant mutations in rpoB (rifampicin), KatG/inhA/ahpC (isoniazid), rpsL/rrs (streptomycin), embB (ethambutol), and gyrA (fluoroquinolones). Phenotypic DST was performed using the MicroDST™ system (Yinke AUTOBIO Diagnostics Co., Ltd., Zhuhai, China), testing susceptibility to 14 antituberculosis drugs: first-line drugs including isoniazid (0.2–1.6 μg/mL), rifampicin (1–8 μg/mL), ethambutol (2.5–20 μg/mL), and streptomycin (1–8 μg/mL); and second-line drugs including rifapentine (0.5–2 μg/mL), levofloxacin (2–8 μg/mL), amikacin (1–4 μg/mL), capreomycin (2.5–10 μg/mL), protionamide (10–40 μg/mL), pasiniazide (0.5–2 μg/mL), moxifloxacin (0.5–2 μg/mL), p-aminosalicylic acid (2–8 μg/mL), clarithromycin (4–16 μg/mL), rifabutin (0.75–3 μg/mL), kanamycin (2.5–10 μg/mL), and clofazimine (2–8 μg/mL). Quality control was maintained using the H37Rv reference strain (ATCC 25618).

External quality assessment (EQA) was conducted for smear microscopy, culture, and DST through the Innovation Alliance on Tuberculosis Diagnosis and Treatment (Beijing, China). Additionally, about 10% of isolates were randomly selected for blinded retesting at a superior provincial-level Centers for Disease Control and Prevention laboratory.

All statistical analyses were performed using STATA 17.0 (StataCorp, College Station, TX). Continuous variables with normal distribution were expressed as mean ± standard deviation (SD) and compared using analysis of variance (ANOVA). Non-normally distributed variables were analyzed using nonparametric tests. Categorical variables were presented as frequencies (percentages) and compared using χ² tests or Fisher’s exact tests as appropriate. A two-tailed p value <0.05 was considered statistically significant.

This study protocol was reviewed and approved by the Ethics Committee of PHCC (Approval No. YJ-K2023-08-01). Since all patient data were obtained through routine clinical care and the mandatory national TB reporting system, the requirement for individual informed consent was waived by the ethics committee.

Among the 2,208 pediatric TB patients in this study, 90 cases were diagnosed with pediatric DR-TB. In pediatric DR-TB patients, 43.3% (39/90) were male and 56.7% (51/90) were female, resulting in a male-to-female ratio of 0.76:1. The mean age of the 90 pediatric DR-TB patients was 10.94 ± 3.52 years (range: 279 days to 14 years). Among the patients, 6.7% (6/90) were aged 0–4 years, 21.1% (19/90) were aged 5–9 years, and the highest proportion was in the 10–14 years group (72.2%, 65/90). Among the 90 pediatric DR-TB cases, 74.4% (67/90) were primary patients, and RR-TB and MDR-TB accounted for 36.7% (33/90) and 30.0% (27/90), respectively. The number of retreated cases accounted for 25.6% (23/90) of the total cases, and the MDR-TB in the retreated cases was 16.7% (15/90). Regarding population distribution, the Tibetan ethnic group had the highest proportion of DR-TB cases in children (63.3%, 57/90) (Supplementary Table 1; World Health Organization, 2009).

Out of 90 cases of pediatric DR-TB, 84.4% (76/90) showed abnormal chest x-rays, displaying varying degrees of the following findings: patchy shadows (68.9%, 62/90), nodules (65.6%, 59/90), cavity (21.1%, 19/90), hydrothorax (20.0%, 18/90), and hydropericardium (3.3%, 3/90). The most commonly associated clinical symptoms were cough (54.4%, 49/90), headache (15.6%, 14/90), and vomiting (12.2%, 11/90). The most common co-occurring complications were hyperuricemia (44.4%, 40/90), pulmonary bacterial infection (41.1%, 37/90), and anemia (23.3%, 21/90). Among the 90 children with DR-TB, 94.4% (85/90) were discharged after their condition improved and was controlled, while 5.6% (5/90) of patients requested early discharge for personal reasons (Supplementary Table 1).

Among the pediatric DR-TB patients, 13.33% (12/90) had PTB, 5.56% (5/90) had EPTB, and 81.11% (73/90) had combined TB (Figure 2A) (χ² = 5.687, p = 0.031). The most common form of pediatric DR-TB was TB lymphadenitis (30.00%), followed by pleural TB (20.71%), abdominal TB (19.29%), TB meningitis (14.29%), pericardial TB (10.00%), skeletal TB (2.86%), and other forms of EPTB (2.86%). Cases of disseminated TB involved 2 extrapulmonary lesions in some patients, 16 cases at 3 sites, and 6 cases at 4 sites (Figure 2B).

Among the pediatric DR-TB patients in this study, all were diagnosed as positive for the pathogen, and 3 cases were complicated with NTM infection (including 1 case each of M. lentiflavum, M. fortuitum, and M. simiae). DST results of 90 DR-TB cases were confirmed by mycobacterial culture in 33 cases and molecular diagnosis in 57 cases. Except for one case of female urogenital TB with single fluoroquinolone resistance, cases were divided into four drug resistance groups based on DST results: rifampicin-resistant tuberculosis (RR-TB), isoniazid-resistant tuberculosis (HR-TB), MDR-TB, and XDR-TB. The classification and statistical results showed that RR-TB, HR-TB, MDR-TB, and XDR-TB incidence from 2017 to 2024 showed varying degrees of upward trend (Figure 3A). The proportion of DR-TB in girls was significantly higher than that in boys (Figure 3B), with the majority of cases occurring in 11-14-year-old children (Figure 3C). In the distribution of tissue types in DR-TB cases, PTB/EPTB co-infection accounted for the highest proportion (Figure 3D). In terms of ethnic distribution, the proportion of Tibetans in RR-TB, MDR-TB, and XDR-TB groups was significantly higher than that in other groups (Figure 3E).

Among 2,208 children with TB from 2017 to 2024 included in this study, the incidence rate of pediatric DR-TB increased from 0.42% in 2017 to 5.61% in 2024, with an average incidence rate of 4.08%, showing an upward trend year by year (p = 0.011) (Figure 4). The 90 pediatric DR-TB patients diagnosed and treated at PHCC between January 2017 and December 2024 were matched to a 1:100,000 digital map of China using Python 3.7. The geographical distribution map indicated that over the 8-year period, twelve of the pediatric DR-TB patients were from neighboring cities outside Sichuan Province [including Tibet (8 cases), Qinghai (3 cases), and Gansu (1 case)], while the remaining cases were from within Sichuan Province, primarily from the western Sichuan, Ganzi, Aba, and Liangshan minority areas, with the highest number of cases in the Ganzi Tibetan Autonomous Prefecture. The remaining cases were mainly distributed in Chengdu, Leshan, Bazhong, and other central urban areas (Figure 5).