A 4-year-old girl visited the pediatric gastroenterology outpatient clinic with recurrent abdominal pain and chronic diarrhea for the past year. She had previously been healthy until she developed these symptoms. Her height was 104 cm (25th–50th percentile) and weight was 13 kg (less than 3rd percentile). She did not have any fever or cough. An inflammatory bowel disease and resulting malnutrition were highly suspected as an initial diagnosis. Her white blood cell count was 7,820/μL (neutrophil 59%, lymphocyte 39%), hemoglobin (Hb) level was 7.8 g/dL, and platelet count was 571,000/μL. Her C-reactive protein level was 5.9 mg/dL (normal range, 0–0.3 mg/dL). Stool occult blood was positive and calprotectin in stool specimen was markedly elevated (>1,000 μg/g; normal range, <50 μg/g). Abdominal computed tomography (CT) performed in the evening of the day of admission revealed multiple necrotic lymphadenopathy with peritonitis and wall thickening of colon (Fig. 1A, B). Additional chest CT was performed emergently and the patient was put into isolation. Chest CT showed multiple calcified mediastinal, hilar lymphadenopathy, and tree-in-bud opacity in both lungs (Fig. 1C, D).

After the chest CT, the parents revealed the additional history information that she had not received any vaccination including BCG after birth and had been treated only with herbal medication by a traditional healer for her condition. Of note, a close relative (grandaunt) was diagnosed with pulmonary TB when the child was 15 months old.

Based on this history and radiologic findings, TB was strongly suspected. Acid fast bacilli (AFB) staining with culture was performed for 3 consecutive days with gastric aspirates. No AFB were observed by staining and a polymerase chain reaction test was also negative. Colonoscopy showed skipped lesions of ulcers and pseudopolyps from proximal transverse to descending colon (Fig. 2). Pathologic findings of the biopsy specimens from colon showed active inflamed granulation tissues. Cerebrospinal fluid study was normal and bone scan did not show any abnormal uptake. Her interferon gamma releasing assay (IGRA) test was positive.

After initial evaluations, she immediately started anti-TB medication. After 5 days of treatment the abdominal pain improved and her oral intake increased. She was discharged after 7 days of anti-TB treatment and continued isolation at home. Of note, she was apathetic and did not show any emotional response for communication on the admission day. However, she showed emotional responses with facial expressions including smiling at discharge. Her body weight at discharge was 14.8 kg. Cultures from all 3 gastric juice specimens finally grew M. tuberculosis after 17 days and 2 out of 5 colon biopsy specimens grew M. tuberculosis 16 days later. Isolates were sensitive to all anti-TB medicines.

A whole body magnetic resonance imaging (MRI) scan was performed to investigate other organ involvement. Although a previous bone scan showed negative findings, MRI showed diffuse abnormal bone marrow signal changes on axial and appendicular skeleton highly suggestive of bone marrow involvement of TB.

After 8 months of treatment, the patient showed marked improvement of anemia and nutritional status. Her height was 110.8 cm (50th–75th percentile), weight was 17.6 kg (25th–50th percentile), and Hb level was 13.3 g/dL. She started childhood vaccination with an accelerated schedule according to Korean NIP.

Exposed individuals were divided into 3 groups of household, nosocomial, and kindergarten contacts. Investigations were performed according to the Korean guidelines for TB [5]. Data on kindergarten contacts was provided by Korean Center for Disease Control and Prevention (KCDC).

All exposed individuals underwent chest X-ray and tuberculin skin tests (TSTs) or IGRA tests were also performed. Contacts with known positive results for TST or IGRA were tested only with chest X-ray. (for example, health care workers who had a positive result from regular TB screening).

Household and nosocomial contacts were considered close contacts. Children younger than 5 years received a serial TST (2 TST with 8- to 10-week interval). Adults and children aged ≥5 underwent testing with TB IGRA.

Kindergarten contacts were divided into 2 groups of close and casual contacts. Among children, close contact was defined as sharing the same classroom for more than 8 hours a day or 40 hours in total. Children in the same class or adjacent class sharing the aisle, and those riding the same bus were included in this group. Casual contacts were defined as other children in the same kindergarten. Among adults, teachers of the class of the index patient and the bus driver were classified as close contacts. Casual contacts were defined as all other adult employees. Children were tested with TST and adults were evaluated with IRGA. Pediatric close contacts were evaluated with a serial TST. Pediatric casual contacts were evaluated with one TST after 8 weeks of last exposure. Adult casual or close contacts without previous positive IGRA test results were evaluated with one IGRA test after 8 weeks of last exposure. The Institutional Review Board at Samsung Medical Center, Seoul, South Korea reviewed and approved this study (approval number: 2018-09-028-002). The chi-square test was used to analyze categorical variables. A P value of <0.05 was considered statistically significant. All statistical analysis was performed using the R ver. 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria).

The patient’s parents did not have any evidence of TB infection, with normal chest X-ray and negative IGRA results. Her 12-monthold brother also did not receive the BCG vaccine and was thoroughly evaluated. He was diagnosed with latent TB and was placed on isoniazid treatment for 9 months. Her brother also received childhood vaccination of accelerated schedule according to Korean NIP.

The patient was initially admitted to a shared room with other children on a general ward but was isolated after 19 hours when TB was suspected. Because of this nosocomial exposure, a contact investigation was performed among other pediatric patients (n=4) and adults (n=6) who were considered to have significant exposure. These individuals were followed up for 2 months. Chest X-ray and serial TB-specific IGRA tests were performed at 8- to 10-week intervals. At 3 months, none of these individuals was identified as infected with TB from this exposure. Her TB infection was reported to the health authority and contact investigations in her family and kindergarten were also performed by the KCDC.

A total of 314 exposed individuals were investigated: 44 adults and 270 children. Data on demographic information, extent of exposure, and assessment results of kindergarten children are shown in Table 1.

Of all exposed children, 257 (95.1%) completed LTBI tests and 17 (6.3%) of all exposed were diagnosed with LTBI. LTBI diagnosis was more frequent in children who had close contact with the index case (7 of 41 [17.1%] close contacts compared with 10 of 229 [4.4%] casual contacts; P=0.007).

Among the 229 casual contacts of all ages, LTBI was diagnosed in 4.8% (2 of 42) of children younger than 5 years and 4.3% (8 of 187) of children 5 years or older (P=0.9).

There were no LTBI cases from this exposure among adults. As of 12 months from the investigation, there was no active TB case from this exposure.