TABLE 1 Diagnostic approaches in children History Symptoms: chronic, unremitting symptoms including cough, weight loss (or failure to put on weight as expected), fever, reduced playfulness: Nonspecific, more atypical (even acute) with immature/impaired immunity (infants/younger children or HIV/malnourished) Initial children of CNS are subtle and nonspecific History of exposure to an individual with infectious PTB: Less valuable as children grow older (more sources of contacts) Challenging often, the index case is still undiagnosed (importance of reverse contact tracing) Specific symptoms locating disease to the site of possible EPTB Examination Weight loss or failure to gain weight appropriately (check curves) General: cachectic, lethargic PTB: coughing, wheezing, difficulty breathing, haemoptysis (a rare and late sign, generally seen only in adolescents), crackles or wheeze on auscultation Signs of EPTB dependent on site of disease Immune tests IGRA, TST and novel TST: These tests demonstrate immunological sensitisation to M. tuberculosis and do not discriminate TBI from TB disease (they do not indicate viable bacilli, only previous exposure) however, in a child or adolescent who has symptoms and signs that could be consistent with TB but also consistent with another diagnosis, a positive IGRA/TST may suggest an increased probability that their symptoms/signs are due to TB Conversion period of up to 3 months to become positive following infection TST has low specificity due to false positives for BCG and NTM requires two visits (often with need for boosting at 8 weeks) IGRA requires laboratory infrastructure, phlebotomy (4 mL) and is expensive False negatives in young infants and with immunosuppression, severe malnutrition and severe TB Novel TST has better specificity than original TST but still requires two visits and has imperfect sensitivity Microbiology Sample collection: Respiratory samples include expectorated sputum, induced sputum, nasopharyngeal aspirates, gastric aspirates or washings, and stool samples (stool detects M. tuberculosis from both abdominal TB and from PTB that has been coughed up and swallowed automated processing systems have been developed) Ideally, more than one and diverse extrapulmonary samples might include cerebrospinal fluid to evaluate for TB meningitis, lymph-node aspiration biopsy samples for lymph-node disease, urine samples for renal or bladder TB, peritoneal fluid, biopsy samples from bone or joint samples Sample processing: Smear microscopy: a sample of sputum is smeared onto a microscope slide, stained with Ziehl–Neelsen stain (conventional) or auramine (fluorescence) and visualised under a microscope to detect M. tuberculosis this test is cheap and fast but is associated with poor sensitivity, particularly in those with paucibacillary disease, and has poor specificity with NTM Mycobacterial culture can be on solid or liquid medium and is the most sensitive test to identify M. tuberculosis but is more expensive than a smear or NAAT, and commonly takes many days/weeks for a culture to be positive compared with solid culture, liquid culture is more expensive and has higher contamination rates but is more sensitive once cultured, either phenotypic or genotypic DST can be carried out Continued 216 https://doi.org/10.1183/2312508X.10025322 ERS MONOGRAPH |THE CHALLENGE OF TB IN THE 21ST CENTURY