The clinical presentation of tuberculosis (TB) in the paediatric age group differs from that in adults, which makes TB a particular challenge in the former: a) the risk of developing disease following infection is greater in children, and the latency period is shorter compared to adults; b) children are more likely to developing extrapulmonary TB and severe disease; c) clinical and radiological manifestations are heterogeneous and often nonspecific; d) collection of samples for microbiological testing is more difficult in children, and the disease is often paucibacillary.1–4

Tuberculosis disease in children is an epidemiological indicator of recent transmission of Mycobacterium tuberculosis in the community, revealing a public health system failure in controlling TB.3–5 Screening of contacts of index cases and administration of prophylaxis to those susceptible of developing the disease (specially children) are recommended to minimise its spread.4,6

In Portugal, the incidence and notification of cases of TB has declined by approximately 40% in the last decade, and Portugal has been a low-incidence country since 2015.7 A public health policy of universal vaccination with Bacillus Calmette-Guérin (BCG) was in effect until 2016, when it shifted to selective vaccination of individuals in at-risk groups.8 In 2017, its incidence was estimated at 5.8 cases per 100 000 children aged 0–4 years and 1.8 cases per 100 000 children aged 5–14 years.9 This year, imported TB cases amount to 19.5% of all TB cases notified in Portugal, a percentage that is lower compared to other countries in the European Union and European Economic Area (33.1%).9

Given the epidemiological context, we considered the importance of improving health care facilities where TB patients are diagnosed and treated, particularly children. In Portugal, there is a network of TB Outpatient Centres within the National Health Service where patients with suspected or confirmed TB are evaluated and treated. These centres are linked to different hospitals to allow a full investigation until a diagnosis is made. Only 2 of these centres have paediatricians and a multidisciplinary team specialized in TB. The first one is the Reference Centre for TB in Children in North Portugal, which has been operative since 2010. The aim of our study was to describe the cases of TB in children and adolescents managed in this reference centre during a 10-year-period.

We conducted a retrospective cross-sectional study at a TB centre, where children are managed by paediatricians specialized in TB. This centre managed 233 initial visits and 485 follow-up visits in 2017. We identified cases of TB disease in patients aged less than 18 years by consulting the records of the national TB surveillance system from 2007 through June 2017.10

We reviewed health records and collected data in an electronic database. We retrieved information on demographic characteristics (sex, age, enrolment in childcare facility or school), the medical history (previous diseases, vaccination status, use of medication), the reason and source of referral for consultation at the TB centre and the index case, when available. When it came to the cases of suspected TB, we collected data on clinical manifestations (symptoms and their duration), immunological test results (tuberculin skin test [TST] and interferon gamma release assay [IGRA]), chest radiographs, chest computed tomography (CT) scan, other imaging tests and bronchoscopy if performed. We considered a TST positive if the induration had a diameter of 5mm or greater in children that were immunocompromised or aged less than 5 years and not vaccinated, or of 10mm or greater in vaccinated children aged less than 5 years or children aged more than 5 years independently of vaccination status.6 We also recorded information on the collection of microbiological specimens (gastric secretion aspirate, sputum, bronchoalveolar lavage [BAL], cerebrospinal fluid [CSF], pleural fluid, biopsy), the results of acid-fast bacilli (AFB) smears, nucleic acid amplification tests (NAATs) and/or cultures and the presence of granulomas in the histological examination.

We classified TB by type as pulmonary or extrapulmonary, and cases as ‘possible’, ‘probable’ and ‘confirmed’ applying the definitions of the European Centre for Disease Prevention and Control (ECDC).11

Lastly, we collected data on the pharmacological treatment, including the agents used, dosage and duration of treatment, adverse events and adherence to treatment.

We collected data and performed the statistical analysis with the software IBM SPSS Statistics® version 24. We have expressed continuous variables as median and interquartile range (IQR) and categorical variables as absolute and relative frequencies. The study protocol was approved by the Ethical Committee and Administrative Council of the centre.

Paediatric tuberculosis reference centres play a key role in the early identification, screening and diagnosis of children exposed to tuberculosis. This specialized care allows children to be diagnosed in the early stages of disease or even when they are still asymptomatic, as was the case of 10 of our patients (21.7%).1,4

Twelve of the cases in our sample (26.1%) were confirmed. In the group aged less than 5 years, 23.8% of patients had a positive culture, which was consistent with the proportion reported in a study conducted in Northern Portugal between 2000 and 2009.16 The percentage of confirmed cases in patients aged 0–14 years reported for the same period in Europe was slightly lower (19.2%).17 There is evidence that collection of different microbiological specimens in addition to gastric aspirate samples is associated with a higher yield for TB confirmation, and can also guide the management.18,19

In our sample, the diagnostic yield of culture of BAL samples for detection of M. tuberculosis was higher compared to the yield for gastric aspirate samples (in one third of confirmed cases, cultures of BAL samples were positive and cultures of gastric aspirate samples negative). Most of the studies in the literature report a higher overall yield with the combination of both.18

Bronchoscopy should be considered as an additional tool to collect microbiological samples (BAL, endobronchial tissue, transbronchial needle-aspiration biopsy samples) when less invasive methods are ineffective.1,19 Bronchoscopic images can provide additional clues about TB, of which extrinsic compression of airway, identified in 12.5% of our cases, is the one reported most frequently (19%-24%). Other macroscopic changes described in the literature and observed in our sample are the presence of intraluminal caseous material, tissue granulation, endobronchial mucosal inflammation and mucosal erosion and ulceration.19

Despite positive detection in 85% of immunological tests, a negative IGRA or TST result does not rule out disease, and the sensitivity of these tests is reduced in immunocompromised patients.1,4,18,20 In almost half of the patients in our sample that had both tests done, the results of IGRA and TST were discordant, which was consistent with the literature.18

At present, there is no accurate test for diagnosis of paediatric TB, so while it is important to carry out the microbiological investigation, it is equally important to take the clinical manifestations into account and consider other available methods.4 An European survey of doctors/researchers with experience in paediatric TB found that the criterion used most frequently to initiate antituberculosis treatment without microbiological confirmation was the epidemiological context (89%), followed by signs and symptoms (86%), abnormal imaging findings (77%), the age of the patient (48%) and the results of immunological tests (45%).21 These percentages are similar to those found in our study (73.5% – 85.3%).

The selection of the pharmacotherapy regimen was based on the results of antimicrobial susceptibility testing of the M. tuberculosis isolate of the patient or the index case and the guidelines of the World Health Organization (WHO).13 Rapid molecular tests were not performed, as the study refers to years when they were not used routinely, but rather in case of risk of multidrug-resistant TB. In Portugal, the estimated prevalence of isoniazid resistance in 2018 was 5.3% (according to the most recent report, pending publication). The decision to prescribe a three-drug regimen (HRZ) was based in the known M. tuberculosis susceptibility of the index cases.

Corticosteroids were needed in cases with paradoxical reactions.12 These responses are thought to occur when treatment reverses a process of immune dysregulation, with worsening of clinical or radiological findings following initiation of appropriate antituberculosis drugs in the absence of evidence of recurrence or another diagnosis. These reactions are not rare, and the proportion of HIV-negative children with paradoxical reactions found in our study was within the range reported in larger studies (3.3%-10.2%).22,23

The regimens used in children have an excellent safety profile, which explains the low incidence of adverse reactions in our sample; routine biochemical monitoring is not necessary in children, unless clinically justified.1,12

The non-existent mortality and low morbidity observed in our sample could be explained by several factors: there were no children infected with HIV; only one patient had central nervous system disease; the potential impact of low socioeconomic status was minimised by free health care and medication, and the use of daily observed therapy and a client-centred approach.

Other important factors were the follow-up by a team of paediatricians specialized in TB and the advantages offered by a paediatric TB reference centre fully integrated in the national TB programme. However, it is important to continue educating the broader community of clinicians and the general population to guarantee early diagnosis of TB in children.

The identification of children at risk of TB infection, early diagnosis of the disease and successful treatment of paediatric cases are key for the purpose of reducing future cases in adults, community spread and the development of sequelae. Although many cases are not confirmed by testing, in the presence of clinical features suggestive of tuberculosis, efforts should be made to collect specimens for microbiological testing before initiating treatment. The pharmacotherapy regimen should be selected based on current clinical guidelines and the clinical response of the patient. Follow-up by specifically trained paediatricians in outpatient TB clinics in the community allows more adequate management of paediatric TB, while links to referral hospitals are important to allow further investigation.