The role of the Xpert MTB / RIF assay among adolescents suspected of pulmonary tuberculosis in Rio de Janeiro , Brazil

Introduction: The teste rápido molecular para tuberculose (TRM-TB) was introduced in 2014 in Brazil for tuberculosis screening. However, its role in adolescents in Brazil has not been studied. Methods: A descriptive study of adolescents with suspected tuberculosis using National Laboratory software. Results: Of 852 (15.4%) suspected cases, 131 were positive by TRMTB and 2% were resistant to rifampicin. Among TRM-TB-positive cases, 105 (91.4%) were culture-positive. Sixty-four of 96 samples were sensitive to rifampicin by TRM-TB; 11 were resistant to other drugs by drug sensitivity test (DST). Conclusions: Among suspected cases, 16% were diagnosed by TRM-TB, of which 17% were drug-resistant by DST.

According to World Health Organization (WHO) data, of the 30 highest tuberculosis (TB)-burden countries accounting for 87% of cases worldwide, Brazil ranks 20 th in the number of new TB cases annually 1 .In 2015, 68,000 new cases were recorded in Brazil, an incidence rate of 33.6/100,000.In the City of Rio de Janeiro, the TB incidence rate was 91/100,000 inhabitants, one of the highest nationwide ( ECC Soares: unpublished data).
Molecular diagnosis of TB was introduced in Brazil through the Xpert MTB-Rif system(Cepheid, Sunnyvale, U.S.A.).[Teste rápido molecular para tuberculose (TRM-TB)], which is based on real-time polymerase chain reaction (PCR).It is an automated, rapid, and easy laboratory test that simultaneously detects Mycobacterium tuberculosis (Mtb) and rifampicin resistance (RIF) in approximately two hours.In October 2009, the TRM-TB was authorized for use in Brazil, and the WHO recommended its use in TB-endemic countries in December 2010 [2][3][4] .In Brazil, bacteriological confirmation increased with use of the TRM-TB; however, its incorporation in the public health network since 2014 has not yet affected the global notification of TB cases in Rio de Janeiro 5 .
While the performance of the TRM-TB has been evaluated in adults and children, to the best of our knowledge, no studies have focused on its use in adolescents.In general, children have paucibacillary TB, which allows bacteriological confirmation in only 20% of cases.After the age of 10, patients begin to have bacilliferous forms of TB similar to those found in adults, making bacteriological diagnosis more feasible 3 .Therefore, we hypothesized that the TRM-TB would detect a higher proportion of adolescent TB cases compared to that in younger children but a lower proportion compared to that in adults.
We performed a cross-sectional study of adolescents, defined as individuals between 10 and 18 years of age, meeting the following inclusion criteria: 1) suspected of having pulmonary TB (PTB), 2) receiving care in the Primary Health Care Clinics in the municipality of Rio de Janeiro, Brazil, between January 1 and December 31, 2015, and 3) had sputum, induced sputum, and/or gastric aspirate samples analyzed by TRM-TB.We reviewed the detection of Mtb and RIF resistance by the TRM-TB in these samples.Eligible adolescents were identified from a database created from the Gerenciador de Ambiente Laboratorial (GAL), national laboratory software for the monitoring and controlling of laboratory exams implemented by the Ministry of Health of Brazil (MoH) 6 .
We described the frequency of the detection of Mtb and RIF resistance by the TRM-TB.We also report the Mycobacteria Growth Indicator Tube (MGIT) culture results for Mtb and conventional drug sensitivity test (DST) to anti-TB drugs.The data are presented as frequencies and percentages.

Ethical considerations
This study was approved by the Instituto de Puericultura e Pediatria Martagão Gesteira da Universidade Federal do Rio de Janeiro Research Board (Proceeding 961.452 of 02/24/2015) and by the Municipal Health Secretariat of Rio de Janeiro (Proceeding 1.629.126de 07/08/2016).
In 2015, in the municipality of Rio de Janeiro, 852 TRM-TB exams of spontaneous sputum, induced sputum, or gastric aspirate in adolescents were registered in the GAL platform.Of these, 824 were spontaneous sputum, 22 were induced sputum, and six were gastric aspirates.Mtb was not detected in 716 (84%) of the 852 samples and the test was non-conclusive in two samples (0.23%).Mtb was detected in 134 (15.7%) samples by TRM-TB, of which 131 (98%) and 3 (2%) were RIF-sensitive and RIF-resistant, respectively.Of the three adolescents with samples resistant to RIF by TRM-TB, one had a negative culture [microbial growth <20 urine flow cytometry (UFC)] and the other two had positive cultures.One of the two was also resistant to isoniazid by conventional DST (MGIT).
Culture was performed on 105 samples from among 131 adolescents with RIF-sensitive Mtb detected by TRM-TB.Ninety-six of these cases were culture-positive (91.4%).Conventional DST was performed on 64 of the culture-positive samples.Eleven (17%) were resistant to other drugs: 6 (9%) to streptomycin, three (5%) to isoniazid, and two (3%) to RIF and isoniazid.The distributions of the TRM-TB results, Mtb culture, and DST are illustrated in Figure 1.
TRM-TB is indicated for adolescents in Brazil suspected of having PTB.The study of TB among adolescents is difficult because disaggregated data for this age group seldom exist.The literature suggests that childhood TB comprises individuals <15 years of age; older individuals are grouped with adults.The lack of adolescent-specific reporting has led to the invisibility of the needs of this age group.
Most adolescents (≥10 years) have bacilliferous PTB, unlike the paucibacillary PTB seen in children <10 years 3,7 .A high sensitivity (around 95%) of the Xpert is only reached in patients with bacteriologically confirmed TB 8 .The detection of Mtb in 16% of suspected adolescent PTB cases by TRM-TB is similar to the findings of a study of adults in Rio de Janeiro 9 .Unsurprisingly, our Mtb detection rate was higher than the 10.5% Xpert rate reported by Raizada et al. 10 in India among 4,600 children (0-14 years of age) suspected of having TB.In this paper, Xpert was performed under optimum collection and transportation conditions and its sensitivity was double that of smear microscopy.Similarly, in a hospital-based study, Nicol et al. 11 assessed children and adolescents <15 years, of whom 24% were HIV-infected.The Xpert positivity among the hospitalized TB suspects was 13%.In contrast, Bunyasi et al. 12 investigated children <4 years of age in South African hospitals, reporting an Xpert positivity rate of 1.2% among subjects with suspected TB.This finding is unsurprising because TB forms are predominantly paucibacillary in this age group.In another study of adult TB suspects, the Xpert positivity ranged from 13-30% and increased with TB prevalence in the study setting 13 .
In the present study, it was not possible to compare the TRM-TB and smear microscopy results.The guidelines from the MoH indicate that smear microscopy should only be requested for diagnosis in individuals who had been previously treated for TB 14 .However, a study in Rio de Janeiro did not show concordance between the smear microscopy and TRM-TB results in children 5 .
In our study, 98% of samples from adolescents with detectable TRM-TB results were sensitive to RIF, consistent with the low level of resistance of Mtb to RIF in Brazil.This finding was confirmed by the high detection of sensitivity to RIF by conventional DST.Of the 64 of 105 samples that were RIF-sensitive by TRM-TB, conventional DST showed resistance to other drugs in eleven (17%) samples.The most common was resistance to streptomycin, a drug widely used previously in Brazil.Among the 64 suspicious cases in which TRM-TB genotypically detected RIF-sensitivity and in which conventional DST was performed, two were phenotypically  13,15 .
Among the limitations of this study were the lack of clinical data and HIV test results of the adolescents with suspected TB.Therefore, we are unable to match the TRM-TB results with clinical diagnoses.In addition, we do not have conventional DST data for all samples in which Mtb was detected by the TRM-TB.
The TRM-TB may improve the diagnosis of PTB in adolescents.Since 25-30% of adolescents with PTB in Brazilian cities are negative by smear microscopy 7 , the use of TRM-TB may increase the number of TB-confirmed-cases.However, a negative TRM-TB result in adolescents does not eliminate the possibility of PTB.Clinical, radiological, and epidemiologic data are still very important in pediatric practice 3,8 .
Such individuals would go undetected by the TRM-TB.A possible reason for the discrepancy between conventional DST and Xpert is a mutation in sites not yet assessed by the TRM-TB