High diagnostic yield of endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) in the diagnosis of adolescent pulmonary tuberculosis

The global burden of infections with M. tuberculosis (MTb) remains high [1]. Furthermore, in the European Union (EU), MTb infections constitute a substantial public health issue [1]. Although Tb incidence in Germany was at a stable, low level in the years prior to 2013, a significant increase in Tb was seen thereafter, with numbers peaking in 2016 [2]. Largely, this rise was driven by the increase of migrants seeking asylum in Germany during this period. As in many other low-incidence countries, Tb incidence in Germany reflects international migration dynamics: in 2018, 69.8% of newly diagnosed Tb patients in Germany were foreign-born. This translates to an incidence approximately 18 times higher among foreign nationals than among German citizens (37.3 vs. 2.1 cases per 100,000 population, respectively). Patients with a migration background were of younger in age as compared to Germans (age median 28 vs. 59 years). For patients with or without migration background, the pulmonary Tb form prevailed (72.9%) with a low detection rate of multiple drug resistance (3.1%; MDR) [2].

In children, Tb diagnostics are limited both by the lower performance of immunodiagnostic tests, such as the interferon-gamma-release assays (IGRA), and by the low yield of microbiological cultures, due to the paucibacillary nature of the disease [3]. In adolescents, a bacteriological confirmation of pulmonary tuberculosis with positive smear, PCR or culture can be obtained more frequently. However, paucibacillary disease remains a common finding [4]. For several additional reasons, refugees with suspected tuberculosis represent a particularly challenging patient cohort for Tb diagnostics. First, the Tb index patient, (and therefore the drug sensitivity of the infectious MTb), usually is unknown. Second, Tb symptoms, which are often non-specific in children, are even more difficult to elicit due to a variety of factors: language barriers and absence of patient history, as well as the manifestation of symptoms such as weight loss, which may directly result from the poor circumstances of the patient’s migration [5]. Third, empirical Tb treatment based upon mycobacterial resistance patterns in the country of origin is not reliable, because the infection may have occurred in one of the transit countries on the migration route. Fourth, in refugee centers visited by the URMs on their migration route, a significant proportion of patients undergo incomplete treatment [6, 7]. Altogether, these factors underline the utmost importance of culture confirmation of suspected Tb. This should include drug sensitivity testing (DST) for especially vulnerable minor refugees entering the EU.

Among adults, endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) has emerged as a valuable procedure for the culture confirmation of suspected MTb intrathoracic lymphadenopathy. This minimally invasive technique allows for ultrasound-guided direct sampling of intrathoracic lymph nodes and it has been shown to increase the sensitivity and culture yield among adult patients with pulmonary Tb [8‐11]. However, there is only limited information regarding the utility and safety of EBUS-TBNA in children and adolescents for Tb diagnostics [12‐15]. Because many pediatric and adolescent Tb patients initially present with isolated hilar lymphadenopathy and without pulmonary infiltrates [3], sampling techniques such as collection of induced sputum or gastric aspirates typically provide only a low yield of bacterial culture-positivity. The implementation of PCR-based diagnostics have significantly increased the case detection rate of suspected tuberculosis. It has been shown that the pooled sensitivity of GeneXpert in children reaches 62% with a specificity of 92% (with culture as reference standard), and a 82–94% pooled sensitivity in smear positive children aged 0–4 years [16]. However, as culture remains the current diagnostic gold standard, minimal-invasive, direct sampling of enlarged lymph nodes may offer an attractive method for testing pediatric patients with suspected Tb in situations where DST is essential.

Here, we present the results of a retrospective cohort of 34 unaccompanied minor refugees (URM), who, over a two-year period (January 2016–July 2018), were referred to our tertiary care center for workup of suspected pulmonary Tb and who required related invasive diagnostic procedures. Of note, lymph node samples acquired through EBUS-TBNA had the highest yield of all analyzed specimens.

Among children and adolescents, diagnostic yields from microbiologic techniques from bronchoalveolar lavage and sputum provide only a low yield. As a result, many patients have to be treated with empirical anti-tuberculous drugs. EBUS-TBNA is a minimally invasive, diagnostic procedure with relatively high diagnostic yield. Among adults, it is broadly used [18, 19], but in the pediatric population it is rarely applied [15].

The first case of EBUS-TBNA in a 13-year-old male child with sarcoidosis was published in 2009 [20]. To date, four larger, retrospective studies evaluating EBUS-TBNA use in the diagnosis of mediastinal lymphadenopathy, including Tb in children, have been published [12, 14, 21, 22]. Of note, in most of these studies, EBUS-TBNA was not performed for specific Tb diagnostics, but for the workup of enlarged mediastinal lymph nodes of non-infectious origin. Additionally, individual case reports have documented its utility in Tb diagnostics for children and adolescents [13, 23]. Because the use of EBUS-TBNA is limited by the size of airway diameter, the number of children under 12 years old who have been reported to have received EBUS-TBNA is low. In younger patients, an alternative method that does not impede ventilation, e.g., endoscopic ultrasound bronchoscope-guided fine needle aspiration (EUS-B-FNA), needs to be applied [12, 14, 21, 24]. In the literature, the youngest patient described for whom TBNA with a 4 mm scope has been used to evaluate mediastinal lymphadenopathy was a nine-month-old infant [22]. Because the patients in our study were ≥ 14 years of age, there was no anatomical hindrance to using EBUS-TBNA.

Out of the 34 URM eligible for our study, EBUS-TBNA was perfomed in 21 patients. The diagnostic yield of the fine-needle aspiration (14/21; 66.7%) was higher as compared to other pediatric studies (36.0–56.9%) [12, 14, 21, 22], and also as compared to the EBUS-TBNA diagnostic yield reported in adult patients (64.6–96.6%) [19], as well as to meta-analysis with pooled sensitivity of the procedure of 80% [25]. Of the 14/21 patients who were positive through EBUS, culture was positive in 9/14 samples (64.3%), PCR in 13/14 samples (92.9%) and microscopy in 1/14 samples (7.1%). This finding illustrates that at least genomic material of mycobacteria is present in the great majority of enlarged intrathoracic lymph nodes. However, the lower rate of culture positivity compared to PCR detection suggests that lymph node enlargement may persist after successful immunological clearance of replication competent mycobacteria from a lymph node. Of note, our PCR detection rate of 27% from BAL samples was lower than in published pediatric cohorts, reporting a sensitivity of up to 78% [26, 27], We hypothesize this may be due to the mostly paucibacillary disease of our patients, where the majority of pathogen burden is contained within lymph nodes. Overall, EBUS was the only positive diagnostic method via PCR or culture in 7/14 patients, illustrating the benefit of this diagnostic approach.

In our evaluation of several Tb detection methods, the molecular genetic approach using PCR performed best. PCR provided a positivity of 70.6% (24/34), followed by culture proof (19/34; 55.9%) and detection with ZNS (5/34; 14.7%). These findings correspond to already published data for adults [10, 19]. Using a combination of invasive and non-invasive diagnostic approaches allowed us to detect MTb and evaluate drug sensitivity in 93.5% (29/31) active Tb cases. Importantly, a molecular or cultural diagnosis of Tb was possible in 6/7 lymph note specimens without detectable granulomas and a negative Ziehl-Nehlsen staining, illustrating the insufficient negative predictive value of negative histology.

Our detection of drug resistance in 17.2% (5/29) of patients with active Tb (monoresistance in 10.3% (3/29); MDR in 6.9% (2/29) patients) underlines the importance of sample acquisition for drug sensitivity testing. This drug resistance level is higher than has been reported elsewhere (MDR in 3.3.% pediatric and 5.0% adult patients, respectively) [10, 21].

Contrary to study of Gulla et al. [21], where approximately one-fourth of patients were treated with empiric anti-tuberculous medication, none of our patients received any tuberculostatic drugs during the period of microbial sampling. This fact may have contributed to the higher diagnostic yield, especially with respect to culture-positivity.

Six of our patients had a history of tuberculostatic treatment. In two of them, MTb detection failed. For another two, MTb could be detected in sputum (1 x ZNS and PCR; 1 x PCR) and in pleural effusion or BAL (PCR). In one patient, MTb was found in BAL (ZNS, culture, PCR) and in pleural effusion (ZNS and PCR). In the last patient, MTb detection was only possible in a LN sample gained via EBUS-TBNA (PCR).

In our cohort, no complications occurred following EBUS-TBNA. Others have reported rare (i.e., sepsis, pneumothorax, significant hemorrhage) or minor (transient hypoxemia, hypotension or tachycardia, and excessive coughing) complications [10, 12, 14, 21, 28]. EBUS-TBNA is considered to be a safe diagnostic procedure in adults as well as in pediatric patients. Nevertheless, the procedure requires additional training and experience. Due to the low number of pediatric Tb cases pediatric pulmonologists may see, such experience may be difficult to obtain [29]. Our team included both pediatric and adult pneumologists.

Our study demonstrates the utility, safety and feasibility of EBUS-TBNA in a relatively large pediatric population. Its high diagnostic yield may facilitate more accurate diagnostic follow-up for patients with mediastinal lymphadenopathy and suspected Tb. It also may help physicians more clearly identify when tuberculostatic therapy is needed.