Factors Associated With Successful Diagnosis of Infections With Endobronchial Ultrasound-Guided Transbronchial Needle Aspiration Performed for Mediastinal or Hilar Lymphadenopathy

Abstract

SESSION TITLE: Rapid Session: Interventional Procedures SESSION TYPE: Original Investigation Slide PRESENTED ON: Sunday, October 29, 2017 at 03:15 PM - 04:15 PM PURPOSE: Although endobronchial ultrasound (EBUS) with transbronchial needle aspiration (TBNA) of lymph nodes is frequently performed to diagnose cancer or sarcoidosis, the yield of this procedure for diagnosing infectious causes of mediastinal lymphadenopathy is unknown. Reported use of EBUS-TBNA in diagnosing infections other than tuberculosis is limited to case reports and case series, and diagnostic sensitivity of EBUS-TBNA for diagnosing infections has not been evaluated in large-scale studies. It is unclear if the low incidence of infectious diagnoses by this approach is the result of TBNA samples not being submitted for microbiologic studies, a low incidence of infection in the population being examined or a true limitation of the procedure. We aimed to identify pre-procedure factors that were associated with an ultimate diagnosis of infection. METHODS: All patients undergoing EBUS-TBNA performed by pulmonary and critical care medicine physicians for hilar and/or mediastinal lymphadenopathy at Los Angeles County Medical Center and Keck Medical Center of the University of Southern California from 2010-2016 were included. We report counts with percentages and means with standard deviations; multivariate logistic regression results are reported as odds ratio (OR) with 95% confidence interval (CI). All analyses were performed with SAS 9.4. RESULTS: 166 patients who underwent EBUS TBNA were analyzed. Patients tended to be male (55%) with an average age of 60. Half the patients had a history of cancer, while 2.4% of patients were seropositive for human immunodeficiency virus. Final diagnoses were cancer, sarcoidosis, and infection (mainly mycobacterial and coccidioidal) in 49%, 14%, and 10%, respectively. In multivariate analyses, only the clinician’s pre-procedural indication of “rule out infection” was associated with an infectious diagnosis (OR 5.51, 95% CI 1.40 - 21.74). Patient-specific variables such as age, past medical history, and lymph node or lung parenchymal radiographic features were not associated with final infectious diagnoses. The EBUS procedure was non-diagnostic in 3 cases that were subsequently diagnosed with infections by other means. CONCLUSIONS: In a retrospective review in an endemic region, Shah et al found that only 5 of 119 diagnostic EBUS-TBNA cases yielded a diagnosis of coccidioidomycosis, while Harris et al reported only 3 infections in 82 EBUS-TBNA sampled patients. The diagnosis of infections with EBUS-TBNA is uncommon, and often unexpected. Our small study shows that this diagnosis cannot be predicted from any patient demographic, lymph node or parenchymal radiographic patterns. However, a pre-procedural suspicion for infection significantly increased the odds of an infection being ultimately diagnosed with EBUS-TBNA. CLINICAL IMPLICATIONS: In patients undergoing EBUS-TBNA for mediastinal or hilar lymphadenopathy, subjective clinical suspicion for an infectious etiology was the only factor that correlated with an eventual infectious diagnosis. We suggest that clinicians consider sending EBUS-TNBA samples in such patients also for microbiological studies. Whether this practice would lead to more infectious diagnoses with EBUS-TBNA remains to be determined. DISCLOSURE: The following authors have nothing to disclose: Udit Chaddha, Andrew Morado, Ching-Fei Chang, Alex Balekian, Ramyar Mahdavi No Product/Research Disclosure Information