To the Editors:
We read with significant interest the report from Haas 1 in a recent issue of the European Respiratory Journal, which described two cases of infectious complications following endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA). This technique has been widely adopted for the evaluation of mediastinal and hilar lesions for several reasons, including its high-diagnostic accuracy, minimally invasive nature and excellent safety profile. As stated by Haas 1, no significant complications have previously been reported.
We recently experienced a similar complication in our interventional bronchoscopy centre (Royal Melbourne Hospital, Parkville, Australia). A female with small cell lung cancer and a moderate-sized retrosternal thyroid lesion underwent EBUS-TBNA sampling of the thyroid lesion for the purpose of staging. EBUS imaging demonstrated a cystic space and cytology performed on the TBNA specimen revealed normal thyroid colloid. She presented to her local medical officer with erythema over the sternal notch 48 h after the procedure. Despite initial therapy with flucloxacillin this progressed and on re-presentation to our centre she had developed a spontaneous purulent discharge from this site, with subsequent resolution over the following 6 days. Penicillin-sensitive Streptococcus pneumoniae was isolated from multiple culture specimens of purulent material.
Bacteraemia following conventional TBNA is extremely rare 2. The sole description of bacteraemia following TBNA reported Streptococcus viridans as the causative agent, suggesting an oral source 3. Patients described by Haas 1 and Epstein et al. 4 developed bacterial infections complicating TBNA with microbiological and clinical features, highly consistent with inoculation of oropharyngeal bacteria into the relatively avascular pericardial space 1, 4. Infection resulting from direct inoculation of relatively avascular necrotic tissue following endoscopic fine needle aspiration has also been reported 5. With respect to our patient, S. pneumoniae is a common nasopharyngeal commensal and we believe infection was similarly caused by bronchoscope contamination during insertion through the upper airway, with subsequent direct inoculation of bacteria by the TBNA needle into an avascular cystic space.
EBUS-TBNA appears to be the optimal procedure for sampling of mediastinal and hilar lesions due to its safety in comparison to formal surgical procedures. However, our experience, and that of Haas 1, indicates that a complication rate of 0% for a novel technique almost certainly means insufficient procedures have been performed to fully inform commentary on complications rates, or possibly that complications remain unreported. It is important for proceduralists performing EBUS-TBNA to be aware that, while rare, significant infection is a recognised complication. Contamination of the bronchoscope working channel during introduction into the airway is possible and direct inoculation of bacteria appears the likely cause of infective complications 1, 4, 5.
We commend Haas 1 for the timely report on the potential for infection complicating EBUS-TBNA, but feel the issue of antibiotic selection in such cases should be emphasised. Further studies are required to elucidate the organisms involved in infections following EBUS-TBNA. However, our experience, and that reported by Haas 1, suggests empiric antibiotic use in patients with clinical suspicion of infection must include agents with activity against indigenous oral and nasopharyngeal organisms. Prophylactic antibiotics for EBUS-TBNA do not appear to be indicated on the basis of risk of bacteraemia; however, we feel our experience, and previously reported experiences 1, 4, 5, indicate further evaluation is required to determine the role of prophylactic antibiotics in specific patient groups. Consideration of use of antibiotic prophylaxis may be appropriate in patients with relatively avascular lesions such as cysts or necrotic lymph nodes, or in immunocompromised patients at high risk of local infective complications.
Statement of interest
None declared.
- © ERS Journals Ltd