Complications associated with endobronchial ultrasound-guided transbronchial needle aspiration: a nationwide survey by the Japan Society for Respiratory Endoscopy

Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) performed using a convex probe ultrasound bronchoscope is a minimally invasive diagnostic method for peritracheal and peribronchial lesions [1]. EBUS-TBNA enables the accurate collection of samples from the lesion since the puncture site can be confirmed under real-time ultrasound guidance, allowing cytological and histological diagnosis. In the lymph node staging of lung cancer, EBUS-TBNA has a cumulative sensitivity of 88%-93% and a cumulative specificity of 100% [2, 3]. For the diagnosis of mediastinal lymph node metastasis of lung cancer, mediastinoscopy is conventionally considered as the gold standard [4]. However, compared with surgical diagnosis using a mediastinoscope or thoracoscope, EBUS-TBNA and transesophageal endoscopic ultrasound with fine needle aspiration (EUS-FNA) followed by surgical diagnosis have been reported to increase the diagnostic rate [5]. EBUS-TBNA alone has a diagnostic rate similar to that of mediastinoscopy [6, 7]. In addition, the introduction of EBUS-TBNA has resulted in cost benefit in terms of minimally invasive lung cancer staging [8, 9]. Hence, EBUS-TBNA has effectively been replacing mediastinoscopy for the staging of lung cancer and in establishing a definitive diagnosis of mediastinal and hilar lymphadenopathy owing to its high diagnostic rate, minimal invasiveness, and cost benefit. EBUS-TBNA has also an extremely low complication rate and is therefore considered to be safe and minimally invasive [2, 10].

However, there have been occasional reports on complications with the recent widespread use of EBUS-TBNA [11‐14]. Moreover, previous reviews on EBUS-TBNA have been limited to studies by skilled operators, and thus the results may not always be applicable to recent clinical practice. Previous reviews particularly on the cost benefit of EBUS-TBNA have not included the repair cost for damaged EBUS-TBNA instruments. As EBUS-TBNA will increasingly be used for the staging and diagnosis of lung cancer, accompanying severe complications may also increase. Damage to EBUS-TBNA instruments not only poses a problem with regard to the cost benefit, but also threatens the safety and effectiveness of the procedure. To assess the safety of EBUS-TBNA for the staging and diagnosis of lung cancer in Japan, a nationwide survey on its current usage status and complications associated with its use was conducted by JSRE.

This study was a nationwide, retrospective questionnaire survey approved by JSRE. In July 2012, a JSRE-approved questionnaire was mailed to 520 JSRE-accredited facilities and affiliated facilities throughout Japan. The questionnaire (an additional data file shows this in more detail [see Additional file 1]) included specific questions regarding the number of cases examined by EBUS-TBNA, the characteristics of the cases, and the complications of EBUS-TBNA performed using a convex probe ultrasound bronchoscope during the preceding 18 months (January 2011 through June 2012). Details regarding the complications were obtained by asking questions about patient background, procedural factors, and outcomes using an exclusive case report form. A mark sheet selection system was used, except for some items that were to be described using free text. As a rule, a JSRE-accredited senior fellow representing each institution provided the responses to the questionnaire. The number of cases, number of cases with complications, and details of cases showing complications were reported by the same senior fellow, based on the medical records in each institution. The investigated complications included hemorrhage that required treatments other than aspiration, compression and cold physiological saline injection (massive hemorrhage was defined as a blood loss of ≥ 300 mL or that requiring blood transfusion), pneumonia, mediastinitis, pericarditis, other infectious complications, pneumothorax requiring bed rest or thoracic drainage, lidocaine intoxication requiring special intervention because of convulsions/loss of consciousness, bronchial asthma, respiratory failure (excluding patients treated by oxygen administration alone), and complications related to the circulatory system requiring special treatment, among others. Complications due to procedures performed simultaneously with EBUS-TBNA, such as forceps biopsy or bronchoalveolar lavage, were excluded.

Results were automatically collected using a computer. For multiple-choice questions permitting only 1 response, no response to the question was regarded as an invalid response. All proportions were expressed as a percentage of the number of valid responses. Results were analyzed using a standard statistical method with the SAS system for Windows release 9.3 (SAS Institute Inc., Cary, NC, USA).

Responses were obtained from 455 facilities (response rate, 87.5%).

To the best of our knowledge, this is the first nationwide survey describing in detail the complications associated with the use of EBUS-TBNA. As this survey was administered by a scientific society, and the obtained response rate to the questionnaire was considerably high (87.5%), the results may effectively reflect the present status of the use of and complications associated with EBUS-TBNA in Japan. A previous questionnaire survey performed in Japan in 2010 showed that 143 facilities nationwide owned and used EBUS-TBNA devices [15]. This number increased by 67 during the subsequent 1.5-year period. The number of cases examined using EBUS-TBNA per year, as calculated from the number of questionnaires recovered, was 4,108 cases in a previous survey and 5,596 in the present survey, indicating an approximately 1.4-fold increase in the number of EBUS-TBNA performed per year.

Previous meta-analyses of conventional TBNA showed a major complication rate of 0.3% [16]. Massive hemorrhage, pneumomediastinum, pneumothorax requiring drainage, cardiac tamponade, and hemomediastinum have all been reported as TBNA complications [16, 17]. A meta-analysis of EBUS-TBNA showed a complication rate of 2 in 1,299 cases (0.15%), with only 1 case showing pneumothorax requiring drainage as a major complication [2]. However, the studies evaluated by this meta-analysis had been performed in facilities with skilled operators. With the recent widespread use of EBUS-TBNA, complications have occasionally been reported [11‐14]. A survey in 2010 showed a complication rate of 0.46% for EBUS-TBNA [18]. On the other hand, a prospective study by Eapen et al. revealed a complication rate of 1.44% [19]. The incidence of complications differs and depends not only among medical examination/treatment environments, but also on the definitions of the complications. The definitions of the different complications assessed in this study were the same as those in the 2010 survey, except for hemorrhage. In the present survey, a broader definition of hemorrhage was used compared with the previous survey, and hemorrhage that required treatment was also included in the list of complications. This possibly accounted for the higher complication rate in the present study. Furthermore, the incidence of the different types of complications in the present survey differed from that reported by Eapen et al. [19]. In the present survey, the incidence of pneumothorax was low but that of infectious complications was high. This may have been due to not only differences in the medical examination/treatment environments between Japan and the U.S. as well as differences in the survey methods, but also the evaluation of only complications developing within 24 hours and the inclusion of complications due to other procedures such as transbronchial biopsy that were simultaneously performed in the survey by Eapen et al. The complication rate in this survey was lower than that using mediastinoscopy (2.5%) [20] or other bronchoscopic diagnostic methods [18], demonstrating the safety of EBUS-TBNA.

However, in the present study, death or life-threatening consequences were observed in 5 cases. In addition, the incidence of infectious complications was high. Regarding infectious complications, mediastinitis [12, 13], pericarditis [11], and tumor infection [11] have been reported, each of which is a severe condition requiring broad-spectrum antibiotics and surgical procedures such as drainage. Two cases in the present study were considered to have life-threatening conditions. As for the possible causes of infectious complications, cross-contamination, leading to mediastinal inoculation by oral and nasopharyngeal commensals [11, 13], and nosocomial infection due to inappropriate bronchoscope disinfection or mishandling of devices [12] have been reported. Previous studies have suggested that these infectious complications are not caused by hematogenous seeding due to bacteremia, but by spread from the puncture site [13]. When the puncture site is necrotic or cystic, blood flow is slight, resistance is decreased, and bacterial attachment clearance is decreased [11]. To prevent this condition, the location of the needle tip should be confirmed by ultrasonography at the time of puncture, and the lesion should be accurately punctured [11]. As in EUS-FNA, the puncture of necrotic [21] or cystic areas [22] should be avoided. Unlike conventional TBNA, lesions are punctured with a puncture needle containing a stylet in EBUS-TBNA. Since infection may be caused by operating the stylet, it should be handled aseptically when multiple punctures are applied. Disinfection of devices according to the guidelines [23] and careful observation of the postprocedure course for signs of infection are important. However, prophylactic antibiotic administration is not recommended in routine diagnostic bronchoscopy, except for certain diseases [24]. In the present study, there were no clearly immunocompromised patients with infectious complications, and prophylactic antibiotics were not administered in 11 of the 14 cases showing infectious complications. Further studies are necessary to evaluate the need for prophylactic antibiotic administration and to identify the cases requiring such prophylaxis. In particular, Eapen et al. reported patients who died of hemorrhage [19]. In the present study, only 1 case showed massive hemorrhage caused by perforation of a minor blood vessel in the lymph node. Botana-Rial et al. reported a case of erroneous puncture of the pulmonary artery [14] without symptoms and detected by echography or contrast CT. Thus, the incidence of erroneous puncture of a major vessel may be higher than that of massive hemorrhage as a complication. Hence, it is important to know the anatomy around the puncture area and to accurately puncture only the lesion. Bronchial asthma, cerebral infarction, and lidocaine intoxication were considered to be complications of the bronchoscopic procedure. In patients at risk, procedures following the guidelines are necessary, including prophylactic bronchodilator administration, continuation of antiplatelet drugs/anticoagulants or drug replacement depending on the underlying diseases and drug composition, adequate assessment of the indications for examination, and observance of the maximum lidocaine dose that can be safely administered [24].

In the present study, the complication rate was slightly higher in the procedures performed by JSRE-accredited non-fellows than in those performed by JSRE-accredited fellows. Regarding the diagnostic rate, a prospective study previously showed that high-volume hospitals had high diagnostic yields, and that biopsy was performed on more nodes including smaller nodes [25]. Although Eapen et al. stated that safety is not associated with experience [19], their data were, however, obtained from tertiary care hospitals. Therefore, we cannot exclude the possibility that the complication and diagnosis rates differ among facilities and according to the operator’s experience. In EBUS-TBNA, the learning speed varies even among skilled bronchoscopists [26]. Moreover, the number of cases that must be performed to be able to carry out EBUS-TBNA safely and to achieve an adequate diagnostic rate is difficult to determine. The guidelines of the American College of Chest Physicians [27], European Respiratory Society/American Thoracic Society [28], and Thoracic Society of Australia and New Zealand [29] indicate that 50 (radial probe), 40 (convex and radial probe), and 20 cases, respectively, are necessary for this purpose, and that EBUS-TBNA must be performed in 25, 20, and 20 cases per year, respectively, to maintain EBUS-TBNA competency. The present survey did not include the number of patients undergoing EBUS-TBNA by operators with no experience of complications, and thus the complication rate according to the number of cases experienced by the operators could not be determined. Accordingly, the number of cases experienced which are necessary to acquire the skill to safely perform EBUS-TBNA is unclear. Most of the operators in the present study had experienced less than 50 EBUS-TBNA cases at the time of the occurrence of complications, and about 50% of the operators had never attended any hands-on training. At present, concrete instructions and educational methods are left to the judgment of each institution, and they vary among facilities. Experienced bronchoscopists, such as JSRE-accredited fellows in each institution, should receive hands-on training in the course authorized by JSRE in order to become supervisors of EBUS-TBNA. A common learning system [30] and an evaluation method across facilities using objective parameters of skills [31] should be immediately established and applied to enable EBUS-TBNA operators to receive optimal training under the instruction of qualified supervisors. In addition, useful simulation trainings [32] and their active incorporation into the learning system should also be considered.

With regard to device breakage, this has been infrequently reported [33]. In the present study, the breakage rate for EBUS-TBNA bronchoscopes and devices was 1.54%, which was significantly higher (p < 0.001, Fisher exact test) than that in all bronchoscopy procedures (0.77%) in the 2010 survey [18]. Breakage most commonly occurred in the working channel mostly due to the needle perforation of the working channel. Proper placement of the puncture needle in the sheath and its blockage should be confirmed before insertion and removal of the puncture needle. Tremblay et al. reported that the repair cost of EBUS-TBNA instruments for each patient is higher than that of the bronchoscope used for conventional flexible bronchoscopy. In the present survey, EBUS-TBNA was not performed in 245 facilities, possibly because of the lack of EBUS-TBNA instruments in most of these facilities [15]. Unlike conventional TBNA, EBUS-TBNA is complex and requires hands-on training. Moreover, EBUS-TBNA instruments are expensive, and their maintenance and repair costs are high because of their high damage rate, which may be obstacles to the introduction of EBUS-TBNA.

The present study has several limitations. Recall bias was highly likely because the study was a retrospective questionnaire survey. Some data may have been provided based on the memory of the persons who completed the questionnaire and thus some values may be inaccurate. Moreover, the surveyed subjects may have hesitated to report complications and deaths. Although the response rate to the questionnaire was high, the facilities with a high complication rate may not have participated. In Japan, chest radiography is usually performed following bronchoscopy to confirm the presence of pneumothorax [15]. However, the follow-up method to determine the occurrence of complications is entrusted to each facility and is not standardized. Accordingly, it is possible that asymptomatic complications and severe cases treated at other facilities were overlooked. Moreover, the present survey involved facilities certified by JSRE, and expert operators were included. Therefore, the complication rate may have been underestimated compared with the actual situation. Furthermore, when the complication and damage rates were combined, some problems occurred in more than 3% of the patients who received EBUS-TBNA.

Although the complication rate for EBUS-TBNA was found to be low, severe complications and a high incidence of device breakage were observed during the study period. With the rapid and widespread use of EBUS-TBNA as a diagnostic method, an educational system providing clear instructions on how to properly perform the procedure safely needs to be immediately established.