Endobronchial ultrasound (EBUS) is a minimally invasive technique that has a high yield in the diagnosis and staging of mediastinal disease and lung cancer.1,2 The ultrasound bronchoscope combines a flexible bronchoscope with a convex ultrasound transducer on the tip for simultaneous endoscopic vision and capture of ultrasound images of structures in contact with the bronchial wall. This modification results in an instrument with a thicker, rigid tip that must be inserted orally, an approach that is known to be more distressing for patients.3 Moreover, the endoscopic image is captured at a slightly different angle from a conventional bronchoscope because the optical system is not located on the tip of the instrument.

Bronchoscopy causes anxiety and intolerance in the patient, making it difficult for the bronchoscopist to perform their task. Performing the procedure in distressed patients reduces yield and can provoke hemodynamic and even respiratory changes. Sedation improves patient tolerance, allowing the bronchoscopist to work more easily,4 but the level of sedation applied can range from conscious sedation to general anesthesia, depending on the agents and the dose used. Conscious sedation is generally preferred, particularly for procedures requiring rapid recovery.3 However, in a recent study5 comparing the effect of light sedation and deep sedation on the number of lymph node territories aspirated and the histological yield of the samples, benefits were found for deep sedation. When sedation was deeper, patient tolerance was better, movements and cough during aspiration were reduced, and yield improved. Several groups recommend conscious sedation monitored by a bronchoscopist experienced in the use of drugs such as midazolam, propofol, and narcotics used alone or in combination, at low doses.6–8 Midazolam is an anxiolytic benzodiazepine that causes anterograde amnesia; propofol is a hypnotic agent with amnesic and antiemetic effects, and rapid onset and elimination, depending whether it is administered in intermittent boluses or by infusion. Narcotics – fentanyl and remifentanil – can be combined with either midozalam or propofol. They have useful analgesic and antitussigenic properties, and act in synergy, so the final dose of midazolam or propofol can be reduced.

Few groups have studied the level of sedation in endobronchial ultrasound, and the best drug or combination of drugs for this procedure remains undefined. Studies comparing patient satisfaction with different levels of sedation and the presence or absence of an anesthesiologist during the procedure are scarce.9 Studies that evaluate complications associated with the use of drug combinations during EBUS are also rare.

The objective of this study was to evaluate different sedation models and associated complications during EBUS, and to establish which is the safest drug or combination of drugs that gives most patient satisfaction.

During the study period, 342 consecutive patients who had undergone endobronchial ultrasound completed the questionnaire. Thirty-five questionnaires were excluded because they were incomplete (Fig. 1). The remaining 307 questionnaires were completed by 226 (73.6%) men with a mean age of 63.4±12.4 years (range 29–87 years), and 81 women, mean age 61.47±11.8 years (range 37–85 years). Mean duration of the procedure was 45.27±16.9min (range 15–90min). Twenty-four patients received Model A sedation, 37 received Model B, 107 received Model C, 62 received model D, and 77 received model E. There were no statistically significant differences in patient age among the sedation models. The duration of the procedure differed significantly among the different models (P<.001), and the post hoc analysis revealed 2 homogeneous subgroups, namely A, C and E, versus E, B, and D (P=.82) (Table 1). Reasons for performing EBUS were: study of mediastinal lymphadenopathies (156 cases); staging (71); simultaneous lymphadenopathy study and staging (35); study of mediastinal (22) and hilar (9) mass; and other indications (14). Sedation Models A, C and E were performed without the presence of an anesthetist. The level of sedation of all patients was monitored by at least 1 of the available systems, most often the VAS (62.1%). Model C was the only regimen monitored with BIS.

Fig. 1.

Study protocol flow chart. Types of sedation: Model A, midazolam bolus; Model B, propofol infusion; Model C, midazolam bolus and propofol infusion; Model D, propofol infusion and remifentanil infusion; and Model E, midazolam bolus and fentanyl bolus. Patients screened for the study.

(0.15MB).

Patients’ perception of the various aspects of the procedure, and scores for each Likert-type question are shown by sedation model in Table 2. (Responses to the other 2 questions – what was the worst moment of the procedure and if they would repeat the EBUS – are shown in Table 3.) Patients’ response scores – overall and by sedation model – are given in Table 2. Overall scores for each question show that patient-perceived sensations of recall, pain, cough, dyspnea, and prolonged procedure (0.65±1.11, 0.3±0.73, 0.46±0.9, 0.29±0.73, and 0.59±0.96, respectively) are scored lower than the sensations of fear and nervousness before the procedure (1.26±1.37 and 1.5±1.41, respectively). The higher scores given to indifference toward repetition (1.49±1.3) and pleasant sensation during the procedure (1.23±1.17), together with low scores for sensation of distress (0.49±0.85) and discomfort (0.62±1.1) during the procedure indicate that the EBUS was well tolerated with the various sedation models. When we compare the responses to each question by sedation model, significant differences are found for each question among the various models. Post hoc tests for perception of cough, pain, dyspnea, and recall revealed 2 significant homogeneous groups formed by sedation Models A and E, and Models B, C, and D. Two groups were also obtained for other parameters, consisting of Model E versus the other sedation models.

When asked about the worst moment during the procedure, 46.6% of patients did not report any bad moment during the procedure, and the worst moments for the remaining patients were when the local anesthesia was applied to the upper airways (17.5%), and when the ultrasound bronchoscope was introduced into the trachea (15.4%) (Table 3). When asked if they would repeat the examination, 89.6% answered that they would be willing, if necessary.

Complications occurring during the EBUS procedure are listed in Table 4. Models E and C produced fewer complications (87.01% and 68.2%, respectively, of complication-free procedures). A higher percentage of patients had complications with Models B and D (86.5% and 69.4%, respectively). All complications were resolved with simple therapeutic measures taken in the endoscopy room.

Our study compares different sedation models in patients undergoing EBUS with an ideally moderate level of sedation. All sedation methods were well accepted by the patients and the few complications that occurred were successfully managed in the endoscopy room.

Although the need for satisfactory sedation in bronchoscopy has been under discussion for some time,11 it is still a controversial area,12,13 partly because half of complications during bronchoscopy are a consequence of sedation.14 It is accepted practice to offer sedation, mainly in the form of a benzodiazepine, such as midalozam, to patients who request it,15 but some groups systematically use sedation,16 unless it is contraindicated. EBUS, however, with its oral approach, is a more complicated and less tolerated procedure that generates greater patient discomfort. Appropriate sedation is essential for the bronchoscopist to be able to complete the examination and for the patient to tolerate it.

International guidelines, such as BTS15 or ACCP,17 have established recommendations for bronchoscopy and ultrasound bronchoscopy procedures. However, they do not provide any clear criteria for the degree of sedation necessary or which drugs to use. A combination of 2 sedative agents has been proposed as a safe option for patients undergoing diagnostic EBUS.9 However, very few studies have evaluated the best route of administration or the safest method of monitoring sedation. In most pulmonology departments, the most commonly used combination for bronchoscopy is probably fentanyl and midazolam, which is why it is also the most commonly used in EBUS, and why we included it in our analysis. Other groups have introduced propofol, either in boluses or continuous infusion. Sarkiss et al.18 used intravenous propofol in EBUS, and did not encounter any significant complications. When we compared cough, pain and dyspnea among our patients who received the different regimens, we found that the model that included propofol was more successful in controlling these symptoms than those using midazolam. Thus, while the degree of sedation of our patients was similar, irrespective of the method used to determine it, propofol provided better symptom control.

When patients sedated and not sedated during bronchoscopy are compared, patients who do not receive sedation are more reluctant to repeat the test.4 In our patients, another of the benefits of using sedation for performing EBUS was that 89.6% of the patients were willing to repeat the procedure if necessary, irrespective of the type of sedation administered. These findings coincide with those of Steinfort and Irving.19 Sedative-induced lack of recall following the procedure is one of the variables associated with the patient's willingness to repeat it. Moreover, 46.6% of patients did not report any disagreeable moment, and the most poorly tolerated moments were when the upper respiratory tract was anesthetized and when the bronchoscope was inserted. Again, this finding reflects those of other studies.4 Sedation begins before local anesthesia is applied and the ultrasound bronchoscope is introduced, but the bronchoscopist may have to initiate the procedure before the patient is fully sedated, as some degree of patient collaboration is required for the local anesthesia. The level of sedation measured by BIS before starting the EBUS also deepens when the ultrasound bronchoscope is inserted in the airway.

The availability of several sedation models is due, in part, to the potential adverse effects of the drugs used. Complications occurred with all the sedation models analyzed in this study, all of which were resolved in the endoscopy room. The most common complication of sedation was hypotension, most frequently in patients receiving propofol and remifentanil infusion. Hypotension is one of the known side effects of propofol, so physicians using this agent must be skilled in managing this event. Cough is another common complication. Lower levels of local anesthesia are often used in sedated patients during endoscopy, and this sometimes leads to cough, both during and after completion of the procedure. Other complications included desaturation or bradypnea, possible complications in moderately sedated patients. We found that these complications were more common in the models which used propofol infusion and propofol with remifentanil. Remifentanil was not used by the pulmonologists in this study, and propofol infusion was used by only a few. This would explain why we found fewer complications in our study when the sedation was administered by the pulmonologist. In contrast, anesthetists are more accustomed to using different drugs and can find the best combination for the patient, according to the type of procedure and its duration. Remifentanil and propofol infusion have more potential adverse events, as we experienced.

The ability of the pulmonologist for managing these complications may be questioned, but in our opinion, this is not an issue: a pulmonologist's training includes mechanical ventilation, treatment of desaturation, and even the use of sedatives for ventilation in critical or intensive care units. Accordingly, in bronchoscopy units or interventional pulmonology, sedation is often administered by the pulmonologist. The models which presented fewest complications were E, C, and A, respectively, while the shortest durations were for Models A and E, respectively. The duration of the procedure under Model C was similar to that of Models B and D, but for B and D, the percentage of patients with complications was higher. Although the duration of the procedure differed significantly, we believe that the complications were due more to the choice of drug than to the length of the procedure, since the dose was adjusted according to the level of sedation at any given moment, and the duration of the procedure is unrelated to the percentage of complications.

Our study is limited insofar as we did not analyze the diagnostic yield of the EBUS according to the sedation model applied. However, EBUS yield for mediastinal staging and for the study of mediastinal lymphadenopathies and masses is well established, and is generally high,20 and we do not believe that this would differ among the different sedation models included in our study. Another limitation of the study is that the satisfaction of the bronchoscopists was not explored. Evaluating the presence of an anesthetist during the procedure was not a study objective, but it is unlikely that this factor has biased the results, since the responses of the patients regarding their perceptions were not influenced by the presence of an anesthetist.

The main conclusion of our study is that the different sedation models under study were all well tolerated by patients, to the extent that most patients would agree to repeat the procedure if necessary. In sedation models using bolus administration, patients were more poorly controlled, according to recall responses. Complications were rare and managed in the endoscopy room itself, so we are satisfied that some types of sedation can be administered by a bronchoscopist who is familiar with these drugs and capable of minimizing side effects.

None of the authors has any direct or indirect conflict of interest with the contents of this manuscript.