1. IntroductionAlthough the use of rigid bronchoscopy has declined since the introduction of flexible bronchoscopy, it is an indispensable procedure for therapeutic intervention in the field of interventional pulmonology because it has several advantages over flexible bronchoscopies, such as facilitating more complex procedures, better airway control, and greater capability for suction [1,2]. Rigid bronchoscopy is generally performed in the operating room (OR) under general anesthesia to facilitate the procedure and reduce patient discomfort [3]. In the OR, an anesthesiologist monitors the condition of the patient and adequately adjusts the levels of anesthesia and ventilation. However, there is no standard method for ventilation or level of anesthesia for rigid bronchoscopy [4].In a critically ill patient, it can be harmful to perform rigid bronchoscopy electively in the OR due to delay in the intervention [5]. Consequently, rigid bronchoscopy sometimes has to be completed at the intensive care unit (ICU) bedside. In these emergency situations, the intervention cannot be performed with an anesthesiologist or with devices to assess the degree of anesthesia. If emergent rigid bronchoscopy could be performed safely at the ICU bedside, it would allow improved management of severe patients who require urgent airway intervention.Previous studies have reported various anesthetic techniques for rigid bronchoscopy [4,6]. However, these studies have focused on effective induction, placing preference on a smooth intervention rather than its safety from the anesthesiologist’s perspective. Some studies have evaluated the safety of rigid bronchoscopy but did not consider the degree of urgency, location of bronchoscopic intervention, or presence of an anesthesiologist [7,8]. Therefore, we aimed to investigate the safety and outcomes of emergent rigid bronchoscopy for therapeutic intervention at the ICU bedside without an anesthesiologist using tertiary hospital-based retrospective data. 2. Materials and Methods 2.1. Study PopulationWe conducted a retrospective cohort study by reviewing medical records of patients who underwent rigid bronchoscopy while admitted to the ICU from January 2014 to December 2020 in Samsung Medical Center (a 1960-bed university-affiliated tertiary referral hospital in Seoul, Republic of Korea). During the study period, we had a hospital-wide medical emergency team (MET) for the rapid response system, which consisted of dedicated intensivist physicians, including critical care fellows and attending intensivists [9]. Activation of MET was based on several criteria (respiration, circulation, and neurology), and respiratory criteria include a respiratory rate ≥30/min, oxygen saturation 10]. When the MET decided to transfer the patient to the ICU, the interventional pulmonologists (H. Kim and B.-H. Jeong) carefully decided whether an emergency procedure at the ICU bedside was necessary or an elective procedure in the OR was possible according to the patient’s condition on a case-by-case basis. In general, we made decisions for an emergent procedure based on the following criteria: (a) patients with respiratory failure (PaO2/FiO2 ratio 2 > 50 mmHg, or arterial pH

According to the location of rigid bronchoscopic intervention, patients were classified into (a) the ICU group, cases emergently performed at the ICU bedside without anesthesiologists, or (b) the OR group, cases electively performed in the OR with anesthesiologists.

2.2. Anesthesia and Airway Intervention Techniques

We used different anesthetic techniques according to the location of intervention as follows. In the OR, the anesthesiologist performed general anesthesia using target-controlled infusion (Orchestra®; Fresenius Vial, Brezins, France) with propofol and remifentanil under bispectral index (Aspect Medical Systems, Norwood, MA, USA) monitoring. During the procedure, the anesthesiologist monitored vital signs, and a reversal agent for neuromuscular block agents (NMBA) was administered at the end of the procedure as soon as extubation was possible. At the ICU bedside, anesthesia was induced by the operator with a bolus dose of sedatives, opioids, and NMBA. The assistant monitored vital signs and administered additional drugs depending on the status of the patient. A reversal agent of NMBA was not routinely used in the ICU because the drug was not generally prepared.

Rigid bronchoscopy was performed following standard techniques [11]. After anesthesia, the patient was intubated using a rigid bronchoscope tube (Bryan Co., Woburn, MA, USA, or Karl-Storz, Tuttlingen, Germany). Patients were manually ventilated by an anesthesiologist in the OR and by an assistant at the ICU bedside. Various airway intervention methods were used, such as silicone stent insertion, mechanical debulking, bougienation with a balloon and rigid scope, and laser cauterization, based on the characteristics of the lesion and the reasons for intervention [12]. 2.3. Adverse EventsWe investigated procedure-related complications to assess the safety of rigid bronchoscopy. We divided complications into intra- and post-procedural complications. Intra-procedural complications consisted of hypertension, hypotension, tachycardia, and hypoxia. Hypertension was defined as episodes of systolic blood pressure >160 mmHg [13]. Hypotension was defined as episodes of systolic blood pressure 14,15]. Tachycardia was defined as episodes of heart rate >110 beats per min [16]. Hypoxia was defined as episodes of saturation of oxygen 17]. Post-procedural complications consisted of respiratory failure, atelectasis, pneumonia, newly developed arrhythmia, massive bleeding, pleural effusion, and pneumothorax within three days after the intervention [14]. Respiratory failure was defined as unplanned mechanical ventilation. Pneumonia was defined as the presence of new lung infiltrates on the chest radiograph with fever (body temperature ≥38 °C). Newly developed arrhythmias were defined as a new presentation of atrial fibrillation, bundle branch block, sustained ventricular tachycardia, supraventricular tachycardia, or ventricular fibrillation. Atelectasis, pleural effusion, and pneumothorax were defined using formal readings reported by a chest radiologist. Major bleeding was defined as procedure-related bleeding requiring transfusion [18].

Intra- and post-procedural complications were classified by severity into (a) mild, no additional management; (b) moderate, need for unexpected non-invasive management such as medical therapy; and (c) severe, need for unexpected invasive management.

2.4. Data CollectionWe retrospectively collected the following data: demographics, patient status before the intervention, disease characteristics, bronchoscopic findings, procedure details, anesthetics used, and intra- and post-procedural findings. Performance status was assessed using the American Society of Anesthesiologists (ASA) physical status classification [19]. In short, ASA class 3 indicates severe systemic disease with substantive functional limitations, class 4 indicates severe systemic disease that is a constant threat to life, and class 5 is moribund, for which survival is not expected without intervention. The severity of stenosis was defined using the Myer–Cotton stenosis grading system: Grade II, 51% to 70% luminal stenosis; Grade III, 71% to 99% luminal stenosis; and Grade IV, no lumen [20]. Patients who underwent intervention for fistula (n = 7) or stent removal without airway stenosis (n = 1) were classified as Grade II. 2.5. Statistical Analysis

Data are expressed as numbers (%) for categorical variables and median (interquartile range, IQR) for continuous variables. Differences between the two groups were analyzed using Pearson’s chi-square test or Fisher’s exact test for categorical variables and the Mann–Whitney test for continuous variables.

Multivariable logistic regression analysis was used to adjust for potential confounding factors in intra- and post-procedural complications according to the location of bronchoscopic intervention, with the OR group as the reference category. Four models were constructed: Model 1 was adjusted for selected variables with p2/FiO2 ratio was input after being converted into the following categorical variables: (a) 21] with patients who had no plans for extubation and would remain on a home ventilator or were transferred to another hospital immediately after the procedure classified as ≥61 min.

All tests were two-sided, and a p < 0.05 was considered significant. All statistical analyses were performed using SPSS (IBM SPSS Statistics ver. 27, Chicago, IL, USA).

4. Discussion

Although the ICU group had a lower PaO2/FiO2 ratio, a higher proportion of ECMO support, more grade IV stenosis, and a longer intervention duration than the OR group, the overall intra-procedural complication rate was similar between the OR and ICU groups. There was only a difference in the incidence of each intra-procedural complication between the two groups. Of note, there was no severe intra-procedural complication, regardless of the location of the bronchoscopic intervention. Post-procedural complications, including those that were severe, occurred more frequently in the ICU group than in the OR group. However, the odds of severe post-procedural complications for the ICU group were not statistically significant after adjusting for all possible confounding factors.

The rigid bronchoscope was a unique diagnostic and therapeutic tool that allowed direct observation of the inside of the airway until the flexible bronchoscope was introduced in the late 1960s [2]. Flexible bronchoscopy is less invasive, does not require general anesthesia, and is easier to observe the peripheral airway than rigid bronchoscopy, so rigid bronchoscope was virtually abandoned [22]. However, Dr Dumon pioneered modern rigid bronchoscopy with the development of silicone stent insertion to treat central airway obstruction in the late 1980s [22,23]. Because rigid bronchoscopy has several advantages over flexible bronchoscopies, such as facilitating mechanical debulking and silicone stent insertion, better airway control, and greater capability for suction, rigid bronchoscopy is irreplaceable in the field of interventional pulmonology [2]. In particular, silicone stents are generally believed to be less injurious to the airway and easier to modify and remove the stent itself than metallic stents [24]. Therefore, rigid bronchoscopy is still used despite the disadvantages of having difficulties in use and requiring general anesthesia. However, there is a limitation to applying the rigid bronchoscopy as an emergency procedure because general anesthesia is required with an anesthesiologist in the operating room, so we analyzed the safety of emergent rigid bronchoscopy at the ICU bedside without an anesthesiologist.Several studies have reported complications associated with rigid bronchoscopy [8]. However, there is no standardized tool to assess the complications of rigid bronchoscopy, and the rate of complications varies [25,26,27,28]. The subjects of this study were in acute states, such as being ASA class 5 in 52.1% and receiving invasive respiratory supports in 50.6% of total 336 patients, while being in chronic conditions, such as having a median 21.1 kg/m2 of body mass index and many comorbidities. In particular, a higher ASA score before the intervention is a risk factor for procedure-related complications [28]. A recent study reported that the rates of intra-procedural hypoxemia and hypotension were 67% and 77% in patients with poor performance status, respectively [7]. In our report, the incidences of intra-procedural hypoxemia and hypotension in patients who underwent rigid bronchoscopy at the ICU bedside were lower at 18% and 11%, respectively. Instead, the incidence of hypertension and tachycardia was high in our study, and the overall incidence of intra-procedural complications was 86%. Although the incidences of these intra-procedural complications were high in our study, they were usually transient and rarely led to serious outcomes. Considering that ICU patients are not easy to transport and may require prompt intervention in an emergency situation, rigid bronchoscopy for therapeutic intervention at the ICU bedside can be a reasonable option for critically ill patients.Although the overall incidence of intra-procedural complications was similar between the two groups, the characteristics were markedly different. These differences may be due to the differences in anesthetic agents and reversal agents used in the two groups. The occurrence of hypertension, hypotension, and tachycardia was correlated with the known side effects of the drugs mainly used in the two groups [29]. Significantly, more caution should be observed when using ketamine in patients with underlying heart failure, significant arrhythmias, and ischemic cardiomyopathy. Unlike in those complications, the occurrence of intra-procedural hypoxia may be associated with the oxygenation status of the patient before intervention. Our findings showed that the PaO2/FiO2 ratio was lower in the ICU group than in the OR group, and this may have resulted in more frequent intra-procedural hypoxia in the ICU group. Importantly, although the ICU group had worse conditions than the OR group, all intra-procedural complications were transient and did not require an escalation in management.The risk of post-procedural complications was higher in the ICU group than in the OR group, possibly due to the use of reversal agents. Delayed extubation after surgery increases the risk of post-procedural complications in various types of surgery [21]. Reversal agents were rarely used in the ICU group, which resulted in delayed extubation. Therefore, it could be postulated that the use of reversal agents may affect the development of post-procedural complications. However, since our study enrolled only a few patients, further studies are needed to prove our hypothesis.Severe post-procedural complications were associated with poor clinical outcomes, such as longer length of ICU stay and the need for additional interventions (Table S1), whereas mild to moderate complications did not show a statistical correlation with these clinical outcomes (data not shown). So, we investigated the factors related to the occurrence of severe complications that require unexpected invasive management. Eventually, we noted that stent change or repositioning and longer intervention duration may be associated with severe complications (Table S1). Stent change or repositioning is usually required when complications occur, such as stent migration or stenosis at both tips of the stent. The procedure is generally performed by widening the stenosis site, removing the granulation tissue and the existing stent, and inserting a new stent. Therefore, a stent change generally takes longer than other procedures. In addition, prolonged intervention duration could be associated with physiological changes that predispose one to respiratory failure [21]. Therefore, careful post-procedural monitoring and management are necessary when the procedure is complicated and takes a long time.

To the best of our knowledge, this is the first study investigating the safety of rigid bronchoscopy at the ICU bedside among critically ill patients. Our findings could serve as a reference for the management of critically ill patients requiring immediate rigid bronchoscopic intervention. However, some limitations should be considered. First, our data were from a single center, where two skilled interventional pulmonologists performed approximately 2700 rigid bronchoscopy cases during the study period. This can limit generalization. Second, differences between the two groups could not be fully adjusted due to the small number of cases. Third, a causal relationship between intervention and the occurrence of complications cannot be asserted due to the retrospective design. However, we meticulously defined the complications to overcome this limitation. Finally, the median time elapsed from ICU admission to the intervention was 18 h in the ICU group, which is hard to see as an emergency situation. Considering the limitation that it was difficult to collect the time points when the operators were contacted retrospectively, we had no choice but to collect data on ICU admission time. However, since 49.1% of the ICU group received the procedure on the weekends or nighttime, and only 13.5% received the procedure during the weekdays in the elective time, we can assume that we performed an emergency procedure with a quick decision in the ICU group.