Authors: Akiva Nachshon, Avishai Shapiro, Smadar Goldfarb, Natalia Kuzmina, Marc Romain, Asaf Schwartz, Avraham Abutbul, Ido Vilchik, Michael Beil, Peter V. van Heerden
Categories: Research, Percutaneous tracheostomy, Bronchoscopy, Randomized controlled trial, Intensive care, Complication
Source: Critical Care
Authors: Akiva Nachshon, Avishai Shapiro, Smadar Goldfarb, Natalia Kuzmina, Marc Romain, Asaf Schwartz, Avraham Abutbul, Ido Vilchik, Michael Beil, Peter V. van Heerden
Bronchoscopic guidance is commonly used during percutaneous dilational tracheostomy (PCT) despite limited and conflicting evidence; this study aimed to evaluate whether bronchoscopy improves the safety of PCT in mechanically ventilated patients.
This prospective, randomized controlled study compared two approaches of percutaneous tracheostomy; with and without bronchoscopy assistance. The primary outcome measured was a composite of procedure-related complications at 24 hours, classified as minor (e.g., bleeding, desaturation) or major (e.g., airway loss, tracheal rupture, severe pneumothorax, major bleeding requiring transfusion and death). Secondary outcomes included procedure duration, 30-day mortality, and hospital length of stay.
PCT performed with or without bronchoscopic guidance.
307 adults requiring PCT between 2016 and 2024 were randomized to PCT with bronchoscopy (n=154) or without (n=153). Minor complications, particularly intra - procedural hypertension, were more frequent in the bronchoscopy group versus the non bronchoscopy (37.6% vs. 22.7% respectively p = 0.005). No significant differences were found in 24-hour mortality (1% for both), 30-day mortality (24.8% vs. 26.6%; p = 0.72), or hospital length of stay (40.5 ± 48.7 days vs. 38.9 ± 33.1 days; p = 0.73). Mean procedure duration was longer with bronchoscopy (5.46 vs. 4.17 minutes; p = 0.003). Major complication rates were identical (5.2% p = 0.8).
Bronchoscopy during PCT does not reduce major complications or mortality, but increases procedure time, complexity, and costs. For experienced proceduralists, PCT without bronchoscopy is safe and may represent a more efficient approach. Multicenter studies are needed to assess generalizability of these findings.
This study was registered at ClinicalTrials.gov (identifier: NCT02802527) on April 13 2016.
The online version contains supplementary material available at 10.1186/s13054-026-05912-2.
Many centers perform percutaneous dilational tracheostomy (PCT) under bronchoscopic guidance, to reduce the risk of airway injury and confirm cannula placement [1–3].
Some guidelines, such as those published in 2017 recommend that existing evidence does not support the routine use of bronchoscopy during PCT but reflects the persistent lack of high-quality data demonstrating clear benefit of bronchoscopic guidance during the procedure [4]. While bronchoscopy may improve procedural confidence by visualization of tracheal anatomy and possible abnormalities, it also introduces additional complexity—prolonging procedure time, requiring extra staff and equipment, and increasing costs related to endoscope maintenance and disinfection. Moreover, bronchoscopy may worsen gas exchange, leading to hypercarbia, hypoxemia, and elevated intracranial pressure [5, 6].
Reported complication rates, both minor and major, for PCT range from 4 to 12% [7–9]. However, evidence of a reduction in complication rates by the routine use of bronchoscopy remains inconclusive. Several retrospective studies have shown no difference in outcomes with or without bronchoscopic guidance [10–13], while others have suggested fewer complications when bronchoscopy was employed [11, 12]. More recent reviews remain divided—some cautioning against routine bronchoscopy due to physiologic risks such as hypercarbia and desaturation due to temporary narrowing of the airway by the presence of the bronchsoscope, and others endorsing its use for improved visualization and perceived procedural safety [6, 14]. Notably, these recommendations are largely based on expert opinion rather than robust comparative data.
As a result, practice patterns vary bronchoscopy-assisted PCT remains routine in many U.S. centers with no current surveys measuring real life utilization, whereas in Europe, there is a greater variation in practice from place to place [15, 16]. Despite decades of practice, no consensus has emerged regarding the optimal approach to PCT, and large-scale prospective randomized trials addressing this question are still lacking.
The objective of this study was to evaluate the impact of bronchoscopy on the safety of PCT. Demonstrating a benefit of bronchoscopy would support routine integration during PCT in the future. While showing no benefit—or increased risk—could justify avoiding its use, thereby reducing costs and staffing demands.
This prospective, randomized, controlled, open-label trial aimed to evaluate the impact of bronchoscopy on the safety of percutaneous tracheostomy (PCT) by comparing two PCT performed with bronchoscopy guidance and PCT performed without bronchoscopy. The study was conducted in the general (surgical) and medical intensive care units of a single tertiary/quaternary academic center, Hadassah–Hebrew University Medical Center, between 2016 and 2024.Throughout the study period, procedural technique, equipment, and institutional protocols for percutaneous tracheostomy remained unchanged.
After institutional review board (IRB) approval (HMO-0684–15) and patient or surrogate informed consent, patients were randomized 1 using a computer-generated allocation table into two PCT without bronchoscopy (guided by direct laryngoscopy) and PCT with bronchoscopy assistance.
This study was registered at ClinicalTrials.gov (identifier: NCT02802527) on April 13 2016.
The study population included ventilated patients in the ICU for whom PCT was clinically indicated, mainly prolonged mechanical ventilation (over 10 days) and improved patient comfort and to aid weaning from the ventilator.
Patients were excluded if they were < 18 years old; had an uncontrolled coagulopathy; had anatomic conditions precluding percutaneous access and requiring open surgical tracheostomy (e.g., short neck obscured anatomy, restricted cervical extension, cervical spine instability, cervical hematoma, tumor, enlarged thymus, or dense scarring); had local soft tissue infection at the planned insertion site; the attending physician objected for a technical or for a substantive reason to enrolling the patient.
All procedures were performed by a three-operator team including at least one senior ICU attending physician and two anesthesia residents or ICU fellows. An experienced operator was defined as a senior ICU attending physician credentialed by the institution to independently perform both PCT and flexible bronchoscopy. At least one experienced attending was present at the bedside for every procedure. Operator identity was recorded for subgroup analysis. Sedation and neuromuscular blockade were administered as clinically indicated. Neuromuscular blockade was systematically administered in all patients using weight-based rocuronium. Deep sedation was achieved in all patients to a Richmond Agitation–Sedation Scale (RASS) target of − 5 according to the institutional ICU sedation protocol, using propofol 2% (1000 mg in 50 mL 0.9% NaCl) infusion in combination with opioid analgesia (fentanyl 1 mg in 50 mL 0.9% NaCl or remifentanil 5 mg in 50 mL 0.9% NaCl infusions). Etomidate was administered when clinically indicated, such as in cases of hemodynamic instability or lightening of sedation. Continuous monitoring included continuous electrocardiography (ECG), invasive blood pressure and pulse oximetry.
Patients were positioned supine with a shoulder roll and neck extension unless contraindicated.
The PCT procedure involved local anesthesia to the operative site (lignocaine 1% with 100 000 adrenaline), a 1 cm horizontal skin incision was used as the standardized institutional approach, as recommended in prior cohort studies [17, 18]. Tracheal puncture, guidewire insertion, progressive dilation, and cannula placement, followed by confirmation of ventilation (chest auscultation and end-tidal CO2 monitoring) followed. In the bronchoscopy group, real-time airway, wire, and tracheostomy visualization were performed throughout the procedure by a separate member of the ICU team.
The primary outcome was a composite of procedure-related complications occurring within 24 h, classified according to established percutaneous tracheostomy safety frameworks.
Minor complications Desaturation (< 90%), minor bleeding not requiring transfusion, transient blood-pressure changes, arrhythmias, and cuff-related events.
Major complications Airway loss, cannula malposition, tracheal or esophageal injury, pneumothorax requiring drainage, major bleeding requiring transfusion, or death within 24 h.
Secondary outcomes included procedure duration, 30-day mortality, and hospital length of stay [9–11, 19].
Demographics, co-morbidities, SOFA score, and pre-procedure arterial blood gases were recorded. Data accuracy was cross-verified using the Hadassah Hospital proprietary electronic medical record (EMR) in the Medical ICU and the MetaVision EMR (iMDsoft Ltd., Tel Aviv, Israel) in the General ICU. Patients were followed for 30 days post-procedure.
To accurately assess the duration of the procedure, two-time intervals were defined. Preparation time was measured from the administration of anesthesia until the skin incision. Procedure time was measured from the beginning of the skin incision until the end of the operation, defined as the time of inflating the tracheostomy tube cuff.
All time intervals were recorded using a stopwatch or the monitor clock by a dedicated staff member.
Sample size was calculated based on complication rates reported in previously published percutaneous tracheostomy (PCT) cohorts available at the time of study design, which reported overall complication rates of approximately 3–5% for bronchoscopy-assisted PCT and 8–12% for PCT performed without bronchoscopy. Assuming 150 patients per group, a two-sided α of 0.05 and 80% power, the study was powered to detect a difference between complication rates of approximately 3.5% versus 10–11% [6,7,8,9,10]. The composite endpoint was selected to increase statistical power and to reflect overall procedural safety.
Continuous variables were compared using the t-test or Mann–Whitney U test, as appropriate, and categorical variables using the chi-square or Fisher’s exact test. Multivariable logistic regression was used to assess the association between bronchoscopy and complications. A two-sided p-value < 0.05 was considered statistically significant.
From August 4, 2016, to December 24, 2024, a total of 313 patients underwent randomization, 819 were excluded, being performed by attending physicians not participating in the trial. The participants were randomly assigned to two 156 patients who received PCT without bronchoscopy and 157 patients who received PCT with bronchoscopy assistance (Fig. 1).Fig. 1Patient flow diagram
Baseline characteristics, including age, gender, indication for ICU admission, comorbidities, and severity of illness (SOFA score), were comparable between groups. The proportion of procedures performed by experienced physicians also did not differ between groups. Although a small difference in baseline PaCO₂ was statistically significant, it was not clinically relevant (Table 1).Table 1Demographic data and pre procedure statusCharacteristicWithout Bronchoscope (n = 153)With Bronchoscope (n = 154)Total (n = 307)P ValueSexMale104 (68%)111 (72.1%)215 (70%)0.43Age, Mean (SD) (years)60.64 (19.89)59.39 (19.65)60.01 (19.75)0.58Indication for ICU admission0.70Medical admission73 (47.7%)70 (45.5%)143 (46.6%)Respiratory failure (including COVID-19)39 (25.5%)38 (24.7%)77(25%)COVID-1915 (9.8%)11 (7.1%)26 (8.5%)Neurological17 (11.1%)16 (10.4%)33 (10.7%)Other17 (11.1%)16 (10.4%)33 (10.7%)Surgical admission80 (52.3%)84 (54.5%)164 (53.4%)Trauma38 (24.8%)41 (26.6%)79 (25.7%)Abdominal30 (19.6%)33 (21.4%)63 (20.5%)Other12 (7.8%)10 (6.5%)22 (7.2%)ComorbiditiesSmoking50 (32.7%)50 (32.5%)100 (32.6%)0.97Diabetes Mellitus49 (32%)48 (31.2%)97 (31.6%)0.87Cardiac42 (27.5%)34 (22.1%)76 (24.8%)0.26Hypertension67 (43.8%)64 (41.6%)131 (42.7%)0.69Malignancy31 (20.3%)31 (20.1%)62 (20.2%)0.98Respiratory35 (22.9%)31 (20.1%)66 (21.5%)0.56Neurological45 (29.4%)43 (27.9%)88 (28.7%)0.77SOFA, Mean (SD)7.85 (3.02)7.86 (3.16)7.86 (3.09)0.98Experienced Physician63 (42%)50 (32.5%)113 (37.2%)0.09Blood gasespH Mean (SD)7.39 (0.08)7.41 (0.07)7.4 (0.07)0.07PaO2 Mean (SD) – mmHg96.1 (40.1)101.7 (37.0)98.9 (38.6)0.2PaCO2 Mean (SD) – mmHg46.2 (10.3)42.7 (8.9)44.4 (9.8)0.002HCO3- Mean (SD) – mmol/l27.4 (5.6)26.28 (4.7)26.83 (5.2)0.06^^ P value refers to the comparison of overall medical versus surgical admission indications between the bronchoscopy and non-bronchoscopy groups (Chi-square test)Standard Deviation (SD). The sequential organ failure assessment score (SOFA score), Partial pressure of oxygen (PaO₂), Partial pressure of carbon dioxide (PaCO₂)Patient characteristics, including age, sex distribution, Indication for ICU admission, comorbidities, and SOFA scores, were similar between groups. The proportion of procedures performed by experienced physicians also did not differ. Although baseline PaCO₂ levels differed statistically, the variation was not clinically significant. Standard Deviation (SD)
The rate of the composite primary outcome did not differ significantly between groups. The overall rate of major complications was identical between groups (5.2% p = 0.8), and no differences were found when analyzed by subtype. Among minor complications, only intra - procedural hypertension was significantly more frequent in the bronchoscopy group (34 [22%] vs 56 [38%]; p = 0.005) (Fig. 2, Table 2). Three patients (1%) died within 24 h, with one death directly related to the procedure; mortality did not differ between groups (Fig. 2).
Fig. 2Complications. Forest plot displays relative risk (RR) with 95% confidence intervals (CI) for each complication type. Circles represent point estimates, and horizontal lines denote CIs. The vertical red dashed line indicates an RR of 1 (no difference between groups). 'All major/minor complications' denotes number of patients with complications, not total events
Table 2Vital signs during the procedureCharacteristicWithout Bronchoscope (n = 153)With Bronchoscope (n = 154)Total (n = 307)P ValueMinimal HR Mean (SD) –beats per minute84.78 (19.54)82.86 (17.35)83.82 (18.48)0.37Maximal HR Mean (SD) – beats per minute101.76 (22.59)101.17 (20.84)101.47 (21.7)0.81Minimal MAP Mean (SD) – mmHg81.36 (19.73)84.98 (21.32)83.15 (20.58)0.13Maximal MAP Mean (SD) – mmHg103.6 (25.88)111.6 (24.18)107.6 (25.32)0.006Minimal SAT Mean (SD)—%97.04 (5.22)96.81 (5.94)96.93 (5.58)0.73Heart Rate (HR), Mean arterial pressure (MAP), Standard Deviation (SD). O₂ Saturation (SAT)No significant differences were observed between groups in heart rate (HR) or oxygen saturation (SAT) throughout the procedure. MAP was significantly higher in the bronchoscopy group, while minimum blood pressure values did not differ
No differences were observed between groups in post-procedure outcomes. Both preparation time (mean[SD]; 8.15 [6.92] vs 11.77 [8.99]; p < 0.001) and procedure duration (4.17 [3.56] vs 5.46 [3.97]; p = 0.003) were significantly longer in the bronchoscopy group (Table 3; Fig. 3). These differences remained statistically significant when stratified by operator experience (experts vs. trainees). No significant differences were observed in the distribution of complications between specialists and trainees (Supplementary Table 4).
Table 3Procedure DurationCharacteristicWithout Bronchoscope (n = 153)With Bronchoscope (n = 154)Total (n = 307)P value Overall procedure time Sedation to skin incision, mean (SD), min8.15 (6.92)11.77 (8.99)9.92 (8.19)0.0004Skin incision to cuff inflation, mean (SD), min4.17 (3.56)5.46 (3.97)4.81 (3.82)0.003 Procedure time by operator experience
Sedation to skin incision, mean (SD), min Non-experienced physician8.83 (8.24)11.61 (6.51)—0.001Experienced physician7.06 (4.63)12.10 (12.76)—0.016 Skin incision to cuff inflation, mean (SD), min Non-experienced physician4.59 (2.16)6.02 (3.91)—0.002Experienced physician3.56 (4.81)4.26 (3.85)—0.016Mean arterial pressure (MAP), Standard Deviation (SD)Both preparation time (sedation to skin incision) and procedural time (skin incision to cuff inflation) were significantly longer in the bronchoscopy group, with findings consistent across operator experience strata Fig. 3Procedure durations. Violin plots display the distribution of times for each phase of the total duration (A), preparation time from sedation to skin incision, and operative time from incision to balloon inflation (B). Horizontal lines represent medians and interquartile ranges. Both preparation (p = 0.003) and operation (p < 0.001) times were significantly longer in the bronchoscopy group
Thirty-day mortality was similar between groups (24.8% vs. 26.6%; p = 0.72), as was in-hospital length of stay (40.5 ± 48.7 days vs. 38.9 ± 33.1 days; p = 0.73). The proportion of patients alive at hospital discharge was also comparable (61.2% vs. 61.8%; p = 0.91).
This study found that performing PCT with the assistance of a bronchoscope increases the procedure time by about 30%. Bronchoscopy was associated with higher mean arterial pressure during the procedure, likely reflecting airway manipulation or transient sympathetic stimulation; this finding may be incidental rather than a direct surrogate of patient stress. However, no significant differences were found in major complication rates, mortality, or length of hospitalization between the groups. Since the use of a bronchoscope requires additional staff, higher skill levels, and expensive equipment, and can sometimes cause delays due to limited equipment availability, this study shows that PCT may be performed safely without routine use of a bronchoscope.
The results of the study support the approach commonly used in many countries, where PCT is performed without a bronchoscope [4], and do not support the method used in other intensive care units [6, 14]. However, despite the study's results showing that the procedure is safe to perform without a bronchoscope, they also indicate that choosing to perform the procedure with a bronchoscope, despite possibly increasing minor complications such as hypercapnia [5], is still safe in terms of major complications and mortality.
Differences in the approaches to performing the procedure between medical centers arise from various and contradictory retrospective studies published over the years [10–13]. This current study, being a large, randomized, and prospective trial, adds significant value to the existing literature. Its results may influence the establishment of guidelines for treating patients requiring tracheostomy in intensive care units, especially when considering whether to perform the procedure with or without a bronchoscope.
It should be noted that in recent years (during the course of the current study), two smaller randomized prospective studies were conducted simultaneously, concluding that bronchoscope-guided procedures have a lower complication rate. One study where 60 subjects were randomly assigned to those who underwent PCT with and without bronchoscopic guidance found that the complication rate was higher in the group without bronchoscope guidance (60% vs. 13.3%, P < 0.05), and the procedure time was longer in the bronchoscope-guided group (14 vs. 7 min, P < 0.05) [20].
In 2019, another prospective study included 90 patients [21]. In 45 patients, the PCT was performed with bronchoscope assistance and in 45 patients the procedure was done without bronchoscopic guidance. The complication rate was higher in the group without bronchoscopic guidance (40% [18/45] vs. 20%[9/45]; P < 0.05); mainly minor bleeding; 26.7% (27/90) overall, with 1 pneumothorax and 2 major bleeding in the non bronchoscopy group. The average procedure duration was longer in the group without bronchoscope guidance (12.9 vs. 9.8 min; P < 0.05). The conclusions of these two studies regarding procedure duration remain contradictory, and it is not clear how procedure duration was defined in them. It is important to note that the complication rate in both studies was driven by the gap between minor complications, and it was significantly higher than what was reported in previous studies in the literature. This raises questions about the sample size calculation and the validity of the study findings. The current study aims, among other things, to address the question of procedure duration, and therefore, it defined the duration precisely. Additionally, due to the relatively low rate of major complications in the literature, the study was designed with a larger sample size in an attempt to provide more statistically valid information about these complication rates.
Although an approximately 8% increase in blood pressure (a minor complication) was observed in the bronchoscope group, this finding does not have clinically significant meaning, and no difference in late complication rates was recorded. However, despite the obvious conclusions of our study that a procedure without a bronchoscope should be preferred, it should be noted that the procedure requires individual tailoring to the patient, and the use of a bronchoscope may be particularly useful in cases of complex anatomy or when more precise monitoring is required during the procedure.
Additionally, it was found that although procedure time is statistically significantly longer between residents and specialists, this is clinically insignificant.
It is noted that in previous studies, the distinction between residents and specialists was not made, likely because in many places worldwide, especially in North America, it is not common for a resident to initially perform the procedure.
This study also has several limitations. First, it was conducted at a single center, which may limit generalizability. Centers with less experience in PCT might derive greater benefit from bronchoscopic guidance. Second, outcomes were limited to the short term; long-term complications such as tracheal stenosis were not assessed. Third, some data on later complications were collected retrospectively from patient records, which may underestimate their true incidence. Given the low incidence of major complications observed, the study may not have been powered to detect small between-group differences; however, any such differences are unlikely to be clinically relevant in experienced centers.
Nonetheless, our study has many strengths. This prospective, randomized controlled design minimizes bias and allows direct comparison between techniques. The inclusion of a large sample across both medical and surgical ICUs enhances generalizability. Uniform complication classification and standardized procedural protocols strengthen the reliability of outcome assessment.
The results of this study demonstrate that in experienced centers, routine bronchoscope use during PCT offers no advantage in reducing major complications or improving outcomes. However, bronchoscopy prolongs the procedure and increases costs. Its selective use in anatomically complex cases remains reasonable. Larger multicenter studies with extended follow-up are needed to clarify generalizability of our findings regarding the role of bronchoscopy in PCT.
Additional file 1.