Flexible bronchoscopy (FB) in intubated subjects on mechanical ventilation (MV) has a good safety profile although it has known physiological consequences such as increased airway resistance, hypoxemia, and hypercarbia.1–6 Suctioning through an endotracheal tube with a catheter or during FB exacerbates hypoxemia and causes a reduction in lung volumes.7–9
During FB, two ventilatory strategies are possible: maintaining tidal volume (VT) to maintain baseline CO2 or allowing reduction of VT with the consequence of hypercapnea. Meduri and Chastre10 proposed that the former strategy be applied whereby the ventilator is adjusted to maintain adequate ventilation during FB. Not addressed by this recommendation is that dynamic hyperinflation is a potential consequence due to expiratory flow limitation with FB. Lawson et al11 defined risk of dynamic hyperinflation during FB using a lung model. Hyperinflation leads to an increase in functional residual capacity which can be detected by measuring end expiratory lung volume (EELV). We studied changes in EELV during FB of intubated subjects while limiting VT.
Subjects enrolled were mechanically ventilated patients in the medical intensive care unit (MICU) of Beth Israel Medical Center (BIMC) who required diagnostic FB. Subjects were enrolled consecutively, and the need for FB was determined by the clinical management team. Subjects were excluded if they had contraindications to sedation and paralysis. This study was approved by the Committee on Scientific Activities of BIMC (COSA Project # 126-01) and written consent was obtained from the appropriate health care surrogate of each subject, as no patient was able to participate in the informed consent process due to their critical illness.
Patient Preparation and Ventilator Settings
All subjects were deeply sedated and paralyzed with a neuromuscular blocking agent at the time of the FB. Subjects were hemodynamically stable at the time of the procedure. Setup and calibration of the measurement devices was performed by the study team immediately before performance of the FB.
A commercially available swivel adapter with a rubber diaphragm was used for bronchoscope insertion to reduce air leak (Intersurgical Inc., 417 Electronics Pkwy, Liverpool, NY). Endotracheal tubes ranged from 7.0 to 8.0 mm internal diameter with FB outside diameter of 5 mm (Olympus BF-P30, Center Valley, PA).
Puritan Bennett 840 or 7200 series ventilators (Nellcor Puritan Bennett Corp., Carlsbad, CA) were used for MV on volume-cycled assist control mode. Ventilator settings at the time of the FB were standardized with positive end expiratory pressure (PEEP) level (5 cm H2O), VT (7 to 8 mL/kg) ideal body weight, respiratory rate (12 per minute), pressure limit (60 cm H2O), and peak inspiratory flow rate (60 L/min) with ramp wave. Fifteen minutes prior to commencement of the procedure, the fraction of inspired oxygen was set to 1.0 and was maintained at that setting throughout the procedure. During FB, all ventilatory parameters were held constant.
Performance of FB
The research team had no input into the decision to perform FB. The clinical care team performed the FB as indicated by the patient’s clinical condition. All FB were performed under the direct supervision of an attending physician by fellows in training who had limited experience with FB. At the beginning of the FB procedure, before the clinical team performing the diagnostic portion of the examination, 20 seconds of suction was applied with the FB at the level of the carina. No suction was applied before the initial suction maneuver.
EELVs were measured with respiratory inductance plethysmography (RIP) (Respitrace Data Acquisition System 204, NIMS Inc., Miami Beach, FL) in the DC mode using a supine only posture double band method with Respiband transducers (Ambulatory Monitoring Systems Inc., Ardsley, NY) as previously described and validated.12–14 To ensure thermal stability and eliminate drift, the system was allowed to warm up for four hours before the procedure. Qualitative diagnostic calibration sequences were used over five minutes to calibrate the RIP system as described and validated previously.13,15 Precalibration and postcalibration sequences were plotted against the volume output from the ventilator to confirm linearity and reproducibility in each patient. Signals were collected using an analog adapter board (National Instruments #777270-01, Austin, TX) connected to a laptop computer through a 12-bit analog to digital input card (National Instruments DAQCard #777230-01). BioBench 1.0 (National instruments, 1997) was used for signal storage and output to spreadsheet software for analysis.
Patient vital signs were monitored during the procedure using the MICU monitoring system (Hewlett Packard Viridia Information Center, Model HPM153A, Andover, MA) including blood pressure by arterial line or an automatic oscillometric cuff, pulse oximetry, heart rate, respiratory rate, and electrocardiogram. Arterial blood gases were obtained in subjects immediately before and after FB.
Calibration sequences were constructed using 3 to 5 different VT measurements versus the millivolt (mV) output of the RIP system (Fig. 1). The ventilator was allowed to iterate VT for at least three breaths after each VT change. The inverse slope establishes the conversion factor in mL/mV. The plots were checked for linearity while the inverse slopes were compared before and after FB for reproducibility. One to three separate FB insertions were performed during each procedure as clinically indicated. EELV before and after each insertion were compared with determine changes resulting from each individual FB.
Respiratory system compliance (Crs) was measured immediately preprocedure and postprocedure and calculated using the following formula: Crs (mL/cm H2O)=VT (mL)/[plateau pressure (cm H2O)-PEEP (cm H2O)]. End inspiratory airway resistance (Rei) was measured immediately preprocedure and postprocedure with square wave flow at 60 L/min using the following formula: Rei (cm H2O/L/s)=[Peak inspiratory pressure (cm H2O)-plateau pressure (cm H2O)]/end inspiratory flow (L/s).
Means and SDs are reported for all data.
Sixteen consecutive subjects were enrolled in the study (mean age, 60±17.8 y; male:female, 11:5). For all subjects, the presence of pneumonia and the need for MV was the indication for FB (Table 1). The length of FB procedures ranged from 2 to 8 minutes. All subjects had measurement of changes in EELV; 13 had complete arterial blood gases data.
On insertion of the FB, all subjects had immediate activation of the pressure limit alarm on the ventilator with prompt reduction in VT. Inserting the FB into the subject’s airway decreased EELV in 16/25 (64%) insertions by −325±371 mL (mean±SD). In 8/25 (32%) FB insertions EELV increased by 65±59 mL (mean±SD). Suctioning for 20 seconds reduced the EELV (Fig. 2); resulting in the EELV decreasing in 19/25 (76%) instances by −120±104 mL (mean±SD). In 4/25 (16%) FB insertions with suctioning EELV increased by 29±33 mL (mean±SD). Prebronchoscopy and postbronchoscopy, PaO2 decreased by −61±96 mm Hg (mean±SD) and PaCO2 increased by 15±7 mm Hg (mean±SD) (Figs. 3 and 4, respectively). In subjects with diffuse lung disease, loss of EELV coincided with marked desaturation. This was followed by prompt resaturation in the majority of subjects following the procedure; however, two subjects required a recruitment maneuver to improve oxygenation. Neither Crs nor Rei changed significantly prebronchoscopy and postbronchoscopy [Crs mean change, −3.27±2.43 (mL/cm H2O); Rei mean change, 1.58±4.03 (cm H2O/L/s)]. No patient had a peri-procedural change in blood pressure or cardiac rhythm.
We performed FB on intubated, sedated, and mechanically ventilated MICU subjects with the intention of avoiding dynamic hyperinflation during the procedure by pressure limiting each breath. This resulted in reduction in VT, but protected subjects from dynamic hyperinflation during FB. This was demonstrated by the majority of subjects having a decrease in EELV during FB and no subjects having a clinically significant increase in EELV. As a result of the reduction in VT, all subjects had a rise in PaCO2, which seemed to be well-tolerated. As previously reported, FB and suctioning resulted in desaturation and reduction in EELV in most subjects.1,2,7–9
Data and recommendations for ventilatory management during FB are sparse and are largely based on expert opinion. Meduri and Chastre10 propose setting the ventilator to maintain VT and baseline CO2. That strategy puts patients at risk for dynamic hyperinflation. This study provides evidence to change the recommendations from a ventilatory strategy which seeks to maintain VT and eucapnea to a strategy which allows for a reduction in VT while permitting hypercapnea. In our subjects, the reduction in PaO2 and elevation in PaCO2 were well-tolerated. The reduction in PaO2 was transient as we observed return to baseline SaO2 following resumption of ventilatory support. We assumed that PaCO2 returned to the pre-FB baseline after ventilatory support was reinstituted. To minimize the extent and duration of these consequences of FB, the bronchoscopist performing the procedure is reminded to minimize the duration of FB and to remove the bronchoscope periodically to allow the patient to be ventilated.
This study has several limitations. It is a small prospective study of a single ventilator strategy during FB and thus, our recommendation is based on the toleration of the studied strategy rather than a direct comparison with a strategy that does not limit VT. This ventilator strategy during FB was studied on MICU patients undergoing FB for evaluation of pneumonia; the generalizability of our findings to other populations of patients is not known. It was expected that all FB insertions and all suction maneuvers would result in a decrease in EELV. This occurred the majority of the time, however, there were instances where the EELV increased slightly. These instances of paradoxical increases in EELV may have occurred because of flow limitation related to underlying obstructive airways disease. RIP has inherent problems with drift of signal in the DC mode. This was not a problem with the system used in this study based upon bench testing for baseline drift. RIP is a validated method for determination in changes of EELV.8,14 We did not make formal measurement of VT during FB insertion, but rather focused on the EELV; all subjects by visual assessment of the RIP tracing had marked reduction of VT during FB insertion.
On the basis of our data and the dynamic hyperinflation during FB that was described by Lawson et al,11 we recommend that FB performed on intubated patients use a ventilatory strategy which allows for a reduction in VT to reduce the risk of dynamic hyperinflation. Limiting suctioning time reduces the risk of desaturation, especially in patients vulnerable to derecruitment. Regular interruption of FB with a period of unimpeded ventilation will attenuate the predictable rise of PaCO2 when limiting VT.
1. Albertini RE, Harrell JH, Kurihara N, et al. Arterial hypoxemia induced by fiberoptic bronchoscopy
. JAMA. 1974;230:1666–1667.
2. Dubrawsky C, Awe RJ, Jenkins DE. The effect of bronchofiberscopic examination on oxygenation status. Chest. 1975;67:137–140.
3. Lindholm CE, Ollman B, Snyder JV, et al. Cardiorespiratory effects of flexible fiberoptic bronchoscopy
in critically ill patients. Chest. 1978;74:362–368.
4. Trouillet JL, Guiguet M, Gibert C, et al. Fiberoptic bronchoscopy
in ventilated patients—evaluation of cardiopulmonary risk under midazolam sedation. Chest. 1990;97:927–933.
5. Montravers P, Gauzit R, Dombret MC, et al. Cardiopulmonary effects of bronchoalveolar lavage in critically ill patients. Chest. 1993;104:1541–1547.
6. Pue CA, Pacht ER. Complications of fiberoptic bronchoscopy
at a university hospital. Chest. 1995;107:430–432.
7. Brochard L, Mion G, Isabey D, et al. Constant-flow insufflation prevents arterial oxygen desaturation during endotracheal suctioning. Am Rev Respir Dis. 1991;144:395–400.
8. Maggiore SM, Lellouche F, Pigeot J, et al. Prevention of endotracheal suctioning-induced alveolar derecruitment in acute lung injury. Am J Respir Crit Care Med. 2003;167:1215–1224.
9. Lindgren S, Odenstedt H, Erlandsson K, et al. Bronchoscopic suctioning may cause lung collapse: a lung model and clinical evaluation. Acta Anaesthesiol. 2008;52:209–218.
10. Meduri GU, Chastre J. The standardization of bronchoscopic techniques for ventilator-associated pneumonia. Chest. 1992;102:557S–564S.
11. Lawson RW, Peters JI, Shelledy DC. Effects of fiberoptic bronchoscopy
during mechanical ventilation in a lung model. Chest. 2000;118:824–831.
12. Watson HL, Poole DA, Sackner MA. Accuracy of respiratory inductive plethysmographic cross-sectional areas. J Appl Physiol. 1988;65:306–308.
13. Sackner MA, Watson H, Belsito AS, et al. Calibration of respiratory inductive plethysmography
during natural breathing. J Appl Physiol. 1989;66:410–420.
14. Dall’Ava-Santucci J, Brunet F, Nouira S, et al. Passive partitioning of respiratory volumes and time constants in ventilated patients. Eur Respir J. 1992;5:1009–1017.
15. Leino K, Nunes S, Valta P, et al. Validation of a new respiratory inductive plethysmograph. Acta Anaesthesiol Scand. 2001;45:104–111.