This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by O’Connor, C. J. Articles by Sandler, R. H. Search for Related Content PubMed PubMed Citation Articles by O’Connor, C. J. Articles by Sandler, R. H.

---

TECHNOLOGY, COMPUTING, AND SIMULATION

Identification of Endotracheal Tube Malpositions Using Computerized Analysis of Breath Sounds via Electronic Stethoscopes

Christopher J. O’Connor, MD^*, Hansen Mansy, PhD, Robert A. Balk, MD, Kenneth J. Tuman, MD^*, and Richard H. Sandler, MD

*Department of Anesthesiology, Department of Pediatrics, Department of Pulmonary Medicine and Critical Care, Rush Medical College at Rush University Medical Center, Chicago, Illinois

Address correspondence to Christopher J. O’Connor, MD, Department of Anesthesiology, Rush Medical College at Rush University Medical Center, 1653 W. Congress Parkway, Chicago, IL 60612. Address electronic mail to: christopher_oconnor{at}rush.edu.

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
Endotracheal tube (ETT) malpositioning into a mainstem bronchus or the esophagus may result in significant hypoxemia. Current methods to determine correct ETT position include auscultation, radiography, and bronchoscopy, although the current acceptable standard procedure for proper endotracheal (versus esophageal) intubation is detection of end-tidal carbon dioxide (ETco₂) by capnography, capnometry, or colorimetric ETco₂ devices. Unfortunately, capnography may be unavailable or unreliable in nonhospital/emergency settings or in low cardiac output states, and it does not detect endobronchial intubation. The purpose of this study was to quantify and assess breath sound characteristics using electronic stethoscopes placed over each hemithorax and epigastrium to determine their ability to detect ETT malposition. We recorded breath sounds in 19 healthy, non-obese adults before general surgical procedures. After intubation of the trachea, the ETT was bronchoscopically positioned 3 cm above the carina, after which 3 breaths of 500 mL were given and breath sounds were recorded. A second ETT was placed in the esophagus and the same series of breaths and recordings were performed. Finally, the tracheal ETT was advanced into the right mainstem bronchus and breath sounds were recorded. Using computerized analysis, breath sounds were digitized and filtered to remove selected frequencies, and acoustic signals and energy ratios were obtained for all 3 positions. Total energy ratios using band-pass filtering of the acoustic signals accurately identified all esophageal and endobronchial intubation (P < 0.001). These preliminary results suggest that this technique, when incorporated into a 3-component, electronic stethoscope-type device, may be an accurate, portable mechanism to reliably detect ETT malposition in adults when ETco₂ may be unavailable or unreliable.

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
Endobronchial intubation (EBI) may produce significant hypoxemia or atelectasis, and unrecognized esophageal intubation can cause brain injury or death. The importance of these complications is heightened by the finding that 25% of tracheas intubated in the prehospital setting have improperly placed endotracheal tubes (ETTs) on arrival to the hospital (1); other studies have shown EBI rates from 5% to 28% in intensive care units and cardiac arrest patients, respectively (2,3). Accidental EBI accounted for 42% of events related to the ETT in the Australian Incident Monitoring Study (4), and more recent analysis of that study demonstrated that capnography remained normal in 88.5% of these cases (5). EBI may also occur during emergency transport of patients (6) and with changes in patient or neck position (7). Less commonly, but no less important, is the occurrence of EBI in the operating room (8–10). Effective monitors for EBI, though lacking, are clearly necessary.

Current methods to detect esophageal intubation include radiography, bronchoscopy, auscultation, capnography/capnometry, pulse oximetry, sonomatic confirmation of tracheal intubation (SCOTI device) (11,12), acoustic reflectometry (13), the esophageal detector bulb (14), electronic esophageal detectors (15), colorimetric detection of end-tidal carbon dioxide (ETco₂) (16), and direct visualization of ETT passage through the vocal cords. All of these methods have documented limitations and, except for radiography and bronchoscopy, none can reliably detect EBI.

The purpose of this study was to measure breath sound (BS) characteristics associated with tracheal, endobronchial, and esophageal intubations and to assess the utility of these characteristics for ETT malposition detection. Identifying a variable (such as the acoustic energy ratio found in the current study) that has a precise accuracy of differentiating tracheal, bronchial, and esophageal intubation would be a useful addition to current techniques of detecting proper ETT position.

	Methods

Top Abstract Introduction Methods Results Discussion References

 
After IRB approval and informed consent, BS were recorded in 19 healthy, non-obese adults undergoing general surgery. The required sample size was calculated for a Wilcoxon’s signed rank test to estimate the difference between the left and right (L/R) ratio for the tracheal versus endobronchial comparison. Based on previous observations, means for the two groups were estimated at 25% and 75%, with a 25% standard deviation, generating an effect size of 50%. For this power analysis, a difference of 20% in the standard deviations was used. Using the UnifyPow SAS macro (SAS, Cary, NC), a sample size of 14 patients gave 90% power with an of 0.05. We determined, based on the sample size calculations, experimental design, and feasibility, that a total sample size of 19 patients would be optimal. While patients were supine, 2 electronic stethoscopes (Labtron 1060; Grahm-Field, Hauppage, NY) were placed at the L/R intersections of the axillary and nipple lines and a third stethoscope was placed over the epigastrium. After induction of anesthesia and tracheal intubation with a cuffed 7.0-mm inner diameter ETT, BS from the three stethoscopes were recorded on a digital audio recorder (MD8; Yamaha, Japan). Three breaths (500 mL each, every 2 s) were delivered via positive-pressure bag ventilation. A second 7.0-mm inner diameter ETT was then placed in the esophagus and its balloon was inflated. Stomach contents were evacuated using a gastric tube and another three breaths were delivered into the esophagus while acoustic signals were recorded. The stomach contents were again evacuated and the esophageal tube was removed. Subjects were ventilated through the ETT for 2 min. The ETT cuff was then deflated and the ETT advanced 2–3 cm into the right mainstem bronchus under fiberoptic guidance. After reinflating the balloon, BS were again recorded during 3 similar breaths. The ETT was then promptly repositioned into the trachea and the balloon was reinflated.

The acoustic signals were converted into digital format using laptop-assisted computerized analysis. The acoustic energy in each BS was then calculated (17), and the energy ratio between stethoscope pairs was calculated before and after filtering out specific acoustic frequencies (passband = 300–600 Hz). The choice of the appropriate frequency band is discussed elsewhere (17). The ratio of the acoustic energy between the L/R stethoscopes was calculated to assess BS asymmetry. The energy ratio between the epigastrium and the right (E/R) stethoscope was also calculated. Energy ratios for the tracheal, endobronchial and esophageal intubations were compared using the Wilcoxon signed-rank sum test with P 0.05 considered significant.

	Results

Top Abstract Introduction Methods Results Discussion References

 
Figure 1 shows the L/R energy ratio of the filtered BS for the tracheal and EBI. Dashed lines connect data of the same subject. The L/R ratio was smaller for EBI (range, 9% to 32%) than for tracheal (range, 57% to 133%) intubation (P < 0.00001), indicating diminished BS over the left hemithorax during right EBI. Figure 1 also shows that BS were more symmetrical (closer to 100%) for tracheal intubation and that the two states may be completely separated using a L/R ratio threshold of 45% ± 10%. Sensitivity and specificity values were both 100% for L/R energy ratio threshold values between 35% and 55% (95% confidence interval, 79.1%–100%). Separation accuracy was 100% (P < 0.00001) using this threshold when the acoustic signals were filtered as described.

View larger version (25K): [in this window] [in a new window]   Figure 1. A, Energy ratio of filtered breath sounds (passband, 300-600 Hz) between the left (L) and right (R) chest sensors for the tracheal and mainstem bronchus intubation states. The L/R ratio was smaller for endobronchial (range, 9% to 32%) than tracheal (range, 57% to 133%) intubation (P < 0.00001), indicating diminished breath sounds over the left hemithorax during right endobronchial intubation. Dashed lines connect data of the same subject. The dotted line shows an example threshold (45% ± 10%) that can separate the 2 states with 100% sensitivity and specificity (P < 0.0001). B, Energy ratio of unfiltered breath sounds between the left and right chest sensors for the tracheal and right mainstem bronchus intubation states. Dashed lines connect data of the same subject. No clear threshold can separate the 2 states and no threshold value that provides 100% sensitivity and specificity simultaneously can be found.

The E/R energy ratios for filtered BS are shown in Figure 2 for all 3 intubation states. This ratio was larger for esophageal intubations in all patients (P < 0.00001). The E/R during esophageal intubation was always 64% (range, 64% to 565%), whereas EBI and tracheal intubation ratios were always 38% (range, 4% to 38%). Because there is no overlap in the E/R value between esophageal and tracheal intubation/EBI, the former may be separated from the latter with a 100% accuracy using the E/R ratio threshold of 50% ± 10% (dotted line in Figure 2).

View larger version (32K): [in this window] [in a new window]   Figure 2. A, Energy ratio of filtered BS is shown for all 3 intubation states. This ratio was larger for esophageal intubations in all patients (P < 0.00001). Dashed lines connect data of the same subject. A threshold value of 50% ± 10% is in the middle of the range of threshold values that yield a sensitivity and specificity of 100% simultaneously. B, Energy ratio of unfiltered breath sounds between the epigastrium and right chest sensors for the tracheal, esophageal, and right mainstem bronchus intubation states. Dashed lines connect data of the same subject. The shown threshold (45%) can separate the esophageal intubation state from the other 2 states in most cases.

The importance of filtering BS before calculating energy ratios is shown in Figure 1B, which displays the L/R energy ratio of unfiltered BS with dashed lines connecting data of the same subject. The values of L/R ratio in the two intubation states overlapped, suggesting that separation of the two states may not be feasible without baseline tracheal intubation measurements when sounds are not filtered. The unfiltered L/R ratio was smaller for EBI in 13 of the 19 patients but was larger in 4 patients and changed only slightly in 2 subjects. This further suggests that EBI may not be reliably detected using unfiltered L/R ratios, even when baseline measurements are available for comparison.

The E/R ratio of unfiltered BS is shown in Figure 2B. The unfiltered E/R ratio was larger for esophageal intubation in 18 of the 19 subjects, suggesting that if a baseline state (tracheal or EBI) measurement of the E/R ratio is available, then esophageal intubations may be detectible with a sensitivity of 95% (18/19). When baseline measurements are unavailable and a threshold value of 45% to 50% is used to separate esophageal and tracheal ETT position, 2 tracheal and 3 esophageal intubations would be incorrectly identified for a sensitivity and specificity of 84% and 89%, respectively. This highlights the importance of filtering, which increases the sensitivity and specificity to 100%.

	Discussion

Top Abstract Introduction Methods Results Discussion References

 
Our results suggest that computerized analysis of BS may be useful for the assessment of proper ETT positioning. The results are significantly improved when BS are filtered. The data show that if the simple (unfiltered) output of the electronic stethoscope is used, baseline (tracheal intubation) measurements are necessary for correct identification of EBI. However, when filtered energy ratios are used, the computer could properly identify 100% of endobronchial positioned ETTs without the need for baseline measurements (P < 0.00001). The data also suggest that if only unfiltered data are used (as essentially occurs with clinical auscultation), then relatively symmetric BS (unfiltered L/R around 100%) may occur during EBI, and conversely, relatively asymmetric BS (with a L/R different from 100%) may occur during tracheal intubation. This may lead to false identification of ETT position when routine auscultation or unfiltered electronic stethoscope methods are used. In addition, although there was a tendency of the unfiltered E/R ratio to be larger for esophageal intubation, there was overlap in that ratio between esophageal intubation and the other two intubation states, which leads to less than perfect separation accuracy if unfiltered signals are used. However, with the filtering technique used in this study, no overlap was encountered and all esophageal intubations were correctly identified (100% sensitivity and specificity).

A computerized approach to assessing proper ETT position has potential advantages over simple auscultation and other currently available methods. The L/R hemithorax and epigastrium can be compared simultaneously, avoiding reliance on operator memory. Monitoring can be continuous and computerized analysis may more accurately identify ETT malposition than current technology, especially for EBI. Finally, computerized analysis can perform filtering and energy ratio comparisons that are not possible with simple auscultation. Devices such as used in this study could be used for continuing monitoring of ETT position in the intensive care unit, in the operating room, during patient transport, and they may be especially useful when radiographs are unavailable, impractical, or unreliable.

Although capnography is considered the standard procedure for confirmation of endotracheal intubation, a meta-analysis of 2192 patients determined a false-negative and false-positive rate of 7% and 3%, respectively, for emergency capnography use (18). False-negative results may result from severe airway obstruction, low cardiac output states such as cardiac arrest, pulmonary embolism, and severe pulmonary disease, whereas false-positive capnography findings may result from excessive bag-mask ventilation before intubation, antacids or carbonated beverages in the stomach, or the ETT tip located in the oropharynx (18). Moreover, capnography is not useful for detecting EBI. Combining multiple techniques can improve the sensitivity and specificity for confirmation of proper endotracheal position, but this approach is cumbersome and still inadequate to detect EBI (19,20). Our technique theoretically avoids many of the limitations of using ETco₂ for proper ETT position, although it has yet to be tested under conditions such as severe lower airway obstruction or cardiac arrest.

The most reliable clinical sign of ETT (versus esophageal) placement is direct visualization of the ETT passing through the vocal cords. However, direct vocal cord visualization may not be possible, particularly during suboptimal clinical settings such as field emergencies or during attempted intubations by inexperienced emergency personnel. Physiologic methods such as pulse oximetry and ETco₂ detection are often used to verify ETT placement. Unfortunately, oxygen desaturation is a nonspecific and late manifestation of esophageal intubation in most patients undergoing surgery. Our approach of assessing BS alone is not affected by the intrinsic limitations of these physiologic methods.

Although our study showed that quantitative measurements and analysis of BS facilitated accurate identification of EBI and esophageal intubation, it has several limitations. The computer-assisted/automated chest auscultation was assessed only in non-obese adults without lung pathology. Chest or upper abdominal wound dressings, bandages, and chest tubes that would make access to the chest wall difficult may impact the device efficacy. Results in patients with underlying lung disease where BS may be difficult to reliably obtain, as well as in obese adults and infants, may differ from our study population. Finally, our study was performed in a relatively quiet operating room environment. The impact of loud ambient noise on the accuracy of computerized BS analysis remains unstudied, although addition of a simple ambient microphone for noise canceling purposes could be a potential method to eliminate confounding ambient noise if required. These limitations will need to be addressed in future studies of this device.

In summary, preliminary data from this pilot study in patients with no lung pathology suggest precise accuracy (100% sensitivity and specificity) for the detection of esophageal intubation and EBI using this novel computer-assisted chest auscultation device. Further evaluation is necessary in a broader population of patients with a variety of pathologic states and with different conditions and environments. If successful, this technique may be useful in the design of a 3-component, electronic-type stethoscope device that would allow for a rapid, easy, portable, and radiation-free method of determining and monitoring proper ETT position.

The authors wish to express their appreciation to Mario Moric, PhD, for his assistance with the statistical analysis.

	Footnotes

 
Accepted for publication February 28, 2005.

Supported, in part, by grant R44 HL-61108 from the National Heart Lung Blood Institute and the National Institutes of Health.

	References

Top Abstract Introduction Methods Results Discussion References

 

1. Katz SH, Falk JL. Misplaced endotracheal tubes by paramedics in an urban emergency medical services system. Ann Emerg Med 2001;37:32–7.[Web of Science][Medline]
2. Brunel W, Coleman DL, Schwartz DE, et al. Assessment of routine chest roentgenograms and the physical examination to confirm endotracheal tube position. Chest 1989;96:1043–5.[Abstract/Free Full Text]
3. Dronen S, Chadwick O, Nowak R. Endotracheal tip position in the arrested patient. Ann Emerg Med 1982;11:116–7.
4. Szekely SM, Webb RK, Williamson JA, Russell WJ. Problems related to the endotracheal tube: an analysis of 2000 incident reports. Anaesth Intensive Care 1993;21:611–6.[Web of Science][Medline]
5. McCoy EP, Russell WJ, Webb RK. Accidental bronchial intubation. Anaesthesia 1997;52:24–31.[Web of Science][Medline]
6. Kapadia FN, Bajan KB, Raje KV. Airway accidents in intubated intensive care unit patients: an epidemiological study. Crit Care Med 2000;28:659–64.[Web of Science][Medline]
7. Salem MR. Verification of endotracheal tube position. Anesthesiol Clin North Am 2001;19:813–39.[Medline]
8. Caplan RA, Posner KL, Ward RJ, Cheney FW. Adverse respiratory events in anesthesia: A closed claims analysis. Anesthesiology 1990;72:828–33.[Web of Science][Medline]
9. Morray JP, Geiduschek JM, Caplan RA, et al. A comparison of pediatric and adult anesthesia closed malpractice claims. Anesthesiology 1993;78:461–7.[Web of Science][Medline]
10. Schwartz DE, Matthay MA, Cohen NH. Death and other complications of emergency airway management in critically ill adults: a prospective investigation of 297 tracheal intubations. Anesthesiology 1995;82:367–76.[Web of Science][Medline]
11. Haridas RP, Chesshire NJ, Rocke DA. An evaluation of the SCOTI device. Anaesthesia 1997;52:453–6.[Medline]
12. Petroianu G, Maleck W, Bergler W, Rufer R. Sonomatic confirmation of tracheal intubation using the SCOTI. Prehospital Disaster Med 1997;12:78–82.
13. Rapheal DT. Acoustic reflectometry profiles of endotracheal and esophageal intubation. Anesthesiology 2000;92:1293–9.[Web of Science][Medline]
14. Zalesky L, Abello D, Gold MI. The esophageal detector device. Anesthesiology 1993;79:244–7.[Web of Science][Medline]
15. Wolfe TR, Kimball EJ, Ogden LL, et al. Evaluation of an electronic esophageal detector device in patients with morbid obesity and pulmonary failure. Prehosp Emerg Care 2002;6:59–64.[Medline]
16. Rabitsch W, Nikolic A, Schellongowski P, et al. Evaluation of an end-tidal portable ETco₂ colorimetric breath indicator (COLIBRI). Am J Emerg Med 2004;22:4–9.[Medline]
17. Mansy HA, Royston TJ, Balk RA, Sandler RH. Pneumothorax detection using computerized analysis of breath sounds. Med Biol Engl Comput 2002;40:526–32.
18. Li J. Capnography alone is imperfect for endotracheal tube placement confirmation during emergency intubation. J Emerg Med 2001;20:223–9.[Web of Science][Medline]
19. Grmec S, Mally S. Prehospital determination of tracheal tube placement in severe head injury. Emerg Med J 2004;21:518–20.[Abstract/Free Full Text]
20. Grmec S. Comparison of three different methods to confirm tracheal tube placement in emergency intubation. Intensive Care Med 2002;28:701–4.[Web of Science][Medline]

This article has been cited by other articles:

				P. Rudraraju and L. A. Eisen Analytic Review: Confirmation of Endotracheal Tube Position: A Narrative Review J Intensive Care Med, September 1, 2009; 24(5): 283 - 292. [Abstract] [PDF]

				S. Tejman-Yarden, A. Zlotnik, L. Weizman, J. Tabrikian, A. Cohen, N. Weksler, and G. M. Gurman Acoustic Monitoring of Lung Sounds for the Detection of One-Lung Intubation Anesth. Analg., August 1, 2007; 105(2): 397 - 404. [Abstract] [Full Text] [PDF]

				A. Mahajan, N. Hoftman, A. Hsu, R. Schroeder, and S. Wald Continuous Monitoring of Dynamic Pulmonary Compliance Enables Detection of Endobronchial Intubation in Infants and Children Anesth. Analg., July 1, 2007; 105(1): 51 - 56. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by O’Connor, C. J. Articles by Sandler, R. H. Search for Related Content PubMed PubMed Citation Articles by O’Connor, C. J. Articles by Sandler, R. H.