Share this article on:

The Kepler Intubation System

Hemmerling, Thomas M. MSc, MD, DEAA; Wehbe, Mohamad MASc; Zaouter, Cedrick MD; Taddei, Riccardo MD; Morse, Joshua

doi: 10.1213/ANE.0b013e3182410cbf
Technology, Computing, and Simulation: Technical Communication
Chinese Language Editions

Our goal in this study was to develop a robotic intubation system and to conduct a feasibility pilot study on the use of a robotic intubation system for endotracheal intubations. The Kepler Intubation System was developed, consisting of a remote control center (joystick and intubation cockpit) linked to a standard videolaryngoscope via a robotic arm. Ninety intubations were performed by the Kepler Intubation System on an airway trainer mannequin by a single operator. The first group of 30 intubations was performed with the operator in direct view of the mannequin (direct view group). The second group of 30 intubations was performed with the operator unable to see the mannequin (indirect view group). Thirty semiautomated intubations were also performed during which the robotic system replayed a trace of a previously recorded intubation maneuver (semiautomated group). First-attempt success rates and intubation times for each trial were recorded. Trends were analyzed using linear regression. Data are presented as mean (SD). All intubations were successful at first attempt. The mean intubation times were 46 (18) seconds, 51 (19) seconds, and 41 (1) seconds for the direct view, indirect view, and semiautomated group, respectively. Both the direct and indirect view groups had a negative slope, denoting that each successive trial required less time. The semiautomated group had a slope of 0 and a low SD of 1 second, illustrating the high reproducibility of automated intubations. We concluded that a robotic intubation system has been developed that can allow remote intubations within 40 to 60 seconds.

Published ahead of print December 20, 2011 Supplemental Digital Content is available in the text.

From the Department of Anaesthesiology, McGill University, Montreal, Quebec, Canada.

Funding: Departmental funding.

Conflict of Interest: See Disclosures at the end of the article.

Reprints will not be available from the authors.

Address correspondence to Professor Thomas M. Hemmerling, MD, DEAA, Department of Anaesthesiology, McGill University, Montreal General Hospital, 1650 Cedar Avenue Montreal, H3G 1A4, Quebec, Canada. Address e-mail to

Accepted November 4, 2011

Published ahead of print December 20, 2011

Endotracheal intubation is a complex procedure of inserting an endotracheal tube (ETT) into the trachea of an anesthetized patient to allow artificial ventilation. Endotracheal intubations, commonly performed with high success rates in anesthesia departments, can be associated with lower success rates in emergency departments, where intubation success rates can be as low as 70% at first attempt and 89% at second attempt.1 The prehospital success rates of endotracheal intubations have been reported to be as low as 49.1%,2 performed by rural basic emergency medical technicians.

A study conducted by Tighe et al. demonstrated the possibility of using a multipurpose surgical robot to assist in airway management3; however, the expense of such a system prohibits its widespread adoption for this application. In addition, urologists used the da Vinci surgical system for either oro- or nasopharyngeal fiberoptic intubation, which is not the standard practice in anesthesia. However, recently widespread use of videolaryngoscopic devices has increased, and they have been introduced into guidelines of difficult airway, with some anesthesiologists advocating their standard use in everyday practice.4

This article presents the development and feasibility pilot study of the Kepler Intubation System® (KIS), a robotic system developed using low-cost, commercially available components for performing endotracheal intubations. This system is named in honor of the German mathematician and astronomer Johannes Kepler, best known for his laws of planetary motion.

Back to Top | Article Outline


A remote-controlled robotic system was developed to intubate using a standard videolaryngoscope. This system is composed of 4 main components: a ThrustMaster T.Flight Hotas X joystick (Guillemot Inc., New York, NY), a JACO robotic arm (Kinova Rehab, Montreal, Quebec, Canada), a Pentax AWS videolaryngoscope (Ambu A/S, Ballerup, Denmark), and a software control system. Both the Pentax AWS videolaryngoscope and the JACO robotic arm are medical devices currently available on the market. The software control system was developed in C sharp (Microsoft Corporation, Redmond, WA), and the graphical user interface was developed in LabVIEW (National Instruments, Austin, TX). The setup also included a commercial Webcam (Microsoft Corporation, Redmond, WA) placed laterally to the mannequin, providing a video feed of the videolaryngoscope going into the mannequin's mouth (Fig. 1, right panel).

Figure 1

Figure 1

The software control system of the KIS comprises a client and a server to allow intubations to be performed remotely from the operating room. The client and server communicate with each other via the Internet. The main interface of this system is the KIS® Intubation Cockpit (shown in Fig. 1).

Ninety endotracheal intubations were performed using an airway trainer mannequin (Laerdal Airway Management Trainer, Laerdal Medical, Stavanger, Norway) by the same anesthesiologist (T.H.) via a robotically mounted standard videolaryngoscope to determine first- attempt success rate and the average intubation time. The intubation time was defined as the start of moving the KIS from its resting position (Fig. 2A) towards successful insertion of the ETT into the trachea. The operator had no experience with the system before this study.

Figure 2

Figure 2

The base of the robotic arm was fixed on the operating table above the mannequin's head (Fig. 2B), providing a support point for the system; the videolaryngoscope was fixed to the robotic arm's wrist, and the ETT was inserted into the blade of the videolaryngoscope. The positioning of the robotic arm and the videolaryngoscope is a fairly easy, straightforward process, allowing it to be done by minimally trained personnel. Thirty endotracheal intubations were performed with the operator in direct view of the mannequin (group direct view); 30 more were performed with the operator unable to directly see the mannequin—he was sitting in a remote location and had no direct view on the KIS. In the 30 indirect endotracheal intubations, camera feeds provided by the KIS® Intubation Cockpit (Fig. 1) were used by the operator to guide the laryngoscope into the mouth of the mannequin and then to perform the intubation (group indirect view). Thirty semiautomated endotracheal intubations were also performed: in these 30 trials, the robotic arm replayed a trace of an intubation maneuver that was performed and recorded by the operator (group semiautomated). The videolaryngoscope was placed at the tip of the mouth with the mouth open. After pressing a play button, the system then automatically moved the videolaryngoscope in place so that the green crosshair crossed the vocal cords without any operator-guided assistance. Insertion of the ETT was then attempted by the operator. The operator was again in a remote location. All times were measured automatically using video playbacks by a research assistant not involved in this study. All 90 intubations were executed in the following sequence: group direct view, group indirect view, and group semiautomated.

For all 90 endotracheal intubation sessions, the airway mannequin was secured to a standard operating room table and the videolaryngoscope was positioned in order for the green crosshair of the videolaryngoscope to be in the middle of the vocal cords (Fig. 1, left panel). Once this orientation was achieved, the ETT was manually inserted by the operator into the trachea of the mannequin along the track of the blade by simply pushing the ETT forward without additional maneuvering. Success at first attempt was defined as placement of the green crosshair at the crossroads of the vocal cords and insertion of the ETT by the operator simply pushing the ETT forward without additional manipulation. First attempt was only determined if the videolaryngoscope stayed within the airway mannequin's throat without ever leaving the mouth.

The percentage of first attempts was recorded as well as the time it took to perform them. Intubation time was defined as the time duration from the beginning of manipulation of the robotic arm to the insertion of the ETT in the mannequin's trachea.

Data are presented as mean (SD). Trend lines were identified for each of the 3 groups via linear regression. Parameters among the 3 groups were compared using analysis of variance (ANOVA) test and Levene test. P values <0.05 were considered statistically significant. Data analysis was performed using SPSS Statistics (IBM Corporation, Armonk, NY).

Back to Top | Article Outline


All 90 endotracheal intubations performed using the KIS on the airway trainer mannequin were successful at first attempt (100% success rate).

The mean intubation times were 46 (18) seconds, 51 (29) seconds, and 42 (1) seconds for the direct view, indirect view, and semiautomated groups, respectively, without being significantly different. Semiautomated intubations were successfully attempted with a significantly smaller variance of times in comparison to the 2 other groups (P < 0.0001). Figure 3 shows the intubation times measured in seconds for all 3 groups for all trials.

Figure 3

Figure 3

In the direct group and indirect view group only, there was a negative trend for the intubation time (Fig. 4): in the direct view group, the slope was calculated as −1.3 seconds per trial (indicating that with each successive trial, a reduction of intubation time by 1.3 seconds was achieved on average); in the indirect view group, the slope was slightly lower at – 0.9 seconds per trial. There was no improvement in the semiautomated group, with a slope of 0 seconds per trial.

Figure 4

Figure 4

Back to Top | Article Outline


We successfully developed a robotic intubation system, remotely controlling a standard videolaryngoscope with the option of semiautomated intubation. Using the KIS, endotracheal intubation was feasible with mean intubation times of 42 to 51 seconds. Semiautomated intubation was also successfully performed with minimal manual intervention; the times showed minimal SD of 1 second.

Intubation times of 40 to 60 seconds are approximately 2-fold the reported times for manual intubation using a videolaryngoscope.5 This time difference is comparable with similar time differences when laparoscopic procedures are compared with open surgical procedures,6,7 or endoscopic procedures are compared with robotic surgery. The times noted in this study are significantly less than fiberoptic intubations performed using a multipurpose surgical robot.3

The controller used by the KIS is a standard gaming joystick and provides an intuitive control scheme that could reduce the learning curve for performing endotracheal intubations. A negative slope was obtained for the trend lines for both the direct and indirect view groups, denoting that the intubation time decreases with the amount of experience the operator has with the system. The KIS is designed to allow for remote intubations and could be used as part of a teleanesthesia system. Additionally, the KIS could allow for advanced virtual reality simulation training for difficult intubation procedures.

Although the KIS still requires manual input, the ultimate goal is to fully automate the intubation process. The first step toward this goal is to automate the navigation of the blade of the videolaryngoscope through the mouth and pharynx, avoiding damage to the teeth and tissue. At this point, the operator could easily align the videolaryngoscope with the trachea and manually insert the ETT. We therefore tried to automate the intubation process by recording a joystick-controlled intubation and replaying it. The 30 semiautomated trials highlight the strength of automating this process: the SD of 1 second of the semiautomated group demonstrates that the intubation procedure can be automatically performed with very high reproducibility in terms of both time and movement.

A recent meta-analysis investigated intubation success rates of physicians and nonphysician clinicians.8 Average intubation success rates varied from 69.8% to 86.8%, depending on the environment, the patient condition, and the training level of the subjects. That article also highlights the problem that the individual success rate is not only influenced by initial skill attainment or subsequent training sessions, but also by the subsequent opportunity to use these skills in real patients to maintain proficiency. Therefore, we believe that robotic aids, such as our system, might have a place in the future development of systems that can either help or replace human manual skills.

There are several limitations to this study. At present, the robot needs to be installed at the operating table before placing the patient on it. This usually takes approximately 5 minutes before the system is ready to be used. This obviously is longer than using a standard laryngoscope, or even a modern videolaryngoscope. There are therefore plans for installing the KIS on a mobile platform. The KIS is also an initial prototype with future technical focus being placed on making the system smaller and eventually portable. In comparison with manual use of the Pentax videolaryngoscope, times achieved with the KIS were significantly longer; at present, the KIS was also only used by 1 operator who, however, had no prior experience with manual use of the Pentax videolaryngoscope. The semiautomated mode delivered successful intubations more quickly than when KIS was used by the human operator; however, the same airway mannequin was used for all trials. Future developments will focus on testing the KIS, both with manual operator and in semiautomated use, in difficult airway situations. In general, human manual skills can adapt more readily to unexpected airway situations. Future developments are also planned to provide force feedback to the operator. In the mannequin, a standard mouth opener was applied to open the mouth by the operator. In human tests, this might have to be provided by a human aid. The Pentax videolayrngoscope has a bigger size than other videolarygnoscopes because the ETT is attached to it. It is therefore more difficult to use with smaller mouth openings or restricted head inclinations, whether used together with the KIS or directly, manually. However, it offers the advantage that, once it is inserted into the mouth and the vocal cords are within the view of the crosshair, the ETT can easily be pushed into the trachea, without stylet or additional manipulation. Human tests will also focus on the effects of KIS on cardiovascular and respiratory vital signs during its use.

In conclusion, we present the first robotic intubation system for oral endotracheal intubation. Future studies will include trials on advanced simulator mannequins with difficult intubation scenarios and human clinical tests.

Back to Top | Article Outline


Name: Thomas M. Hemmerling, MSc, MD, DEAA.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Thomas M. Hemmerling approved the final manuscript.

Conflict of Interest: Thomas M. Hemmerling is the inventor and patent holder of the Kepler Intubation System and stands to benefit financially.

Name: Mohamad Wehbe, MASc.

Contribution: This author helped conduct the study and analyze the data.

Attestation: Mohamad Wehbe approved the final manuscript.

Conflict of Interest: This author has no conflict of interest to declare.

Name: Cedrick Zaouter, MD.

Contribution: This author helped with the revision of the manuscript.

Attestation: Cedrick Zaouter approved the final manuscript.

Conflict of Interest: This author has no conflict of interest to declare.

Name: Riccardo Taddei, MD

Contribution: This author helped conduct the study and revise the manuscript.

Attestation: Riccardo Taddei approved the final manuscript.

Conflict of Interest: This author has no conflict of interest to declare.

Name: Joshua Morse.

Contribution: This author helped conduct the study, analyze the data, and write the manuscript.

Attestation: Joshua Morse approved the final manuscript.

Conflict of Interest: This author has no conflict of interest to declare.

This manuscript was handled by: James G. Bovill, MD, PhD, FCARCSI, FRCA.

Back to Top | Article Outline


The authors would like to express their gratitude to Shantale Cyr, PhD, for her invaluable help in preparing the illustrations and Lingshan Tang, MD, during the conduct of the study.

Back to Top | Article Outline


1. Choi YF, Wong TW, Lau CC, Siu AYC, Lo CB, Yuen MC, Tung WK, Ng P, Kam CW, Mui TK, Yuen WL, Lim B, Lit AC. A study of orotracheal intubation in emergency departments in five district hospitals in Hong Kong. HKJEM 2003;10:138–45
2. Bradley JS, Billows GL, Olinger ML, Boha SP, Cordell WH, Nelson DR. Prehospital oral endotracheal intubation by rural basic emergency medical technicians. Ann Emerg Med 1998;32:26–32
3. Tighe PJ, Badiyan SJ, Luria I, Lampotang S, Parekattil S. Robot-assisted airway support: a simulated case. Anesth Analg 2010;111:929–31
4. Saxena S. The ASA difficult airway algorithm: is it time to include video laryngoscopy and discourage blind and multiple intubation attempts in the nonemergency pathway? Anesth Analg 2009;108:1052
5. Teoh WH, Saxena S, Shah MK, Sia AT. Comparison of three video laryngoscopes: Pentax Airway Scope, C-MAC™, Glidescope® vs the Macintosh laryngoscope for tracheal intubation. Anaesthesia 2010;65:1126–32
6. Kang SB, Park JW, Jeong SY, Nam BH, Choi HS, Kim DW, Lim SB, Lee TG, Kim DY, Kim JS, Chang HJ, Lee HS, Kim SY, Jung KH, Hong YS, Kim JH, Sohn DK, Kim DH, OH JH. Open versus laparoscopic surgery for mid or low rectal cancer after neoadjuvant chemoradiotherapy (COREAN trial): short-term outcomes of an open-label randomised controlled trial. Lancet Oncology 2010;11: 637–45
7. McGreevy JM, Goodney PP, Birkmeyer CM, Finlayson SR, Laycock WS, Birkmeyer JD. A prospective study comparing the complication rates between laparoscopic and open ventral hernia repairs. Surg Endosc 2003;17:1778–80
8. Hubble MW, Brown L, Wilfong DA, Hertelendy A, Benner RW, Richards ME. A meta-analysis of prehospital airway control techniques: part I. Orotracheal and nasotracheal intubation success rates. Prehosp Emerg Care 2010;14:377–401
© 2012 International Anesthesia Research Society