The importance of subjective facial appearance on the ability of anesthesiologists to predict difficult intubation

Emergency medicine updates: Endotracheal intubation

INTRODUCTION

Airway management including endotracheal intubation (ETI) is a key skill for emergency clinicians. Therefore, it is important for emergency clinicians to be aware of the current evidence regarding the identification and management of patients requiring ETI.

OBJECTIVE

This paper evaluates key evidence-based updates concerning ETI for the emergency clinician.

DISCUSSION

ETI is commonly performed in the emergency department (ED) setting but has many nuanced components. There are several tools that have been used to predict a difficult airway which incorporate anatomic and physiologic features. While helpful, these tools should not be used in isolation. Preoxygenation and apneic oxygenation are recommended to reduce the risk of desaturation and patient decompensation, particularly with noninvasive ventilation in critically ill patients. Induction and neuromuscular blocking medications should be tailored to the clinical scenario. Video laryngoscopy is superior to direct laryngoscopy among novice users, while both techniques are reasonable among more experienced clinicians. Recent literature suggests using a bougie during the first attempt. Point-of-care ultrasound is helpful for confirming correct placement and depth of the endotracheal tube.

CONCLUSIONS

An understanding of literature updates can improve the ED care of patients requiring emergent intubation.

Development and validation of a screening method for difficult tracheal intubation based on geometric simulation and computer technology

BACKGROUND

The anatomical characteristics of difficult airways can be analysed geometrically. This study aims to develop and validate a geometry-assisted difficult airway screening method (GADAS method) for difficult tracheal intubation.

METHODS

In the GADAS method, a geometric simulated model was established based on computer graphics. According to the law of deformation of the upper airway on laryngoscopy, the expected visibility of the glottis was calculated to simulate the real visibility on laryngoscopy. Validation of the new method: Approved by the Ethics Committee of Yijishan Hospital of Wannan Medical College. Adult patients who needed tracheal intubation under general anaesthesia for elective surgery were enrolled. The data of patients were input into the computer software to calculate the expected visibility of the glottis. The results of tracheal intubation were recorded by anaesthesiologists. The primary observation outcome was the screening performance of the expected visibility of the glottis for difficult tracheal intubation.

RESULTS

The geometric model and software of the GADAS method were successfully developed and are available for use. We successfully observed 2068 patients, of whom 56 patients had difficult intubation. The area under the receiver operating characteristic curve of low expected glottis visibility for predicting difficult laryngoscopy was 0.96 (95% confidence interval [CI]: 0.95-0.96). The sensitivity and specificity were 89.3% (95% CI: 78.1-96.0%) and 94.3% (95% CI: 93.2%-95.3), respectively.

CONCLUSIONS

It is feasible to screen difficult-airway patients by applying computer techniques to simulate geometric changes in the upper airway.

On the Horizon: Specific Applications of Automation and Artificial Intelligence in Anesthesiology

Purpose of Review

The purpose of this review is to summarize the current research and critically examine artificial intelligence (AI) technologies and their applicability to the daily practice of anesthesiologists.

Recent Findings

Novel AI tools are developed using data from electronic health records, imaging, waveforms, clinical notes, and wearables. These tools can accurately predict the perioperative risk for adverse outcomes, the need for blood transfusion, and the risk of difficult intubation. Intraoperatively, AI models can assist with technical skill augmentation, patient monitoring, and management. Postoperatively, AI technology can aid in preventing complications and discharge planning. While further prospective validation is needed, these early applications demonstrate promise in every area of perioperative care.

Summary

The practice of anesthesiology is at a precipice fueled by technological innovation. The clinical AI implementation would enable personalized and safer patient care by offering actionable insights from the wealth of perioperative data.

Artificial Intelligence for Perioperative Medicine: Perioperative Intelligence

The anesthesiologist's role has expanded beyond the operating room, and anesthesiologist-led care teams can deliver coordinated care that spans the entire surgical experience, from preoperative optimization to long-term recovery of surgical patients. This expanded role can help reduce postoperative morbidity and mortality, which are regrettably common, unlike rare intraoperative mortality. Postoperative mortality, if considered a disease category, will be the third leading cause of death just after heart disease and cancer. Rapid advances in technologies like artificial intelligence provide an opportunity to build safe perioperative practices. Artificial intelligence helps by analyzing complex data across disparate systems and producing actionable information. Using artificial intelligence technologies, we can critically examine every aspect of perioperative medicine and devise innovative value-based solutions that can potentially improve patient safety and care delivery, while optimizing cost of care. In this narrative review, we discuss specific applications of artificial intelligence that may help advance all aspects of perioperative medicine, including clinical care, education, quality improvement, and research. We also discuss potential limitations of technology and provide our recommendations for successful adoption.

Usefulness of lateral cephalometric radiography for successful blind nasal intubation: a prospective study

Background

This study aimed to investigate the relationship between pharyngeal morphology and the success or failure of blind nasotracheal intubation using standard lateral cephalometric radiography and to analyze the measurement items affecting the difficulty of blind nasotracheal intubation.

Methods

Assuming a line perpendicular to the Frankfort horizontal (FH) plane, the reference point (O) was selected 1 cm above the posterior-most end of the hard palate. A line passing through the reference point and parallel to the FH plane is defined as the X-axis, and a line passing through the reference point and perpendicular to the X-axis is defined as the Y-axis. The shortest length between the tip of the uvula and posterior pharyngeal wall (AW), shortest length between the base of the tongue and posterior pharyngeal wall (BW), and width of the glottis (CW) were measured. The midpoints of the lines representing each width are defined as points A, B, and C, and the X and Y coordinates of each point are obtained (AX, BX, CX, AY, BY, and CY). For each measurement, a t-test was performed to compare the tracheal intubation success and failure groups. A binomial logistic regression analysis was performed using clinically relevant items.

Results

The items significantly affecting the success rate of blind nasotracheal intubation included the difference in X coordinates at points A and C (Odds ratio, 0.714; P-value, 0.024) and the ∠ABC (Odds ratio, 1.178; P-value, 0.016).

Conclusion

Using binomial logistic regression analysis, we observed statistically significant differences in AX-CX and ∠ABC between the success group and the failure group.

Facial Emotion Recognition in Obesity and in Fibromyalgia: A Systematic Review

Facial emotion recognition (FER) is extensively investigated in psychological sciences in healthy individuals and clinical conditions. In this paper, we analyzed those studies in which FER was assessed in the case of obesity or fibromyalgia, in relation to the levels of alexithymia. Crucially, these two conditions frequently co-occur; however, no study has explored FER considering both fibromyalgia and obesity. Studies were identified using the electronic search engine of PubMed. The last research was run on 23 July 2021. Two independent lists were generated for the two clinical conditions. Six records were reviewed about obesity, while three records about fibromyalgia. The evidence relative to FER in obesity was not conclusive, whereas the evidence about an altered FER in fibromyalgia seemed more straightforward. Moreover, the role of alexithymia on FER in these clinical conditions was not extensively investigated. In our discussion, we highlighted those factors that should be carefully addressed in investigating FER in these clinical conditions. Moreover, we underlined methodological criticisms that should be overcome in future research.

Identification of difficult to intubate patients from frontal face images using an ensemble of deep learning models

Failure to identify difficult intubation is the leading cause of anesthesia-related death and morbidity. Despite preoperative airway assessment, 75-93% of difficult intubations are unanticipated, and airway examination methods underperform, with sensitivities of 20-62% and specificities of 82-97%. To overcome these impediments, we aim to develop a deep learning model to identify difficult to intubate patients using frontal face images. We proposed an ensemble of convolutional neural networks which leverages a database of celebrity facial images to learn robust features of multiple face regions. This ensemble extracts features from patient images (n = 152) which are subsequently classified by a respective ensemble of attention-based multiple instance learning models. Through majority voting, a patient is classified as difficult or easy to intubate. Whereas two conventional bedside tests resulted in AUCs of 0.6042 and 0.4661, the proposed method resulted in an AUC of 0.7105 using a cohort of 76 difficult and 76 easy to intubate patients. Generic features yielded AUCs of 0.4654-0.6278. The proposed model can operate at high sensitivity and low specificity (0.9079 and 0.4474) or low sensitivity and high specificity (0.3684 and 0.9605). The proposed ensembled model outperforms conventional bedside tests and generic features. Side facial images may improve the performance of the proposed model. The proposed method significantly surpasses conventional bedside tests and deep learning methods. We expect our model will play an important role in developing deep learning methods where frontal face features play an important role.

Artificial Intelligence and Machine Learning in Anesthesiology

Commercial applications of artificial intelligence and machine learning have made remarkable progress recently, particularly in areas such as image recognition, natural speech processing, language translation, textual analysis, and self-learning. Progress had historically languished in these areas, such that these skills had come to seem ineffably bound to intelligence. However, these commercial advances have performed best at single-task applications in which imperfect outputs and occasional frank errors can be tolerated.The practice of anesthesiology is different. It embodies a requirement for high reliability, and a pressured cycle of interpretation, physical action, and response rather than any single cognitive act. This review covers the basics of what is meant by artificial intelligence and machine learning for the practicing anesthesiologist, describing how decision-making behaviors can emerge from simple equations. Relevant clinical questions are introduced to illustrate how machine learning might help solve them-perhaps bringing anesthesiology into an era of machine-assisted discovery.

Mallampati test with phonation, tongue protrusion and supine position is most correlated with Cormack-Lehane test

Many modified Mallampati tests have been developed to date. Samsoon's modified Mallampati test (standard Mallampati test) is currently widely used. We newly designed seven types of assessment protocol of Mallampati test, in addition to standard Mallampati test. In this study, we studied the correlation between eight types of protocol (standard and seven alternative protocols) of Mallampati test and Cormack-Lehane test. We newly designed assessment protocols as new Mallampati test. These are different protocols depending on the presence or absence of phonation, those of protrusion of tongue, and sitting position or supine position. The oropharyngeal structures visualized by these eight types of Mallampati test for total of 145 patients undergoing dental oral surgery were evaluated. The scores derived via eight types of Mallampati test were recorded. The influence of phonation, tongue protrusion and body position on Mallampati test score was analyzed, respectively. The relationships between eight types of Mallampati test and Cormack-Lehane test were analyzed. Tongue protrusion, phonation and sitting position tended to lower the score of Mallampati test (p < 0.001, respectively). The standard Mallampati test was not correlated with Cormack-Lehane test. In the new Mallampati tests, assessment protocol with tongue protrusion, phonation and sitting position, that with tongue protrusion and supine position, or that with tongue protrusion, phonation and supine position were significantly correlated with Cormack-Lehane test, respectively. (p = 0.020, p = 0.007 and p = 0.004, respectively). The standard Mallampati test did not correlate with Cormack-Lehane test. Mallampati test with phonation, tongue protrusion and supine position were most correlated with Cormack-Lehane test.

Can thyromental distance be measured accurately?

Using the thyromental distance (TMD) measured based on the ultrasonographic location of the thyroid cartilage prominence as the criterion, we investigated the accuracy of TMD measurement by surface landmark identification of the thyroid cartilage prominence. Twenty-nine anesthetist resident volunteers were recruited, including 10 first-year residents, 9 second-year residents and 10 third-year residents. Each volunteer measured the other 28 volunteers’ TMD. Then, the thyroid cartilage prominence of each volunteer was identified by ultrasonography of the junction of the vocal cord and thyroid cartilage, and the TMD was measured precisely. The error of the TMD measurement was determined by the minimal detectable difference (MDD) compared to the ultrasound measurement. A difference of greater than 5.4 mm between the TMD measured by volunteers and that based on ultrasound localization was defined as a measurement error. The measurement error rate of females’ TMD was significantly higher than that of males’ (50 vs 10%, P < 0.001). The error rates of anesthetist residents of first-year, second-year and third-year were 34, 27, and 31%, respectively, and were not significantly different. The error of TMD measurement by surface landmark identification is often, especially for women. More clinic experience don’t improve it.

Incorporating airway examination photography into the electronic record

BACKGROUND

Photography of the airway has been used in research to validate preoperative airway assessment and the likelihood of identifying the difficult-to-mask ventilate and/or intubate patient. Up till now, no study has demonstrated the perceived utility of incorporation of airway photographs into the anesthesia preassessment.

METHODS

The University of Florida Health Presurgical Clinic routinely incorporates three photographs of all adult patients during their preanesthesia visit. The first is a head-on view of the patient opening the mouth widely as part of a Mallampati examination, and the second and third are side views of the patient prognathing and with the neck in maximal extension, respectively. After IRB approval, providers of anesthesia were surveyed regarding their opinions on the perceived value of the new process. Chi-square tests were used to determine if the responses to each question significantly differed from the distribution that would be predicted by chance. P < 0.05 was considered statistically significant.

RESULTS

The survey was emailed to 180 individuals, with 145 responding. The responses significantly (P < 0.0001) indicated that the photographs helped the providers plan care for their patients and improved their satisfaction with the preoperative assessment. Technical and educational barriers were overcome using iterative Plan-Do-Study-Act cycles and coaching, respectively.

CONCLUSIONS

Photographs of the airway assessment can successfully be taken and incorporated into an electronic medical record in a busy presurgical clinic. The pictures provide additional perceived value to the traditional written assessment of a patient's airway examination by someone else.

The effects of dynamic airway photographs on preoperative airway planning among a panel of anesthetists

STUDY OBJECTIVE

To determine whether having preoperative airway photographs will change the preanesthetic airway plan.

DESIGN

Questionnaire.

SETTING

American academic medical center (Brigham and Women's Hospital, Boston MA).

SUBJECTS

Twenty-five test subjects (American Society of Anesthesiologists 1-4) were enrolled to have their preoperative airway photographs taken as well as to have a customary preoperative history and physical examination. In addition, 15 anesthetists were enrolled to review the subjects' preoperative history, physical examination, and preoperative airway photographs.

MEASUREMENTS

All 15 anesthetists were asked to fill out a survey for airway management for each test subject.

MAIN RESULTS

All 15 anesthetists completed the survey. Across all providers, plans were changed a median of 24% (95% confidence interval [CI], 12.7-38.6). Among attending anesthesiologists, airway management plans were changed 30% of the time (95% CI, 12.4-40.0), whereas among nonattending level providers, plans changed 24% of the time (95% CI, 12.0-38.8). χ² Tests found no difference between the percent change of airway plans between attending and nonattending level providers (P=.306).

CONCLUSIONS

Our findings suggest that the addition of dynamic airway photographs to preoperative airway reports affects airway management plans among a variety of anesthesia care providers. In general, dynamic airway photographs can aid preoperative airway management planning.

The diagnostic validity of clinical airway assessments for predicting difficult laryngoscopy using a grey zone approach

Objectives

The diagnostic validity of clinical airway assessment tests for predicting difficult laryngoscopy in patients requiring endotracheal intubation were evaluated using receiver operating characteristic (ROC) curve analysis and a grey zone approach.

Methods

In this prospective observational study, patients were evaluated during a pre-anaesthetic visit. Predictive airway assessment tests (i.e. Modified Mallampati [MMT] classification; upper lip bite test [ULBT]; mouth opening; sternomental distance; thyromental distance [TMD]; neck circumference; neck mobility; height to thyromental distance [HT/TMD]; neck circumference-to-thyromental distance [NC/TMD]) were performed on each patient and LEMON, Naguib, and MACOCHA scores were also calculated. In addition, laryngeal images were acquired and assessed for percentage of glottic opening (POGO) scores. A POGO score of zero was categorized as difficult laryngoscopy.

Results

The incidence of difficult laryngoscopy was 14.4% (35/243). Although seven predictive airway assessments (i.e. MMT classification, ULBT, mouth opening, HT/TMD, NC/TMD, and the LEMON and Naguib models) predicted difficult laryngoscopy by ROC analyses, a grey zone approach showed that the parameters were inconclusive in approximately 70% of patients. From all the tests, the HT/TMD ratio showed the highest sensitivity (80.0%) and ULBT had the highest specificity (95.2%).

Conclusion

Using the grey zone approach, all predictive airway assessment tests showed large inconclusive zones which may explain previous inconsistent results in the prediction of difficult laryngoscopy. Our results suggest that the usefulness of clinical airway evaluation tests for predicting difficult laryngoscopy remains controversial.

Clinical trial registration

ClinicalTrials.gov (NCT01719848)

[Post analysis simulated correlation of the El-Ganzouri airway difficulty score with difficult airway]

BACKGROUND

Difficult airway (DA) occurs frequently (5-15%) in clinical practice. The El-Ganzouri Risk Index (EGRI) has a high sensitivity for predicting a difficult intubation (DI). However difficult mask ventilation (DMV) was never included in the EGRI. Since DMV was not included in the EGRI assessment, and obstructive sleep apnea (OSA) is also correlated with DMV, a study correlating the prediction of DA and OSA (identified by STOP-Bang questionnaire, SB) seemed important.

METHODS

We accessed a database previously collected for a post analysis simulation of the airway difficulty predictivity of the EGRI, associated with normal and difficult airway, particularly DMV. As secondary aim, we measured the correlation between the SB prediction system and DA, compared to the EGRI.

RESULTS

A total of 2747 patients were included in the study. The proportion of patients with DI was 14.7% (95% CI 13.4-16) and the proportion of patients with DMV was 3.42% (95% CI 2.7-4.1). The incidence of DMV combined with DI was (2.3%). The optimal cutoff value of EGRI was 3. EGRI registered also an higher ability to predict DMV (AUC=0.76 (95% CI 0.71-0.81)). Adding the SB variables in the logistic model, the AUC increases with the inclusion of "observed apnea" variable (0.83 vs. 0.81, p=0.03). The area under the ROC curve for the patients with DI and DMV was 0.77 (95% CI 0.72-0.83).

CONCLUSIONS

This study confirms that the incidence of DA is not negligible and suggests the use of the EGRI as simple bedside predictive score to improve patient safety.

Post analysis simulated correlation of the El-Ganzouri airway difficulty score with difficult airway

BACKGROUND

Difficult airway (DA) occurs frequently (5-15%) in clinical practice. The El-Ganzouri Risk Index (EGRI) has a high sensitivity for predicting a difficult intubation (DI). However difficult mask ventilation (DMV) was never included in the EGRI. Since DMV was not included in the EGRI assessment, and obstructive sleep apnea (OSA) is also correlated with DMV, a study correlating the prediction of DA and OSA (identified by STOP-Bang questionnaire, SB) seemed important.

METHODS

We accessed a database previously collected for a post analysis simulation of the airway difficulty predictivity of the EGRI, associated with normal and difficult airway, particularly DMV. As secondary aim, we measured the correlation between the SB prediction system and DA, compared to the EGRI.

RESULTS

A total of 2747 patients were included in the study. The proportion of patients with DI was 14.7% (95% CI 13.4-16) and the proportion of patients with DMV was 3.42% (95% CI 2.7-4.1). The incidence of DMV combined with DI was (2.3%). The optimal cutoff value of EGRI was 3. EGRI registered also an higher ability to predict DMV (AUC=0.76 (95% CI 0.71-0.81)). Adding the SB variables in the logistic model, the AUC increases with the inclusion of "observed apnea" variable (0.83 vs. 0.81, p=0.03). The area under the ROC curve for the patients with DI and DMV was 0.77 (95% CI 0.72-0.83).

CONCLUSIONS

This study confirms that the incidence of DA is not negligible and suggests the use of the EGRI as simple bedside predictive score to improve patient safety.