Comparative Study of Preoperative Airway Assessment by Conventional Clinical Predictors and Ultrasound-Assisted Predictors.

AIMS: The aim of this study is to evaluate the effectiveness of airway sonographic parameters as the predictors of difficult laryngoscopy and to evaluate the validity of combined sonographic and clinical tests.
MATERIALS AND METHODS: This prospective and observational study analyzed a sample of 200 patients who were categorized as having easy (Grades 1 and 2) or difficult (Grades 3 and 4) laryngoscopy based on the laryngoscopic criteria of Cormack-Lahane (CL). Sonographic parameters, including the anterior neck soft-tissue thickness at the level of vocal cord (ANS-VC), ANS tissue thickness at the level of hyoid, and ratio of depth of pre-epiglottic space to distance from epiglottis to midpoint of the distance between vocal cords and clinical parameters, including modified Mallampati class, thyromental distance, and hyomental distance ratio (HMDR), were analyzed. Univariate and multivariate regression analysis was used for the statistical analysis.
RESULTS: Twenty patients (10%) were categorized as having difficult laryngoscopy. Statistically significant differences between patients with difficult and easy laryngoscopy were noted for 5 of 6 parameters. The diagnostic validity profiles showed variable sensitivity (26.5%-87.5%) and good specificity (58.9%-94.2%) and negative predictive value (88.8%-97.03%). ANS-VC has the highest sensitivity (87.50%) and area under curve value (0.887), whereas HMDR showed highest specificity (94.2%) and highest accuracy (89.60%) which means it has low false-positive prediction rate. The combination of tests improved the diagnostic validity profile (highest area under the curve, 0.897).
CONCLUSION: Sonographic predictors can help in identifying difficult laryngoscopy. Combined clinical and ultrasonographic parameters showed better validity profiles in comparison to the individual tests. Copyright:

INTRODUCTION

Endotracheal intubation is one of the most important skills for anesthesiologists in securing the airway during general anesthesia and resuscitation. Difficulties in handling airways optimally can lead to severe adverse effects and failure can also lead to mortality. Many clinical screening tests have been used to classify patients at risk of difficult laryngoscopy, but none have been shown to predict this complication with adequate sensitivity and specificity.[1]

Airway sonography is a noninvasive diagnostic device for bedside visualization and measurement of dimensions that may be linked to the prediction of difficult airway management.[234] Researchers had speculated that increased anterior soft-tissue thickness of neck could impede the forward mobility of the pharyngeal structures and that an increase in the preepiglottic space (PES) or a decrease distance from the epiglottis to the vocal cords may be correlated with increasingly difficult laryngoscopy and intubation.[56]

There is minimal literature available that compares the parameters of ultrasound (US) to the Cormack-Lehane (CL) grade and physical parameters. Hence, the present study was undertaken to find out the usefulness of ultrasonographic airway assessment in predicting difficult laryngoscopy. Physical parameters: Modified Mallampati (MMP) class, thyromental distance (TMD) and hyomental distance ratio (HMDR), and US parameters anterior neck soft-tissue thickness at the level of hyoid (ANS-hyoid), ANS tissue thickness at the level of vocal cord (ANS-VC), and ratio of depth of preepiglottis space to distance from epiglottis to midpoint of the distance between vocal cords (Pre-E/E-VC) were measured and sensitivity, specificity, and accuracy of these test were calculated and compared.

MATERIALS AND METHODS

This study was conducted in the anesthesia department of a tertiary care center during January 2018 to June 2019 after approval from the Institutional Ethical Committee (RIMSandR/DEAN/2018-19/E.C. NO: 16 dated June 4, 2018). It was registered prospectively with Clinical Trial Registry of India (www.ctri.nic.in) and registration no for this trial is CTRI/2018/06/014467. Two hundred consecutive patients older than 18 years undergoing elective surgery under general anesthesia with endotracheal intubation were enrolled in this study after taking written informed consent. The patients who require rapid sequence intubation, with cervical spine pathology, scheduled for fiber-optic tracheal intubation, uncooperative patients, and pregnant patients were excluded from the study. During general anesthesia, patients were graded as Group 1 = Easy Laryngoscopy (CL Grade 1 and 2) or Group 2 = Difficult laryngoscopy (CL Grade 3 and 4) with endotracheal intubation based on CL laryngoscopic criteria, as depicted in Figure 1.[7]

Figure 1

Modified Mallampati class

Before data collection, all the investigators were trained in the proper technique of performing difficult airway clinical and US-guided screening tests. All measurements were obtained by primary investigator who was not involved in tracheal intubation. Airway assessment of patients was performed in two phases. MMP was measured with the patient in an upright sitting position, head in a neutral position, mouth wide open, and tongue protruding to its maximum without phonation [Figure 2].

Figure 2

Cormack Lehane grading

Class 0 = Ability to see any part of the epiglottis upon mouth opening and tongue protrusion

Class I = Soft palate, anterior and posterior pillars, and uvula visualized

Class II = Soft palate and uvula visualized

Class III = Soft palate and base of the uvula visualized

Class IV = Only hard palate visualized.

MMP III or IV were considered as the predictors of difficult laryngoscopy.[8] Then, in the supine position, with the head on a firm-operating table. The patients were told to look straight ahead, keeping the head neutral, closing mouth, and not swallowing. The distance from the hyoid to the mentum was measured with a measuring tape, and this length was Hyomental distance (HMDn) in the neutral position. HMD was again measured in full extension, without lifting the shoulders and this variable was defined as the HMDe at the extreme of head extension. The HMDR was determined as the ratio of the HMDe/HMDn, i.e., at extreme of head extension to that in the neutral position. HMDR ≥ 1.2 was taken cutoff value for difficult laryngoscopy.[9] In the same position, the straight distance from the anterior-most part of the mentum to the thyroid notch was measured and defined as the TMD at the extreme of head extension. A value of TMD <6.5 cm was considered as difficult laryngoscopy.

Airway assessment using ultrasonic parameters was done using a Sonosite Fujifilm ultrasound system with HFL38/13-6 MHZ transducer. The ANS tissue thickness was measured with the patient in the sniffing position, as the distance from the skin to the anterior part of the trachea at the hyoid level (ANS-hyoid)[5] Hyoid bone appears as a hyperechoic line with posterior acoustic shadow. ANS-VC was measured as the distance from the skin to the anterior commissure of true vocal cords. A cutoff value of ANS-VC > 0.23 was taken as difficult laryngoscopy.[10] True vocal cords appear as triangular hypoechoic structure with hyperechoic vocal ligaments at their medial border. Anterior commissure is the point where true vocal cords meet anteriorly. The high-frequency linear probe was mounted in midline at the submandibular region. The linear array of the US probe was rotated from cephalad to caudad in the transverse plane without adjusting the direction of the probe, until the epiglottis and posterior part of vocal folds with arytenoids were visualized simultaneously on the screen. The epiglottis appears as a hypoechoic curvilinear structure through thyrohyoid membrane in the oblique-transverse view. The anterior boundary was demarcated by a vivid linear mucosal air interface with the hyperechoic PES and its posterior boundary. The Pre-E and the E-VC both were measured, and the ratio calculated [Figure 3]. Ratio of pre-E/E-VC 2-3 was considered as difficult laryngoscopy.[7]

Figure 3

Preepiglottic space (red arrow) and distance between epiglottis to vocal cords (yellow arrow)

The patients were induced with an intravenous anesthetic injection propofol 2 mg.kg −1 and a nondepolarizing muscle relaxant injection vecuronium 0.1 mg.kg −1. They were ventilated with 100% oxygen for 3 min. Laryngoscopy and intubation were done after full relaxation by a senior anesthesiologist with 10 years' experience who was blinded to the findings of preoperative airway assessment. Direct laryngoscopy was performed using appropriately sized Macintosh blade, and the CL grade was noted. Intubation was graded as easy (CL Grade 1 and 2) or difficult (CL Grade 3 and 4). Appropriately sized endotracheal tube (ETT) was inserted, and anesthesia was maintained. The number of intubation attempts, the need for alternative difficult intubation methods, and invasive access to airway or cancellation of the procedure due to inability to intubate was also noted.

A total sample size of 200 was arrived at for a power of 80% to detect a change in sensitivity from 0.645 to 0.75 using a two-sided binomial test and 97% power to detect a change in specificity of MMP from 0.824 to 0.95 using a two-sided binomial test. The sensitivity and specificity values for calculation were based on a previous study by Adamus et al.[11] The target significance level is 0.05. The MS Excel and Epi Info software packages were used for the data entry and analysis. For the comparison of continuous variables between the groups, the Student's t-test was used, and the results were presented as mean ± standard deviation. The comparison of categorical variables between the groups was assessed by the Chi-square test, and the results were reported as absolute numbers and percentages. Univariate logistic regression was used to assess the accuracy, sensitivity, specificity, positive and negative predictive values, and accuracy of the collected data. Multivariate logistic regression analysis was used to evaluate the validity of different combinations of sonographic and clinical parameters of difficult intubation under the study. The predictive ability of the final score was assessed by measuring the area under the receiver operating characteristic curve [Figures 4 and 5].

Figure 4

Receiver operating characteristic curve for ultrasound measurements Difficult laryngoscopy was taken as Cormack-Lehane Grade >2. Yellow line was taken as the reference line and various values of area under curve are given in Table 2

Figure 5

Receiver operating characteristic curve for bedside clinical measurements. Difficult laryngoscopy was taken as Cormack-Lehane Grade > 2. Yellow line was taken as the reference line and various values of area under curve are given in Table 2

Table 2

Diagnostic validity of ultrasonography and physical parameters predicting difficult laryngoscopy

Parameters	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	Accuracy (%)	AUC
ANS-VC >0.23	87.5	58.9	23.48	97.03	62.50	0.887
ANS-hyoid	45.3	62.5	14.83	88.80	60.33	0.515
Ratio of (PES/E VC)	58.3	75.6	23.9	91.8	72.8	0.660
MMP ≥3	74.1	85.4	42.25	95.81	83.98	0.763
TMD <6.5	26.5	91.2	30.27	89.59	83.05	0.580
HMDR	57.9	94.2	59.0	93.9	89.6	0.691

PPV=Positive predictive value, NPV=Negative predictive value, AUC=Area under curve, ANS-VC=Anterior neck soft tissue thickness at vocal cord level, ANS hyoid=Anterior neck soft tissue thickness at hyoid level, PES=Depth of preepiglottic space, EVC=Epiglottis to vocal cord distance, MMP=Modified Mallampati, TMD=Thyromental distance, HMDR=Hyomental distance ratio

RESULTS

The incidence of difficult laryngoscopy was found to be 10%. All patients of Group 2 were finally intubated successfully either with fiber-optic scope or video laryngoscope. Demographic variables such as age and sex were comparable between the two groups [Table 1]. Body mass index in Group 2 patients was higher than Group 1, and the difference was statistically significant (P = 0.010). In the univariate analysis, statistically significant difference was noted in MMP of both the groups having low MMP in patients with difficult intubation (P = 0.0051) [Table 1]. TMD and HDMR were found significantly low in patients with difficult intubation (P = 0.0004 and 0.002, respectively).

Table 1

Characteristics of patients and clinical screening tests with easy and difficult laryngoscopy

Parameters	Group 1 (n=180) Easy laryngoscopy	Group 2 (n=20) Difficult laryngoscopy	P
Age (year)	38.66±6.49	41.35±8.36	0.080
Sex male/female	97/83	10/10	0.810
BMI (kg.m−2)	23.61±3.43	25.63±2.80	0.010
ASA grade	2.49±0.50	2.50±0.51	0.963
MMP
1	68 ( 37.7)	3 ( 15)	0.0051
2	80 ( 44.4)	8 ( 40.0)
3	28 ( 15.5)	7 ( 35)
4	3 ( 1.66)	2 ( 10)
TMD (cm)	6.86±0.25	6.52±0.52	0.0004
HMDR	1.12±0.07	1.07±0.08	0.002
ANS-VC (cm)	0.221±0.008	0.259±0.013	0.0001
ANS-HYOID	0.578±0.012	0.581±0.013	0.195
Ratio of PES/EVC	1.244±0.033	1.273±0.093	0.00073

BMI=Body mass index, ASA=American Society of Anaesthesiologist, MMP=Modified Mallampati, TMD=Thyromental distance, HMDR=Hyomental distance ratio, ANS-VC=Anterior neck soft tissue thickness at vocal cord level, ANS hyoid=Anterior neck soft tissue thickness at hyoid level, PES=Depth of preepiglottic space, EVC=Epiglottis to vocal cord distance

Patients with easy laryngoscopy had mean ANS-VC of 0.221 ± 0.008, while those having difficult intubation had mean ANS-VC of 0.259 ± 0.013, Thus, ANS-VC was significantly higher in the group of patients having difficult laryngoscopy (P = 0.0001) [Table 1]. On the other hand, there was no significant difference in the values of ANS hyoid between the two groups. Ratio of PES/epiglottis to vocal cord (EVC) was significantly higher in patients with difficult intubation (P = 0.00073) [Table 1].

Validity profiles [Table 2] showed high sensitivity and accuracy of ANS-VC. It has the highest sensitivity (87.50%) and area under the curve (AUC) value (0.887) among all other predictors of difficult airway under the study. HMDR showed highest specificity (94.2%) and highest accuracy (89.60%) which means it has low false-positive prediction rate, and thus, this test reduces the risk of subjecting many patients to unnecessary advance intubation methods. Although, HMDR is not a very good screening tool due to its average sensitivity. MMP ≥3 is the second-most sensitive tool after ANS-VC, with sensitivity of 74.1 5 and AUC 0.763. ANS hyoid is not a useful test for the predictions of difficult laryngoscopy because of its low sensitivity and accuracy. Similarly, TMD also has low sensitivity, but it has good accuracy comparable to HMDR and MMP ≥3. Ratio of PES/EVC is also useful test in the prediction of difficult laryngoscopy, but it has low sensitivity (58.3) in comparison to ANS VC and MMP ≥3.

Table 3 shows the multivariate logistic regression results of combined tests encompassing the mixed sonographic and clinical parameters. Combinations of tests had definite improved diagnostic value in comparison to individual tests. Highest sensitivity and highest AUC on ROC analysis were seen in the combination of ANS-VC, ratio of PES/EVC, HMDR, MMP, and TMD followed by combination of all of the above excluding ANS-VC.

Table 3

Diagnostic validity profiles of sonographic and clinical tests combination for predicting difficult laryngoscopy

Combination	AUC	Sensitivity	Specificity
ANS VC + MMP	0.834	85.6	90.1
ANS VC + TMD	0.725	82.5	45.5
ANS VC + HMDR + MMP + TMD	0.861	92.3	78.3
Ratio (PES/E VC) + MMP	0.812	80.2	55.2
Ratio (PES/E VC) + HMDR + MMP + TMD	0.862	78.4	81.3
ANS VC + Ratio (PES/E VC) + HMDR + MMP + TMD	0.897	93.2	86.9

ANS-VC=Anterior neck soft tissue thickness at vocal cord level, MMP=Modified Mallampati, TMD=Thyromental distance, HMDR=Hyomental distance ratio, PES=Depth of preepiglottic space, EVC=Epiglottis to vocal cord distance, AUC=Area under curve

DISCUSSION

Despite recent advances in the field of anesthesiology, unpredicted difficult airway remains a great challenge to medical fraternity. Sonographic assessment of upper airway is an emerging tool for the prediction of difficult airway, particularly when anatomic landmarks are difficult to palpate, emergency situations and unconscious patients, although there is a paucity of literature in this regard. Our study demonstrates that sonographic parameters of upper airway may be of great help in the prediction of difficult airway. ANS-VC, ANS hyoid, Ratio of PES, and EVC in cases of easy and difficult laryngoscopy were studied and compared with clinical predictors. The combination of sonographic and clinical parameters showed improved values for the identification of unanticipated difficult laryngoscopy. Strong positive linear correlation was observed between ANS-VC and difficult airway. The AUC for ANS-VC was 0.887, indicating that it is an excellent parameter for the prediction of difficult laryngoscopy with very high sensitivity. The results were supported by a similar study conducted by Reddy et al. concluding that the ANC-VC is a sensitive tool for the screening of difficult airway, and its value of more than 0.23 cm was associated with difficult airway.[10] HMDR has recently shown to be a descent predictor of decreased extension capacity of occipitoatlantoaxial complex and a helpful parameter for the preoperative assessment of airway.[12131415] A smaller hyomental distance results in lesser space for the tongue to be displaced during laryngoscopy leading to more difficult and challenging intubation. Wojtczak studied the role of US-guided HMDR in the prediction of difficult airway and found that it is a good predictor of CL grading.[12] They found that the mean HMDR in six patients who had a history of difficult intubation was 1.02 ± 0.01, and the ratio in six patients whose were easy to intubate was 1.14 ± 0.02 (P = 0.002). They noted that the ultrasound-guided HMDR in the difficult airway group was in the range of 1–1.05, and those in the easy airway group were in the 1.12–1.16.[12] Huh et al. evaluated the value of surface HMDR measurements for the prediction of difficult intubation in 213 adult patients undergoing surgery under general anesthesia with endotracheal intubation.[9] They observed that the HMDR alone had the highest predictive validity for difficult laryngoscopy with an optimal cutoff point of 1.2 with sensitivity of 88% and specificity of 60% for predicting difficult airway. Mean value of HMDR in easy laryngoscopy group was 1.12 ± 0.07 and in difficult laryngoscopy was 1.07 ± 0.08 with P = 0.002. Thus, lesser values of HMDR were related to difficult laryngoscopy. HMDR showed highest specificity (94.2%) and highest accuracy (89.60%) which indicates low false-positive prediction rate. Although HMDR is not a very good screening tool due to its low sensitivity.

In the present study, it was observed that ANS-hyoid is not a good predictor of difficult airway. Difference in the mean values of ANS-Hyoid among easy and difficult laryngoscopy patient groups was statistically insignificant with P = 0.195. The results in this regard are in agreement with the findings of the similar study by Reddy et al. who also concluded that ANS-hyoid did not prove to be significant predictor of difficult intubation.[10]

Gupta et al. performed preanesthetic sonographic assessment of airway and correlated the findings with CL grading.[7] They observed that PES was positively correlated with CL grades while E-VC had negative correlation with the same. However, they noted that ratio of PES/EVC had strong significant positive correlation with CL grading. Contrary to above study, Soltani Mohammadi et al. concluded a weak correlation between PES/E-VC and CL grade.[16] The results of the present study in terms of ratio of PES/E-VC are similar to that of Gupta et al. Statistically significant difference in PES/EV-C ratio was observed between group of patients with easy and difficult laryngoscopy having positive correlation with later group. Differences in the findings among various studies can be explained by anatomical variations in the study population, racial difference, and different sample sizes.

Although MMP class is one of the most commonly used clinical tools to predict a difficult laryngoscopy, it mainly depends on patient compliance and position of body, which can significantly affect the accuracy of difficult airway prediction. Therefore, despite its good predictive value, MMP has been criticized to be of limited value by many studies.[1617] A recent study including 177,088 patients observed that MMP is inadequate as a single test to predict difficult tracheal intubation, but it may well be a part of a combined model of clinical and sonographic tests for the prediction of a difficult airway.[2] It was observed in this study that MMP ≥3 is the second most sensitive tool after ANS-VC with sensitivity of 74.1 5 and AUC 0.763 with statistically significant difference between the two groups. It was also observed that when MMP is combined with other sonographic tests, it improves the validity profile combined model in comparison to individual tests.

Short TMD is one of the clinical parameters used for the prediction of difficult laryngoscopy, although there is debate regarding it's cutoff value in the literature. Tripathi and Pandey advocated that the normal value of the TMD in adults is ≥6.5 cm.[18] If the TMD is 6.0–6.5 cm without other anatomical abnormalities, laryngoscopy and intubation are difficult but usually possible. A TMD of <6 cm indicates that laryngoscopy may not be possible. In similar study, Reddy et al. observed that TMD was not a good predictor of difficult intubation.[10] In the present study, mean TMD in easy laryngoscopy group was 6.86 ± 0.25 and difficult laryngoscopy group was 6.52 ± 0.52, and there was significant between the two groups (P = 0.0004). It was also observed that TMD has good accuracy and specificity but poor sensitivity in the prediction of difficult intubation.

Validity profiles of combined tests were improved in comparison to individual sonographic tests. Stepwise multivariate logistic regression analysis was used to attain diagnostic validity profile of various test combinations, and we found that the combination of ANS-VC, ratio of PES/EVC, HMDR, MMP, and TMD were the most valuable predictor with the highest AUC followed by the combination of ratio of PES/EVC, HMDR, MMP, and TMD. Thus, the sonographic assessment of airway has practical utility in the prediction of difficult intubation. The combination of sonographic and clinical tests even showed improved the validity profiles in difficult airway prediction. Thus, the use of preoperative sonographic assessment of airway is helpful to limit the number of airway intervention attempts.

Limitations

The operator of ultrasound machine has to master the technique of sonographic anatomy of airway to delineate acoustic artefacts from that of the structure of interest and interpretation is operator-dependent. We do assume some interindividual variability in the laryngoscopic grading.

CONCLUSION

Sonographic upper airway evaluation may help to identify patients having difficult airway. Sonographic parameter ANS-VC is a good predictor of difficult laryngoscopy. ANS-VC is a better predictor for difficult airway than clinical parameters such as MMP and TMD. On the other hand, HMDR has high specificity and accuracy which can reduces the risk of subjecting many patients to unnecessary advance intubation techniques. Combined models of sonographic and physical tests further improve diagnostic value to identify the cases of difficult intubation.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.