Evaluation of different airway tests to determine dicult intubation in apparently normal adult patients: A prospective observational study

We conducted this study to determine which airway test, or combination of airway tests, most clinically useful. One hundred sixty adult patients underwent preoperative assessments of the mandible protrusion test (MPT), thyromental (TMD) and sternomental (SMD) distances, inter-incisor gap (IIG), and the modied Mallampati tests with tongue protrusion (MMT-TP) and without tongue protrusion (MMT-NTP). Grade C on the MPT, TMD ≤ 6 cm, SMD ≤ 12 cm, and MMT grades III and IV were considered to be predictors of dicult tracheal intubations. A modied Cormack-Lehane grading (MCLG) of laryngoscopic views with backward, upward, and right-sided pressure on the thyroid and cricoid cartilages (BURP) maneuver was also documented, with grades 2B, 3, and 4 considered to be dicult airways for intubation. positive predictive value, and good sensitivity may be used as a routine screening test for preoperative prediction of dicult tracheal intubations.

Unanticipated di cult intubations remain a major concern for anesthesiologists due to the potentially serious consequences of failed tracheal intubations [4]. The identi cation of patients with di cult airways is crucial during preoperative evaluations [5]. A variety of tests are used to evaluate for a potentially di cult intubation in advance of the procedure [6,7]. It is not clear; however, which test has the best predictive ability.
Therefore, we conducted this prospective study to evaluate the accuracies of the mandibular protrusion test (MPT), thyromental distance (TMD), sternomental distance (SMD), inter-incisor gap (IIG), and the modi ed Mallampati test (MMP) for prediction of di cult intubations relative to the modi ed Cormack-Lehane grading (MCLG) with backward, upward, and right-sided pressure on the thyroid and cricoid cartilages (BURP) maneuver for di cult laryngoscopies. The main goal of the study was to determine which airway assessment test and/or combination of tests was best at predicting di cult intubations.

Materials And Methods
After the institutional research Ethics Committee approval of this observational, prospective study (20190210), we obtained written informed consent from all patients.
We studied 160 patients with American Society of Anesthesia physical statuses (ASA) of I to III who required tracheal intubation for elective surgeries. Patients were excluded from the study if they met any of the following criteria: 1) age < 18 years; 2) pregnancy; and patients scheduled for cesarean section; 3) increased risk of pulmonary aspiration; 4) body mass index of 35 kg/m 2 or greater; or 5) inability to communicate (e.g. confusion, poor hearing, or language barrier).
Patients were premedicated with 5 mg diazepam orally on the evening before surgery.
Upon arrival at the anesthetic room, all patients received an intravenous catheter. Routine monitoring included electrocardiography (ECG), pulse oximetry, noninvasive blood pressure, and end-expiratory gas analysis. After preoxygenation, anesthesia was induced in all patients with 2 µg/kg of fentanyl and 2 to 3 mg/kg of propofol. Neuromuscular blockade was achieved with 0.5 mg/kg of atracurium besilate. Patient lungs were hand-ventilated via facemask with 1% sevo urane and 100% oxygen till the neuromuscular block was completed.
Direct laryngoscopy was performed, and a MCLG of laryngoscopic views with the BURP maneuver was documented. This 5-grade scoring system involves the subdivision of the original grade 2 into 2A (partial view of glottis is visible) and 2B (only the arytenoids are visible) [7,8]. Grades 2B, 3, and 4 were considered to be di cult intubations.
Controlled ventilation through an endotracheal tube was maintained with 40% O2 in air and 1 to 1.2 minimum alveolar concentration of sevo urane. Mechanical ventilation was set to maintain an end-tidal carbon dioxide (CO2) between 32 and 40 mmHg. At the end of surgery, residual neuromuscular blockade was reversed with neostigmine and atropine. The sevo urane was discontinued 3 to 5 minutes before completion of the surgical procedure.

Measurements
One day prior to surgery, measurements were obtained by anesthesiologists not involved in tracheal intubations of the study participants. Data was documented in an allocated data sheet.
We documented the grades of the MMT, according to the Samsoon and Young [9] airway classi cation. This measurement was performed in a sitting posture with a neutral head position and the tongue maximally protruded from the mouth without phonation (MMT-TP). Results were categorized into four classes, in which class III (only the soft palate could be seen) and class IV (the soft palate was not visible) were considered to be predictors for di cult tracheal intubations. Patients with class I (soft palate, fauces, uvula, and pillars could be seen) and class II (soft palate, fauces, and uvula could be seen) were predicted to have easier intubations.
The MPT was assessed based on the classi cation system of UlHaq et al. [10], and is described as follows: class A = lower incisors can be protruded anterior to the upper incisors; class B = lower incisors can be brought edge-to-edge with the upper incisors; and class C = lower incisors cannot be brought edge-to-edge with the upper incisors. Class C was considered to be predictive of a di cult tracheal intubation.
The thyromental distance (TMD) was measured from the tip of thyroid cartilage to the tip of mentum. A di cult intubation was predicted in patients with a thyromental distance of 6 cm or less. The sternomental distance (SMD) was measured from the sternal notch to the tip of insight of the mentum. TMD and SMD were measured with the neck fully extended and the mouth closed, using a ruler approximated to the nearest 0.5 cm [ Figure 1]. TMDs of ≤ 6 cm and SMDs of ≤ 12 cm were considered to be predictors for di cult visualizations of the larynx and di cult tracheal intubations.
On the day of surgery, the inter-incisor gap (IIG) and the modi ed Mallampati test without phonation (MMT-NTP) were performed in the supine position and documented. Measurements of IIG were obtained at the midline between the upper and lower incisors while the patient was in the supine position, with maximum mouth opening and a neutral head position [11]. The MMT-NTP was performed in the supine posture with a neutral head position, without tongue protrusion and without phonation. Its classi cation schema was the same as that of the MMT-TP.

Statistical analysis
The sample size was calculated with a precision error of 5% and type I error of 5%. We assumed an incidence of 11% for di cult laryngoscopies, based on a previously published study [12]. According to Eq. 1, the desired number of patients was 151. In anticipation of losses, and for more adequate control of potential confounding effects of variables, we enrolled 160 patients in this study. n = (Z1-α/2) 2 *P (1-P)/E 2 (1) n = number of patients in the sample. Z1-α/2 = 1.96. P = expected proportion in population based on previous study. E = precision error of 5%.
Continuous demographic data and continuous predictors of di cult intubation were presented as means ± SD. Number of patients was analyzed with a t-test. Pearson chi-square tests were used for categorical variables. Logistic regression analysis was performed to determine the predictors for di cult intubation in patients. Data analyses were performed using the Statistical Package for Social Sciences version 18 (SPSS Inc., USA). Statistical signi cance was considered as a P-value of 0.05 or less.

Results
A total of 160 patients who were scheduled for elective surgical procedures requiring general anesthesia with endotracheal intubation were enrolled in this study. Patient characteristics are shown in Table 1.  Fifteen patients (9.38%) were found to have airways that were di cult to intubate during laryngoscopy. This incidence of di cult intubations represents the sum of the true-positive (TP) and false-negative (FN) cases.
This study had no occurrences of failed intubations. The tracheas of 11 patients were intubated using a standard endotracheal tube introducer (a so-called gum elastic bougie). Fiberoptic intubation was necessary in the remaining four patients.
MMT-NTP had the highest sensitivity (60%) and the lowest positive predictive value (PPV) (21.95%) and speci city (77.93%). The sensitivity and the Youden's index of MMT-TP were found to be the lowest (40% and 0.29, respectively).
The MPT was the most accurate and speci c test (90.63% and 95.17%, respectively). This test also had the highest PPV  Table 3. Receiver operating characteristic curves (ROC) and AUC were used to identify the predictive abilities of the clinical tests (Fig. 2). The highest AUC was for MPT and the lowest AUC was for SMD (0.781 and 0.310, respectively) ( Table 3).
Using the t-test, continuous variables, including weight, height, and BMI, were not signi cantly associated with the MCLG with BURP (P-values of 0.674, 0.387, and 0.263, respectively). Patient age, however, was signi cantly associated with the MCLG with BURP (P = 0.028). Associations between different airway tests and the di culty of intubations obtained by bivariant analysis for preoperative variables are shown in Table 4. The combination of various airway assessment tests are shown in Table 5. The categorical variables of sex and ASA were not strongly associated with the di culty of tracheal intubations.

Discussion
Unanticipated di cult endotracheal intubations are the most common cause of anesthesia-related morbidity and mortality [13,14], and are a major source of concern for anesthesiologists. As a result, it is important to identify a clinical test that is quick and easy to perform during a preoperative evaluation in order to accurately predict potentially di cult tracheal intubations with high sensitivity and speci city [15].
In a study by Prakash and Ravi, no test could be identi ed that reliably predicted the majority of di cult intubations with a low false-positive rate [16]. The incidence of di cult intubations in the present study was identi ed to be 9.38%. In accordance with our results, Iohom et al. [17] reported an incidence of di cult intubations of 9%. Domi [18] encountered a di cult tracheal intubation in 40 out of 426 patients (9.38%). The incidence of di cult intubations varied in other studies from between 3.4 to 23% [19,20]. Differences in reported incidences may have been due to the diversity of de nitions for di cult intubations [2,7] or differences in anatomical structures of the patients [4,21]. The amount of clinical experience of the anesthetists who are performing the tracheal intubations may also have played an important role in previous assessments of the di culty of a tracheal intubation.
The incidence of di cult laryngoscopies may be improved by use of the BURP maneuver. Even in pediatric patients and with usage of a glidescope, Hirabayashi et al. [22] found that the BURP maneuver provided better glottis views. In contrast, Lee et al. [23] used the Clarus Video System and found that the BURP maneuver actually worsened the laryngeal view compared with the conventional maneuver. They also found that the MCLG was improved with the modi ed jaw thrust maneuver compared with the conventional maneuver.
For predicting di cult intubations, the MPT is a well-established and relatively simple grading system [10]. Savva [24] reported that protrusion of the mandible was too insensitive for routine use, with a sensitivity of 29.4%, speci city of 85%, and PPV of 9.1%. In that study, no patients were classi ed as grade C.
On the other hand, Yildiz et al. [25] found the incidences of di cult intubations in patients with mandibular protrusion grades of B or C were signi cantly lower than in patients with MMT scores of III or IV, with a lower sensitivity than observed in our study (31% vs. 46.67%, respectively). That study identi ed a higher number of patients with grade C than we did (32 vs. 14, respectively).
The reported speci city and accuracy, however, were similar to the values identi ed in our study. The differences in the reported ndings may have been attributable to inter-observer variability, inability of some patients to protrude the lower incisors anterior to the upper incisors, the diversity of de nitions of di cult intubation, and the use of different patient populations [24,26,27].
Previous studies reported various cut-off points for TMD that could predict a di cult airway for intubation. Honarmand et al. [28] reported a TMD of ≤ 7.1 cm as a cut-off value for a di cult intubation. Badheka et al. [29] suggested 6 cm as the cut-off point for di cult intubations, and reported a sensitivity, speci city, PPV, and NPV of 70.59%, 68.63%, 84%, and 50%, respectively, using that value. In our study we considered a TMD of ≤ 6 cm as a predictor for di cult tracheal intubations. We found a TMD sensitivity, speci city, PPV, NPV, and Youden's index of 46.67%, 88.97%, 30.43%, 94.16%, and 0.36, respectively. The AUC for TMD was 0.343 (CI 0.198-0.488). Differences in TMD-related ndings could be explained by factors that might in uence the measurement of TMD, including limitation of head extension, shortness and depth of the mandible, and the height of the larynx [28]. As a result, some authors have doubted the reliability of TMD as an isolated predictive test for di cult laryngoscopies and intubations [30,31]. On the other hand, Benumof [32] found that both a large and small TMDs could predict di cult intubations.
In our study, the sensitivity, speci city, and Youden's index of 40%, 88.9%, and 0.29, respectively (PPV = 27.3, NPV = 93.5), for MMT-TP supports ndings of Shiga et al. [33], whose meta-analysis was comprised of 41,193 patients. That metaanalysis identi ed an overall sensitivity and speci city for MMT-TP of 49% and 86%, respectively. Similar results were reported by Iohom et al. [17]. Our results differed from some studies that have reported a higher sensitivity [34,35,36], and from those of Hashim et al. [27], who evaluated ve airway tests in 60 patients of both genders, and found a 23% sensitivity, 68% speci city, 58% accuracy, and 16% PPV of the Mallampati test, which were smaller in comparison to our study. The wide variations in the reported sensitivities and speci cities of the MMT may be due to the considerable interobserver variability found during this assessment, which related to the performance of the test with or without phonation, patient cooperation, or patient position [16,37].
In the present study, we performed the MMT-NTP in the supine position before the induction of anesthesia. We found an increase in the sensitivity (60%) using this technique; however, the PPV, speci city, and AUC were reduced in comparison to the results obtained using MMT-TP in the sitting position. On the other hand, the number of false-positives for MMT-NTP in the supine position was two times higher than those of MMT-TP in the sitting position ( SMD is anatomically easy to measure and is commonly used in clinical practice [40]. Previous studies have reported different cut-off points for SMD, with consistent values ranging from 12.5 to 13.5 cm [3,6,29]. In the present study, SMD values of ≤ 12 cm were considered to be predictors of di cult tracheal intubations. In our study, SMD sensitivity was found to be 53.3%, speci city was 86.2%, PPV was 28.6%, NPV was 94.7%, and accuracy was 83.1%. These ndings are consistent with the results of Palczynski et al. [40], who found a sensitivity of 60% and a PPV of 19% for SMD. A poor sensitivity and PPV for this test (8.3% and 3.4%, respectively) were observed by Khatiwada et al. [41] and, in a study by Shobha et al. [42], SMD sensitivity was found to be 3.3% and PPV was 6.25%.
Although repeatedly reported to be a good measure of head extension, previous studies have reported that the SMD has limited clinical value and fails to adequately and solely predict di cult intubations [33,41,42].
This study had several limitations, including its exclusion of pregnant women, obese patients with a BMI of ≥ 35 kg/m2, and emergency cases. Also, TMD and SMD were measured by different persons. Finally, we did not assess neck mobility or neck circumference, which might also be important factors in predicting di cult laryngoscopies.
In conclusion, we found an incidence of di cult intubations of 9.38%, with signi cant increases noted with increasing age. Ideally, any clinical test that is used for prediction of these di cult airways should be quick, simple, convenient, and practical. Unfortunately, there is still no individual test, or combination of tests, with 100% sensitivity (i.e., no false negatives) and 100% speci city (i.e., no false positives). While the Mallampati score is an established method for predicting di cult intubations, its relatively low sensitivity and speci city limit the practical value of the test.

Conclusions
The mandibular protrusion test (MPT), with its high accuracy, speci city, positive predictive value, and good sensitivity, The institutional research Ethics Committee approved this observational, prospective study (20190210). A written informed consent from all patients was obtained.

Consent for publication
Not Applicable.
Availability of data and materials