Preoperative Evaluation of Extension Capacity of the Occipitoatlantoaxial Complex in Patients with Rheumatoid Arthritis: Comparison between the Bellhouse Test and a New Method, Hyomental Distance Ratio

Forty patients with rheumatoid arthritis who were scheduled to undergo total hip or knee arthroplasty requiring general anesthesia were studied. The study was approved by the institutional ethical committee (Nippon Steel Yawata Memorial Hospital, Kitakyushu, Japan), and written informed consent to participate in the study was obtained from all patients. Patients with severely symptomatic instability were excluded from the study. We did not always choose general anesthesia with a conventional laryngoscopy when airway difficulty was predicted.

All testing was performed by two observers (I. T. and T. I.), who had the appropriate training in accordance with the original method,1in the following manner (fig. 1). The goggles on which a goniometer was mounted were secured snugly on the patient's head. The patient sat upright adjacent to the film cassette, looking straight ahead and with the head in the neutral position. The observer (I. T.) supported the patient's shoulders to minimize leaning. Another observer (T. I.) measured the angle of the goggles and the HMD using a specially made ruler. The ruler was made out of a metal stylet with a diameter of 2 mm for a pediatric-sized endotracheal tube. A rounded tip of the thin stylet was pressed on the skin surface just above the hyoid bone, and the distance from the tip to the most anterior part of the mentum was measured using a Vernier micrometer, which was defined as the HMD (fig. 1).8,9The patient was then asked to extend the head maximally while attempting to move the neck as little as possible. The observer was careful not to move the neck by inspecting surface contours. The angle of the goggles and the HMD were again measured in this position. Lateral cervical radiographs were taken at the same time that the angle of the goggles and the HMD were measured in the two head positions. The HMDR was defined as the ratio of the HMD at the extreme of head extension to that in the neutral position. The Bellhouse angle was defined as a difference in the angle of the goggles between the two head positions.

The radiographs were analyzed by two experienced radiologists who were blinded to the purpose of the study and who did not know the Bellhouse angle, the HMD, or the HMDR. As a reference for the position of the occiput relative to the cervical spine, the McGregor line was used. This line connected the most dorsal edge of and caudal portion of the occiput and the dorsal edge of the hard palate (fig. 2). The reference line for the axis was drawn through the basal plate of the vertebral body (fig. 2). Because reference lines did not intersect on the radiograph in some patients, the occiput or axis angle was defined as a difference in angle between the occiput or axis reference line and the common line that was the ventral vertical edge on each radiograph, respectively. The occipitoatlantoaxial angle was calculated by the difference between the occiput and axis angles. The occipitoatlantoaxial extension angle was defined as the difference between the occipitoatlantoaxial angle in the neutral position and that at the extreme of head extension. The subaxial extension was assessed by the difference between the axis angles in these positions.

The HMD on the radiograph was measured in each head and neck position using the same definition stated above, and the HMDR was calculated (fig. 2). In addition, the distances from anteroinferior border of the hyoid body to the center of the external acoustic meatus and to the line drawing tangential to anterior border of the vertebral body adjacent to the hyoid bone were measured on the radiograph in each head and neck position (fig. 2).

The occipitoatlantoaxial extension angle required to expose the glottis during conventional laryngoscopy has been reported as varying from 12° to 23°.3,11Because there seemed no doubt that at least 12° was needed for laryngoscopy,3the occipitoatlantoaxial extension capacity of less than 12° was defined as “reduced.” We compared the degree of the subaxial extension between patients with and without a reduced occipitoatlantoaxial extension capacity.

The difference was determined using the Wilcoxon signed rank test or the Mann–Whitney U test and was considered significant when a P  value was less than 0.05. The relation between paired measurements was analyzed by linear regression analysis. A high correlation had an r  value of greater than 0.85, a poor correlation had an r  value of less than 0.5, and a moderate correlation lay between these two values. Sensitivity was the ratio of the true-positive number to the sum of true-positive plus false-negative numbers. Specificity was the ratio of the true-negative number to the sum of true-negative plus false-positive numbers. Positive and negative likelihood ratios were calculated as sensitivity/(1 − specificity) and (1 − sensitivity)/specificity, respectively. Likelihood ratios above 10 and below 0.1 were considered to provide strong evidence for ruling in or out diagnoses, respectively.12The accuracy of the Bellhouse test and the HMDR in identifying a reduced occipitoatlantoaxial extension capacity was assessed by building the receiver operating characteristic (ROC) curve.13The ROC curve was constructed from a set of (x,y) points, where x = the proportion of false positive results (1 − specificity) and y = the proportion of true positive results (sensitivity). The area under the ROC curve equals the probability of correctly identifying a reduced occipitoatlantoaxial extension capacity. An area under the ROC curve ranges from 0.5 to 1.0. An area of 1.0 indicates perfectly accuracy, whereas an area of 0.5 means a completely uninformative test. Statistical comparison of areas under the ROC curves was based on the method described by Delong et al.  14 

A total of 40 patients were enrolled in the study. There were 3 men and 37 women. Mean (± SD) age, weight, and height were 62 ± 7 yr (range, 51–78 yr), 51 ± 9 kg (37–73 kg), and 153 ± 6 cm (142–172 cm), respectively.

Median distances between the hyoid bone and the basiocciput (external acoustic meatus) in the neutral position and in head extension position were 8.8 cm (lower and upper quartiles: 8.3, 9.6 cm) and 8.9 cm (8.3, 9.5 cm), respectively. There was no statistically significant difference between them. The median value of the distance between the hyoid bone and the adjacent cervical vertebra was 3.3 cm (3.1, 3.6 cm) in the neutral position, significantly increased to 3.4 cm (3.1, 3.7 cm; P < 0.05) in head extension position. But the median difference in the distance between the two positions was only 0.0 cm (−0.05, 0.2 cm).

Median hyomental distances measured externally in the neutral position and in head extension position were 3.7 and 4.7 cm, respectively, which were approximately 0.5 cm smaller than those measured radiologically (P < 0.01; table 1). There was a high correlation between hyomental distances measured externally and radiologically (P < 0.0001, r = 0.89 and 0.91 in the neutral and head extension positions, respectively). Median hyomental distance ratios measured externally and radiologically were 1.24 and 1.19, respectively (table 1). External HMDR correlated well with radiologic HMDR (P < 0.0001, r = 0.92; fig. 3).

The median occipitoatlantoaxial extension angle measured radiologically (actual occipitoatlantoaxial extension angle) was 11.2° in the 40 patients (table 2). In 21 of them (53%), the occipitoatlantoaxial extension angle was less than 12° (reduced occipitoatlantoaxial extension capacity). Lateral radiographs at the extreme of head extension showed that near-maximal extension of the occipitoatlantoaxial complex occurred in all patients. Obvious atlantoaxial subluxation was not found on the radiographs. Discrepancies in radiologic angle measurement between two radiologists averaged 1.3°. The median value of the Bellhouse angle was 24.9° (table 2). When the test was performed in patients with rheumatoid arthritis, extension of 12.5° occurred at the subaxial regions despite attempts to move the neck as little as possible, which could not be found by inspecting surface contours of the neck. The extent of the subaxial extension was consistent with a difference between the Bellhouse angle and the actual occipitoatlantoaxial extension angle. Among 40 patients, in patients with an actual occipitoatlantoaxial extension angle of less than 12°, a subaxial extension of median angle of 16.4° (13.0°, 19.5°) occurred, which was greater than that of 8.5° (5.8°, 12.5°) in patients with an actual occipitoatlantoaxial extension angle of 12° or more (P < 0.01). Because a greater degree of subaxial extension occurred when the occipitoatlantoaxial complex was impaired, there was a poor correlation between the Bellhouse angle and the actual occipitoatlantoaxial extension angle (P < 0.01, r = 0.48; fig. 4). In contrast, because the HMDR was not influenced by occurrence of the subaxial extension, the HMDR was highly correlated with the actual occipitoatlantoaxial extension angle (P < 0.0001, r = 0.88; fig. 5). The ROC plots for the Bellhouse test and the HMDR in identifying a reduced occipitoatlantoaxial extension capacity are shown in figures 6 and 7, respectively. The areas under the ROC curves of the test and the HMDR were 0.72 and 0.95, respectively. The area in the latter was greater than that in the former (P < 0.01). Sensitivity, specificity, and likelihood ratios at different cutoff HMDR values for identifying a reduced occipitoatlantoaxial extension capacity are shown in table 3. The cutoff HMDR value of 1.25 was considered to be appropriate.

There was a high prevalence of a reduced occipitoatlantoaxial extension capacity (occipitoatlantoaxial extension angle < 12°) in patients with rheumatoid arthritis requiring total hip or knee arthroplasty. But the subaxial cervical regions were relatively intact in most of them. In this study, when performing the Bellhouse test in patients with an occipitoatlantoaxial extension capacity of 12° or more, the degree of the subaxial extension was similar to that in volunteers with normal cervical spines in our previous study.6In contrast, in patients with a reduced occipitoatlantoaxial extension capacity, a greater degree of compensatory extension occurred at the subaxial regions because most of them had intact regions. Occurrence of even a great degree of the subaxial extension was difficult to find by observing appearance because movement of skin overlying a cervical spinous process did not dependably follow the underlying spinal segment.7Moreover, rheumatoid patients with an impaired occipitoatlantoaxial complex had moved the head using intact subaxial regions for a long time, which made detection of the subaxial extension all the more difficult. Because the degree of the subaxial extension was an error when performing the Bellhouse test, unavoidable and unbalanced occurrence of it caused a poor correlation between the Bellhouse angle and the actual occipitoatlantoaxial extension angle. Therefore, the Bellhouse test did not identify a substantial proportion of patients with a reduced occipitoatlantoaxial extension capacity and was not clinically reliable. This problem can present a potential for missing a prediction of difficult airway in patients with rheumatoid arthritis.

We found a median HMD of 4.7 cm in the head extension position in patients with rheumatoid arthritis. Turkan et al.  9demonstrated that mean hyomental distances in the same head and neck position were 5.9 cm in men and 5.6 cm in women. A reason for this difference may be related to slightly built rheumatoid patients' having a small mandibular space associated with severe rheumatoid arthritis that was long-standing. Radiography revealed that when the occipitoatlantoaxial complex was extended and the head rotated around the basiocciput, the hyoid bone hardly moved anteriorly together with the mandible and was pulled cranially parallel to the cervical vertebra by the basiocciput, which was in accord with the findings of Penning.10As the mandible went away from the hyoid bone by the degree of the head rotation, the HMD increased in proportion to the degree of the occipitoatlantoaxial extension. Therefore, the HMDR was highly correlated with the occipitoatlantoaxial extension capacity and could correctly assess it in rheumatoid patients. We considered that an HMDR of less than 1.25 was a good predictor of a reduced occipitoatlantoaxial extension capacity. To calculate the HMDR, the HMD is measured in the head extension position for evaluating the mandibular space and then is merely measured in the neutral position again. Therefore, the HMDR can serve as a simple and effective method for preoperatively identifying rheumatoid patients with a reduced occipitoatlantoaxial extension capacity.

There are some potential study limitations of our experimental design. First, we did not investigate the relation between the HMDR and difficulty in laryngoscopy. Moreover, only with the HMDR, it may be difficult to predict difficult laryngoscopy completely because several anatomical factors other than a reduced occipitoatlantoaxial extension capacity influence visualization of the glottis during laryngoscopy.5,8,9However, because capacity is one of these factors, we believe that the HMDR may be a useful predictor of difficult laryngoscopy. Second, there was the variation of the skin thickness on the hyoid bone among the patients and between different head and neck positions, which could affect accuracy of the HMD and HMDR measurements. In this study, there were some differences between the hyomental distances measured externally and radiologically. We measured the HMD by pressing the tip of the thin stylet (ruler) on the skin surface just above the hyoid bone. Moreover, because most of patients were considerably thin, the hyoid bone was easily palpable from the skin surface. As a result, a high correlation was established between hyomental distance ratios measured externally and radiologically. Therefore, we believe that the variability was of minor importance in this study. When the hyoid bone is not palpable, such as in morbidly obese patients, the occipitoatlantoaxial extension capacity should be evaluated by other methods, e.g. , radiologic examinations. Although the notch of the thyroid cartilage is more easily identified from the skin surface than the hyoid bone, it should be cautioned that the ratio of the thyromental distances in different head and neck positions is a poorer predictor of a reduced occipitoatlantoaxial extension capacity than the HMDR (our preliminary study). Third, the intrapatient variability of the Bellhouse angle was also possible because the end point for extending the head maximally depended on the voluntary participation of each patient. We did not examine it because of using a radiologic method for measuring the occipitoatlantoaxial extension capacity. However, we do not consider that the variability was important because the observer had the appropriate training in accordance with the original method,1and near-maximal extension at the occipitoatlantoaxial complex without the inclination of the body was confirmed on lateral radiographs at the extreme of head extension. Fourth, despite severe cases of rheumatoid arthritis, we did not find obvious atlantoaxial subluxation. In this study, lateral radiographs were taken in the neutral and head extension positions in which the subluxation was commonly corrected. Finally, we found a 53% prevalence of reduced occipitoatlantoaxial extension capacities because of severe rheumatoid arthritis that was long-standing. It should be cautioned that this is not applicable to all rheumatoid cases.

In summary, in rheumatoid patients requiring total hip or knee arthroplasty, the Bellhouse test did not reflect well the occipitoatlantoaxial extension capacity because of the occurrence of compensatory extension at the subaxial regions when the occipitoatlantoaxial complex was impaired. Therefore, the test missed identifying many pathologic conditions at the occipitoatlantoaxial complex, which can lead to missing a prediction of a difficult airway. However, the HMDR could accurately identify patients with a reduced occipitoatlantoaxial extension capacity despite occurrence of the subaxial extension in this study. Additional studies are indicated to validate this result.