

Validation of age and height based formulae to predict paediatric airway distances – a prospective observational study P Mathew^{1}, V Ashok^{1}, MM Siraj^{1}, V Grover^{1}, D Sethuraman^{2}^{1} Department of Anaesthesia and Intensive Care, Post Graduate Institute of Medical Education and Research (PGIMER), Chandigarh, India ^{2} Department of Anaesthetics, Royal Liverpool and Broadgreen University Hospital Trust, Liverpool, United Kingdom
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/jpgm.JPGM_545_18
Keywords: Airway assessment, general anesthesia, pediatric
Perioperative airwayrelated complications are an important cause for morbidity and mortality in pediatric patients.^{[1]} Although the incidences of unanticipated difficult airway in otherwise healthy children is less common than in adults,^{[2]} unexpected difficulties may still arise leading to serious adverse events if not managed properly. A number of anatomical tests and measurements like the modified Mallampati score, thyromental distance, and sternomental distance have been used in adults to predict the likelihood of a difficult airway. Although their discriminatory power is poor when used alone, a combination of these assessment parameters yield a significant predictive value for a potentially difficult airway.^{[3]} However, unlike in adults, there is very little literature available regarding the usefulness of these factors in predicting a difficult airway in pediatric patients. Additionally, surface measurements such as thyrohyoid, thyromental, and sternomental distances are growth dependent and there is limited information on how age, height, and weight affect these measurements in growing children. An earlier study conducted in our institute attempted to assess the efficacy of Mallampati classification, mentohyoid, thyromental, and sternomental distances in predicting difficult laryngoscopy and intubation in pediatric patients.^{[4]} At the end of the study, the investigators had derived simple formulae for correlating the height and age of children with the aforementioned distances. The purpose of the current study was to validate the formulae derived from the previous study to predict pediatric airway measurements such as thyromental distance, sternomental distance, and mentohyoid distance in relation with the height and age of pediatric patients.
The study was a prospective crosssectional observational single arm study and was initiated after approval by the ethics committee of our institution (INT/IEC/2015/648) and appropriate permission from the author group of the original study. The study was conducted in accordance to the Declaration of Helsinki and reporting is consistent with the STROBE guidelines for observational studies.^{[5]} The study was registered with the Clinical Trials Registry of India. Children in the age group of 315 years scheduled for elective surgery under general anesthesia were included in the study after informed written consent from parents or legal guardians. The children were divided into six age groups—34, 56, 78, 910, 1112, and 1315 years. A minimum of 35 children were assessed for enrolment in each subgroup with a target of enrolling at least 30 children in each subgroup. Children with obvious anomalies of the airway, neck, or face and those requiring emergency surgery were excluded. All the patients were examined prior to surgery and the assessment of airway was recorded as follows:
These distances were measured with a Vernier caliper—an extremely precise measuring instrument with a resolution of 0.01 mm and a reading error of 0.05 mm. Height of the patient was measured with normal measuring tape while the patient was asked to stand against a wall and the back of head, back, buttocks, calves, and heels were touching the wall and the feet placed together. After measuring the airway distances and height of patient, we then calculated the thyromental distance, sternomental distance, and mentohyoid distance according to the height and age of the patient with the help of the following formulae derived by Nikhar et al.^{[4]} On the basis of height:
On the basis of age:
We then statistically compared the measured values with the values derived from the formulae. Sample size was estimated based on the correlation coefficients derived from the earlier study by Nikhar et al.^{[4]} At a power of 95% and confidence interval of 95%, the sample size was calculated to be 25 and 29 children for height and agebased parameters, respectively, in each of the six age subgroups considered. Hence, we decided to assess a total of 210 children, i.e., 35 children in each age subgroup, to enroll a minimum of 30 children in each subgroup. Descriptive statistical parameters such as mean, median, and standard deviation were used for calculating average distances. The test for significance agreement was done using the BlandAltman test. BlandAltman analysis compares two values measured by two methods of measurement in terms of bias (mean difference) and limits of agreement by plotting the average of the two measurements against their differences between them in a graph.^{[6]} Positive bias indicates the underprediction of the new method (in our study, it was measured value). Similarly, negative bias indicates the overprediction of the new method. In this way, bias and limits of agreements were used to interpret the degree of correlation between the measured and predicted distances.
The flowchart representing patient recruitment is described in [Figure 1]. Similar numbers of children were enrolled in each of the 6 subgroups (range 3235). Of the 202 children enrolled in the study, 162 were male and 40 were female children. The mean weight and height were 25.14 ± 11.01 kg and 125.93 ± 21.74 cm, respectively. The measured and predicted values of mentohyoid, thyromental, and sternomental distances are summarized in [Table 1]. We found moderate to good correlation between the measured and predicted values of airway distances in our study [Table 2].
BlandAltman analysis to describe the bias between measured and predicted values is described in [Figure 2] and [Table 3]. The bias between measured and predicted mentohyoid distance for age was 0.6 cm with a 95% CI of 0.510.69 and the bias between measured and predicted mentohyoid distance for height was 0.14 cm with a 95% CI of 0.040.23.
The bias between measured and predicted thyromental distance for age was 0.4 cm with a 95% CI of 0.280.58, whereas the bias between measured and predicted thyromental distance for height was 0.4 cm with a 95% CI of 0.210.49. For sternomental distance, the bias between measured and predicted values for age and height were 0.4 cm with a 95% CI of 0.120.59 and –0.6 cm with a 95% CI of −0.860.38, respectively. All the predicted distances were found to be underpredicted than measured with both agebased and weightbased equations (i.e., bias was positive), except for heightbased sternomental distance calculation (bias was –0.6). Percentage of bias was calculated using the formula: Bias precentage = [Bias (cm)/Measured distance (cm)] ×100 The bias percentages for mentohyoid, thyromental, and sternomental distances are summarized in [Table 3]. Although the bias percentages were found to be similar (5.5%) for age and heightbased equations for thyromental distance, they were lower with heightbased equations for mentohyoid distance (3% vs 13%) and agebased equations for sternohyoid distance (3% vs 4.5%).
Remarkable advances have been made in the management of difficult intubation in adults as well as children. However, an unanticipated difficult airway can still occur in anesthesia practice and is a combinational result of variant patient anatomy, skills of the anesthetist, and the clinical setting of care.^{[7]} Therefore, identification of a potentially difficult airway with simple to use tests can significantly improve patient outcome. Several studies done in adult patients undergoing general anesthesia have shown that easy to perform bedside tests can reasonably predict a potentially difficult airway, particlarly when used in combination.^{[3]} The thyromental distance, sternomental distance, modified upper bite test, and the Mallampati score have all found to be useful and reliable in this regard.^{[8],[9],[10]} However, very little information is available regarding the same for pediatric patients undergoing surgery. A study by Rafique et al.^{[11]} demonstrated that the Mallampati score, thyromental distance, and distance between the tragus and the nares could be useful in evaluating the airway for potential difficulty in children less than 5 years of age. While performing these tests may be difficult in very young children owing to their lack of comprehension and cooperation, older children can be easily evaluated in the preanesthetic clinic prior to surgery. However, interpretation of such airway distances in children is complicated by the fact that they are growth dependent and hence change with age, weight, and height. The earlier prospective doubleblind study in children carried out by Nikhar et al.^{[4]} in our insitu te led to the derivation of simple formulae to predict the sternomental, thyromental, and mentohyoid distances from the height and age of children. When we attempted to validate the formulae derived by Nikhar et al.,^{[4]} we found good agreement between the calculated and measured values. BlandAltman analysis of the sample patients showed a mean difference (bias) between measured and calculated values ranging from 0.14 to 0.60, which corressponds to a bias percentage range of 313%. The bias for all three distance prediction were found to be statistically significant, in both height and agebased equations, as the confidence intervals did not contain zero. Overall agreement in terms of bias were found to be more with heightbased equation for mentohyoid distance and thyromental distance and agebased equation for sternomental distance. We also demonstrated that the percentages of bias to the measured values were sufficiently small to be clinically acceptable. In our study, we took a sample of children with normal airways, i.e., normal anatomy of head and neck as a representative cohort of the general population. Such normal children do come for elective surgery and in fact constitute a large proportion of children undergoing general anesthesia and surgery. Since the logistics of doing such a study by anesthesia personnel was more feasible in preoperative patients than in the general population, we included preoperative patients for this study. The gender imbalance observed in this study probably originates from the pattern of gender distribution observed in the overall statistics of children undergoing surgery. Additionally, puberty causes growth spurts in children that could have an implication in airway distances. Puberty usually occurs in girls between the ages of 10 and 14 while it occurs slightly later in boys between 12 and 15 years. The sample for the study covered children up to the age of 15 to take into account the difference in the onset times of puberty in both sexes. Moreover, the bias and bias% were not significantly higher in these specific pubertal age groups and hence the measurements do not seem to be affected by growth spurts occurring during puberty. To our best knowledge, this study is the first to attempt to validate the relationship between age and height with airway measurements in pediatrics patients, which was earlier proposed by the study of Nikhar et al.^{[4]} An important strength of the study was that our study design ensured equal sampling of children across the age groups of 34, 56, 78, 910, 1112, and 1315 years. Age and heightbased formulae can help in determining what is the expected thyromental, sternomental, and mentothyroid distance for a child of a particular age and height which can be used as a yardstick to compare the patient's actual distances measured directly to determine possible deviation from normal. Our finding is expected to serve as a reference for these airway distances and hence be useful in identifying abnormal values to predict airway difficulty in children with craniofacial syndromes. The study was, however, not without limitations. Ours was a single center study involving Indian children. The growth curve of children varies across countries and depends on genetic, cultural, and environmental factors.^{[12]} However, we hope that the heightbased formulae may be applicable across the world irrespective of these factors. Secondly, our study did not involve obese children and those with obvious craniofacial anomalies. Nikhar et al.^{[4]} had derived weightbased airway distance formulae in their study as well. Further studies are warranted in these special pediatric populations as well across other ethnic groups. To conclude, the formulae derived by Nikhar et al.^{[4]} have been validated so that they can be used as reference to define normal and abnormal pediatric airway distances. The extrapolation of information from this study along with data of airway distances in children with difficult airways can be expected to identify the cutoff for these distances to predict difficult airway in children. This would help warn emergency physicians and anesthesiologists to formulate a safe airway plan for children with an “atrisk” airway. Acknowledgements The authors would like to thank all the children and their parents who consented to participate in the study. Financial support and sponsorship This work was supported by departmental funds. Conflicts of interest There are no conflicts of interest.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3]


