Airway sizes and proportions in children quantified by a video-bronchoscopic technique

In pediatric respiratory medicine airway malacia disorders and disorders of size and shape are common [1] yet there are no in-vivo quantitative measurement studies of these disorders. The semi-quantitative photogrammetric techniques that have been published regarding these disorders [2, 3] are limited by a number of factors [4], and at best, have tended to underestimate the change being measured. Some of these limitations have now been overcome by our new quantitative technique for measurement of airway lumen: the colour histogram mode technique [5]. However when malacia is identified it can only have relevance if there is a reference point. Similarly as bronchoscopic techniques evolve roles in research of areas such as wheezy infants and even endobronchial challenge testing [6], the ability and capacity to measure luminal changes and an understanding of airway proportions and an appropriate comparator site for measurements becomes essential.

The cricoid with its easily accessible position, relative resistance to distortion with pressure change, relatively constant shape at its outlet [7] and its ease of identification all contribute to make it the ideal site for distal airway comparisons. Indeed its size being similar to that of age appropriate endotracheal tube (ETT) potentially provides an additional way for comparisons [5]. However there are no in-vivo data: the available data on these subjects being limited to autopsy work [8‐12]. When lung function tests have been used to assess airway size and associated governing factors, they only offer unquantified nominal statements or inferences on airway cross-sectional area within compartments of the airways as exemplified by the gender differences in lung function which have been almost universally reported as having resulted from small airways [13‐17]. In addition to making bronchoscopic assessments of size, the ability to understand factors contributing to specific airway site sizes from anthropometric factors would obviously be very useful to the clinician and researcher alike.

As there is little data dealing with the issues of actual quantified airway sizes in terms of cross-sectional area measurements and airway proportions, the aims of this study were to: (i) define the sizes and the relationship of the large airways to their respective cricoids and (ii) define factors that influence these relationships. We hypothesized that that the large airways undergo proportionate growth when compared to the cricoid as a reference point and that gender and the usual childhood anthropometric factors (age, height or body length, weight, gender) influence airway size.

One hundred and twenty five children were enrolled of whom 87 were males and 38 females. The median (range) age of the overall group was 2.05 (0.13–9.80) years and there were no significant differences between the median ages of males 1.91 (0.13–9.25) years and females 2.51 (0.18–9.80) years (p = 0.30) or within any age grouping with p values = 0.59, 0.66 and 0.19 for the 3 age groups I, II and III respectively. All children in this cohort had chronic cough and none had significant airway lesions.

This is the first study reporting detailed in-vivo airway cross-sectional area measurements from cricoid to lobar bronchial divisions and area based airway proportion measurements from children using a validated bronchoscopic methodology. We found that body length and weight but not gender influenced the size of major airways measured. In addition we found that there were no gender differences in large airway sizes as assessed by cross sectional area measurements and the cricoid to be most suitable for comparative measurements as ETT derived areas underestimated comparative assessments.

In children there are no in-vivo studies that have determined major airway sizes and proportions to the level of the lobar bronchi. Existing data on this subject is limited to autopsy studies involving small numbers of subjects and not surprisingly there is a wide scatter of values [9‐12]. Despite the considerably larger numbers of subjects in our study we also found a wide scatter in sizes of the large airways but have affirmed that the LMS is significantly smaller that the RMS (79% of RMS area) and that this relationship remains constant across childhood. This latter issue has not been previously described even though there have been suggestions of such from autopsy and computerized tomography (CT) based studies [10, 21, 22]. With respect to the anthropometric factors that influence airway size, the multivariate analyses (Table 3) revealed significant but only very weak predictive associations with the low adjusted r² and β values. This indicates that either additional factors or factors other than those examined are likely to be important. The children in this study were relatively healthy without histories of nutritional state disorders, prematurity or growth-retarded birth, severe neonatal and acquired lung disease that are known to interfere with somatic lung growth. We did not analyse for other potential confounders such as drug and or cigarette smoke exposure in pregnancy. However it is more likely that these effects would influence airway size when there is somatic growth effects such as growth retardation, whereas in contradistinction lung function assessments might be affected by exposure without somatic growth effects [23, 24].

Airway proportion measurements from cadaver studies suggest that the airways progressively and proportionately reduce size from the central to peripheral airways in both children and adults [8‐10, 25]. However the numbers in these studies were small. Our study with far greater numbers of subjects confirms these findings and indeed shows the large airway proportions remain constant across the whole infancy and childhood periods. Combining these factors with the lack of gender differences in size of the large airways and therefore proportions, lends support to the small airways being the likely sites of gender related differences in lung function in children [13‐17]. This raises the possibly of gender differences in factors governing growth of the small airways. However these interpretations require some caution, as even though we accounted for the dynamic respiratory cycle changes by measuring the cross sectional area at end expiration and the proportions remained constant across age groups, we did not control for lung volume, and we did not make independent confirmatory assessments of growth of the large airways using volumetric and length measurements. In addition to this new information, these data provide the clinician and researcher with both absolute values and greater perspective of the range of changes that could be expected during transitions from the trachea to mainstem and to lobar bronchi respectively.

In relating airway proportions a suitable denominator, as the comparator is necessary. We used the cricoid as the comparator as it is easily accessible and identifiable, has a relative resistance to distortion with pressure change, and has a relatively constant shape at its outlet [7]. The importance of a comparator is also evident when one describes measurements of tracheomalacia, bronchomalacia and or airway stenoses. The significant differences between the ACR and AER indicate or support the use of the cricoid area as a reasonable comparative point for airway lesions in that it is less likely to underestimate proportions than the ETT as a comparator. This level of underestimation is however expected as Cole's formula was designed for safety of placement of the ETT through the larynx. It is therefore likely that these differences are of no clinical consequence and in fact, would equate to approximately one half of an age appropriate ETT size. Consequently, the ETT size assessment could provide a useful and acceptable adjunct to measurements as has been described for ETT use in upper-airway lesions in the past and particularly when the cricoid is itself involved in pathology [26].

A potential limitation of this study pertains to patient selection and using symptomatic patients as "normals". While it is unlikely that it will ever be ethically acceptable for normal children to undergo an invasive bronchoscopic procedure to resource this type of data for comparative purposes, it is also equally unlikely that airways' sizes measured in these children are appreciably different to that of normal children. Indeed, to date there are no data dealing with the large airways to suggest otherwise. In addition, the children in our cohort were otherwise well (those with intercurrent illness were excluded); the cricoid areas were larger than the area predicted from ETT size and our data is similar to the values obtained from cadaver samples.

Another potential limitation relates to the methodology of measurement, in particular our using a 10 mm viewing distance to measure angulated sites such as the RML and RUL and our ability to precisely define this 10 mm distance. Indeed we excluded these sites from our measurements because using even shorter viewing distances to overcome angulation difficulties results in even greater magnification, thus rendering the image outside the limits of measurement. The use automated distortion correction along with repeated and sequential assessments while approaching and withdrawing the bronchoscope from the object of interest using the applications of optic flow techniques [27] and its recent advances [28, 29] that could be applied to bronchoscopic assessments, might overcome these issues and measurement failures and offer an even more detailed assessment of airway anatomy. This would potentially include a three-dimensional (3-D) reconstruction image of the accessible airways. Despite these current limitations, the advantages of this technique exceed those of reconstruction CT and or CT "virtual bronchoscopy" where their capacity to localize specific sites such as the distal end of the cricoid and the origins of lobar bronchi are limited. In addition when used in this context, the radiation doses for extensive, localised and particularly for repeated assessments remain a considerable safety concern to both clinical and research work in children [30, 31].

We conclude that large airways of children grow progressively with increasing age but their proportions with respect to the cricoid remain constant until at least 10 years of age. Gender does not influence cricoid or RMS or LMS size in childhood. The regression equations indicate that anthropometric factors of body length and weight are significantly related to but only have weakly predictive influences major airway size. The cricoid area is a suitable comparator or reference point for comparisons of lower airway sizes and lesions.