Determination of Endotracheal Tube Size in a Perinatal Population: An Anatomical and Experimental Study

The larynx and trachea from 150 postmortem examinations of fetuses and infants were prospectively collected, between January 1999 and February 2002. Legal and ethical considerations were consistent with our institution’s rules.

Inclusion criteria required that the fetuses and infants were free of malformative syndromes affecting the laryngotracheal structures and had never been intubated. All anatomical structures, including the larynx, the trachea, and the thyroid gland, that were free of malformation on macroscopic examination were included.

The study was performed on anatomical pieces during postmortem examination. The pieces were studied within a delay of 6 h after death. No fixative agent or conservative agent was used.

All biometrical, clinical, and pathologic data were collected during autopsy. For each fetus, gestational age (GA) was determined based on biometric data.

Each larynx was intubated to estimate clinically the largest tube size (clinical ID) that could be introduced without glottic friction or cricoid deformation. Portex Blue Line® endotracheal tubes (SIMS Portex, Hythe, United Kingdom) with ID size ranging from 2.5 to 4.0 and Vygon® endotracheal tube (Vygon, Ecouen, France) ID size 2.0 were used. The largest endotracheal tube that could be introduced without deforming the cricoid ring was considered to be the largest size tube that would not result in injury to the upper airway. In the case where an endotracheal tube ID size 2.0 could not be introduced without cricoid deformation, the patient was considered to have a “high injury risk” intubation.

In a second stage, each larynx was intubated successively with different calibrated inextensible balloons of increasing size. Balloons were made with thermoformed polyvinyl chloride fixed on an 18-gauge catheter. Balloon diameters corresponded, when inflated with air at 20 cm H²O, to the OD of an endotracheal tube. The balloon was connected to a low-pressure manometer (Mallincrodt, Hazelwood, MO) (fig. 1), and inflated with 20 cm H²O pressure. It was positioned through the laryngotracheal structures with the proximal extremity placed 1 cm over the glottis and with the distal one sticking out under the fifth tracheal ring.

Pressure changes were evaluated during the intubation. The optimal tube size estimated by this method (pressure ID) corresponded to the largest one that could pass through the laryngotracheal structures without increasing the balloon pressure or with a pressure normalization time that was less than 15 min.

Intubation was considered to have a high injury risk if the pressure in the smallest balloon remained increased for at least 15 min.

Finally, each larynx was dissected for anatomical estimation. The larynx was separated from the trachea by section along the inferior edge of the cricoid cartilage. A precision digital caliper was used to measure the anteroposterior and lateral diameters of the cricoid lumen, introducing the caliper by the inferior face of the cricoid lumen, and the tracheal lumen. The anteroposterior and lateral diameters of the cricoid lumen were measured using a precision digital caliper (Mitutoyo, Roissy, France; precision = 0.02 mm). The caliper was introduced into the cricoid ring through the caudal end of the lumen. Measurements were obtained without deformation of the anatomical structures (fig. 2). The interarytenoid distance (IAD) was obtained by introducing the caliper between the vocal processes of the arytenoids and opening the glottic plane until the maximal passive abduction movement of the arytenoid cartilage without friction was achieved.

By these measurements, two different tube size estimations were calculated:

• The mean subglottic and tracheal perimeters were calculated for each specimen, considering the lumen as an ellipse (P = 2π√(a²+ b²/2)). These perimeters were compared with the endotracheal tube perimeters calculated based on the external diameter of each tube size. The optimal tube size based on the laryngotracheal perimeter (SG ID) corresponded to the tube whose external perimeter was less than or equal to the subglottic or tracheal perimeter. The intubation was considered to have a high injury risk when the laryngotracheal perimeter was less than the external perimeter of the smallest endotracheal tube.

• The maximal tube size that could be passed through the glottic plane was determined by the maximal OD that was less than or equal to the IAD (IAD ID).

Intubation was considered to have a high injury risk when the IAD was less than the OD of the smallest endotracheal tube.

The population was divided into five groups according to GA (A: ≤ 28 weeks GA; B: 29–32 weeks GA; C: 33–36 weeks GA; D: 37 weeks GA to term; and E: term to 3 months) to analyze the data statistically.

Statistical analysis was performed using StatView 6.0 software (SAS Institute, Cary NC). A descriptive analysis of the population was realized. Continuous variables were compared between groups stratified for age and sex. A nonparametric Wilcoxon test was used for the comparison of variables in each group. The comparison of variables between groups was performed using a Mann–Whitney test. A nonparametric correlation test (Spearman) was performed between the intubation data (clinical ID, pressure ID, SG ID, and IAD ID) and clinical data (theoretical GA, corrected GA, weight, and height) collected for each patient. Differences were considered statistically significant at P < 0.05.

Based on these results, the recommended OD was calculated in each larynx, corresponding to the external perimeter of the tube size determined by pressure estimation less than or equal to the maximal IAD. The corresponding ID according to different trademarks was reported.

The larynxes and tracheas from 122 fetuses aged 25–41 weeks GA and 28 infants from newborn to 3 months of age were included (table 1). The causes of death in the fetus population were termination of pregnancy for neurologic malformations (43 cases), cardiac malformations (31 cases), urogenital malformations (12 cases), polymalformative syndromes (11 cases), intrauterine death (17 cases), and stillbirth (8 cases). The cause of death in the infant population was sudden death syndrome (24 cases) and evolution of nonreanimated neurologic disease (4 cases). No infant was intubated. In 9 cases, a laryngeal mask was used during resuscitation.

Anatomical measurements were characterized by a quasi-linear evolution until birth and followed by an inflection after the birth (fig. 3).

The narrowest part of the airway was the cricoid area in the general population and in each age group (Wilcoxon, P < 0.001). Interindividual variations of measurements were important.

There was no statistical difference between males and females for subglottic and tracheal measurements. The interarytenoid distance in females was significantly less than that in the population (Mann–Whitney, P < 0.001). This difference was statistically significant only in age group E (Mann–Whitney, P < 0.01).

The results of the intubation study are illustrated in table 2.

Intubation based on clinical estimation was always possible.

Intubation based on pressure estimation was considered with a high injury risk in only 2 cases; both were included in group A (7.4%).

Intubation based on laryngotracheal estimation was considered to have a high injury risk in 46 cases: 25 cases in group A (92.5%), 16 cases in group B (43.2%), and 5 cases in group C (20.8%).

The value of the optimal ID varied according to the method of measurement. In premature babies before 37 weeks GA, pressure modification occurred more frequently. The mean difference between pressure estimation of ID and laryngotracheal estimation of ID (pressure ID − SG ID) (fig. 4) was statistically significant in groups A, B, and C (Wilcoxon, P < 0.001). This observation disappeared progressively after 37 weeks GA (group D) and was no longer valid after term (group E) (Wilcoxon).

The mean difference between pressure ID and IAD ID (pressure ID − IAD ID) (fig. 5) is statistically significant only in groups C–E (Wilcoxon, P < 0.001). In 8 of 24 cases in group A, 2 of 24 cases in group C, and 6 of 40 cases in group D, the IAD ID was inferior to the pressure ID.

The clinical ID was always significantly greater than the pressure ID (Wilcoxon, P = 0.02).

There was no statistical difference for the different ID estimations between males and females in all age groups.

The correlations between ID obtained by the different methods of measurement and clinical parameters were always statistically significant (Spearman test, P < 0.0001). The highest correlation coefficients were obtained with corrected GA and birth weight (respectively, clinical ID r = 0.841, r = 0.834; SG ID r = 0.825, r = 0.808; IAD ID r = 0.704, r = 0.663; pressure ID r = 0.770, r = 0.773).

Recommended ODs by weight evaluated by the results of this study are illustrated in fig. 6.

For the purposes of conducting the current study, we obtained fresh autopsy specimens to ensure that the airway dimensions were close to the dimensions in live patients. The use of fresh specimens allows preservation of physicochemical characteristics of the cartilaginous structures necessary to obtain an evaluation similar to physiologic conditions, especially for pressure evaluation. In embedded specimens, because of tissue shrinkage, measurements are modified27or impossible to realize (e.g. , IADs which correspond to the abduction passive movement of the cricoarytenoid join).

The major finding of this study concerns premature babies at less than 37 weeks gestation. In this age group, we observed that IDs based on anatomical measurements were systematically lower those obtained with other methods (pressure, clinical, and IAD estimations). In all age groups, we observed correlations between ID and both corrected GA and weight.

Few studies have been published on the anatomical measurements of the airway in either pediatric populations13,27–32or premature populations.

The dimensions of the cricoid lumen in preterm infants in the current study were less than those published previously in the same age group30–32but greater than those published by Tucker et al.  13The dimensions of the cricoid lumen in term neonates in the current study are consistent with those published in some studies27,28,33–35but less than those published in others.30,32Observed differences between our data and the literature may be attributed to the interindividual variability due to small sample sizes in other studies as well as methodologic differences in measurement. In the current study, the narrowest diameter of the cricoid rings was based on measurements obtained from the entire cricoid cartilage. Differences between the results of the current study and others may be attributed in part to the latter studies measuring luminal diameters at predetermined planes in the cricoid ring. Serial section studies may be influenced by the level of the section and the orientation of the section plane.

Ultrasound measurements of tracheal diameter36in fetuses reported a smaller diameter than our results. An experimental study37demonstrated that tracheal measurements obtained by ultrasound are smaller than anatomical measurements. For newborns and babies, the tracheal lumen observed in our study is similar to those observed by Pracy33and less than those observed by Fearon and Whalen.29 

At least two relations have been reported for the diameter of the cricoid lumen and age. On the one hand, Fishman and Pashley31described a hyperbolic relation between cricoid ring diameter and GA. On the other hand, Schild,32along with our evidence, supports a quasi-linear relation.

The absence of difference between males and females was also observed by some authors27,30,34for laryngeal and tracheal lumens, but no observation was reported regarding the IAD.

Some studies have reported that the anteroposterior dimension of the glottis exceeds the same dimension at the level of the subglottis and tracheal lumen.32,33This does not constitute a limiting factor in intubation. Only IAD was considered as a potential limiting factor at the level of the glottis; the OD of the tube should be less than or equal to the maximal IAD determined for each larynx. In our study, IAD was significantly less in females older than 40 weeks GA. Even if no significant difference was found for IAD ID estimation, the use of a smaller size tube in this population should probably be considered.

The pressure study was based on the capillary pressure estimated between 20 and 30 mmHg16,17,19in adults with a pressure of the venous end of a capillary of approximately 12 mmHg17and on the study of Nordin et al. ,16which demonstrated ischemic lesions as a result of excessive pressure after 15 min. The normal pressure within the tracheal mucosa in preterm infants has not been reported. Nonetheless, many clinicians accept a maximum inflation pressure of 20 cm H²O (13 mmHg) to minimize tracheal mucosal injury. The observations of Nordin et al.  justified criteria used in our study for a pressure normalization shorter than 15 min in case of pressure increase. The role played by the laryngeal surrounding tissues, dissected for the study, is difficult to evaluate in pressure variations. The absence of circumferential muscular structures and presence of large adipoconnective tissue should not significantly affect pressure variations.

This pressure study concerned the whole laryngotracheal lumen and did not define the level of the obstacle in case of pressure increase. The anatomical study confirms that the thinner level of the lumen corresponds to the subglottic area. Moreover, the pressure test was performed directly in the larynx and did not consider the potential pressure induced by the inclination of the tube by the upper structures (nasal fossa, oropharynx) and the position of the head.8,9 

The analysis of the intubation study shows a significant difference between the anatomical estimation of ID and the pressure estimation of ID in the premature population with a higher number of patients considered to have a high injury risk according to anatomical estimation compared with pressure estimation. These results suggest that premature laryngeal structures possess an elasticity allowing the passage of a tube with a higher size than that predicted by anatomical measurements. This observation could explain the better tolerance of intubation in premature infants than in newborns as suggested by Hawkins.6In this case, although the cricoid is anatomically the narrowest part of the airway, its elasticity allows the passage of a higher diameter endotracheal tube, and the limiting factor becomes the interarytenoid distance.

Therefore, the risk of injury is most likely evident in the posterior glottis as demonstrated in clinical reports.7–12 

This elasticity disappears around 37 weeks GA, and the limiting parameter becomes the cricoid area with a risk of induced subglottic stenosis.

Clinical estimation of ID was greater than the ID evaluated by pressure estimation, suggesting that clinical estimation does not seem to be a valid criterion to determine tube size. The clinical parameter reported in the literature that best correlated with tube size was GA23,38–40or birth weight.1,2,26In our study, the highest correlation was observed with corrected GA and weight. The correction of the GA based on biometric parameters seems unrealistic in clinical practice. Guidelines based on birth weight seem more relevant.

Based on these results, a recommended OD table could be proposed to limit a potential injury risk due to intubation. The relation between OD and ID is different according to different trademarks because of different tube thicknesses. For similar ODs, the difference of IDs should reach 1 mm. This is an important point for tube choice for ventilation assistance because respiratory resistance is correlated with the internal endotracheal tube size.

Recommended OD guidelines proposed in this study are intended to provide a safe intubation and to limit laryngeal injury. This does not always correspond to efficient ventilation, principally when high-pressure ventilation is required.

This study demonstrated that laryngeal structures possess a certain degree of elasticity in the premature that permits a tracheal intubation with a tube size that is greater than that predicted by the anatomical measurement. In the premature population, injury risks predominate in the posterior part of the glottis. This elasticity disappears around 37 weeks GA, and injury risk then predominates in the subglottis.

Endotracheal tube size in neonates should be based on anatomical and experimental data to limit these injury risks.