Ellingsen, Ivar MD, PhD; Holmedahl, Nils H. MD
During bronchoscopy, some patients show marked pressure-dependent compression of the trachea and the main bronchi during coughing and at forced expiration. When the posterior wall is pushed forward approaching the anterior wall, it is termed excessive dynamic airway collapse (EDAC). Tracheobronchomalacia (TBM) is used when the airway cartilage is weakened.1–5 The severity of the obstruction is judged through the bronchoscope at quiet breathing and cough,1,6 and at forced expiration, but the expiratory effort to achieve collapse has not been standardized.3 Bronchoscopy is regarded the “gold standard” for diagnosing central airway obstruction,3 and the smallest cross-sectional area should be estimated during some sort of expiratory effort. However, in clinics without video recording or morphometric equipment, the 1-dimensional shortest anteroposterior distance (AP-distance) might be easier to estimate by eye than the 2-dimensional cross-sectional area.
A classification of tracheobronchial obstruction based on the clinical experience of experts states that <50% reduction of cross-sectional area is within normal limits, while further reduction indicates pathology.7 However, this classification is not followed by data showing the flow resistance created by the obstruction or data suggesting the luminal reduction necessary to make pulmonary function suffer.
The aims of this study were: (1) by a model to find the relations between the cross-sectional area, flow resistance, and the AP-distance of the trachea; and (2) applying these data in a prospective chronic obstructive pulmonary disease (COPD)-patient study to evaluate the severity of EDAC during cough and forced expiration, and the impact of EDAC on pulmonary function tests.
MATERIALS AND METHODS
According to Weibel,8 the normal adult trachea has a cross-sectional area of about 2.54 cm2 and the length of about 12 cm. Simulating this we used a garden hose (Hozelock Ltd., Birmingham, UK) with wall thickness of 3.5 mm and inner diameter of 18 mm (cross-sectional area about 2.54 cm2). To simulate EDAC, the hose was inserted into a 12-cm long brass pipe with an inner diameter of 24 mm, of which about 1/2 of the circumference was cut off. Another pipe with external diameter of 10 mm was placed over the “naked” part of the hose. By 4 screws this pipe increasingly compressed the “naked” part of the hose making it protrude into the lumen as shown in Figure 1B, simulating the posterior tracheal wall protrusion. This choking was done in 8 steps, starting at step 1 with the 10-mm pipe just resting on the hose without pressure, and step 8 giving an AP-distance of 1 mm, representing a luminal reduction of 80%. Flow was measured downstream to the choking device with ASV Flowmeter model DFM-350-1″-1000 (ASV Stubbe GmbH & Co. KG, Vlotho, Germany), a tank supplied pressurized air through the hose approximately 2 m upstream to the choking device (Fig. 1A). To measure intraluminal pressure, 2 cannulas with external diameter of 2 mm was inserted through the hose wall, one near each end of the choking device. The cannulas were blinded when uncoupled from the barometer. Pressure was measured by Maximal Respiratory Mouth Pressure Monitor (Pollard Electronic Design, Stockport, Cheshire, UK) with a measurement range of 0 to 769 cm H2O. The airflow was adjusted stepwise from 1 L/s (60 L/min) through 9 L/s (540 L/min). At each step the pressure was measured upstream and downstream of the choking area and the “driving pressure” over the obstruction was calculated.
The test series was performed twice: (1) with the 10-mm pipe being 3 cm long simulating EDAC of the lower trachea; and (2) with the 10-mm pipe being 12 cm long simulating obstruction of the trachea in full length. Gas temperature was registered during the series with 12 cm obstruction with a fast responding probe introduced through the cannulas for pressure measurements using Thermometer ETI 2001 (Electronic Temperature Instruments Ltd., Worthing, UK).
Measurements of Cross-Sectional Area Reduction
The garden hose was cut at one end of the brass pipe, the cut surface painted white to contrast the dark lumen. For each choking step, a digital photo of this end together with a measure of millimeters (Fig. 1B) was transferred to a computer to calculate the AP-distance and the cross-sectional area, using the software Canvas version 6 (Deneba Systems Inc., Miami, FL).
A total of 104 (64 male) COPD patients admitted to Glittreklinikken AS for a 4-week rehabilitation program participated after written information and consent. The study was approved by the regional committee for medical research ethics (Oslo, Norway), and conducted from June 2004 to March 2006.
Bronchoscopy was carried out with 2 aims, firstly to map intrabronchial premalignant lesions by autofluorescence (data not published), and secondly to look for tracheal obstruction during cough and forced expiration. Inclusion criteria: current or ex-smokers, no current signs, symptoms or history of malignant illness, no signs of tumor on chest x-ray taken at inclusion, and no current COPD-exacerbation. Table 1 shows the study population data.
Pulmonary function tests were performed as recommended by European Respiratory Society/American Thoracic Society9 with the European Coal and Steel Community reference equations,10 using Jaeger Master Screen/Master Lab (Erich Jaeger GmbH, Hoechberg, Germany).
Bronchoscopy was performed through the mouth with the patient in a semirecumbent position premedicated 30 to 60 minutes ahead with diazepam 5 mg by mouth, hydrocodone 5 mg by mouth, atropinsulphate 0.3 mg subcutaneously, alfentanil 0.5 mL (0.5 mg/mL) intravenously, and lidocaine 2 mL (40 mg/mL) from a dispenser was given at the start of the bronchoscopy and lidocaine 1 mL (4 mg/mL) through the bronchoscope. When the inspection of the bronchial tree to subsegmental level and sampling were finished, the bronchoscope was positioned about 5 cm proximal to the carina. The patient was asked to cough and after a few breaths to make a forced expiration, and the shortest AP-distance in millimeters was estimated during both the maneuvers.
Data from both model and patient studies were assessed for normality of distribution, linearity, and homogeneity of variance. Correlations were investigated by Spearman ρ, relationship between variables by standard linear regression. Two sided P-values of ≤0.05 were considered significant. All analyses were performed using MS Excel 2003 and IBM SPSS Statistics version 19.
Measurement of Cross-Sectional Area Reduction
A strong relationship was found between cross-sectional area and AP-distance (R2=0.99, P<0.0005). Figure 2 shows the cross-sectional area being reduced by approximately 50% at an AP-distance of 6 mm. The points were well fitted by a polynomial curve; its regression formula used to calculate the reduction in cross-sectional areas corresponding to AP-distances 1 to 6 mm (Table 2). This table was used in the patient study to estimate the tracheal cross-sectional area based on the observed AP-distances.
Figure 3A shows how the driving pressure increases as the cross-sectional area is reduced for an obstruction length of 3 cm, and Figure 3B for an obstruction length of 12 cm. Because the upstream pressures exceeded the range of the transducer, choke step no. 8 was not measured for the flows 8 and 9 L/s in Figure 3A, and in Figure 3B the results are limited to 70% cross-sectional area reduction.
When the obstruction length increased 4-fold, the driving pressure increased approximately 3.5-fold.
The gas temperature upstream was approximately 21°C and downstream between 10°C and 16°C. For the calculations below we regarded the gas temperature at the obstruction to be 17°C.
Resistance calculations depends on whether the airflow is laminar or turbulent determined by calculating Reynolds number, which indicates laminar flow when <2000, transient when 2000< number <4000 and turbulent when >4000. Reynolds number for an obstruction length of 3 cm with an expiratory gas mixture at 37°C was calculated to 3596 for 1 L/s flow through 80% luminal obstruction, 4513 for 2 L/s through 70% obstruction, and 3825 for 3 L/s through 50% obstruction. Higher flows and more severe obstructions further increased Reynolds number.
The bronchoscopy was interrupted for 11 of the 104 patients examined, nevertheless we estimated the AP-distance in all as they had spontaneous cough and forced expiration during the procedure.
During forced expiration, AP-distance was reduced to 6 mm in 1 patient, the rest had greater distances. During cough, 5 patients showed deformation/obstruction with the lateral walls approaching each other, whereas 18 patients showed posterior wall protrusion (EDAC) like the deformation simulated in the model (with AP-distances ≤6 mm). Of these, 1 had 2 mm, 2 had 3 mm, 3 had 4 mm, 12 had 5 mm, whereas none had 6 mm. Using the graph in Figure 2 and assuming that all the patients had a tracheal cross-sectional area of 2.54 cm2, the cross-sectional area reductions during forced expiration and cough were calculated for every patient. The correlation coefficient between these 2 sets of values was 0.83 (P<0.0005), indicating that posterior wall protrusion is most often expressed both during cough and forced expiration. However, no significant correlations were found between flow test parameters [peak expiratory flow (PEF) and forced expiratory volume in 1 second (FEV1)] and the cross-sectional area reduction during forced expiration and cough for all patients as a group and for the group of 18 patients who had >50% cross-sectional area reduction during cough.
Cross-Sectional Area Reduction
The posterior wall protrusion is the deformation most often seen both in our study and by others.5,11 In the “model,” the AP-distance was reduced to 1/3 of original diameter when the area was reduced by 50%. However, the AP-distances applied in vivo most often underestimated the obstruction, the reasons being 2-fold. Firstly, the model “posterior wall” protrusion is a bit different from the tracheal posterior wall protrusion, as the latter seemed to approximate the side walls more closely. This appears in the pictures of the cross-sections (Fig. 1B) and is expressed in the equation in Figure 2 as the area y>0 for x=0. Consequently, a model AP-distance of 1 mm corresponds to an 80% obstruction (Table 2), while in vivo assuming uncompressed cross-sectional area of 2.54 cm2, 1 mm corresponds to probably 85% to 90% (but not 100%). Secondly, the uncompressed cross-sectional area of 2.54 cm2 in our model is based on the mean value of Weibel.8 If the cross-sectional area of the garden hose had been larger, our values from Table 2 would overestimate the obstruction (the graph in Fig. 2 would be pushed upwards), while we would underestimate the obstruction if the cross-sectional area had been smaller. Hence, in most patients with ≤6 mm AP diameter the obstruction is ≥50%.
In Figure 2, the uncompressed hose had the AP-distance of 16.3 mm (and not 18 mm), which means that the hose is not perfectly circular. Besides, the cross-sectional area is estimated to 2.63 cm2 (and not 2.54 cm2). This may imply inaccurate measurement of the area and/or that the garden hose was not uniform in every part. The difference of approximately 3% was considered to have minor influence on our results.
Figures 3A and B simulates tracheal resistance which increases during maximum effort pulmonary function tests. For low airflows (1 to 2 L/s) the resistance increases sharply at 70% to 80% obstruction. This is in accordance with Nishine et al12 who found a marked increase in the resistance when cross-sectional area was reduced by 60% to 70%, using a tracheal model with a concentric stenosis and flows up to 2 L/s. Brouns et al13 found substantially increased resistance at 75% obstruction for an airflow of 0.5 L/s and at 80% to 85% obstruction for an airflow of 0.25 L/s, in a model based on computed tomography scans. The results of Nichine and colleagues, Brouns and colleagues, and ours support the expert’s opinion7 that luminal reduction of <50% is to be considered normal, 50% to 75% reduction regarded as mild, 76% to 90% reduction as moderate, and 91% to 100% is regarded as a severe obstruction.
The model gas was air at 17°C and density 1.14 kg/m3. In vivo, expiratory gas is a mixture of air, 5% CO2 (pressure 5.5 kPa) and water vapor (pressure 6.25 kPa), with a calculated density of 1.22 kg/m3 at 37°C. Our calculations of Reynolds number indicate that most of our observations represent turbulent airflow in which the flow resistance depends on gas density and not viscosity. The difference in density between air at 17°C and expiratory gas at 37°C at pressure 101 kPa is 0.08 kg/m3, which is approximately 7% of the density at 17°C. The impact of this difference on the flow resistance (and thus on flow), is uncertain. However, assuming a linear relationship between density and resistance in this interval, resistance and flow would change <10%. This implies that if the model study was done with expiratory gas at 37°C the curves in Figures 3A and B would be 5% to 10% nearer to the x-axis. Nevertheless, we find the model measurements useful in simulating the airflow passing through obstructions in the trachea.
In clinics lacking video recording and morphometric equipment, the bronchoscopist has to estimate changes in the cross-sectional area by eye; the estimation of EDAC often to be done in less than a second. This method is inaccurate. As discussed above, there are limits in applying the relation between cross-sectional area and AP-distance found in the model to AP-distances observed during bronchoscopy. The AP-distances observed will underestimate the obstruction in most cases, to what degree is uncertain. However, in most patients an AP-distance of ≤6 mm will correspond to a cross-sectional area reduction of ≥50%, thus representing an EDAC. This may be a guide for a clinician using visual estimation.
This study investigated EDAC during forced expiration and cough. Generally, estimations during patient maneuvers are uncertain as they depend on the patients’ ability and effort. During cough, the observation may additionally be disturbed by body movements.
In our COPD patients no relation between EDAC and PEF or FEV1 is found. This is in agreement with Loring et al14 who evaluated 80 COPD patients for TBM/EDAC by bronchoscopy and transtracheal pressure to determine flow limitation during quiet breathing, and found no correlation between FEV1 and airway collapse during quiet breathing.
Boiselle et al11 examined 51 healthy subjects of both sexes aged 25 to 75 years with normal pulmonary functions tests and found EDAC in 78%, 2 subjects having 80% to 90% reduction. Figure 3A shows that an 80% obstruction requires a driving pressure of >300 cm H2O (hardly achievable for any person) to give a flow >4 L/s (a pathologically low PEF). So even in healthy persons EDAC seems to have no influence on PEF.
For an EDAC/TBM to influence for example PEF and even give breathing symptoms, we will point at 3 elements of probable importance: (1) the length of the obstruction; (2) the severity of obstruction; and (3) the duration of the obstruction.
As to (1): Figures 3A and 3B show the importance of the length of the obstruction, implying that the impact of a severe, but short obstruction is likely to be small.
As to (2): At the start of a PEF maneuver the intrathoracic pressure may be ~200 cm H2O. Assuming this creates an EDAC of 85%, Figure 3A shows that the maximum flow should be 4 L/s. Later on when pressure drops, a trachea with intact elastic properties will widen. At a pressure of 100 cm H2O, Figure 3A shows that the flow may be normal (9 L/s) if EDAC is reduced to “mild” with 70% obstruction. Hence, in a trachea with normal/nearly normal wall elasticity a moderate or severe EDAC occupying only a fraction of the expiratory time will not influence PEF.
As to (3): The considerations made in (2) provide nearly normal elastic properties of the tracheal wall. If this elasticity is reduced (due to eg, chronic inflammation, trauma, inborn defect) the trachea will be highly susceptible to pressure. Hence, the obstruction will be severe (90% to 100%) even for low intrathoracic pressures during expiration, and EDAC/TBM may be seen during quiet ventilation. Figure 3A shows a maximum flow of 4 L/s at 70% obstruction for a driving pressure of 25 cm H2O, and further flow reduction with increasing obstruction or dropping pressure. In such cases the EDAC/TBM will be present during a longer part of the expiratory time. If present in otherwise healthy airways we expect such EDAC/TBM to influence PEF and FEV1 and even give respiratory symptoms, but probably not in COPD patients.
Our model study of the most usual type of EDAC shows that an obstruction <75% of the luminal area has negligible effect on flow resistance. We find 3 elements important when considering whether EDAC will affect PEF and even give respiratory symptoms: the duration, the length, and the severity of the obstruction, in that order of importance.
The authors thank Grethe Dahle and her staff at the endoscopic laboratory for enthusiastic cooperation. The autofluorescence system DAFE (Ricard Wolf GmbH, Germany) was kindly provided by Didrik Mehn-Andersen AS, Ulsmågvn 29 b, 5892 Bergen, Norway.
1. Jokinen K, Palva T, Sutinen S, et al..Acquired tracheobronchomalacia.Ann Clin Res.1977;9:52–57.
2. Bryant LR, Houck GR, Loughrin JR, et al..Tracheo-bronchial collapsibility, an in-vivo and ex-vivo study.Respiration.1970;27:74–85.
3. Carden KA, Boiselle PM, Waltz DA, et al..Tracheomalacia and tracheobronchomalacia in children and adults: an in-depth review.Chest.2005;127:984–1005.
4. Kandaswamy C, Balasubramanian V.Review of adult tracheomalacia and its relationship with chronic obstructive pulmonary disease.Curr Opin Pulm Med.2009;15:113–119.
5. Murgu SD, Colt HG.Expiratory central airway collapse: a concise review.Egypt J Bronchol.2007;1:87–99.
6. Nuutinen J, Leinonen A.Acquired tracheobronchomalacia. A cineradiographic study with bronchologicl correlations.Ann Clin Res.1977;9:365–368.
7. Ernst A, Majid A, Feller-Kopman D, et al..Airway stabilization with silicone stents for treating adult tracheobronchomalacia: a prospective observational study.Chest.2007;132:609–616.
8. Weibel ER.Morphometry of the Human Lung.1963.Berlin-Gøttingen-Heidelberg:Springer-Verlag.
9. Miller MR, Hankinson J, Brusasco V, et al..Standardisation of spirometry.Eur Respir J.2005;26:319–338.
10. Quanjer PH, Tammeling GJ, Cotes JE, et al..Lung volumes and forced ventilatory flows. Report working party standardization of lung function tests, European Community for Steel and Coal. Official Statement of the European Respiratory Society.Eur Respir J Suppl.1993;16:5–40.
11. Boiselle PM, O’Donnell CR, Bankier AA, et al..Tracheal collapsibility in healthy volunteers during forced expiration: assessment with multidetector CT.Radiology.2009;252:255–262.
12. Nishine H, Hiramoto T, Kida H, et al..Assessing the site of maximal obstruction in the trachea using lateral pressure measurement during bronchoscopy.Am J Respir Crit Care Med.2012;185:24–33.
13. Brouns M, Jayaraju ST, Lacor C, et al..Tracheal stenosis: a flow dynamics study.J Appl Physiol.2007;102:1178–1184.
14. Loring SH, O’Donnell CR, Feller-Kopman DJ, et al..Central airway mechanics and flow limitation in acquired tracheobronchomalacia.Chest.2007;131:1118–1124.
© 2014 by Lippincott Williams & Wilkins.