Table 1.
Study design.
Table 2.
Details of the animals studied in in vivo.
Fig 1.
Illustrative images of a needle puncture experiment (NPE).
Note the echo produced by the needle inserted longitudinally to the long axis of the bronchus in B (white arrow), and its corresponding histological localization as identified in C (black arrow). Dotted lines represent needles positioned transversally to the longitudinal axis of the bronchus.
Fig 2.
Measurements made on EBUS (A) and histological (B) images.
Only a part of L2 area and ASM area have been encircled in yellow, to allow the reader appreciate the rest of the image. L: US layer; D1 and D2: perpendicular diameters (blue dotted lines); LA: lumen area (filled light green area); Pi: airway perimeter (continuous green line).
Fig 3.
Scheme of a distended bronchial wall.
In our mathematical model, the bronchial wall was treated as a structure composed of multiple concentric annuli. Excluding the inner annulus, representing the epithelial layer (in purple, negligible in thickness), we considered the submucosa as composed by two annuli representing the ECM (yellow) and ASM (pink). Pi corresponded to the yellow circumference in our model.
Fig 4.
Comparison of EBUS and histological composition of equine airways.
In absence of cartilages, L3 was represented by the adventitia.
Table 3.
Mixed linear models and Bland-Altman tests for comparison between EBUS and histologic measurements.
Fig 5.
Effect of group and airway size on L1 thickness (A), L2 thickness (B), ratio L2 area/Pi (C), and ratio L2 area/Pi2 (D) in ex vivo EBUS.
Raw data (before statistical corrections) are presented as mean per animal in the graphs on the left, and as mean per airway the graphs on the right. L1 and L2 thickness was similar in horses with heaves and controls across all bronchi (p = 0.4 and p = 0.1, respectively). Contrarily, L2 area/Pi and L2 area/Pi2 ratios were greater in horses with heaves than controls across all bronchi (p = 0.004 and p = 0.02, respectively). L1 values were similar in intermediate (Pi<31mm, p = 0.82) and large bronchi (Pi>31mm, p = 0.32) analyzed separately. L2 tended to be thicker in intermediate bronchi of horses with heaves than in controls (p = 0.06) but not in large bronchi (p = 0.5). Both L2 area/Pi and L2 area/Pi2 ratios were greater in intermediate bronchi of horses with heaves then controls (p = 0.006 and p = 0.003, respectively), but only L2 area/Pi ratios were greater also in large bronchi (p = 0.04; p = 0.3 for L2 area/Pi2 ratios). Overall, L1, L2 and L2 area/Pi values were greater in large than intermediate bronchi (p = 0.0003, p = 0.001, and p = 0.002, respectively), while L2 area/Pi2 values were greater in intermediate than in large bronchi (p = 0.02). L1, L2, and L2 area/Pi increased significantly with Pi in both groups (p<0.0001 for all), while L2 area/Pi2 decreased significantly with increasing Pi (p<0.0001). The slope of the relationship did not differ between the two groups for L1, L2 and L2 area/Pi (p = 1, p = 0.14, p = 0.66, respectively), but it was significantly greater in the heaves group than in controls for L2 area/Pi2 (p = 0.01). #: significantly different from controls. *: significant difference between values of intermediate and large bronchi (pooling the two groups).
Table 4.
Variance component analysis.
Fig 6.
Effect of group and airway size on L1 thickness (A), L2 thickness (B), ratio L2 area/Pi (C), and ratio L2 area/Pi2 (D) in in vivo EBUS.
Raw data (before statistical corrections) are presented as mean per animal in the graphs on the left, and as mean per airway the graphs on the right. L1 and was similar in horses with heaves and controls across all bronchi (p = 0.27). Contrarily, L2 thickness, L2 area/Pi and L2 area/Pi2 ratios were greater in horses with heaves than controls across all bronchi (p = 0.03, p = 0.005 and p = 0.04, respectively). L1 values were similar in intermediate (Pi<31mm, p = 0.13) and large bronchi (Pi>31mm, p = 0.63) analyzed separately. L2 tended to be thicker in intermediate bronchi of horses with heaves than in controls (p = 0.03, non-significant after correction) but not in large bronchi (p = 0.13). Both L2 area/Pi and L2 area/Pi2 ratios were greater in intermediate bronchi of horses with heaves then controls (p = 0.003 and p = 0.02, respectively), but only L2 area/Pi ratios were greater also in large bronchi (p = 0.04; p = 0.22 for L2 area/Pi2 ratios). Overall, L1, L2 and L2 area/Pi values were greater in large than intermediate bronchi (p = 0.002, p = 0.005, and p = 0.01, respectively), while L2 area/Pi2 only tended to be greater in intermediate than in large bronchi (p = 0.05). L1, L2, and L2 area/Pi increased significantly with Pi in both groups (p<0.001 for all), while L2 area/Pi2 decreased significantly with increasing Pi (p<0.001). The slope of the relationship did not differ between the two groups for L1 and L2 area/Pi2 (p = 0.86 and p = 0.49, respectively), and it was significantly greater for L2 and L2 area/Pi (p = 0.02 and p = 0.04, respectively). Filled dots represent heaves and open dots represent controls. #: significantly different from controls. *: significant difference between values of intermediate and large bronchi (pooling the two groups).
Table 5.
Power analysis for in vivo EBUS.
Fig 7.
Effect of group and airway size on AASM/Pi2 (A) and AECM/Pi2 (B) measured on histological sections.
Raw data (before statistical corrections) are presented as mean per animal in the graphs on the left, and as mean per airway the graphs on the right. Histological analysis confirmed in vivo and ex vivo EBUS findings concerning ASM remodeling. The slope of the AASM/Pi2 relationship was similar between the 2 groups (p = 0.26), but heaves-affected horses had significantly higher values compared to controls (p = 0.04). The slope of the AECM/Pi2 relationship tended to be different between the 2 groups (p = 0.07), but heaves-affected horses had similar values compared to controls (p = 0.26). #: significantly different from controls. *: significant difference between values of intermediate and large bronchi (pooling the two groups).
Table 6.
Data used for determining L2 composition.
Fig 8.
Effect of the method used for measuring ASM (A) and ECM (B) thickness in histologic bronchial samples.
Results are shown as mean values per horse. 1Mean value of 5 measures of ASM or ECM thickness performed manually and randomly around the bronchial circumference (ECM was measured from the basal membrane to the ASM inner border, avoiding regions where obvious collagen fiber shredding occurred). 2Thickness of the ASM or ECM calculated as function of the ASM or ECM area and airway Pi, as if it was a continuous and homogeneous layer for an airway completely distended (Pi corresponds to the inner circumference of an annulus). Statistical differences were calculated with paired t-tests.
Fig 9.
Composition of the submucosal tissues at histology.
Values are presented as percentage of the submucosal tissue occupied by ECM and ASM for each airway. N = 14 for controls and 20 for heaves. On average, a greater percentage of submucosal thickness was occupied by ASM in smaller airways of horses with heaves compared to controls. This tendency reverses as the airway size increases. However, as the submucosal thickness is increased in horses with heaves compared to controls, the absolute thickness of the ASM layer still remains greater in horses with heaves than in controls for airways with Pi<26mm.
Fig 10.
Composition of L2 is similar in asthmatic and control horses.
The inner 25% is occupied by ECM while the remaining outer 75% is occupied by ASM. Notice that most of the ECM thickness lies behind the echo of L1.