INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chronic lung disease of early infancy, often called bronchopulmonary dysplasia (BPD), is a major cause of morbidity and mortality in premature infants (1, 2). The pathology of this condition includes nonuniform patterns of inflation, saccular appearance of the distal air spaces, accumulation of elastic fibers in the walls of the distal air spaces, and lung inflammation and edema (3, 4). The causes of these pathologic changes have not been identified, but they are associated with the use of prolonged mechanical ventilation and supplemental oxygen.

Considerable insight into the causes of BPD has been provided by studies with primates. Coalson and colleagues (5) found that alveolar formation was impaired in premature baboons that were ventilated with 100% oxygen for at least 1 wk, followed with approximately 80% oxygen for two more weeks, compared with other premature baboons that were initially ventilated with 100% oxygen and subsequently maintained with sufficient inspired oxygen to keep normal arterial oxygen tension (60 to 80 mm Hg). They also found nonuniform inflation, accumulation of extracellular matrix components, fibrosis, metaplasia and dysplasia of airway epithelial cells, and inflammation and edema. These experiments emphasized the important contribution of sustained hyperoxia to alterations in lung development.

A goal of our study was to see whether alterations in pulmonary development occur after prolonged mechanical ventilation of immature lungs in an experimental model designed to achieve normal respiratory gas exchange without excessive oxygen exposure. Another goal was to test the hypothesis that differences in tidal volume (VT) affect the alterations in pulmonary development. Our idea was that prolonged positive-pressure ventilation with a small VT might reduce lung injury by minimizing lung stretch and overexpansion. Two mechanical ventilation strategies were compared: slow, deep ventilation versus rapid, shallow ventilation. Prolonged mechanical ventilation, in the absence of arterial hyperoxia, resulted in structural abnormalities in the lungs of prematurely delivered lambs. Ventilation strategy had modest effects on structural abnormalities, and surfactant replacement at birth did not prevent disordered lung development.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Groups

Two groups of lambs were delivered prematurely at 125 ± 4 d of gestation (term = 147 d) and were mechanically ventilated for 3 to 4 wk, either at 20 breaths/min (n = 5; 2.62 ± 0.56 kg body weight at birth) or 60 breaths/min (n = 5; 2.64 ± 0.38 kg body weight at birth; Table 1). The timing of elective delivery coincided with the transition from the saccular stage to the alveolar stage of lung development. This transition begins at approximately 120 d gestation in sheep (6, 7).

Controls for our study were (1) fetal lambs that were killed without prior breathing at 126 ± 6 d of gestation (2.63 ± 0.68 kg body weight; n = 5) and whose lungs were taken without prior air breathing; (2) term lambs that were killed when they were < 1 d old (3.86 ± 0.73 kg body weight; n = 5); and (3) lambs that were born at term and killed when they were 3 to 4 wks old (6.50 ± 1.87 kg body weight; n = 5). The three groups of control lambs were selected to allow comparison of the chronically ventilated preterm lambs with lambs at the same stage of lung development at delivery (fetuses), and with lambs of the same postconception age (term, < 1 d old) or the same postnatal age (term, 3 to 4 wk old) at death.

Preparation of Chronically Ventilated Preterm Lambs

On the day of delivery, the ewes received ketamine hydrochloride (10 to 20 mg/kg intramuscularly; Fort Dodge Laboratories, Fort Dodge, IA), followed by either spinal anesthesia with 1% tetracaine hydrochloride (Sanofi Winthrop Pharmaceuticals, New York, NY) or inhalation anesthesia with 1% halothane (Halocarbon Laboratory, River Edge, NJ), and sterile insertion of catheters into a maternal hindlimb artery and vein. We opened the uterus through a midline abdominal incision and placed catheters into a carotid artery and jugular vein of the fetus. Next we inserted either a 4.0- or 4.5-mm cuffed endotracheal tube, withdrew 10 ml of lung liquid, and replaced it with 10 ml of calf lung surfactant (Infasurf, 35 mg/ml; generous gift of ONY Inc., Amherst, NY). After operative delivery and initial stabilization, the chronically ventilated preterm lambs were managed for 3 to 4 wk at a respirator rate of either 20 breaths/min or 60 breaths/min with a time- cycled, pressure-limited infant respirator (model IV-100B; Sechrist, Inc., Anaheim, CA), with sufficient supplemental oxygen to keep the arterial partial pressure of oxygen (Pa_O2) between 60 and 90 mm Hg. End-expiratory pressure was kept at 5 to 6 cm H₂O, and peak inspiratory pressure (PI) was adjusted to keep the arterial partial pressure of carbon dioxide (Pa_CO2) between 35 and 45 mm Hg.

The preterm lambs received intravenous injections of buprenorphine hydrochloride (0.01 mg/kg; Reckitt & Colman Pharmaceuticals, Richmond, VA) for analgesia, and phenobarbital sodium (5 to 10 mg/ kg; Wyeth Laboratories, Philadelphia, PA) and pentobarbital sodium (5 mg/kg; Anpro Pharmaceuticals, Arcadia, CA) for sedation at frequent intervals, and they received intravenous injections of penicillin G (100 mg/kg every 12 h; Marsam Pharmaceuticals, Cherry Hill, NJ) and gentamicin (2.5 ml/kg daily; Elkins-Sinn, Cherry Hill, NJ) beginning immediately after birth. If signs or symptoms indicative of infection occurred thereafter, alternative broad-spectrum antibiotics were given. Continuous intravenous infusion of solutions containing glucose and protein, supplemented with enteral feedings of ewe's milk delivered through an orogastric tube, provided calories for growth (Table 1). During the week after birth, each lamb had two thoracotomies under general anesthesia (25 µg/kg fentanyl hydrochloride, intravenously; Elkins-Sinn, Cherry Hill, NJ) for surgical ligation of the ductus arteriosus and placement of vascular catheters and a silicone rubber balloon catheter in the right pleural space (8). Vascular and pleural pressures were continuously recorded with a physiologic recorder (model 7; Grass Instruments, Quincy, MA). Vascular catheters were flushed intermittently with saline that contained 10 U/ml of heparin.

Respiratory variables were assessed at weekly intervals using a calibrated pneumotachograph incorporated into a PEDS Pulmonary Evaluation and Diagnosis System (Medical Associated Services, Hatfield, PA), as described by Bhutani and coworkers (9). The variables that we measured were VT, minute ventilation (E), dynamic lung compliance, and expiratory resistance. Calibrated pressure transducers (Statham Instruments, Oxnard, CA) were connected via catheters to a side-port on the endotracheal tube adapter and to the silicone rubber balloon catheter to measure peak PI, positive end-expiratory pressure (PEEP), mean airway pressure (), and pleural pressure. End-expiratory lung volume (functional residual capacity [FRC]) was estimated using a closed-circuit helium-dilution technique (10). Pa_O2, Pa_CO2, and pH were measured with a calibrated blood gas analyzer (model 158; Ciba-Corning, Medfield, MA), and values were corrected for the prevailing body temperature.

At the end of the study, after 3 to 4 wk of continuous mechanical ventilation, the lambs received an intravenous injection of pentobarbital sodium (35 mg/kg), after which the chest was opened to remove the lungs for subsequent microscopy and biochemical measurements.

Preparation of Control Lambs

Fetal lambs. We operatively delivered five fetal lambs that received an anesthetic dose of pentobarbital sodium intravenously prior to lung resection. The fetal lambs were not allowed to breathe, so their lungs were fixed with the lung liquid in situ.

Term lambs < 1 d old and 3 to 4 wk old. We used 10 spontaneously born term lambs that were either < 1 d old (n = 5) or 3 to 4 wk old (n = 5). The lambs received pentobarbital sodium anesthesia intravenously, followed by endotracheal intubation with a 5-mm cuffed endotracheal tube and mechanical ventilation for 30 to 60 min at a rate of 30 breaths/min. The inspired oxygen concentration and inflation pressures were adjusted to maintain Pa_O2 at approximately 90 mm Hg and Pa_CO2 at approximately 35 to 45 mm Hg. The lambs then received pentobarbital sodium intravenously before lung resection at the prevailing peak inflation pressure.

All surgical procedures and experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee at the University of Utah School of Medicine.

Postmortem Studies

We decided a priori to evaluate lung histopathology as close as possible to its in vivo state. Therefore, we double-clamped large pieces of the ventral portion of lung lobes (~ 3 to 4 cm³; Reference 11) at the prevailing peak PI. This procedure retained the gas volume of each lung sample, as well as its vascular and air space contents. We placed the piece from the right middle lobe in Carnoy's fixative for light microscopy, the entire cardiac lobe in glutaraldehyde/paraformaldehyde for transmission electron microscopy, and the piece from the right cranial lobe in liquid nitrogen for biochemical analyses. We used the same lung lobe for each fixation protocol so that we could compare results among the groups of lambs, based on the study by Pinkerton and colleagues (12), who showed that the structure of a particular lung lobe in normal sheep is the same. Our clamping procedure is different from that used by other investigators who have fixed the lungs of preterm animals at a constant liquid instillation pressure (5).

The ventral portion of the right middle lobe was used for quantitative histology. The clamped lung tissue was immersed in Carnoy's fixative (20 ml glacial acetic acid in 60 ml of 100% ethanol) because it rapidly hardens the lung, thereby preserving its three-dimensional configuration. The clamp remained attached for at least 24 h. We used ethanol to dehydrate the clamped lung tissue before cutting it into slices for paraffin embedding. We cut each clamped piece into 3-mm-thick slabs along parasagittal planes (13). Large tissue blocks (2 to 4 cm²; 2 to 3 per lamb) were embedded in paraffin and serially sectioned at 4 µm thickness. Two stains were used to reveal lung structural features: Masson's trichrome stain (Sigma kit HT15; St. Louis, MO) and Hart's elastic fiber stain.

We used the cardiac lobe for transmission electron microscopy. We injected 0.1 ml of the fixative in several locations of the clamped lobe (27-gauge needle) before immersion fixation (2.5% glutaraldehyde/1% paraformaldehyde in Millonig's phosphate buffer, pH 7.4, 310 mOsm/kg H₂O, 4° C for 24 h; Reference 11). Small tissue blocks (1 mm³; 8 to 10 per lamb) were collected by the same sampling method as described previously. The tissue blocks were embedded in epoxy resin and sectioned at 80 nm thickness. Thin sections were cut with the aid of a diamond knife and were counterstained with uranyl acetate and lead citrate. We used an Hitachi H-7100 transmission electron microscope (Nissei Sangyo America, Inc., Mountain View, CA) to observe and photograph the thin sections.

The frozen right cranial lobe was used for biochemical measurements of desmosine (for mature, cross-linked elastin) and hydroxyproline (for collagen). Frozen samples of peripheral parts of the lobe, devoid of central airways and blood vessels and visceral pleura, were hydrolyzed. Desmosine content was measured by radioimmunoassay (14). Hydroxyproline content was measured using a Beckman 6300 amino acid analyzer (14). DNA content was measured by a colorimetric assay (15, 16). Lung concentrations of desmosine and hydroxyproline were expressed relative to DNA content.

Quantitative Histology

Quantitative histology was used to estimate the uniformity of inflation, alveolar number, elastin accumulation in lung parenchyma, and distal airway smooth muscle accumulation. The results reflect two- dimensional area measurements. Three-dimensional (stereological) estimates were not made because the entire lung was not available for volume displacement or structural analysis. We used a systematic sampling method to evaluate random, nonoverlapping calibrated fields (13) for each variable described subsequently. We used a Bioquant True Color Windows Image Analysis System (R&M Biometrics, Nashville, TN) to make the measurements. Tissue sections were analyzed without knowledge of the lamb group from which the tissue was taken.

Nonuniform inflation. We assessed atelectasis and the size of the distal air spaces in each of the lambs as markers of nonuniform inflation. Percent atelectasis was estimated as the fractional area of the lung that was collapsed in at least 10 nonoverlapping microscope fields in one random tissue section per lamb. We defined atelectatic areas as regions where the alveolar walls were folded onto themselves. The reference space was the tissue area in the same calibrated field. This analysis was done by automated video thresholding of stain color (Masson's trichrome stain). The fractional area of atelectasis was calculated by dividing the calibrated pixel area over atelectatic regions by the pixel area over the reference tissue space, a procedure that is analogous to determining volume density by standard point-counting morphometry (13, 17). We verified the pixel area results by manual point counting, using a 100-point, coherent square lattice (13). Both approaches gave comparable results for identical fields (data not shown). The automated video method was more efficient because it was faster and minimized observer bias.

The dimensions of the distal air spaces were measured to estimate lung overexpansion. The area of individual air spaces was measured by automated video thresholding of stain color. A threshold was set for the air spaces to obtain their calibrated pixel area, after which all air space profiles that appeared round (that is, profiles that lacked branches) were selected. Their calibrated pixel area, X-projection length, and Y-projection length were automatically measured. For statistical purposes, we included only circular profiles of air spaces, meaning their X-Y projection length ratio was between 0.8 and 1.2 (1.0 would represent a circle). We did this to avoid skewing the results by inclusion of longitudinal or oblique planes of section. We analyzed between 200 and 400 distal air spaces in nonoverlapping fields in one random tissue section per lamb.

Alveolar number. Alveolar number across terminal respiratory units was estimated by the radial alveolar count method described by Emery and Mithal (18). For this method, a line was drawn from the center of a respiratory bronchiole to the nearest interlobular septum, to which an intercept line was drawn perpendicularly. We counted the number of distal air spaces that were transected by the line. We repeated this assessment for 10 terminal respiratory units in one random tissue section per lamb.

Alveolar secondary crest volume density. Alveolar secondary crest volume density was estimated by point counting, using a 100-point coherent square lattice. The reference space for the alveolar secondary crest volume density was the parenchymal tissue volume density. Using this reference space minimized the effect of different levels of lung inflation on the results because the air space volume density was excluded. Ten nonoverlapping, calibrated fields were analyzed in one random tissue section per lamb.

Elastic fiber accumulation in the lung parenchyma. Accumulation of elastic fibers in the lung parenchyma was estimated by automated video thresholding of stain color (Hart's elastic fiber stain). We set separate color thresholds for elastic fibers (stained purple) and parenchyma (counterstained yellow) in the inflated areas of the lung tissue sections. The calibrated pixel area for elastic fibers was divided by the calibrated pixel area for parenchyma (the reference space) to calculate the percent area occupied by elastic fibers. We verified the results by conventional point counting (data not shown). We excluded atelectatic areas because anatomic alveoli could not be identified. This approach minimized the influence of variations in lung inflation among the experiments. We also excluded airways, pulmonary arteries and veins, interlobular septa, and visceral pleura. We evaluated at least 10 nonoverlapping microscope fields in one random tissue section per lamb.

Smooth muscle accumulation around distal airways. Smooth muscle of distal airways was assessed by measuring the area occupied by smooth muscle cells and expressing this area, as a ratio, relative to the measured external diameter of the airway. The latter measurement was useful for verifying that the airways were the same size. We reasoned that the greatest effect of prolonged mechanical ventilation on airway smooth muscle accumulation would be in bronchioles; we therefore analyzed terminal bronchioles by computerized morphometry. The measurements of smooth muscle area and airway external diameter were made on circular profiles of airways that were < 300 µm in external diameter. Circular profiles were defined as having X-Y projection length ratios that averaged 1.0 ± 0.2, as was done for the peripheral air spaces. This criterion avoided inclusion of oblique planes of section, which often had uneven profiles of smooth muscle. Once this shape criterion was satisfied, we selected only those terminal bronchiolar profiles that had a smooth lining of epithelial cells; airways with a folded epithelial lining were excluded to minimize the effect of airway narrowing on the smooth muscle measurements. An average of 15 circular profiles of terminal bronchioles were analyzed per lamb.

Statistical Analysis

Data are presented as the mean ± 1 standard deviation. Respiratory variables for the chronically ventilated preterm lambs were compared within groups using the Wilcoxon signed-rank test, and between groups using the Mann-Whitney U test (19). For the quantitative histologic variables, we used one-way analysis of variance (ANOVA) and Student-Newman-Keuls' multiple comparison tests to identify differences between the experimental and control groups (19). Unpaired t tests were used to identify differences for the quantitative histologic variables between the two ventilation strategies (19). Statistical significance was accepted as p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Respiratory Variables for the Chronically Ventilated Preterm Lambs

The respiratory variables for the chronically ventilated preterm lambs are summarized in Table 1. By experimental design, VT was significantly greater for the lambs that were managed at 20 breaths/min (15 ± 2 ml/kg) than it was for lambs that were managed at 60 breaths/min (6 ± 1 ml/kg; p < 0.05). We adjusted the fractional inspired oxygen concentration (FI_O2) and peak PI to maintain Pa_O2 between 60 and 90 mm Hg and Pa_CO2 between 35 and 45 mm Hg, respectively, for the entire 3- to 4-wk period. There were no differences in FI_O2, peak PI, PEEP, or within either group between postnatal Weeks 1 and 3 (Table 1), nor were there differences in these respiratory variables between the two groups at postnatal Week 3 (Table 1) except for , which was 2 cm H₂O higher in the 20 breaths/min group compared with the 60 breaths/min group. There was a tendency for peak PI to be greater at Week 3 of mechanical ventilation for the 20 breaths/min group compared with the 60 breaths/min group, but because variability was large the difference was not significant. E was the same between the two groups at Week 1 or 3, and it increased for both groups from Week 1 to Week 3 (data not shown). Dynamic lung compliance increased from Week 1 to Week 3 for both groups of lambs (data not shown). Lung resistance (RL) remained constant from Week 1 to Week 3 for the preterm lambs that were managed at 20 breaths/min whereas RL decreased during the same 3-wk period for the preterm lambs that were managed at 60 breaths/min (data not shown). We did not compare compliance or resistance between the two groups because the measurements were made at different ventilation rates and VT. FRC was the same between the two ventilation groups, averaging 13 ± 2 ml/kg and 14 ± 6 ml/kg for preterm lambs that were managed at 20 breaths/min (n = 3) and 60 breaths/min (n = 5), respectively.

Pulmonary Histopathology in Chronically Ventilated Preterm Lambs

Compared with control lambs, the major histopathologic findings in the lungs of the chronically ventilated preterm lambs were: nonuniform inflation, decreased radial alveolar counts and alveolar secondary crest volume density, increased amount and abnormal distribution of elastic fibers, greater smooth muscle around terminal bronchioles, and inflammation and edema (Figures 1-11 and Table 2).

View larger version (139K): [in this window] [in a new window]   View larger version (136K): [in this window] [in a new window]   View larger version (153K): [in this window] [in a new window]   View larger version (144K): [in this window] [in a new window]   Figure 1.   Macroscopic view of lung tissue sections showing the effect of prolonged mechanical ventilation on inflation patterns (trichrome stain). The panels are the same magnification. (a) Chronically ventilated preterm lamb that was managed at 20 breaths/min for 3 wk. (b) Chronically ventilated preterm lamb that was managed at 60 breaths/min for 3 wk. (c) Control term lamb < 1 d old and matched for developmental age. (d ) Control term lamb that was 3 wk old and matched for postnatal age. Nonuniform inflation is present as regions of atelectasis (arrows) mixed with regions of distended distal air spaces in the lungs of the chronically ventilated preterm lambs (a and b) when compared with the lungs of the control lambs (c and d ).

View larger version (95K): [in this window] [in a new window]   View larger version (107K): [in this window] [in a new window]   View larger version (110K): [in this window] [in a new window]   View larger version (95K): [in this window] [in a new window]   Figure 2.   Lung histology illustrating the effect of prolonged mechanical ventilation on the loss of alveoli and alveolar secondary crests in terminal respiratory units (trichrome stain). The panels are the same magnification. (a) Chronically ventilated preterm lamb that was managed at 20 breaths/min for 3 wk. (b) Chronically ventilated preterm lamb that was managed at 60 breaths/min for 3 wk. (c) Control term lamb < 1 d old and matched for developmental age. (d ) Control term lamb that was 3 wk old and matched for postnatal age. The terminal respiratory units (TRU) are distended, and there are few alveolar secondary crests (arrows) in the lungs of the chronically ventilated preterm lambs (a and b) when compared with the lungs of the control lambs (c and d ).

View larger version (14K): [in this window] [in a new window]   Figure 3.   Summary of quantitative histology demonstrating the effect of prolonged mechanical ventilation on lung inflation patterns. Atelectasis (left graph) and distal air space size (right graph) are significantly greater for both groups of chronically ventilated preterm lambs (a) compared with both groups of control term lambs. Atelectasis is significantly greater in the chronically ventilated preterm lambs that were managed at 60 breaths/min (b) when compared with those that were managed at 20 breaths/min. Data are mean ± SD; n = 5 lambs/group. R = ventilator rate (breaths/min). a = Significant difference (p < 0.05) compared to control term by ANOVA and Student- Neuman-Keuls tests; b = significant difference (p < 0.05) compared to R20 group of preterm lambs by unpaired t test.

View larger version (131K): [in this window] [in a new window]   View larger version (102K): [in this window] [in a new window]   Figure 4.   Transmission electron micrographs illustrating the effect of prolonged mechanical ventilation on alveolar secondary crest structure. Both panels are the same magnification. (a) Chronically ventilated preterm lamb that was managed at 20 breaths/min for 3 wk. (b) Control term lamb < 1 d old and matched for developmental age. The alveolar secondary crest of the chronically ventilated preterm lamb is thicker and lacks capillaries (C ) at its tip when compared with the alveolar secondary crest of the control term lamb.

View larger version (15K): [in this window] [in a new window]   Figure 5.   Summary of quantitative histology showing the effect of prolonged mechanical ventilation on alveolar formation. Radial alveolar count (left graph) and alveolar secondary crest volume density (right graph) are significantly less for both groups of chronically ventilated preterm lambs (a) compared with the two groups of control term lambs. Both structural variables for the chronically ventilated preterm lambs are equivalent to the respective variables for fetal lambs that were delivered at the same gestational age as the chronically ventilated preterm lambs. Radial alveolar count and alveolar secondary crest volume density increased from the fetal lambs to the control term lambs (b), as expected. Data are mean ± SD; n = 5 lambs/group. R = ventilator rate (breaths/min). a = Significant difference (p < 0.05) compared to control term lambs by ANOVA and Student-Newman-Keuls tests; b = significant difference (p < 0.05) compared to control fetal lambs by ANOVA and Student-Newman-Keuls tests; c = significant difference (p < 0.05) compared to R60 group of preterm lambs by unpaired t test.

View larger version (63K): [in this window] [in a new window]   View larger version (83K): [in this window] [in a new window]   View larger version (85K): [in this window] [in a new window]   View larger version (79K): [in this window] [in a new window]   Figure 6.   Tissue sections of terminal respiratory units (TRU) showing the effect of prolonged mechanical ventilation on elastic fiber accumulation (stained black with Hart's elastic fiber stain) in the walls of the distal air spaces. The panels are the same magnification. (a) Chronically ventilated preterm lamb that was managed at 20 breaths/min for 3 wk. (b) Chronically ventilated preterm lamb that was managed at 60 breaths/min for 3 wk. (c) Control term lamb < 1 d old and matched for developmental age. (d ) Control term lamb that was 3 wk old and matched for postnatal age. Elastic fibers accumulated as thick, disorganized plates and bands (arrows) in the distal air space walls of the chronically ventilated preterm lambs. In the control lambs, elastic fibers are visible as punctate dots (arrowheads) at the tips of the developing alveolar secondary crests.

View larger version (167K): [in this window] [in a new window]   View larger version (135K): [in this window] [in a new window]   Figure 7.   Transmission electron micrographs illustrating the effect of prolonged mechanical ventilation on elastic fiber accumulation in alveolar secondary crests. (a) Chronically ventilated preterm lamb that was managed at 20 breaths/min for 3 wk. (b) Control term lamb < 1 d old and matched for developmental age. Compared with the alveolar secondary crest of the control term lamb, the alveolar secondary crest of the chronically ventilated preterm lamb has thick, long elastic fibers (arrows) that are distributed parallel to the alveolar epithelial cells and throughout the core of the crest, fields of collagen (arrowhead) that are interposed with the elastic fibers, and interstitial edema.

View larger version (19K): [in this window] [in a new window]   Figure 8.   Summary of quantitative histology showing the effect of prolonged mechanical ventilation on elastic fiber accumulation in the walls of the distal air spaces. Elastic fiber accumulation is significantly greater for both groups of chronically ventilated preterm lambs (a) than for the control fetal and term lambs. Elastic fiber accumulation is significantly greater in the 20 breaths/min group of chronically ventilated preterm lambs (b) compared with the 60 breaths/min group of chronically ventilated preterm lambs. There is progressive accumulation of elastic fibers from the control fetal stage of development to 3 wk after birth at term, as expected. Data are mean ± SD; n = 5 lambs/group. R = ventilator rate (breaths/min). a = Significant difference (p < 0.05) compared to control lambs by ANOVA and Student-Newman-Keuls test. b = significant difference (p < 0.05) compared to R60 group of preterm lambs by unpaired t test.

View larger version (103K): [in this window] [in a new window]   View larger version (91K): [in this window] [in a new window]   View larger version (98K): [in this window] [in a new window]   View larger version (95K): [in this window] [in a new window]   Figure 9.   Tissue sections of lung showing the effect of prolonged mechanical ventilation on airway smooth muscle thickness around terminal bronchioles (Hart's elastic fiber stain). The panels are the same magnification. (a) Chronically ventilated preterm lamb that was managed at 20 breaths/min for 3 wk. (b) Chronically ventilated preterm lamb that was managed at 60 breaths/min for 3 wk. (c) Control term lamb < 1 d old and matched for developmental age. (d ) Control term lamb that was 3 wk old and matched for postnatal age. The width of the smooth muscle layer (opposed arrows in each panel) of terminal bronchioles (TB) is greater for the chronically ventilated preterm lambs (a and b) compared with the control term lambs (c and d ). In the chronically ventilated preterm lambs, the airway smooth muscle is bounded by elastic fibers (stained black).

View larger version (21K): [in this window] [in a new window]   Figure 10.   Summary of quantitative histology showing the effect of prolonged mechanical ventilation on smooth muscle accumulation around terminal bronchioles. The results are expressed as the ratio of airway smooth muscle area to airway external diameter. More airway smooth muscle is present in the terminal bronchioles of the chronically ventilated preterm lambs (a) compared with the control fetal and term lambs. The amount of airway smooth muscle remained constant in the three groups of control lambs. Data are mean ± SD; n = 5 lambs/group. TB = terminal bronchiole; R = ventilator rate (breaths/min). a = Significant difference (p < 0.05) compared to control lambs by ANOVA and Student-Newman- Keuls tests.

View larger version (133K): [in this window] [in a new window]   View larger version (124K): [in this window] [in a new window]   View larger version (92K): [in this window] [in a new window]   Figure 11.   Histopathology showing inflammation, edema, and overexpanded air spaces in lung tissue of preterm lambs that were mechanically ventilated for 3 wk. (a) Chronically ventilated preterm lamb that was managed at 60 breaths/min. Inflammatory cells, such as neutrophils (N), alveolar macrophages (AM), and mononuclear cells (M), are suspended in alveolar edema liquid that is stained dark (asterisk). The dark staining presumably reflects a high concentration of protein in the liquid. (b) Chronically ventilated preterm lamb that was managed at 20 breaths/min. Interstitial edema (asterisk) and dilated lymphatics (L) are present in an interlobular septum. (c) Chronically ventilated preterm lamb that was managed at 20 breaths/min. Some distal air spaces subjacent to the visceral pleura are greatly overexpanded (asterisk).

Inflation is nonuniform. Lung inflation was nonuniform in both groups of chronically ventilated preterm lambs compared with the control lambs that were matched for developmental age and postnatal age (Figure 1). We did not assess atelectasis or distal air space size in the lungs of the control fetal lambs because their potential air spaces were filled with lung liquid. The nonuniformity of lung inflation in the chronically ventilated preterm lambs was visible as atelectasis mixed with overexpansion. Regions of atelectasis were irregularly shaped and clearly demarcated from the surrounding parenchyma, which often appeared overdistended. In the overdistended regions of the lungs, the distal air spaces were dilated and appeared as saccules (Figure 2). Computer-aided image analysis showed that atelectasis and distal air space size were significantly greater in the chronically ventilated preterm lambs that were managed at 60 breaths/min compared with those managed at 20 breaths/min (p < 0.05; Figure 3). The differences in distal air space size between the experimental and control groups, and between the two experimental groups, were more clearly evident when we selectively compared the largest quartile results. This analysis showed that the largest individual distal air spaces in the chronically ventilated preterm lambs (20 breaths/min = 14,667 ± 3,448 µm² area; 60 breaths/min = 24,058 ± 6,190 µm² area) were 3 to 4 times bigger than in the control term lambs (< 1-d-old controls = 4,903 ± 916 µm² area; 3-wk-old controls = 5,752 ± 3,991 µm² area).

Alveolar development is abnormal. The large size of the distal air spaces of the chronically ventilated preterm lambs gave their lungs the appearance of being at the saccular, rather than at the alveolar, stage of development (Figure 2). The walls of the terminal respiratory units were generally straight and flat because the units were not partitioned by long, slender alveolar secondary crests, as noted in the lungs of the control term lambs (Figure 4). Instead, many of the alveolar secondary crests were short and blunt. Some distal air spaces in the chronically ventilated preterm lambs lacked alveolar secondary crests altogether. Capillaries were located only at the base of the rudimentary alveolar secondary crests (Figure 4). They did not extend to the tip, as was the case in the alveolar secondary crests of the developmental and postnatal age-matched controls (Figure 4). Quantitative histologic results showed that radial alveolar counts and alveolar secondary crest volume density were significantly less in the lungs of the chronically ventilated preterm lambs compared with the lungs of the control term lambs (Figure 5; p < 0.05). Our estimates of radial alveolar counts for the control fetal and term lambs are consistent with previous reports (6, 7). The only statistical difference between the two ventilation strategies was the volume density of alveolar secondary crests, which was greater in the 60 breaths/ min group (0.067 ± 0.016) compared with the 20 breaths/min group (0.030 ± 0.023; p < 0.05). This difference appeared to derive from the presence of longer alveolar secondary crests in the 60 breaths/min group (Figure 4) because the number of alveoli spanning terminal respiratory units (radial alveolar counts) was the same for the two ventilation strategies (Figure 5). This effect of prolonged mechanical ventilation on radial alveolar count and alveolar secondary crest volume density was striking: both variables for both groups of chronically ventilated preterm lambs were similar to measurements made in lungs of fetal lambs (Figure 5). This finding suggests that prolonged mechanical ventilation severely inhibited alveolar formation.

Elastin accumulation is abnormal. To determine the effect of prolonged mechanical ventilation on elastic fiber accumulation, we measured the area density of elastic fibers in tissue sections. To assess the amount of total cross-linked, mature elastin content in the lung, we measured the desmosine content in lung hydrolysates. Elastic fibers formed thick, disorganized bands in the walls of the distal air spaces in the lungs of the chronically ventilated preterm lambs (Figure 6). These bands frequently were oriented perpendicular to the air space walls, rather than parallel as occurred in the fetal (not illustrated) and control term lambs (Figure 6). The accumulation of thick bands of elastic fibers was conspicuous in the rudimentary alveolar secondary crests of the chronically ventilated preterm lambs (Figure 7). Thick bands of elastic fibers formed disorganized meshworks along the entire extent of the stunted alveolar secondary crests. Interposed among the meshworks of elastic fibers were bundles of collagen fibers. The interstices were expanded by edema liquid. Despite the disorganized accumulation of elastic fibers, individual fibers retained the normal appearance of an amorphous core surrounded by a mantle of microfibrils. These changes in elastic fiber distribution were striking compared with the distribution of elastic and collagen fibers in the lungs of both groups of control term lambs. In the controls, the fibrous framework of the alveolar secondary crests was composed of a mixture of thin filaments of elastic fibers and narrow bundles of collagen fibers. Fine bands of elastic fibers were concentrated near the tip of the crests. Individual elastic fibers had an amorphous central core surrounded by a mantle of microfibrils. Quantitative histologic analysis (Figure 8) showed that elastic fibers occupied a significantly larger percentage of the parenchymal area in tissue sections from the lungs of chronically ventilated preterm lambs that were managed at 20 breaths/min (0.198 ± 0.051) compared with lungs of lambs that were managed at 60 breaths/min (0.110 ± 0.015; p < 0.05). Both groups of chronically ventilated preterm lambs had significantly more elastic fibers in lung tissue sections than the lungs of the control term lambs (Figure 8; p < 0.05). The latter quantitative histologic difference was supported by measurements of desmosine content per µg DNA in hydrolysates of frozen lung tissue from the same lambs (Table 2). Desmosine content was significantly greater in both groups of chronically ventilated preterm lambs compared with the control fetal and 1-d-old term lambs (p < 0.05). However, neither group of chronically ventilated preterm lambs had significantly more desmosine than the control 3-wk-old term lambs (Table 2). We suspect that this lack of difference, as compared with the quantitative histologic results for elastic fibers in tissue sections, was due to the presence of desmosine in regions other than the distal lung parenchyma, such as the walls of small pulmonary blood vessels and airways, and the interlobular septa, which were not removed before homogenization. The latter compartments were excluded from the quantitative histologic analysis. Desmosine content was not different between the two groups of chronically ventilated preterm lambs, perhaps because of the variability among lambs.

To determine whether the increase in elastin in the lung reflected a generalized increase in extracellular matrix elements, we also assessed the accumulation of collagen by measuring hydroxyproline content per µg DNA in the same samples of lung hydrolysates. However, as shown in Table 2, the only statistical difference was between the preterm lambs that were ventilated at 20 breaths/min (3.1 ± 1.5) compared with the control fetal lambs (1.2 ± 0.3; p < 0.05). Hydroxyproline content was not statistically increased in either group of chronically ventilated preterm lambs compared with both groups of control term lambs. Hydroxyproline content increased during normal development, as others have shown (20).

Smooth muscle accumulation around terminal bronchioles is abnormal. The appearance of the smooth muscle layer of terminal bronchioles is shown in Figure 9. This figure shows, at the same magnification, that the thickness of the airway smooth muscle layer appears greater in the chronically ventilated preterm lambs compared with the control term lambs. This impression was substantiated by quantitative histology (Figure 10). The amount of smooth muscle, expressed as the ratio of airway smooth muscle area/terminal bronchiole external diameter, was significantly greater in both groups of chronically ventilated preterm lambs compared with the three groups of control term lambs (Figure 10; p < 0.05). All three control groups had the same amount of smooth muscle around terminal bronchioles. There was no difference in the amount of smooth muscle around terminal bronchioles between the two ventilation strategies.

Lung inflammation and edema. Inflammation and edema were present in the lungs of the chronically ventilated preterm lambs (Figure 11). Air space inflammation was confined to lung regions that had alveolar edema and therefore were at low volume. Inflammatory cells were alveolar macrophages, neutrophils, and mononuclear cells. The alveolar edema liquid did not contain red blood cells or hyaline membranes. We did not see bacteria in the air spaces. Interstitial edema was prominent, being located in the interlobular septa and the visceral pleura. Dilated lymphatic vessels were notable. We did not see histologic evidence of dysplasia or metaplasia of airway epithelial cells. There were rare clusters of alveolar type II epithelial cells, but there was not evidence of alveolar epithelial cell necrosis or sloughing.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

One goal of our study was to determine whether prolonged mechanical ventilation, without arterial hyperoxia, would result in lung histopathology in preterm lambs that reproduced the histopathologic characteristics of BPD in preterm infants. Each of the structural abnormalities of lung development that we observed in the chronically ventilated preterm lambs is characteristic of the lung histopathology that is seen in preterm infants who die with BPD. For example, Hislop and coworkers (21) showed that preterm infants who required mechanical ventilation for 3 d to 11 wk for hyaline membrane disease had nonuniform inflation, reduced numbers of alveoli, and increased accumulation of elastic fibers in alveolar walls when compared with infants who were born prematurely but did not receive assisted ventilation or have lung disease. Likewise, Chambers and van Velzen (22) showed that extremely immature infants who were mechanically ventilated for 10 d or more had nonuniform inflation, reduced alveolar numbers, greater accumulation of elastic fibers in peripheral lung tissue, and normal airway epithelial cells when compared with lungs of infants who died without ventilatory support. Neither study observed airway epithelial cell dysplasia and metaplasia. Other investigators have made similar histopathologic observations (3, 4, 23). Muscularization of distal airways also has been observed in the lungs of preterm infants who die with chronic lung disease when compared with the lungs of newborn infants who die from other causes (3, 4, 23, 24). Finally, inflammation and edema have been observed in the lungs of immature infants who die with BPD (1, 3, 4, 22, 23, 25). Therefore, our lamb model of chronic lung disease of prematurity reproduced the histopathologic features of BPD.

The pulmonary histopathology in preterm lambs also shares similarities with the histopathology that is seen in the lungs of preterm baboons that are mechanically ventilated for 3 wk (26, 27). The pulmonary histopathologic characteristics in preterm baboons include nonuniform inflation, fewer alveoli, and inflammation and edema. Thus, both animal models reproduce the lung histopathology that is seen in preterm infants who die with BPD.

The pathogenesis of chronic lung disease of prematurity is not known. Major contributors are thought to include pulmonary immaturity, oxygen exposure, and volume trauma induced by mechanical ventilation. Lung immaturity is predisposing. At the time of premature birth, the lungs are at the canalicular or saccular stage of development (5, 21). During both stages of lung development, formation of the gas-exchanging units is initiated. However, the structural and functional development of these rudimentary units is insufficient to sustain life. Interventions, such as oxygen exposure, mechanical ventilation, and surfactant therapy, are required. Of these three interventions, the first two are thought to contribute to chronic lung disease of prematurity.

Prolonged exposure to high concentrations of inspired oxygen is toxic to the lung (28), including to the immature lung of preterm neonates (5). During the organizing, or proliferative, stage of oxygen-induced lung injury there is residual proliferation of alveolar type II epithelial cells, loss of capillary endothelial cells, proliferation of cells in the interstitium, and infiltration of the interstitium by inflammatory cells (25, 29). Accumulation of elastin does not occur. When immature lungs are exposed to high concentrations of inspired oxygen for a prolonged period, there may be dysplasia and metaplasia of airway epithelial cells (1, 5, 29). In addition, alveolar formation is inhibited (5, 21, 22).

Studies using preterm baboons and preterm lambs to investigate the pathogenesis of chronic lung disease of prematurity suggest that prolonged exposure to high concentrations of inspired oxygen may be related to dysplasia and metaplasia of airway epithelial cells. Preterm baboons that were exposed to high concentrations of inspired oxygen (100% oxygen for the first 7 to 10 d of life, followed by approximately 80% oxygen thereafter) subsequently had airway epithelial cell dysplasia and metaplasia (5). On the other hand, preterm baboons that were managed with lower concentrations of inspired oxygen that maintained normal arterial oxygenation had normal airway epithelium (5). Our preterm lambs, which also were managed with lower concentrations of inspired oxygen, had normal airway epithelial cells. Thus, cellular changes in airway epithelial cells in experimental animals with chronic lung disease of prematurity appear to be manifestations of oxygen toxicity. These observations may explain why airway epithelial dysplasia and metaplasia were observed in the original description of BPD by Northway and coworkers (1) but not in more recent studies (21, 22). When BPD was first described, clinical management of preterm infants necessitated high concentrations of inspired oxygen; current clinical practice allows for lower oxygen concentrations, comparable to the levels that we used for the preterm lambs.

The studies using preterm baboons and preterm lambs also provide insight into the role of oxygen in inhibition of alveolar formation. Both species have inhibition of alveolar secondary crest formation, yet the preterm baboons are managed with high concentrations of inspired oxygen whereas the preterm lambs are not. Normally, the saccular air spaces are subdivided into anatomic alveoli by the formation of alveolar secondary crests. Alveolar secondary crests are initially formed as shallow septa that subsequently elongate and attenuate until the air saccules are subdivided into multifaceted alveoli (21, 27). Given that alveolar secondary crest formation is consistently inhibited in preterm baboons and preterm lambs, regardless of the inspired oxygen concentration, high concentrations of oxygen may not be the principal cause of this pathologic change. A contributory role of inspired oxygen cannot be excluded, however, because supplemental oxygen is usually required to keep preterm lambs alive. Nonetheless, these observations for experimental animals provide an explanation for why impaired alveolar formation continues to be a persistent histopathologic feature of BPD in infants, even though the clinical use of supplemental oxygen has changed during the three decades since BPD was originally described (1, 21, 22).

Trauma induced by mechanical ventilation has long been recognized to injure the lung or worsen existing lung injury (32). There is considerable experimental evidence that large lung volume, rather than high intrathoracic pressure, causes acute lung injury (33). Our study was designed to test the hypothesis that differences in VT influence histopathologic outcome of chronic lung injury. Our results showed that large VT causes more lung injury than small VT does. These observations suggest that very small VT may cause even less chronic lung injury. Further studies are needed to test this possibility.

The mechanisms that are responsible for impaired alveolar formation are unknown. One possible mechanism is cyclical stretch and stress caused by mechanical ventilation. There is a growing body of evidence that mechanical stress alters intra- and intercellular signaling of cytokines and growth factors (36). For example, cyclic stretch of pleural mesothelial cells is associated with increased levels of messenger RNA (mRNA) for growth factors (37). Another example is high lung inflation, which is associated with increased mRNA levels for extracellular matrix components, such as procollagens ₁(III) and ₂(IV) and fibronectin, in addition to growth factors (38). An extracellular matrix component that is thought to be especially important during alveolar formation is elastin (39). A number of investigators have suggested that excessive accumulation of elastic fibers in the walls of the peripheral air spaces may play a role in impaired alveolar formation by interfering with the structural formation of alveolar secondary crests (3, 5, 21, 22, 27, 40).

Another interesting observation of our study was that distal airway smooth muscle was significantly increased in both groups of chronically ventilated preterm lambs compared with the control fetal and term lambs. This observation is consistent with a recent report that showed significant muscularization of distal airways in the lungs of preterm infants with chronic lung disease compared with the lungs of newborn infants from 23 wk gestation to term (24). A pathologic increase in airway smooth muscle in preterm infants with chronic lung disease provides a structural basis for the pathophysiologic increase in airways resistance that accompanies BPD (41). We postulate that the increase in smooth muscle around distal airways is a structural response to the prolonged, cyclic stretch that is exerted on the airways by mechanical ventilation. Whether the increase is the result of hyperplasia or hypertrophy has not been determined.

We took several steps to preserve the three-dimensional architecture of the lungs, especially microatelectasis, as closely to the physiologic state as possible. We clamped the lung lobes at the ventilator rate and peak PI that were used for the 3- to 4-wk study period. Furthermore, the lobe that was to be used for quantitative histology was immersed, with the clamp attached, in Carnoy's fixative to rapidly make it rigid, especially the elastic fibers (42). The same steps were taken for the control term lambs. Our fixation approach is different from that used by other investigators who fixed the lungs at a constant liquid instillation pressure to normalize the fixed volume of the lung (5). A limitation of our fixation approach is the potential for more variability in the fixed inflation volume among the lambs. This limitation was not confounding, however, because quantitative histologic differences were detected among the groups. Furthermore, our results overall are consistent with those for preterm baboons that had their lungs fixed at a normalized liquid-filled volume (5, 27). Therefore, our fixation method did not prevent identification of the key histopathologic changes of chronic lung disease of prematurity.

Our hypothesis was that lung stretch, due to overexpansion, was the principal variable that affected lung histopathologic outcome. However, multiple variables over 3 to 4 wk may have been involved, such as infection, nutrition, and influences of sedation and other drugs. Infections were minimized by treating all of the lambs with broad-spectrum antibiotics and using sterile techniques. While infections did occur, they were effectively treated by antibiotic therapy. Nutritional needs were provided both intravenously and enterally, so that each lamb gained weight, albeit at a slower rate than normal for lambs during the 3 to 4 wk of study. Analgesic and sedative drugs were required to conduct these long studies. Such treatment may have influenced outcome; all of the lambs, however, were treated the same. Despite our efforts to standardize the management of our lambs, excluding the mechanical ventilation strategy, it is possible that the myriad of potentially confounding variables influenced at least some of the results.

Yet another methodologic concern is the relevance of our control lambs. Our study design covered a 3- to 4-wk gestational period, starting at approximately 125 d gestation. Therefore, we used fetal lambs that were delivered at the same gestational age, term lambs that were matched for developmental age 3 to 4 wk later, and term lambs that lived for 3 to 4 wk. We did not study normal term lambs that were mechanically ventilated for 3 to 4 wk. We would have preferred to include a group of preterm, mechanically ventilated lambs that required minimal ventilatory support and supplemental oxygen, but this was not possible to accomplish with such immature animals. Thus, the control groups of lambs that we used served as reasonable controls, but they may not have been the optimal control animals.

In summary, the results indicate that we devised a representative animal model of BPD with which to test therapeutic and preventive interventions. The clinical management of preterm lambs was similar to the management of infants who acquire BPD, including prematurity, mechanical ventilation with sufficient oxygen to keep arterial oxygenation normal, surfactant replacement, antibiotic therapy, and intravenous and enteral nutrition. Having established this similarity, we examined the effect of two VT strategies on histopathologic outcome. Prolonged mechanical ventilation with large VT had a greater effect on lung structural abnormalities than did ventilation with smaller VT. We conclude that prolonged mechanical ventilation of preterm lambs disrupts normal lung development, particularly alveolar formation.