Effect of Dexamethasone on Pulmonary Inflammation and Pulmonary Function of Ventilator-dependent Infants with Bronchopulmonary Dysplasia1- 3
MERVIN
c.
YODER, JR., ROSARIO CHUA, and ROBERT TEPPER Introduction
Many premature infants with respiratory distress syndrome fail to resolve this form of acute lung injury and develop bronchopulmonary dysplasia (BPD) (1). In these patients, the inability of the injured lung to heal normally appears to be a primary factor in the development of BPD (2). Healing of injured lung normally requires an influx of inflammatory cells, first neutrophils and later mononuclear cells, to initiate the repair process (3). However, persistence of the acute inflammatory response may interfere with later phases of healing, for example, the persistence of neutrophils has been observed in several pulmonary disorders characterized by progressive lung fibrosis (4, 5). A sustained elevation of both the number of neutrophils and the concentration of neutrophil secretory products has been reported in the tracheobronchial lavage (TBL) fluid of infants with BPD and in ventilatordependent premature infants prior to the onset of BPD (6). Sustained pulmonary inflammation may contribute to ongoing respiratory failure in these patients and further contribute to the progression of BPD. The rationale for adding dexamethasone to the treatment of ventilatordependent premature infants with evolving BPD has been to reduce the inflammatory injury caused by neutrophil products (7). Several controlled studies have demonstrated that dexamethasone can acutely improve pulmonaryfunction and decrease the respiratory support required by ventilator-dependent infants with BPD (8-10). However, these studies have not evaluated whether the observed clinical improvement after dexamethasone was associated with a decline in pulmonary inflammation. Furthermore, changes in respiratory mechanics with dexamethasone treatment have not been adequately evaluated. This placebo-controlled, double-blind 1044
SUMMARY Seventeen ventilator-dependent premature infants with bronchopulmonary dysplasia (BPO) were enrolled In a double-blind, placebo-controlled study to determine the effect of 3 days of Intravenously administered dexamethasone (0.5 mg/kg/day) on pulmonary function, pulmonary inflammation, and the requirement for respiratory support (Flo2, ventilator peak pressure [PPI, and respiratory rate [RRJ). Assessment of pulmonary function Included measurement of FVC, flow at 25% vital capacity (V2S), and static compliance of the respiratory system (Crs), whereas pulmonary inflammation was assessed by the neutrophil count, ratio of elastase/2 x alpha-1-antltrypsln, and the concentrations of albumin and fibronectln in the tracheobronchial lavage (TBl) fluid. After 3 days of placebo treatment there were no significant changes In any of the measured parameters. In contrast, the dexamethasone-treated group demonstrated a significant decrease In respiratory support (F102: 50 versus 36%; PP: 21 versus 16 em H20; RR: 22 versus 14 breaths/min) and improved pulmonary function (ers: 0.63 versus 0.85 mllcm H20/kg; "25: 23 versus 68 mils/kg). In addition, pulmonary inflammation was suppressed in the dexamethasone-treated group (neutrophlls: 23 versus 11 x 104/mg albumin; elastase/2 x alpha-1-antitrypsin: 0.24 versus 0.10; albumin: 7.1 versus 3.5 mg/dl; fibronectin: 33 versus 17 lJ9/mg albumin). We conclude that short-term treatment with dexamethasone Improves pulmonary function and suppresses pulmonary inflammation as well as decreasing the respiratory support required by ventilator-dependent premature infants with BPO. AM REV RESPIR DIS 1991; 143:1044-1 048
study was designed to determine whether the short-term intravenous administration of dexamethasone results in suppression of pulmonary inflammation and improvement in pulmonary mechanics. Pulmonary inflammation was assessed by measuring the concentration of selected cellular and protein constituents in the TBL effluent. Measurements of respiratory mechanics included static compliance and forced expiratory flows as an indirect assessment of respiratory resistance at different lung volumes. Methods Patients with Protocol Seventeen premature infants with a persistent requirement for supplemental oxygen and conventional mechanical ventilation were recruited during an I8-month period from the newborn intensive care units at Indiana University Medical Center. Eligibility criteria were (1) birthweight < 1,500 g; (2) gestational age < 32 wk; (3) postnatal age of at least 30 days; (4) chest radiographic evidence of chronic lung disease (Stage 2 or 3 BPD as described by Northway and coworkers [1]); (5) no evidence of clinical deterioration for at least 5 days prior to entry into the study. This study was approved by the Indiana
University Committee for the Protection of Human Subjects, and parental consent was obtained for all infants enrolled in the study. Seventeen infants were randomly assigned to dexamethasone (n = 9) or placebo (n = 8) treatment groups. For 3 days, infants receivedeither dexamethasone intravenously at a dose of 0.25 mg/kg every 12 h or an equal volume of normal saline (placebo). The study drug was then discontinued. During the treatment period, no major changes were made in the usual standard of care. On the day after the last dose of dexamethasone or placebo, ventilator support and oxygensupplemen-
(Received in original/arm September 17, 1990and in revised form December 26, 1990) 1 From the Department of Pediatrics, Sections of Neonatology and Pulmonology, James Whitcomb RileyHospital for Children, Indiana University School of Medicine, Indianapolis, Indiana. 2 Supported by Research Grants 87-5A,B and 88-6A,B from the RileyMemorial Association and by Clinical Investigator Award HL-01322-04 from the National Institutes of Health. 3 Correspondence and requests for reprints should be addressed to Dr. Mervin C. Yoder, Jr., Department of Pediatrics, Riley Childrens Hospital, 702 Barnhill Drive RR 208, Indianapolis, IN 46223.
EFFECT OF DEXAMETHASONE ON PULMONARY FUNCTION AND INFLAMMATION IN BPD
tation wereweaned as clinicallytolerated. The ventilator settings at the end of this day were recorded for comparison with baseline values. Tracheobronchial lavage effluent was collected and pulmonary functions testing was performed within 24 h prior to initiating dexamethasone or placebo treatment, and again on the day after the last dose of dexamethasone or placebo. Of the nine infants who received dexamethasone, two did not have TBL obtained, and two different infants did not have pulmonary function testing (PFT). Of the eight placebo-treated infants, one did not have sufficient TBL data and one different infant did not have sufficient PFT data for inclusion.
Pulmonary Function 'Jesting Airway function was assessed by forced deflation flow volume curves (11). Flow was measured with a no. 0 Fleisch pneumotachometer (Fleisch, Lausanne, Switzerland) and Validyne differential pressure transducer (± 2 em H 20; Validyne Corp., Northridge, CA) attached between the proximal end of the endotracheal tube and a three-way pneumatically controlled switching valve (Hans Rudolph Co., Kansas City, MO). The flow signal was amplified (Validyne),and the analog signal was electrically integrated to volume (Gould Instruments, Cleveland, OH). The flow and volume signals were digitized and displayed on the CRT of an IBM-AT microcomputer. The infant was manually ventilated with an Ambu bag attached to the second port of the switching valve. The third port of the three-way valve was closed to a IO-L pressure-volume reservoir that had an adjustable negative pressure. Lung volume was inflated to greater than 40 em H 20 pressure (TLC), and then, by activating the pneumatic valve, the endotracheal tube was transiently « 2 s) connected to the negative pressure reservoir to generate the forced expiratory flow volume maneuver. The reservoir was initially set at negative 20 em H 20 pressure, and the forced expiratory flows were repeated at increasing negative pressures until the flow at 25010 of the FVC (\'25) did not increase. This generally did not require pressures more negative than 40 em H 20. The flow-volume curve was quantitated by the FVC and the flows at 50 and 25010 of FVC (\'50 and \'25)' In order to adjust for changes in lung volume during the study period, flow-volume curves obtained on the first and last days of the study were matched at TLC, and maximal expiratory flows were compared at isovolume. In addition, a time constant (Tau) for the respiratory system, which has the units of seconds, was calculated by dividing the FVC (ml) by the flow (\'25 mIls). A static respiratory systemcompliance (Crs) was calculated as the inspiratory volume between passive FRC and TLC, divided by the transrespiratory pressure at TLC. The values for forced expiratory flow,FVC, and Crs were divided by body weight (kg) to normalize for differences in body size.
Lavage Technique Lavage samples were collected as previously described (12). Briefly, 1 ml of sterile saline was instilled into the endotracheal tube, and the infant was manually ventilated for three to fivebreaths. Lavage fluid was aspirated into a sterile Lukens trap. The procedure was repeated three times over a 5-min period for each sample (4 ml total lavage); 1ml of saline was then suctioned directly into the trap to flush the lumen of the catheter, resulting in an average of 3 ± 0.5 ml recovered. The lavage fluid was immediately transferred to a polypropylene tube using a steriledouble-layer gauze pad 2 x 2 to strain mucous material. Additional saline was used as needed to rinse the sides of the Lukens trap. Lavage fluid was then centrifuged at 2,000 rpm for 15min, and cell-free supernatant in 0.25 ml aliquot were frozen at - 70° C for later assays. Cell Studies A total white blood cell count was performed using a hemocytometer. Cell viability was determined by trypan blue exclusion, and a differential count was obtained after WrightGiemsa staining of a cytocentrifuged cell sample. Elastase Assay Effluent elastase concentrations were determined by the hydrolysis of a 0.3 mM synthetic substrate, methoxysuccinyl-Ala-Ala-ProVal-p-nitroanilide (VegaBiotechnologies Inc., Tuscon, AZ) as previouslydescribed (13). Substrate was suspended in 9.8% dimethylsulfoxide and N-2-hYdroxyethylpiperazine-N-2ethanesulfonic acid buffer. Effluent aliquots were incubated at 37° C with the substrate for 30 min prior to the addition of IN acetic acid. Samples were read at 410 nm on a spectrophotometer (Model 25; Beckman Instruments, Irvine, CA) along with a standard curve generated between 1and 6 ug/ml using porcine pancreatic elastase (type III) (Worthington, Freehold, NJ). Standards were suspended in 0.2 M TRIS-O.101o BSA at pH 8. All samples were measured in duplicate. Ftbronectin Assay Fibronectin concentrations in TBL were determined using a modification of a competitive enzyme-linked immunoabsorbent assay as previously described (12).Briefly, effluent aliquots and purified human plasma standards (New York Blood Center, New York, NY) (diluted in PBS-0.05% Tween 20-1% BSA) were incubated with rabbit antihuman fibronectin (Cooper Biomedical, Malvern, PA) for 1 h in a 3-ml polypropylene tube at room temperature. The mixture was then added to polystyrene wells (Dynatech Laboratories, Alexandria, VA)precoated with human plasma fibronectin (diluted in 0.05 M TRIS at pH 9) and blocked with 100 mM glycine diluted in 1% BSA. After a 30-min incubation, the plates were washed with PBS-0.05% Tween, and a 1:1,500dilution of peroxidaselabeled porcine antirabbit antibody (Accurate
1045
Laboratories, Westbury, NY) was added to each test welland allowedto incubate at 20° C for an additional hour. The wells were then rewashed with PBS-0.05% Tween,and a substrate of orthophenylenediamine + 3% hydrogen peroxide + citrate buffer (OPD) was added. The reaction was stopped by adding 0.1 M sodium fluoride to each well, and the optical density of the wells was read at 450 nm on a Microelisa spectrometer (Dynatech). The concentration of fibronectin in the effluent was determined by comparing the optical density of test wells with wells containing known amounts (standards) of fibronectin. All samples were measured in duplicate.
Alpha-I Antitrypsin Assay Alpha-l antitrypsin concentrations were determined using the enzyme-linked immunoabsorbent assay reported by Michalski and coworkers (14). The assay is begun by adding effluent aliquots of human serum alpha-lantitrypsin (Calbiochem-Behring, San Diego, CA) (diluted in PBS-0.05% Tween) to polystyrene plates precoated with rabbit antihuman alpha-l-antitrypsin antibody (Boehringer-Mannheim, Indianapolis, IN) (diluted in NaHC0 3 buffer at pH 9.6) and blocked with 5% casein with 100mM glycine diluted in PBS. After 90 min of incubation at room temperature, the plates were washed with PBS-0.05% Tween,then a 1:2,000dilution of peroxide-labeled goat antihuman alpha-l-antitrypsin antibody (Cooper Biomedical Inc., Malvern, PA) was added to each test welland incubated for another hour at 20° C. The wells were then rewashed with PBS-0.05% Tween and OPD was added to each well. The reaction was stopped with 0.1M sodium fluoride, and the optical density of the wells was read at 450 nm on a Microelisa spectrometer. All samples weremeasured in duplicate. Because 1 mole of alpha-l-antitrypsin neutralizes 1 mole of elastase, the ratio of elastase/2 x alpha-l-antitrypsin was calculated. Albumin Assay Because some variability was encountered in the amount of saline recovered from the amount instilled during the performance of lavage, effluent constituents were expressed in relation to effluent albumin concentration. Use of albumin as a standard for the degree of dilution in lavage techniques is common (15, 16). Albumin concentrations weremeasured by a modification of the direct enzyme-linked immunoabsorbent assay described by Whitfield and Spierto (17). Duplicate effluent samples and human serum albumin standards (Sigma Chemical, St. Louis, MO), all diluted in NaHC03 buffer (pH 9.6)and 0.05% Tween, were incubated with rabbit antihuman albumin antibody (Cooper Biomedical). Polystyrene wells had been precoated with this antibody and blocked with 5% casein and 100 mM glycine in PBS. After allowing to incubate at room temperature for 90 min, the
YODER, CHUA, AND TEPPER
1046
plates were washed with PBS-0.05OJo Tween, A 1:1,500 dilution of peroxide-conjugated goat antihuman albumin antibody (Cappel Laboratories, West Chester, PA) was then added. The plates were rewashed after 1 h at 20° C, and OPD wasadded. The reaction was stopped with 0.1 M sodium fluoride, and the optical density of the wells was read at 450 nm on a Microelisa spectrometer.
Statistical Analysis For pulmonary function data, values analyzed were the highest measurements obtained during the testing period on the first and the last day of the study. Intergroup and intragroup differences in pulmonary function measurements were compared using unpaired and paired t tests, respectively. For tracheobronchial effluent data, values for a givenday were the mean of three samples collected during that day and analyzed separately. The data collected for biochemical and cellular parameters werecompared using the Wilcoxon's nonparametric signed ranks test. A p value < 0.05 was considered statistically significant. Variability about the mean is presented as the standard error of the mean.
TABLE 1 PATIENT DATA" Placebo (n = 8)
p Value
26.9 ± 1.8 39.1 ± 8.9 1.26 ± 0.20
26.0 ± 1.6 41.3 ± 13.2 1.49 ± 0.48
0.3 0.7 0.2
Gestational age, wk Postnatal age, days Weight at time of study, kg • Values are mean ± SEM.
TABLE 2 VENTILATOR SUPPORT" Dexamethasone Pre FI02
Peak inspiratory pressure, cm H20 Respiratory rate, (breaths/min) Positive end-expiratory pressure, cm H 20
50.4 ± 4 21 ± 2 22 ± 3 4 ± 1
Placebo Pre
Post
36 16 14 4
± ± ± ±
4t 4t 4t 1
Post
41 ± 4 20 ± 1 28 ± 6
44 20 27 5
5 ± 1
± ± ± ±
6 2 6 1
• Values are mean ± SEM.
t p < 0.05 post-treatment compared with
pretreatment values in the dexamethasone group.
TABLE 3 PULMONARY FUNCTION DATA" Dexamethasone (n = 7)
Results
On Day 1 of the study, prior to initiating treatment, there were no statistically significant differences in the dexamethasone and placebo-treated infants for gestational age at birth, age at the time of study, study weight (table 1), or severity of disease as indicated by F102 requirement, ventilator inspiratory pressure, ventilator rate, or positive end-expiratory pressure (table 2). After 3 days of treatment, the clinical status of the patients in the dexamethasone-treated group improved. In this group, there was a significant reduction in the inspired oxygen requirement, ventilator inspiratory pressure, and intermittent mandatory ventilator rate (table 2). No significant changes in any of these parameters were observed in the placebo-treated group. There was no significant change in body weight between Days 1 and 4 for the dexamethasonetreated infants (35 ± 56 g); however, the placebo group of infants exhibited a significant weight gain during the study period (83 ± 81 g). On entry into the study, there were no statistically significant differences in the baseline flow, FVC, or Crs of the respiratory system for the dexamethasone and placebo groups (table 3). The dexamethasone group did have a significantly longer time constant (Tau) than the placebo group (1.92 versus 1.01 s, p < 0.01). After 3 days of placebo treatment, there were no significant changes in any of the pulmonary function parameters. In contrast, after dexamethasone treatment, the
Dexamethasone (n = 9)
Post
Pre FVC, ml/kg Crs, mllcm H2O/kg "25' mils/kg "50' mils/kg
44 0.63 23 74
± ± ± ±
Placebo (n = 7)
10 0.15 10 22
56 0.85 68 111
± ± ± ±
17ft 0.16ft 50t§ 54t
Pre
35 0.48 37 92
± ± ± ±
Post
18 0.28 24 59
35 0.50 28 82
± ± ± ±
16 0.29 14 14
• Values are mean ± SEM. 0.05 < p < 0.10 compared with pre-dexamethasone. :t p < 0.05 compared with post-placebo. § p < 0.05 compared with pre-dexamethasone.
t
respiratory system compliance increased on Day 4, and Crs was greater in the dexamethasone group than in the placebo group. In addition, the increase in FVC (+ 11.9 ± 15.8 mllkg) approached statistical significance, and on Day 4 of the study the dexamethasone group had a larger FVC than did the placebo group. The V25 increased after dexamethasone (p < 0.03) and was higher in the dexamethasone than in the placebo group (p < 0.05) on Day 4 of the study. The increase in V50 did not attain statistical significance (p < 0.1) and it was not significantly different for the two groups after intervention. The time constant decreased significantly in the dexamethasone-treated group (1.92 versus 0.87 s, p < 0.025), and after treatment there was no significant difference in Tau for the dexamethasone and placebo groups (0.87 versus 1.27 s). The protein and cellular concentrations of the TBL effluent were not significantly different for the dexametha-
sone and placebo groups when treatment was initiated (table 4). After 3 days of dexamethasone administration, the total neutrophil count and percent neutrophils in the white blood cell differential were significantly decreased in the dexamethasone-treated infants. No significant change occurred for the placebotreated infants. The effluent albumin concentrations decreased significantly in the dexamethasone-treated infants, but it did not change in the placebo-treated group. After 3 days of dexamethasone the elastase concentration in the effluent decreased nearly 50070 from the pretreatment value, and this change in elastase approached statistical significance (p = 0.07). There was a significant elevation in the concentration of alpha-l-antitrypsin in the dexamethasone-treated group and a significant decrease in the elastase/2 x alpha-l-antitrypsin ratio. In contrast, the placebo-treated group had an increase in elastase concentration and decrease in alpha-l-antitrypsin concen-
1047
EFFECT OF DEXAMETHASONE ON PULMONARY FUNCTION AND INFLAMMATION IN BPD
TABLE 4 TRACHEOBRONCHIAL LAVAGE EFFLUENT DATA* Dexamethasone Pre Elastase, Ilg/mg albumin u-1-antitrypsin, Ilg/mg albumin Elastase/2 x a-t-antitrypsin Fibronectin, Ilg/mg albumin Total neutrophil count x 10"/mg albumin Neutrophils, % Albumin, mg/dl
24 40 0.24 33 23 51 7.1
± ± ± ±
Placebo
Post
11 13 0.08 8 ± 8 ± 9 ± 1.5
13 55 0.10 17 11 23 3.5
± ± ± ± ± ± ±
3t 14* 0.02* 3* 6* 7* 0.8*
Pre 18 56 0.16 22 27 41 6.9
± 6 ± 10 ± 0.10 ± 7 ± 11 ± 6 ± 3.7
Post 24 38 0.26 25 25 53 4.3
± 9
± 11 ± 0.06 ± 5 ± 6 ± 8 ± 1.3
• Values are mean ± SEM.
t 0.05 < p < 0.10.
*< P
0.05 post-treatment compared with pretreatment values.
tration, although these changes were not statistically significant. These combined changes in the placebo group did result in a significant increase in the elastase/2 x alpha-2-antitrypsin ratio between Days 1and 4. Fibronectin concentrations were significantly reduced after 3 days of dexamethasone treatment; however, the placebo group had no significant change in fibronectin concentration. No complications of sepsis, hypertension, hyperglycemia, or electrolyte imbalance were seen in the dexamethasonetreated infants. All infants enrolled completed the study. One infant in the placebo group died of peritonitis from intestinal perforation 4 days after completing the study. All infants in the dexamethasone group stopped demonstrating improvement in their respiratory status within several days of drug discontinuation, and four infants reverted back to the pretreatment oxygen and ventilator requirements. Discussion
Three days of intravenous dexamethasone treatment improved respiratory mechanics and decreased the level of respiratory failure in infants with BPD. Dexamethasone also suppressed TBL effluent indices of pulmonary inflammation in treated infants compared with those in placebo-treated control infants. The results of this prospective controlled study confirm the clinical efficacy of dexamethasone in the treatment of infants with BPD and implicate inflammation as an important factor contributing to the amount of respiratory support required by these ventilator-dependent infants (7-9). In our study, the low values for static compliance are in agreement with those previously reported in infants with BPD (8, 18). Treatment with dexamethasone resulted in a significant improvement in the static compliance of the respiratory system and therefore reflects an improve-
ment in the elastic properties of the parenchyma. Previous studies evaluating dexamethasone treatment have measured dynamic pulmonary compliance (Cdyn) with an esophageal catheter (8, 18). In the presence of diffuse airway obstruction, Cdyn is dependent upon the respiratory frequency at which it is measured. With increasing respiratory rates, Cdyn progressivelydeclinesand underestimates the static compliance of the respiratory system. Because of this limitation, the previouslyreported increase in Cdyn with dexamethasone treatment was unable to discriminate whether the observed changes were secondary to a decline in peripheral airway obstruction or to an improvement in the elastic properties of the parenchyma. Our findings of an improvement in the static compliance of the respiratory system imply that the increase in Cdyn is at least partly attributable to changes in the elastic properties of the parenchyma. The infants enrolled in this study had evidence of obstructive airway disease with forced expiratory flows at low lung volumes (V2S: 23 to 37 mIls/kg) that were significantly lower than values reported for "normal" premature infants (95 ± 13 mIls/kg) and similar to previously reported values for ventilator-dependent premature infants with BPD (11). The abnormal flows at low lung volumes and concave shape of the flow volume curves observed for many of the infants suggest the presence of peripheral airway obstruction, a finding consistent with the necrotizing bronchiolitis reported in pathologic specimens from infants with BPD (1). Dexamethasone-treated infants demonstrated a significant improvement in airway function as evidenced by the increase in V2 S • Dexamethasone treatment also produced a significant decline in the time constant (Tau) for the respiratory system. Although calculated from flow and volume, Tau is also the product
of the resistance (R) and the compliance (C) of the respiratory system (Tau = R x C). Because compliance increased after dexamethasone treatment, the decrease in Tau reflects a marked decrease in respiratory resistance. This finding is in agreement with the reported decline in pulmonary resistance after dexamethasone treatment (18). The dexamethasone-treated group had a significant decrease in the concentration of albumin in the TBL effluent. This did not occur in the placebo group. As albumin is a serum protein with a low molecular weight, it has been used as one index of microvascular leak into the alveoli (19). Therefore, the lower albumin concentration in the TBL fluid after dexamethasone treatment may reflect less microvascular leak. Analysis of an undiluted TBL sample to compare lavage albumin/serum albumin would have provided stronger support that dexamethasone had diminished microvascular leak. If dexamethasone decreases pulmonary extravascular fluid, then there should be an association between improved pulmonary mechanics and diuresis as has been reported by Gladstone and coworkers (18).Although wedid not examine fluid balance in our study infants, increased urine output in response to steroid intervention is a plausible explanation for the absence of significant weight gain in our dexamethasonetreated infants in contrast to the significant increase in weight in the placebo group. These findings are consistent with the postulate that dexamethasone treatment reduces pulmonary interstitial fluid by altering membrane permeability and thereby results in improved pulmonary mechanics. In respiratory distress syndromes, neutrophils are believed to contribute to the pathophysiology of pulmonary microvascular injury and altered membrane permeability. Pretreatment TBL effluent obtained from all study infants contained concentrations of neutrophils, elastase, and fibronectin similar to those previously published for infants with BPD (6, 7, 20, 21). The TBL effluent concentration of neutrophils and the protease/antiprotease ratio decreased in our study infants after treatment with dexamethasone. The concentration of neutrophil-derived elastase also decreased in all but two of the treated infants, though no statistical difference in pretreatment and post-treatment concentrations was observed. Our findings are consistent with Gerdes and coworkers (7) who reported that 48 to
1048
72 h of dexamethasone treatment decreases neutrophil elastase concentrations and the elastase/2 x u-l-proteinase inhibitor ratio in the tracheobronchial effluent of patients with BPD. In our study, fibronectin concentrations also decreased in the dexamethasone-treated infants. In contrast, Gerdes and coworkers (7) did not observe an effect of dexamethasone treatment on TBL concentrations of fibronectin. Differences in the results of these two studies may be related to the fact that Gerdes and coworkers (7) enrolled younger infants who had lower baseline concentrations of fibronectin and were treated with a higher concentration of dexamethasone for a shorter period of time. The anti-inflammatory effects of dexamethasone, which result in lower numbers of neutrophils and their secretory products in the TBL effluent of these infants with BPD, may result from effects of dexamethasone on either the phagocytic cellsor the microvascular endothelial cells. Although in vitro or in vivo exposure of neutrophils to dexamethasone inhibits neutrophil adherence to endothelial cells, the predominant in vivo effect of dexamethasone may be to inhibit transendothelial emigration of neutrophils (22). It remains unclear whether inhibition of neutrophil emigration results from decreased local production of inflammatory mediators (plateletactivating factor, arachidonic acid metabolites, complement fragments), which stimulate endothelial cells to enhance neutrophil-endothelial adhesive interactions, or whether the neutrophil is directly inhibited in its capacity to adhere to and then emigrate into the alveolar space (23). Several inflammatory mediators (platelet-activating factor, arachidonic acid metabolites) havebeen identified in the TBL fluid of infants with BPD (24, 25). These factors have been proposed to influence pulmonary airway function and alveolar epithelial permeability as well as neutrophil recruitment in patients with BPD. It has not yet been determined if dexamethasone treatment influences the concentration of these protein and lipid mediators in the TBL fluid of infants with BPD. We conclude that 3 days of intravenously administered dexamethasone in ventilator-dependent premature infants with BPD results in the acute suppres-
YODER, CHUA, AND TEPPER
sion of pulmonary inflammation, improved pulmonary mechanics, and a decrease in the requirement for supplemental oxygen and assisted ventilation. We speculate that the failure to see sustained improvement in these patients after dexamethasone was discontinued is related to return of the inflammatory state and reversion of the lung to the level of pretreatment dysfunction. This would be consistent with the observations of Cummings and coworkers (10) who found that early and long-term steroid treatment was required to maintain clinical benefits in infants with BPD. Because prolonged treatment with dexamethasone results in significant side effects, further evaluations of the mechanisms of steroid action and optimal treatment strategies for BPD prevention in high risk ventilatordependent premature infants are required. Because pulmonary inflammation has been proposed as one of the pathophysiologic factors leading to the development of BPD, perhaps earlier treatment will result in a decrease in the severityof pulmonary dysfunction in susceptible patients (26). References 1. Northway WH, Rosan RC, Porter DY.Pulmonary disease following respirator therapy of hyalinemembrane disease: bronchopulmonary dysplasia. N Engl J Med 1967; 276:357-68. 2. Goetzman BW.Understanding bronchopulmonary dysplasia. Am J Dis Child 1986; 140:332-4. 3. Clark RAE Overview and general considerations of wound repair. In: Clark RAF, Henson PM, eds. The molecular and cellular biology of wound repair. New York: Plenum Press, 1988; 3-33. 4. Haslett C, Henson PM. Resolution of inflammation. In: Clark RAF, Henson PM, eds. The molecular biology and cellular biology of wound repair. New York: Plenum Press, 1988; 185-212. 5. Crystal RG, Bitterman PS, Rennard SI, Hance AJ, Keogh BA. Interstitial lung disease of unknown cause. N Engl J Med 1984; 310:154-66. 6. Merritt TA, Cochrane CG, Holcomb K, et al. Elastase and alpha-1 proteinase inhibitor activity in tracheal aspirates during respiratory distress syndrome. J Clin Invest 1983; 72:656-66. 7. Gerdes JS, Harris MC, Polin RA. Effects of dexamethasone and indomethacin on elastase, n-l proteinase inhibitor, and fibronectin in bronchoalveolar lavage fluid from neonates. J Pediatr 1988; 113:727-31. 8. Mammel MC, Green TP, Johnson DE, Thompson TR. Controlled trial of dexamethasone therapy in infants with bronchopulmonary dysplasia. Lancet 1983; 1:1356-8. 9. AveryGB, Fletcher AB, Kaplan M, Brudno DS. Controlled trial of dexamethasone in respiratordependent infants with bronchopulmonary dysplasia. Pediatrics 1985; 75:106-11. 10. Cummings 11, D'Eugenio DB, Gross SJ. A
controlled trial of dexamethasone in pre-term infants at high risk for bronchopulmonary dysplasia. N Engl J Med 1989; 320:1505-10. 11. Motoyama EK, Fort MD, Klesh K, Mutich RL, Guthrie RD. Early onset of airway reactivity in premature infants with bronchopulmonary dysplasia. Am Rev Respir Dis 1987; 136:50-7. 12. Gerdes JS, Yoder MC, Douglas SD, Paul M, Harris MK, Polin RA. Tracheal lavage and plasma fibronectin: relationship to respiratory distress syndrome and development of bronchopulmonary dysplasia. J Pediatr 1986; 108:601-6. 13. Castillo MJ, Nakajima K, Zimmerman M, Powers JC. Sensitive substrates for human leukocyte and porcine pancreatic elastase: a study of the merits of various chromophoric and fluorogenic leaving groups in assays for serine proteases. Anal Biochem 1979; 99:53-64. 14. Michalski JP, McCombs CC, Sheth S, McCarthy M, DeShago R. A modified double antibody sandwich enzyme-linked immunoabsorbent assay for measurement of alpha-l-antitrypsin in biologic fluids. J Immunol Methods 1985; 83:101-12. 15. Cherniak RM. Proteins in bronchoalveolar lavage fluid. Am Rev Respir Dis 1990; 141(5 Part 2:5183-8). 16. Young KR, Reynolds HY. Bronchoalveolar washings: proteins and cells from normal lungs. In: Bienenstock J, ed. Immunology of the lung and upper respiratory tract. New York: McGraw-Hill, 1984; 157-73. 17. Whitfield W, Spierto E Modified ELISA for the measurement of urinary albumin. Clin Chern 1986; 32:561-3. 18. Gladstone I, Ehrenkranz R, Jacobs H. Pulmonary function tests and fluid balance in neonates with chronic disease during dexamethasone treatment. Pediatrics 1989; 84:1072-6. 19. BlaschkeE, Eklund A, Hembrand R. Extracellular matrix components in bronchoalveolar lavage fluid in sarcoidosis and their relationship to signs of alveolitis. Am Rev Respir Dis 1990; 141:1020-5. 20. Ogden BE, Murphy SA, Saunders OC, Pathak D, Johnson JD. Neonatal lung neutrophils and elastase/proteinase inhibitor imbalance. Am Rev Respir Dis 1984; 130:817-21. 21. Merritt TA, Hallman M. Interactions in the immature lung: protease-antiprotease mechanism of lung injury. In: Merritt TA, Northway WH, Boynton BR, eds. Bronchopulmonary dysplasia. Boston: Blackwell Scientific Publishers, 1988; 117-30. 22. Schleimer RP. Effects of glucocorticoid on inflammatory cells relevant to their therapeutic applications in asthma. Am Rev Respir Dis 1990; 141 (2 Part 2:559-69). 23. Williams TJ, Yarwood H. Effect of glucocorticoid on microvascular permeability. Am Rev Respir Dis 1990; 141(2 Part 2:539-43). 24. Stenmark KR, Eyzaguirre M, Wescott JY, Murphy PM. Potential role of eicosanoids and PAP in the pathophysiology of bronchopulmonary dysplasia. Am Rev Respir Dis 1987; 136:770-2. 25. Mirro R, Armstead W, Leffler C. Increased airway leukotriene levelsin infants with severebronchopulmonary dysplasia. Am J Dis Child 1990; 144:160-1. 26. VanMarter LJ, Leviton A, Kuban KC, Pagano M, Allred KN. Maternal glucocorticoid therapy and reduced risk of bronchopulmonary dysplasia. Pediatrics 1990; 86:331-6.
MERVIN
c.
YODER, JR., ROSARIO CHUA, and ROBERT TEPPER Introduction
Many premature infants with respiratory distress syndrome fail to resolve this form of acute lung injury and develop bronchopulmonary dysplasia (BPD) (1). In these patients, the inability of the injured lung to heal normally appears to be a primary factor in the development of BPD (2). Healing of injured lung normally requires an influx of inflammatory cells, first neutrophils and later mononuclear cells, to initiate the repair process (3). However, persistence of the acute inflammatory response may interfere with later phases of healing, for example, the persistence of neutrophils has been observed in several pulmonary disorders characterized by progressive lung fibrosis (4, 5). A sustained elevation of both the number of neutrophils and the concentration of neutrophil secretory products has been reported in the tracheobronchial lavage (TBL) fluid of infants with BPD and in ventilatordependent premature infants prior to the onset of BPD (6). Sustained pulmonary inflammation may contribute to ongoing respiratory failure in these patients and further contribute to the progression of BPD. The rationale for adding dexamethasone to the treatment of ventilatordependent premature infants with evolving BPD has been to reduce the inflammatory injury caused by neutrophil products (7). Several controlled studies have demonstrated that dexamethasone can acutely improve pulmonaryfunction and decrease the respiratory support required by ventilator-dependent infants with BPD (8-10). However, these studies have not evaluated whether the observed clinical improvement after dexamethasone was associated with a decline in pulmonary inflammation. Furthermore, changes in respiratory mechanics with dexamethasone treatment have not been adequately evaluated. This placebo-controlled, double-blind 1044
SUMMARY Seventeen ventilator-dependent premature infants with bronchopulmonary dysplasia (BPO) were enrolled In a double-blind, placebo-controlled study to determine the effect of 3 days of Intravenously administered dexamethasone (0.5 mg/kg/day) on pulmonary function, pulmonary inflammation, and the requirement for respiratory support (Flo2, ventilator peak pressure [PPI, and respiratory rate [RRJ). Assessment of pulmonary function Included measurement of FVC, flow at 25% vital capacity (V2S), and static compliance of the respiratory system (Crs), whereas pulmonary inflammation was assessed by the neutrophil count, ratio of elastase/2 x alpha-1-antltrypsln, and the concentrations of albumin and fibronectln in the tracheobronchial lavage (TBl) fluid. After 3 days of placebo treatment there were no significant changes In any of the measured parameters. In contrast, the dexamethasone-treated group demonstrated a significant decrease In respiratory support (F102: 50 versus 36%; PP: 21 versus 16 em H20; RR: 22 versus 14 breaths/min) and improved pulmonary function (ers: 0.63 versus 0.85 mllcm H20/kg; "25: 23 versus 68 mils/kg). In addition, pulmonary inflammation was suppressed in the dexamethasone-treated group (neutrophlls: 23 versus 11 x 104/mg albumin; elastase/2 x alpha-1-antitrypsin: 0.24 versus 0.10; albumin: 7.1 versus 3.5 mg/dl; fibronectin: 33 versus 17 lJ9/mg albumin). We conclude that short-term treatment with dexamethasone Improves pulmonary function and suppresses pulmonary inflammation as well as decreasing the respiratory support required by ventilator-dependent premature infants with BPO. AM REV RESPIR DIS 1991; 143:1044-1 048
study was designed to determine whether the short-term intravenous administration of dexamethasone results in suppression of pulmonary inflammation and improvement in pulmonary mechanics. Pulmonary inflammation was assessed by measuring the concentration of selected cellular and protein constituents in the TBL effluent. Measurements of respiratory mechanics included static compliance and forced expiratory flows as an indirect assessment of respiratory resistance at different lung volumes. Methods Patients with Protocol Seventeen premature infants with a persistent requirement for supplemental oxygen and conventional mechanical ventilation were recruited during an I8-month period from the newborn intensive care units at Indiana University Medical Center. Eligibility criteria were (1) birthweight < 1,500 g; (2) gestational age < 32 wk; (3) postnatal age of at least 30 days; (4) chest radiographic evidence of chronic lung disease (Stage 2 or 3 BPD as described by Northway and coworkers [1]); (5) no evidence of clinical deterioration for at least 5 days prior to entry into the study. This study was approved by the Indiana
University Committee for the Protection of Human Subjects, and parental consent was obtained for all infants enrolled in the study. Seventeen infants were randomly assigned to dexamethasone (n = 9) or placebo (n = 8) treatment groups. For 3 days, infants receivedeither dexamethasone intravenously at a dose of 0.25 mg/kg every 12 h or an equal volume of normal saline (placebo). The study drug was then discontinued. During the treatment period, no major changes were made in the usual standard of care. On the day after the last dose of dexamethasone or placebo, ventilator support and oxygensupplemen-
(Received in original/arm September 17, 1990and in revised form December 26, 1990) 1 From the Department of Pediatrics, Sections of Neonatology and Pulmonology, James Whitcomb RileyHospital for Children, Indiana University School of Medicine, Indianapolis, Indiana. 2 Supported by Research Grants 87-5A,B and 88-6A,B from the RileyMemorial Association and by Clinical Investigator Award HL-01322-04 from the National Institutes of Health. 3 Correspondence and requests for reprints should be addressed to Dr. Mervin C. Yoder, Jr., Department of Pediatrics, Riley Childrens Hospital, 702 Barnhill Drive RR 208, Indianapolis, IN 46223.
EFFECT OF DEXAMETHASONE ON PULMONARY FUNCTION AND INFLAMMATION IN BPD
tation wereweaned as clinicallytolerated. The ventilator settings at the end of this day were recorded for comparison with baseline values. Tracheobronchial lavage effluent was collected and pulmonary functions testing was performed within 24 h prior to initiating dexamethasone or placebo treatment, and again on the day after the last dose of dexamethasone or placebo. Of the nine infants who received dexamethasone, two did not have TBL obtained, and two different infants did not have pulmonary function testing (PFT). Of the eight placebo-treated infants, one did not have sufficient TBL data and one different infant did not have sufficient PFT data for inclusion.
Pulmonary Function 'Jesting Airway function was assessed by forced deflation flow volume curves (11). Flow was measured with a no. 0 Fleisch pneumotachometer (Fleisch, Lausanne, Switzerland) and Validyne differential pressure transducer (± 2 em H 20; Validyne Corp., Northridge, CA) attached between the proximal end of the endotracheal tube and a three-way pneumatically controlled switching valve (Hans Rudolph Co., Kansas City, MO). The flow signal was amplified (Validyne),and the analog signal was electrically integrated to volume (Gould Instruments, Cleveland, OH). The flow and volume signals were digitized and displayed on the CRT of an IBM-AT microcomputer. The infant was manually ventilated with an Ambu bag attached to the second port of the switching valve. The third port of the three-way valve was closed to a IO-L pressure-volume reservoir that had an adjustable negative pressure. Lung volume was inflated to greater than 40 em H 20 pressure (TLC), and then, by activating the pneumatic valve, the endotracheal tube was transiently « 2 s) connected to the negative pressure reservoir to generate the forced expiratory flow volume maneuver. The reservoir was initially set at negative 20 em H 20 pressure, and the forced expiratory flows were repeated at increasing negative pressures until the flow at 25010 of the FVC (\'25) did not increase. This generally did not require pressures more negative than 40 em H 20. The flow-volume curve was quantitated by the FVC and the flows at 50 and 25010 of FVC (\'50 and \'25)' In order to adjust for changes in lung volume during the study period, flow-volume curves obtained on the first and last days of the study were matched at TLC, and maximal expiratory flows were compared at isovolume. In addition, a time constant (Tau) for the respiratory system, which has the units of seconds, was calculated by dividing the FVC (ml) by the flow (\'25 mIls). A static respiratory systemcompliance (Crs) was calculated as the inspiratory volume between passive FRC and TLC, divided by the transrespiratory pressure at TLC. The values for forced expiratory flow,FVC, and Crs were divided by body weight (kg) to normalize for differences in body size.
Lavage Technique Lavage samples were collected as previously described (12). Briefly, 1 ml of sterile saline was instilled into the endotracheal tube, and the infant was manually ventilated for three to fivebreaths. Lavage fluid was aspirated into a sterile Lukens trap. The procedure was repeated three times over a 5-min period for each sample (4 ml total lavage); 1ml of saline was then suctioned directly into the trap to flush the lumen of the catheter, resulting in an average of 3 ± 0.5 ml recovered. The lavage fluid was immediately transferred to a polypropylene tube using a steriledouble-layer gauze pad 2 x 2 to strain mucous material. Additional saline was used as needed to rinse the sides of the Lukens trap. Lavage fluid was then centrifuged at 2,000 rpm for 15min, and cell-free supernatant in 0.25 ml aliquot were frozen at - 70° C for later assays. Cell Studies A total white blood cell count was performed using a hemocytometer. Cell viability was determined by trypan blue exclusion, and a differential count was obtained after WrightGiemsa staining of a cytocentrifuged cell sample. Elastase Assay Effluent elastase concentrations were determined by the hydrolysis of a 0.3 mM synthetic substrate, methoxysuccinyl-Ala-Ala-ProVal-p-nitroanilide (VegaBiotechnologies Inc., Tuscon, AZ) as previouslydescribed (13). Substrate was suspended in 9.8% dimethylsulfoxide and N-2-hYdroxyethylpiperazine-N-2ethanesulfonic acid buffer. Effluent aliquots were incubated at 37° C with the substrate for 30 min prior to the addition of IN acetic acid. Samples were read at 410 nm on a spectrophotometer (Model 25; Beckman Instruments, Irvine, CA) along with a standard curve generated between 1and 6 ug/ml using porcine pancreatic elastase (type III) (Worthington, Freehold, NJ). Standards were suspended in 0.2 M TRIS-O.101o BSA at pH 8. All samples were measured in duplicate. Ftbronectin Assay Fibronectin concentrations in TBL were determined using a modification of a competitive enzyme-linked immunoabsorbent assay as previously described (12).Briefly, effluent aliquots and purified human plasma standards (New York Blood Center, New York, NY) (diluted in PBS-0.05% Tween 20-1% BSA) were incubated with rabbit antihuman fibronectin (Cooper Biomedical, Malvern, PA) for 1 h in a 3-ml polypropylene tube at room temperature. The mixture was then added to polystyrene wells (Dynatech Laboratories, Alexandria, VA)precoated with human plasma fibronectin (diluted in 0.05 M TRIS at pH 9) and blocked with 100 mM glycine diluted in 1% BSA. After a 30-min incubation, the plates were washed with PBS-0.05% Tween, and a 1:1,500dilution of peroxidaselabeled porcine antirabbit antibody (Accurate
1045
Laboratories, Westbury, NY) was added to each test welland allowedto incubate at 20° C for an additional hour. The wells were then rewashed with PBS-0.05% Tween,and a substrate of orthophenylenediamine + 3% hydrogen peroxide + citrate buffer (OPD) was added. The reaction was stopped by adding 0.1 M sodium fluoride to each well, and the optical density of the wells was read at 450 nm on a Microelisa spectrometer (Dynatech). The concentration of fibronectin in the effluent was determined by comparing the optical density of test wells with wells containing known amounts (standards) of fibronectin. All samples were measured in duplicate.
Alpha-I Antitrypsin Assay Alpha-l antitrypsin concentrations were determined using the enzyme-linked immunoabsorbent assay reported by Michalski and coworkers (14). The assay is begun by adding effluent aliquots of human serum alpha-lantitrypsin (Calbiochem-Behring, San Diego, CA) (diluted in PBS-0.05% Tween) to polystyrene plates precoated with rabbit antihuman alpha-l-antitrypsin antibody (Boehringer-Mannheim, Indianapolis, IN) (diluted in NaHC0 3 buffer at pH 9.6) and blocked with 5% casein with 100mM glycine diluted in PBS. After 90 min of incubation at room temperature, the plates were washed with PBS-0.05% Tween,then a 1:2,000dilution of peroxide-labeled goat antihuman alpha-l-antitrypsin antibody (Cooper Biomedical Inc., Malvern, PA) was added to each test welland incubated for another hour at 20° C. The wells were then rewashed with PBS-0.05% Tween and OPD was added to each well. The reaction was stopped with 0.1M sodium fluoride, and the optical density of the wells was read at 450 nm on a Microelisa spectrometer. All samples weremeasured in duplicate. Because 1 mole of alpha-l-antitrypsin neutralizes 1 mole of elastase, the ratio of elastase/2 x alpha-l-antitrypsin was calculated. Albumin Assay Because some variability was encountered in the amount of saline recovered from the amount instilled during the performance of lavage, effluent constituents were expressed in relation to effluent albumin concentration. Use of albumin as a standard for the degree of dilution in lavage techniques is common (15, 16). Albumin concentrations weremeasured by a modification of the direct enzyme-linked immunoabsorbent assay described by Whitfield and Spierto (17). Duplicate effluent samples and human serum albumin standards (Sigma Chemical, St. Louis, MO), all diluted in NaHC03 buffer (pH 9.6)and 0.05% Tween, were incubated with rabbit antihuman albumin antibody (Cooper Biomedical). Polystyrene wells had been precoated with this antibody and blocked with 5% casein and 100 mM glycine in PBS. After allowing to incubate at room temperature for 90 min, the
YODER, CHUA, AND TEPPER
1046
plates were washed with PBS-0.05OJo Tween, A 1:1,500 dilution of peroxide-conjugated goat antihuman albumin antibody (Cappel Laboratories, West Chester, PA) was then added. The plates were rewashed after 1 h at 20° C, and OPD wasadded. The reaction was stopped with 0.1 M sodium fluoride, and the optical density of the wells was read at 450 nm on a Microelisa spectrometer.
Statistical Analysis For pulmonary function data, values analyzed were the highest measurements obtained during the testing period on the first and the last day of the study. Intergroup and intragroup differences in pulmonary function measurements were compared using unpaired and paired t tests, respectively. For tracheobronchial effluent data, values for a givenday were the mean of three samples collected during that day and analyzed separately. The data collected for biochemical and cellular parameters werecompared using the Wilcoxon's nonparametric signed ranks test. A p value < 0.05 was considered statistically significant. Variability about the mean is presented as the standard error of the mean.
TABLE 1 PATIENT DATA" Placebo (n = 8)
p Value
26.9 ± 1.8 39.1 ± 8.9 1.26 ± 0.20
26.0 ± 1.6 41.3 ± 13.2 1.49 ± 0.48
0.3 0.7 0.2
Gestational age, wk Postnatal age, days Weight at time of study, kg • Values are mean ± SEM.
TABLE 2 VENTILATOR SUPPORT" Dexamethasone Pre FI02
Peak inspiratory pressure, cm H20 Respiratory rate, (breaths/min) Positive end-expiratory pressure, cm H 20
50.4 ± 4 21 ± 2 22 ± 3 4 ± 1
Placebo Pre
Post
36 16 14 4
± ± ± ±
4t 4t 4t 1
Post
41 ± 4 20 ± 1 28 ± 6
44 20 27 5
5 ± 1
± ± ± ±
6 2 6 1
• Values are mean ± SEM.
t p < 0.05 post-treatment compared with
pretreatment values in the dexamethasone group.
TABLE 3 PULMONARY FUNCTION DATA" Dexamethasone (n = 7)
Results
On Day 1 of the study, prior to initiating treatment, there were no statistically significant differences in the dexamethasone and placebo-treated infants for gestational age at birth, age at the time of study, study weight (table 1), or severity of disease as indicated by F102 requirement, ventilator inspiratory pressure, ventilator rate, or positive end-expiratory pressure (table 2). After 3 days of treatment, the clinical status of the patients in the dexamethasone-treated group improved. In this group, there was a significant reduction in the inspired oxygen requirement, ventilator inspiratory pressure, and intermittent mandatory ventilator rate (table 2). No significant changes in any of these parameters were observed in the placebo-treated group. There was no significant change in body weight between Days 1 and 4 for the dexamethasonetreated infants (35 ± 56 g); however, the placebo group of infants exhibited a significant weight gain during the study period (83 ± 81 g). On entry into the study, there were no statistically significant differences in the baseline flow, FVC, or Crs of the respiratory system for the dexamethasone and placebo groups (table 3). The dexamethasone group did have a significantly longer time constant (Tau) than the placebo group (1.92 versus 1.01 s, p < 0.01). After 3 days of placebo treatment, there were no significant changes in any of the pulmonary function parameters. In contrast, after dexamethasone treatment, the
Dexamethasone (n = 9)
Post
Pre FVC, ml/kg Crs, mllcm H2O/kg "25' mils/kg "50' mils/kg
44 0.63 23 74
± ± ± ±
Placebo (n = 7)
10 0.15 10 22
56 0.85 68 111
± ± ± ±
17ft 0.16ft 50t§ 54t
Pre
35 0.48 37 92
± ± ± ±
Post
18 0.28 24 59
35 0.50 28 82
± ± ± ±
16 0.29 14 14
• Values are mean ± SEM. 0.05 < p < 0.10 compared with pre-dexamethasone. :t p < 0.05 compared with post-placebo. § p < 0.05 compared with pre-dexamethasone.
t
respiratory system compliance increased on Day 4, and Crs was greater in the dexamethasone group than in the placebo group. In addition, the increase in FVC (+ 11.9 ± 15.8 mllkg) approached statistical significance, and on Day 4 of the study the dexamethasone group had a larger FVC than did the placebo group. The V25 increased after dexamethasone (p < 0.03) and was higher in the dexamethasone than in the placebo group (p < 0.05) on Day 4 of the study. The increase in V50 did not attain statistical significance (p < 0.1) and it was not significantly different for the two groups after intervention. The time constant decreased significantly in the dexamethasone-treated group (1.92 versus 0.87 s, p < 0.025), and after treatment there was no significant difference in Tau for the dexamethasone and placebo groups (0.87 versus 1.27 s). The protein and cellular concentrations of the TBL effluent were not significantly different for the dexametha-
sone and placebo groups when treatment was initiated (table 4). After 3 days of dexamethasone administration, the total neutrophil count and percent neutrophils in the white blood cell differential were significantly decreased in the dexamethasone-treated infants. No significant change occurred for the placebotreated infants. The effluent albumin concentrations decreased significantly in the dexamethasone-treated infants, but it did not change in the placebo-treated group. After 3 days of dexamethasone the elastase concentration in the effluent decreased nearly 50070 from the pretreatment value, and this change in elastase approached statistical significance (p = 0.07). There was a significant elevation in the concentration of alpha-l-antitrypsin in the dexamethasone-treated group and a significant decrease in the elastase/2 x alpha-l-antitrypsin ratio. In contrast, the placebo-treated group had an increase in elastase concentration and decrease in alpha-l-antitrypsin concen-
1047
EFFECT OF DEXAMETHASONE ON PULMONARY FUNCTION AND INFLAMMATION IN BPD
TABLE 4 TRACHEOBRONCHIAL LAVAGE EFFLUENT DATA* Dexamethasone Pre Elastase, Ilg/mg albumin u-1-antitrypsin, Ilg/mg albumin Elastase/2 x a-t-antitrypsin Fibronectin, Ilg/mg albumin Total neutrophil count x 10"/mg albumin Neutrophils, % Albumin, mg/dl
24 40 0.24 33 23 51 7.1
± ± ± ±
Placebo
Post
11 13 0.08 8 ± 8 ± 9 ± 1.5
13 55 0.10 17 11 23 3.5
± ± ± ± ± ± ±
3t 14* 0.02* 3* 6* 7* 0.8*
Pre 18 56 0.16 22 27 41 6.9
± 6 ± 10 ± 0.10 ± 7 ± 11 ± 6 ± 3.7
Post 24 38 0.26 25 25 53 4.3
± 9
± 11 ± 0.06 ± 5 ± 6 ± 8 ± 1.3
• Values are mean ± SEM.
t 0.05 < p < 0.10.
*< P
0.05 post-treatment compared with pretreatment values.
tration, although these changes were not statistically significant. These combined changes in the placebo group did result in a significant increase in the elastase/2 x alpha-2-antitrypsin ratio between Days 1and 4. Fibronectin concentrations were significantly reduced after 3 days of dexamethasone treatment; however, the placebo group had no significant change in fibronectin concentration. No complications of sepsis, hypertension, hyperglycemia, or electrolyte imbalance were seen in the dexamethasonetreated infants. All infants enrolled completed the study. One infant in the placebo group died of peritonitis from intestinal perforation 4 days after completing the study. All infants in the dexamethasone group stopped demonstrating improvement in their respiratory status within several days of drug discontinuation, and four infants reverted back to the pretreatment oxygen and ventilator requirements. Discussion
Three days of intravenous dexamethasone treatment improved respiratory mechanics and decreased the level of respiratory failure in infants with BPD. Dexamethasone also suppressed TBL effluent indices of pulmonary inflammation in treated infants compared with those in placebo-treated control infants. The results of this prospective controlled study confirm the clinical efficacy of dexamethasone in the treatment of infants with BPD and implicate inflammation as an important factor contributing to the amount of respiratory support required by these ventilator-dependent infants (7-9). In our study, the low values for static compliance are in agreement with those previously reported in infants with BPD (8, 18). Treatment with dexamethasone resulted in a significant improvement in the static compliance of the respiratory system and therefore reflects an improve-
ment in the elastic properties of the parenchyma. Previous studies evaluating dexamethasone treatment have measured dynamic pulmonary compliance (Cdyn) with an esophageal catheter (8, 18). In the presence of diffuse airway obstruction, Cdyn is dependent upon the respiratory frequency at which it is measured. With increasing respiratory rates, Cdyn progressivelydeclinesand underestimates the static compliance of the respiratory system. Because of this limitation, the previouslyreported increase in Cdyn with dexamethasone treatment was unable to discriminate whether the observed changes were secondary to a decline in peripheral airway obstruction or to an improvement in the elastic properties of the parenchyma. Our findings of an improvement in the static compliance of the respiratory system imply that the increase in Cdyn is at least partly attributable to changes in the elastic properties of the parenchyma. The infants enrolled in this study had evidence of obstructive airway disease with forced expiratory flows at low lung volumes (V2S: 23 to 37 mIls/kg) that were significantly lower than values reported for "normal" premature infants (95 ± 13 mIls/kg) and similar to previously reported values for ventilator-dependent premature infants with BPD (11). The abnormal flows at low lung volumes and concave shape of the flow volume curves observed for many of the infants suggest the presence of peripheral airway obstruction, a finding consistent with the necrotizing bronchiolitis reported in pathologic specimens from infants with BPD (1). Dexamethasone-treated infants demonstrated a significant improvement in airway function as evidenced by the increase in V2 S • Dexamethasone treatment also produced a significant decline in the time constant (Tau) for the respiratory system. Although calculated from flow and volume, Tau is also the product
of the resistance (R) and the compliance (C) of the respiratory system (Tau = R x C). Because compliance increased after dexamethasone treatment, the decrease in Tau reflects a marked decrease in respiratory resistance. This finding is in agreement with the reported decline in pulmonary resistance after dexamethasone treatment (18). The dexamethasone-treated group had a significant decrease in the concentration of albumin in the TBL effluent. This did not occur in the placebo group. As albumin is a serum protein with a low molecular weight, it has been used as one index of microvascular leak into the alveoli (19). Therefore, the lower albumin concentration in the TBL fluid after dexamethasone treatment may reflect less microvascular leak. Analysis of an undiluted TBL sample to compare lavage albumin/serum albumin would have provided stronger support that dexamethasone had diminished microvascular leak. If dexamethasone decreases pulmonary extravascular fluid, then there should be an association between improved pulmonary mechanics and diuresis as has been reported by Gladstone and coworkers (18).Although wedid not examine fluid balance in our study infants, increased urine output in response to steroid intervention is a plausible explanation for the absence of significant weight gain in our dexamethasonetreated infants in contrast to the significant increase in weight in the placebo group. These findings are consistent with the postulate that dexamethasone treatment reduces pulmonary interstitial fluid by altering membrane permeability and thereby results in improved pulmonary mechanics. In respiratory distress syndromes, neutrophils are believed to contribute to the pathophysiology of pulmonary microvascular injury and altered membrane permeability. Pretreatment TBL effluent obtained from all study infants contained concentrations of neutrophils, elastase, and fibronectin similar to those previously published for infants with BPD (6, 7, 20, 21). The TBL effluent concentration of neutrophils and the protease/antiprotease ratio decreased in our study infants after treatment with dexamethasone. The concentration of neutrophil-derived elastase also decreased in all but two of the treated infants, though no statistical difference in pretreatment and post-treatment concentrations was observed. Our findings are consistent with Gerdes and coworkers (7) who reported that 48 to
1048
72 h of dexamethasone treatment decreases neutrophil elastase concentrations and the elastase/2 x u-l-proteinase inhibitor ratio in the tracheobronchial effluent of patients with BPD. In our study, fibronectin concentrations also decreased in the dexamethasone-treated infants. In contrast, Gerdes and coworkers (7) did not observe an effect of dexamethasone treatment on TBL concentrations of fibronectin. Differences in the results of these two studies may be related to the fact that Gerdes and coworkers (7) enrolled younger infants who had lower baseline concentrations of fibronectin and were treated with a higher concentration of dexamethasone for a shorter period of time. The anti-inflammatory effects of dexamethasone, which result in lower numbers of neutrophils and their secretory products in the TBL effluent of these infants with BPD, may result from effects of dexamethasone on either the phagocytic cellsor the microvascular endothelial cells. Although in vitro or in vivo exposure of neutrophils to dexamethasone inhibits neutrophil adherence to endothelial cells, the predominant in vivo effect of dexamethasone may be to inhibit transendothelial emigration of neutrophils (22). It remains unclear whether inhibition of neutrophil emigration results from decreased local production of inflammatory mediators (plateletactivating factor, arachidonic acid metabolites, complement fragments), which stimulate endothelial cells to enhance neutrophil-endothelial adhesive interactions, or whether the neutrophil is directly inhibited in its capacity to adhere to and then emigrate into the alveolar space (23). Several inflammatory mediators (platelet-activating factor, arachidonic acid metabolites) havebeen identified in the TBL fluid of infants with BPD (24, 25). These factors have been proposed to influence pulmonary airway function and alveolar epithelial permeability as well as neutrophil recruitment in patients with BPD. It has not yet been determined if dexamethasone treatment influences the concentration of these protein and lipid mediators in the TBL fluid of infants with BPD. We conclude that 3 days of intravenously administered dexamethasone in ventilator-dependent premature infants with BPD results in the acute suppres-
YODER, CHUA, AND TEPPER
sion of pulmonary inflammation, improved pulmonary mechanics, and a decrease in the requirement for supplemental oxygen and assisted ventilation. We speculate that the failure to see sustained improvement in these patients after dexamethasone was discontinued is related to return of the inflammatory state and reversion of the lung to the level of pretreatment dysfunction. This would be consistent with the observations of Cummings and coworkers (10) who found that early and long-term steroid treatment was required to maintain clinical benefits in infants with BPD. Because prolonged treatment with dexamethasone results in significant side effects, further evaluations of the mechanisms of steroid action and optimal treatment strategies for BPD prevention in high risk ventilatordependent premature infants are required. Because pulmonary inflammation has been proposed as one of the pathophysiologic factors leading to the development of BPD, perhaps earlier treatment will result in a decrease in the severityof pulmonary dysfunction in susceptible patients (26). References 1. Northway WH, Rosan RC, Porter DY.Pulmonary disease following respirator therapy of hyalinemembrane disease: bronchopulmonary dysplasia. N Engl J Med 1967; 276:357-68. 2. Goetzman BW.Understanding bronchopulmonary dysplasia. Am J Dis Child 1986; 140:332-4. 3. Clark RAE Overview and general considerations of wound repair. In: Clark RAF, Henson PM, eds. The molecular and cellular biology of wound repair. New York: Plenum Press, 1988; 3-33. 4. Haslett C, Henson PM. Resolution of inflammation. In: Clark RAF, Henson PM, eds. The molecular biology and cellular biology of wound repair. New York: Plenum Press, 1988; 185-212. 5. Crystal RG, Bitterman PS, Rennard SI, Hance AJ, Keogh BA. Interstitial lung disease of unknown cause. N Engl J Med 1984; 310:154-66. 6. Merritt TA, Cochrane CG, Holcomb K, et al. Elastase and alpha-1 proteinase inhibitor activity in tracheal aspirates during respiratory distress syndrome. J Clin Invest 1983; 72:656-66. 7. Gerdes JS, Harris MC, Polin RA. Effects of dexamethasone and indomethacin on elastase, n-l proteinase inhibitor, and fibronectin in bronchoalveolar lavage fluid from neonates. J Pediatr 1988; 113:727-31. 8. Mammel MC, Green TP, Johnson DE, Thompson TR. Controlled trial of dexamethasone therapy in infants with bronchopulmonary dysplasia. Lancet 1983; 1:1356-8. 9. AveryGB, Fletcher AB, Kaplan M, Brudno DS. Controlled trial of dexamethasone in respiratordependent infants with bronchopulmonary dysplasia. Pediatrics 1985; 75:106-11. 10. Cummings 11, D'Eugenio DB, Gross SJ. A
controlled trial of dexamethasone in pre-term infants at high risk for bronchopulmonary dysplasia. N Engl J Med 1989; 320:1505-10. 11. Motoyama EK, Fort MD, Klesh K, Mutich RL, Guthrie RD. Early onset of airway reactivity in premature infants with bronchopulmonary dysplasia. Am Rev Respir Dis 1987; 136:50-7. 12. Gerdes JS, Yoder MC, Douglas SD, Paul M, Harris MK, Polin RA. Tracheal lavage and plasma fibronectin: relationship to respiratory distress syndrome and development of bronchopulmonary dysplasia. J Pediatr 1986; 108:601-6. 13. Castillo MJ, Nakajima K, Zimmerman M, Powers JC. Sensitive substrates for human leukocyte and porcine pancreatic elastase: a study of the merits of various chromophoric and fluorogenic leaving groups in assays for serine proteases. Anal Biochem 1979; 99:53-64. 14. Michalski JP, McCombs CC, Sheth S, McCarthy M, DeShago R. A modified double antibody sandwich enzyme-linked immunoabsorbent assay for measurement of alpha-l-antitrypsin in biologic fluids. J Immunol Methods 1985; 83:101-12. 15. Cherniak RM. Proteins in bronchoalveolar lavage fluid. Am Rev Respir Dis 1990; 141(5 Part 2:5183-8). 16. Young KR, Reynolds HY. Bronchoalveolar washings: proteins and cells from normal lungs. In: Bienenstock J, ed. Immunology of the lung and upper respiratory tract. New York: McGraw-Hill, 1984; 157-73. 17. Whitfield W, Spierto E Modified ELISA for the measurement of urinary albumin. Clin Chern 1986; 32:561-3. 18. Gladstone I, Ehrenkranz R, Jacobs H. Pulmonary function tests and fluid balance in neonates with chronic disease during dexamethasone treatment. Pediatrics 1989; 84:1072-6. 19. BlaschkeE, Eklund A, Hembrand R. Extracellular matrix components in bronchoalveolar lavage fluid in sarcoidosis and their relationship to signs of alveolitis. Am Rev Respir Dis 1990; 141:1020-5. 20. Ogden BE, Murphy SA, Saunders OC, Pathak D, Johnson JD. Neonatal lung neutrophils and elastase/proteinase inhibitor imbalance. Am Rev Respir Dis 1984; 130:817-21. 21. Merritt TA, Hallman M. Interactions in the immature lung: protease-antiprotease mechanism of lung injury. In: Merritt TA, Northway WH, Boynton BR, eds. Bronchopulmonary dysplasia. Boston: Blackwell Scientific Publishers, 1988; 117-30. 22. Schleimer RP. Effects of glucocorticoid on inflammatory cells relevant to their therapeutic applications in asthma. Am Rev Respir Dis 1990; 141 (2 Part 2:559-69). 23. Williams TJ, Yarwood H. Effect of glucocorticoid on microvascular permeability. Am Rev Respir Dis 1990; 141(2 Part 2:539-43). 24. Stenmark KR, Eyzaguirre M, Wescott JY, Murphy PM. Potential role of eicosanoids and PAP in the pathophysiology of bronchopulmonary dysplasia. Am Rev Respir Dis 1987; 136:770-2. 25. Mirro R, Armstead W, Leffler C. Increased airway leukotriene levelsin infants with severebronchopulmonary dysplasia. Am J Dis Child 1990; 144:160-1. 26. VanMarter LJ, Leviton A, Kuban KC, Pagano M, Allred KN. Maternal glucocorticoid therapy and reduced risk of bronchopulmonary dysplasia. Pediatrics 1990; 86:331-6.
