Aerosolized Surfactant Improves Pulmonary Function ... - ATS Journals

Aerosolized Surfactant Improves Pulmonary Function in Endotoxin-induced Lung Injury CHARLES LUTZ, DAVID CARNEY, CHRISTINE FINCK, ANTHONY PICONE, LOUIS A. GATTO, ANDREW PASKANIK, EDWARD LANGENBACK, and GARY NIEMAN SUNY Health Science Center at Syracuse, Department of Surgery, Syracuse, New York; SUNY Health Science Center at Stony Brook, Stony Brook, New York; and SUNY College at Cortland, Department of Biological Sciences, Cortland, New York

Surfactant dysfunction is a primary pathophysiologic component in patients with adult respiratory distress syndrome (ARDS). In this study we tested the efficacy of aerosolized surfactant (Sf) replacement in a severe lung injury model of endotoxin-induced ARDS. Twenty-one certified healthy pigs were anesthetized, surgically prepared for measurement of hemodynamic and lung function, then randomized into one of four groups: (1) control (n 5 5), surgical instrumentation only; (2) lipopolysaccharide (LPS) (n 5 6), infused with Escherichia coli LPS (100 mg/kg) without positive endexpiratory pressure (PEEP) and ventilated with a nonhumidified gas mixture of 50% N2O and 50% O2; (3) LPS 1 PEEP (n 5 4), infused with LPS, placed on PEEP (7.5 cm H2O), and ventilated with a humidified gas mixture; and (4) LPS 1 PEEP 1 Sf (n 5 6), infused with LPS, placed on PEEP, and ventilated with aerosolized Sf (Infasurf, ONY, Inc.). All animals were studied for 6 h. Arterial PO2 significantly decreased in both the LPS and LPS 1 PEEP groups (LPS 1 PEEP 5 74 6 19 mm Hg; LPS 5 74 6 · · 19 mm Hg, p , 0.05) while venous admixture (Q S/Q T) increased in these groups (LPS 1 PEEP 5 43.3 6 3.9%; LPS 5 47.7 6 11%, p , 0.05) as compared with the control group. PEEP 1 Sf reduced the fall in · · PO2 (142 6 20 mm Hg) and rise in Q S/Q T (15.1 6 3.6%) caused by LPS. Delayed induction of PEEP (2 h following LPS) did not significantly improve any parameter over the LPS group without PEEP in this ARDS model. LPS without PEEP (3.4 6 0.2 cells/6,400 mm2) caused a marked increase in the total number of sequestered leukocytes in the pulmonary parenchyma as compared with the control group (1.3 6 0.1 cells/6,400 mm2). LPS 1 PEEP 1 Sf (2.3 6 0.2 cells/6,400 mm2) significantly decreased while LPS 1 PEEP significantly increased (4.0 6 0.2 cells/6,400 mm2) the total number of sequestered leukocytes as compared with the LPS without PEEP group. In summary, aerosolized surfactant replacement decreased leukocyte sequestration and improved oxygenation in our porcine model of endotoxin-induced lung injury. Lutz C, Carney D, Finck C, Picone A, Gatto LA, Paskanik A, Langenback E, Nieman G. Aerosolized surfactant improves pulmonary function in endoAM J RESPIR CRIT CARE MED 1998;158:840–845. toxin-induced lung injury.

Adult respiratory distress syndrome (ARDS) represents a continuing challenge for the clinician with a 37% mortality despite the advent of numerous anti-inflammatory therapies (1). Ashbaugh and coworkers in 1967 (2) first demonstrated that abnormal surfactant function was a primary component in the pathophysiology of ARDS. This has subsequently been confirmed by others (3–5). Surfactant dysfunction in ARDS can result from altered surfactant composition (4), metabolism of the Type II pneumonocyte (3), and/or inactivation of surfactant along the alveolar surface (5). Surfactant replacement in the treatment of ARDS has

(Received in original form January 22, 1998 and in revised form May 4, 1998) Sponsored in part by a grant from Forest, Inc. Correspondence and requests for reprints should be addressed to Gary F. Nieman, Department of Surgery, SUNY Health Science Center, 750 E. Adams Street, Syracuse, NY 13210. Am J Respir Crit Care Med Vol 158. pp 840–845, 1998 Internet address: www.atsjournals.org

been studied in numerous animal lung injury models (6–13) and clinical trials (14–16) with conflicting results. The variability in success in these studies is attributed to differences in the animal model and species utilized, the severity of the lung lesion, patient selection criteria, inadequate amount of surfactant delivered to the alveoli by the nebulizer, and the type of surfactant used. Our laboratory has confirmed that the effectiveness of surfactant replacement was related to both the severity of lung injury (17) and the addition of positive end-expiratory pressure (PEEP) (18). Our results (17, 18) and that of others (16) suggest that the improved lung function after installation of surfactant is transient, while aerosolized replacement is often ineffective (10– 12). This is due, in part, to inadequate delivery of aerosolized surfactant to the alveolus (14). An efficient aerosol delivery system would enhance the effectiveness of surfactant replacement therapy while eliminating the detrimental effects of repeat surfactant instillation. In this study we used a natural surfactant and an improved nebulizer system in a severe lung injury model of ARDS (17, 18).

841

Lutz, Carney, Finck, et al.: Aerosolized Surfactant Replacement in ARDS

METHODS Surgical Preparation Certified healthy pigs (15 to 20 kg) were pretreated with atropine (0.05 mg/kg, intramuscularly) 10 to 15 min before intubation, and then preanesthetized with ketamine (30 mg/kg, intramuscularly), and xylazine (2 mg/kg, intramuscularly). Anesthesia was induced with sodium pentobarbital administered intravenously. A tracheostomy was established and the animals were ventilated with a mixture of 50% N2O and 50% O2 delivered via an anesthesia ventilator (Narkomed Drager AV, Telford, PA). Sodium pentobarbital (6 mg/kg/h, intravenously) was delivered continuously via a Harvard infusion pump (Model 907; South Natick, MA). Once complete anesthesia was assured by reflex analysis, pancuronium bromide was given (2 mg intravenous bolus over 30 s) to maintain paralysis. A carotid artery cutdown was established with polyethylene tubing (2 mm interior diameter) for analysis of blood gases (Model ABL2; Radiometer Inc., Copenhagen, Denmark) and blood oxygen content (Model OSM3; Radiometer Inc.). A 7-French flow-directed SwanGanz thermodilution catheter was inserted into the femoral vein and advanced into the pulmonary artery for measurement of pulmonary artery pressure (Pap), pulmonary artery wedge pressure (Ppaw), central venous pressure (CVP), mixed venous blood gas and O2 content, and cardiac output (CO) determinations (Explorer; Baxter, Inc., Irvine, CA). Pressures were measured using transducers (Transpac MK404DTNVF; Sorenson, Salt Lake City, UT) leveled at the right atrium and recorded on a Hewlett Packard recorder (7754A; Palo Alto, CA). CO measurements were made in duplicate at end-expiration, every 15 min after LPS infusion, and maintained within 10% of baseline by bolus infusion of 6% dextran 70 (19). A triple lumen catheter was placed into the adjacent jugular vein for intravenous administration of fluid, anesthetic agents, and Escherichia coli lipopolysaccharide (LPS). Peak airway pressure was measured from a side port located 2 cm from the proximal end of the tracheal tube. Static pulmonary compliance (Cstat) was measured by disconnecting the ventilator and injecting twice the calculated tidal volume (VT 5 wt kg ? 12) into the lung with a Collins syringe. Plateau airway pressure was recorded and used to calculate Cstat (injected volume/plateau pressure [ml/mm Hg]). A cystostomy was performed through a low midline incision and a Foley catheter was inserted into the bladder. Base excesses (BE) below normal limits (23 mEq/L) were corrected with intravenous sodium bicarbonate and adjustments were made in tidal volume and ventilatory rate to maintain PaCO2 within normal range (35 to 45 mm Hg). Heating pads and warmed, intravenously administered fluids served to maintain core temperature between 37–398 C. Following instrumentation, a period of approximately 15 min was allowed for the stabilization, at which point values were considered baseline and preintervention measurements were obtained. All pigs were studied from time zero to 6 h (t 5 0–6) while being infused with lactated Ringer’s solution (25 ml/kg/h).

put for the inspiratory line from the ventilator, and an output to the endotracheal tube. Aerosolized surfactant particles which were too large to enter the alveoli impacted on the internal baffles and drained to the bottom of the Plexiglas chamber. This surfactant was pumped from the chamber back into the reservoir for reaerosolization. The nebulizer was triggered by the ventilator so that aerosolization occurred only during inspiration. Ventilatory circuit flow bypassed the nebulization chamber on expiration. Therefore, the pig received only gas containing nebulized surfactant, and nebulizer efficiency was improved. Pulmonary deposition of radiolabeled (technetium-99m) superoxide dismutase (rhSOD) was tested with this nebulizer in piglets using a gamma camera (Picker International, Cleveland, OH). The percent of rhSOD delivered by aerosol to all lung lobes was compared with that of rhSOD delivered by instillation into the airway. Deposition of rhSOD was more evenly distributed throughout all lobes of the lung with aerosolization than with instillation (unpublished observations). Surfactant particle size was measured using 99mTc-diethylene triamine pentaacetic acid (99Tc-DTPA) with an InTox 2 L/min seven-stage cascade impactor (InTox Products, Albuquerque, NM). Approximately 55% of the aerosol surfactant generated had a particle size in the 1- to 5-mm range which is ideal for alveolar deposition (unpublished measurements).

Experimental Protocol Pigs were randomized into one of four groups: (1) control (n 5 5), surgical instrumentation only; (2) lipopolysaccharide (LPS) (n 5 6), infused with LPS (E. coli 111:B4, 100 mg/kg in 500 ml of saline; Sigma Chemical Co., St. Louis, MO) from t 5 0 to t 5 1 h; (3) LPS 1 PEEP (n 5 4), infused with LPS over 1 h, allowed to respond for 1 h, placed on PEEP and ventilated with humidified gas mixture; and (4) LPS 1 PEEP 1 surfactant (Sf) (n 5 6), infused with LPS, allowed to respond for 1 h, placed on PEEP, and ventilated with aerosolized Sf (Infasurf, donated by ONY, Inc., Amherst, NY). Surfactant aerosol was continued for the entire experiment. All measurements of hemodynamic and lung function were recorded during the baseline period and then every 30 min for 6 h.

Calculations Based on the collected data, the following calculations were performed: . . (1) Venous admixture (QS/QT) 5 (CcO2 2 CaO2/CcO2 2 CvO2) where CaO2 and CvO2 are arterial and venous blood oxygen content, . . QS is venous admixture blood flow, and QT is total blood flow. CaO2 and CvO2 were measured with the OSM3. Capillary content values (CcO2) were calculated from the alveolar gas equation with the assumption that pulmonary capillary oxygen saturation (ScO2) is 100%: (2) O2 content (Vol%) 5 [(Hb ? ScO2 ? 1.39) 1 (0.003 ? Po2)]. Ventilatory efficiency index (VEI) was calculated with the formula:

Surfactant Infasurf is a sterile, nonpyrogenic pulmonary surfactant. It is an organic solvent extract of calf lung lavage suspended in 0.9% saline. Infasurf contains phospholipids, neutral lipids, fatty acids, and surfactant-associated proteins. It contains no preservatives and is heat-sterilized. The protein content of Infasurf consists of two hydrophobic, low-molecular-weight, surfactant-associated proteins (SP-B and SP-C). The concentration of SP-B is 0.17 to 0.37 mg/ml with a ratio of SP-B to SP-C at 3:2. Each milliliter of Infasurf contains 35 mg of phospholipids and less than 1.0 mg of protein. The primary lipid is phosphatidyl choline (24 to 30 mg/mg [70 to 80% of the total phospholipid]) with a significant fraction as disaturated phosphatidyl choline (13 to 19 mg/ml) and the remaining portion made up to cholesterol (1.5 to 2.5 mg/ml). The amount of surfactant aerosolized in 4 h was 20.0 6 1.3 ml (1.3 6 0.12 g/m2) and of saline 38.8 6 3.5 ml.

Nebulizer Two modified Omron nebulizers were mounted in a Plexiglas chamber which contained a reservoir for surfactant, internal baffling, an in-

5 ml/kg/min (3) VEI = ---------------------------------------------------( ∆P ⋅ Rf ⋅ Pa CO2 ) ⁄ 760 where DP 5 the difference between peak and end-expiratory pressure (mm Hg) and Rf 5 respiratory frequency. The VEI is described in units analogous to compliance and was calculated assuming that the rate of total CO2 production was constant at 5 ml/kg/min, and PaO2 5 PaCO2. The index allowed comparison of respiratory status among animals whose airway pressures, respiratory rate, and PaCO2 varied throughout the experiment (20).

Histometric Evaluation At the end of the experiment the right caudal lung lobe (RCL) was excised and its bronchus was cannulated. Glutaraldehyde fixative (2.5%, phosphate-buffered) was slowly instilled through the cannula until air was no longer displaced from the airway. The lung was immersed in glutaraldehyde and additional fixative was infused with a syringe until the airway pressure stabilized at 25 mm Hg. The cannula was clamped and the tissue was stored at room temperature for 24 h. In the first four animals in each treatment group, lung tissue was

842

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

studied according to a stratified and random sampling method that assured the unbiased coverage of parenchymal structures lacking a homogeneous distribution. One tissue block from the fixed lobe of each animal was randomly chosen and processed for routine paraffin sections. The blocks were sectioned grossly until the entire profile of the tissue entered the plane of the section, usually constituting a rectangle of approximately 15 mm 3 20 mm. At that point, 10 serial sections made at 7 mm were individually mounted on numbered slides. Then, either even or odd numbered slides (random choice) were stained with hematoxylin and eosin. In this manner, five equidistant sections were studied in each animal. A sampling probe consisting of a vertical line traversing the height of the slide was established for each of the serial sections, as follows: The sampling probe in the first section was drawn close to the left edge of the tissue, the probe on the last was drawn close to the right edge, and the remaining three were drawn at equidistant locations. Ten sampling areas were designated 1 mm apart along each sampling probe, thus establishing 50 unbiased sampling areas in each animal (n 5 200 per treatment condition) prior to microscopic observation. Each sampling area was located blindly according to its x/y coordinates, and then observed at high magnification using a high-resolution video camera. Areas featuring bronchi, connective tissue septa, or blood vessels other than capillaries were discarded by advancing the stage 0.5 mm along the short axis of the section, thus limiting quantification to the alveolar parenchyma. A sampling area of 6,400 mm2 was overlaid with a grid consisting of 64 intersections at 10-mm intervals. Tissue density per sampling area was estimated as the ratio of intersection points falling on tissue versus points corresponding to empty space (21). Alveolar macrophages and leukocytes (polymorphonuclear leukocytes [PMN] and monocytes) were subsequently counted in all focal planes of each sampling area following an unbiased counting procedure (22).

Figure 1. The effect of surfactant aerosol on arterial PO2 over time. See METHODS for group description. LPS 5 the time during infusion of endotoxin; PEEP/PEEP 1 surfactant 5 the time and duration of PEEP (7.5 cm H2O) or PEEP plus surfactant aerosol. Significant lung injury had occurred, indicated by the drop in P O2, by the time PEEP or PEEP 1 surfactant aerosol was initiated (2 h). Only PEEP 1 surfactant reduced the LPS-induced fall in P O2. Two animals died before the end of the study in the LPS and LPS 1 PEEP groups. The reduced number (n) of animals at each time period is depicted adjacent to the group symbol (n 5 2, 3, or 4). Data are mean 6 SE. *p , 0.05 versus baseline; †p , 0.05 versus control group; §p , 0.05 versus LPS 1 PEEP 1 surfactant group.

VOL 158

1998

Statistics Statistical significance between groups was determined using a oneway analysis of variance. Significance between parameters within the same group was determined with a repeat analysis of variance. Whenever the F ratio indicated significance, a Newman-Keuls test was used to identify the individual differences. A significant difference was assumed if the probability of the null hypothesis was , 5% (23).

Animal Care Pigs were euthanized by exsanguination. The experiments described were performed in adherence with the National Institutes of Health guidelines for the use of experimental animals in research. The protocol was approved by the Committee for the Humane Use of Animals (CHUA) at the SUNY Health Science Center, Syracuse, New York.

RESULTS Only PEEP 1 Sf improved the lung dysfunction associated with LPS-induced ARDS; PEEP without surfactant had no beneficial effect on the impaired lung function. Arterial PO2 trended higher and venous admixture was significantly lower in the group treated with surfactant as compared with the other groups exposed to LPS (Figures 1 and 2). VEI tended to be higher in the LPS 1 PEEP 1 Sf group than in both the LPS and LPS 1 PEEP groups (p 5 0.06) (Table 1). VEI fell with time in all groups exposed to LPS (Table 1). VEI was highest at the end of the study in the PEEP 1 Sf treatment group but this improvement was not statistically significant. Static compliance fell significantly from baseline only in the group exposed to LPS without PEEP (Table 1). PEEP, with or without surfactant, did not alter the hemodynamic changes associated with LPS exposure (Table 1, Figure 3). Pulmonary artery pressure increased above that of the control group 3 h after the start of LPS infusion in all three groups ex-

Figure 2. The effect of surfactant aerosol on venous admixture over time. See METHODS for group description. LPS 5 the time during infusion of endotoxin; PEEP/PEEP 1 surfactant 5 the time and duration of PEEP (7.5 cm H 2O) or PEEP plus surfactant aerosol. Only PEEP 1 surfactant reduced the LPS-induced rise in venous admixture. Two animals died before the end of the study in the LPS and LPS 1 PEEP groups. The reduced number (n) of animals at each time period is depicted adjacent to the group symbol (n 5 2,3, or 4). Data are mean 1 SE. *p , 0.05 versus baseline; †p , 0.05 versus control group; §p , 0.05 versus LPS 1 PEEP 1 surfactant group.

843

Lutz, Carney, Finck, et al.: Aerosolized Surfactant Replacement in ARDS TABLE 1 HEMODYNAMIC AND LUNG FUNCTION DATA* Time (h) Control (n 5 5) Baseline 0.5 1 2 4 6 LPS (n 5 6) Baseline 0.5 1 2 4 6 (n 5 4) LPS 1 PEEP (n 5 4) Baseline 0.5 1 2 4 (n 5 3) 6 (n 5 2) LPS 1 PEEP 1 surfactant (n 5 6) Baseline 0.5 1 2 4 6

VEI

Cstat

1.03 6 0.07 1.17 6 0.1 1.11 6 0.08 1.05 6 0.08 0.98 6 0.04 1.05 6 0.09

28 6 4 27 6 3 27 6 3 26 6 3 26 6 3 26 6 3

0.89 6 0.07 0.79 6 0.08 0.76 6 0.11 0.51 6 0.07†‡ 0.32 6 0.09†‡ 0.39 6 0.12†‡

26 6 2 24 6 2 23 6 2 19 6 2 16 6 1†‡ 17 6 1†‡

CO

Ppa

78 6 3 74 6 7 81 6 4 75 6 5 72 6 4 69 6 5

3.9 6 0.6 3.9 6 0.7 4.0 6 0.6 3.7 6 0.6 3.2 6 0.4 2.8 6 0.3

75 6 3 79 6 7 94 6 4† 108 6 4† 108 6 6† 100 6 6†

80 6 3 80 6 2 80 6 2 70 6 3 57 6 3†‡§ 43 6 10†‡§

4.2 6 0.3 4.4 6 0.4 3.8 6 0.3 3.4 6 0.2 3.6 6 0.2 3.5 6 0.5

77 6 6 91 6 8 75 6 5 66 6 5‡ 96 6 10 103 6 15

SvO2

0.87 6 0.11 0.71 6 0.16 0.73 6 0.12 0.59 6 0.05‡ 0.27 6 0.10†‡ 0.19 6 0.03†‡

35 6 6 28 6 4 29 6 3 29 6 4 21 6 4 25 6 0 (n 5 1)

80 6 1 82 6 2 82 6 2 71 6 3† 56 6 3† 56 6 3†

3.7 6 0.5 4.2 6 0.2 3.9 6 0.4 3.6 6 0.2 3.4 6 0.6 3.7 6 0.3

75 6 2 93 6 4 84 6 7 70 6 7‡ 90 6 11 110 6 20

1.13 6 0.10 0.93 6 0.07‡ 0.83 6 0.05‡ 0.58 6 0.07‡ 0.58 6 0.08†‡ 0.54 6 0.11†‡

30 6 3 27 6 3 25 6 2 21 6 2 20 6 3 20 6 3

80 6 1 81 6 2 77 6 3 73 6 3 70 6 3† 72 6 3†

4.6 6 0.3 4.3 6 0.4 4.0 6 0.4 3.9 6 0.4 3.5 6 0.3 3.7 6 0.3

68 6 3 93 6 3† 67 6 7‡ 65 6 4‡ 101 6 5† 107 6 7†

Definition of abbreviations: VEI 5 ventilatory efficiency index; SVO2 5 venous O2 saturation (%); CO 5 cardiac output (L/min); Ppa 5 mean arterial pressure. * Data are mean 6 SE. † p , 0.05 versus baseline. ‡ p , 0.05 versus control group. § p , 0.05 versus LPS 1 PEEP 1 surfactant.

posed to LPS (Figure 3). No difference in cardiac output or mean arterial pressure (Ppa) was measured between any of the groups (Table 1). Histological preparations from control pigs exhibited open and patent alveoli, with occasional areas of lymphoid tissue accumulation, but minimal cellular infiltration. LPS infusion caused a marked thickening of alveolar walls with increased leukocyte (PMN and monocyte) infiltration accompanied by significant red cell congestion. There was increased leukocyte infiltration into the pulmonary interstitium as well (Table 2). LPS infusion was also associated with obvious signs of tissue edema such as dilated lymphatics, enlarged connective tissue septa, and conspicuous cuffs around bronchi and blood vessels. PEEP 1 Sf significantly decreased alveolar macrophage and leukocyte infiltration into the alveolus and pulmonary interstitium (Table 2), which was accompanied by a reduction in peribronchovascular edema and in interstitial congestion. Mortality before the end of the study was 33% in the LPS group and 50% in the LPS 1 PEEP group.

DISCUSSION The most important finding from this study is that aerosolized surfactant generated with our nebulizer improves lung function in a severe model of endotoxin-induced ARDS. The severity of this model was underscored by a 33% mortality in the LPS group and 50% mortality in the LPS 1 PEEP group within the 6-h study protocol. Also, we initiated surfactant aerosol 2 h after endotoxin infusion at a point when lung injury had already occurred. This therapeutic treatment proto-

col most effectively mimics the timing of surfactant replacement therapy in clinical ARDS. Thus, this study suggests that aerosolized surfactant replacement therapy may be an effective adjunct in the treatment of ARDS. Surfactant replacement has been effective in saline-lavage models of ARDS. This may be due to a reversal in the pathophysiologic order of events in the saline lavage lung injury model as compared with those with the endotoxin model. Infusion of endotoxin stimulates the release of inflammatory cytokines which cause pulmonary neutrophil sequestration. Sequestered neutrophils release proteases which damage the alveolar–capillary membrane causing high permeability edema. Edema fluid enters the alveolus and deactivates surfactant. Thus, in endotoxin-induced lung injury surfactant deactivation is the last pathologic event. Endotoxin infusion simulates the pathologic order of events seen with ARDS. In contrast, surfactant deactivation is the initial event with saline lavage. Neutrophil sequestration and high permeability edema occur subsequent to surfactant deactivation. If surfactant replacement is initiated soon after saline lavage, which is the case in most studies, the only lung injury is an inadequate concentration of surfactant. Successful treatment of this lung injury by surfactant replacement would be expected. Using this saline-lavage model, O’Brodovich and Hannam demonstrated that exogenous surfactant significantly improved oxygenation (6). In a similar model, Gommers and coworkers showed that exogenous surfactant improved oxygenation and functional residual capacity in rabbits (7). Lewis and coworkers found that nebulized surfactant improved oxygenation and reduced peak inspiratory pressures (8). In addition to using an

844

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 158

1998

TABLE 2 NUMBER OF CELLS PER SAMPLING AREA (6,400 mm2)* Control

LPS

LPS 1 PEEP

Alveolar macrophages 0.5 6 0.1 1.2 6 0.1† 1.5 6 0.10†‡ Interstitial leukocytes 1.2 6 0.1 2.7 6 0.2† 2.8 6 0.2†‡ Alveolar leukocytes 0.1 6 0.02 0.7 6 0.1† 1.2 6 0.2†‡

LPS 1 PEEP 1 Sf 0.87 6 0.08†‡§ 1.9 6 0.1†‡§ 0.3 6 0.06†‡§

* Data are mean 6 SE (n 5 200 for each group). Leukocytes 5 polymorphonuclear leukocytes (PMN) and monocytes. † p , 0.05 versus control group. ‡ p , 0.05 versus LPS group. § p , 0.05 versus LPS 1 PEEP group.

Figure 3. The effect of surfactant aerosol on pulmonary artery pressure (Ppa) over time. See METHODS for group description. LPS 5 the time during infusion of endotoxin; PEEP/PEEP 1 surfactant 5 the time and duration of PEEP (7.5 cm H 2O) or PEEP plus surfactant aerosol. All groups exposed to LPS demonstrated a similar pattern in Ppa changes. Two animals died before the end of the study in the LPS and LPS 1 PEEP groups. The reduced number (n) of animals at each time period is depicted adjacent to the group symbol (n 5 2, 3, or 4). Data are mean 6 SE. *p , 0.05 versus baseline; †p , 0.05 versus control group; §p , 0.05 versus LPS 1 PEEP 1 surfactant group.

animal model of ARDS which is dissimilar to that in humans, the aforementioned studies made measurements for only 3 h after surfactant replacement. This did not allow sufficient time to assess the beneficial effects of surfactant replacement because duration of therapeutic effect is important to establish the necessity and frequency of repeat dosing. However, in a saline lavage model with an 11-h study period, Kobayashi and coworkers showed that surfactant and PEEP improved oxygenation and compliance to near control levels (9). While Kobayashi’s results are impressive, it can be argued that success was due to use of the saline-lavage model. Human ARDS is more closely simulated, and surfactant efficacy better evaluated, by the more severe lung injury models such as hyperoxia (10, 24), acid aspiration (11), oleic acid (12), N-nitroso-N-methylurethane (NNNMU) (25), or endotoxin (13). Therapeutic efficacy of exogenous surfactant has varied in the more clinically significant models. For example, no improvement in lung edema, oxygenation, or static compliance was observed after surfactant replacement in several of the above studies (10–12). On the other hand Harris and coworkers (25) found that exogenous surfactant improved arterial oxygenation and restored surfactant function to control levels. Similarly, Tashiro and coworkers demonstrated that surfactant replacement significantly improved arterial oxygenation 3 h following airway instillation of endotoxin (13). In contrast to the findings of Huang and coworkers (10), Engstrom and coworkers (24) found that surfactant replacement significantly attenuates the increase in alveolar permeability caused by hyperoxia and improves 72-h survival. These studies demonstrate that surfactant replacement is not always as efficacious

in animal models which more closely simulate the severe lung injury associated with ARDS. Surfactant replacement in ARDS patients has yielded variable results as well. A recent clinical trial testing an aerosolized surfactant failed to demonstrate improved pulmonary function or decreased 28-d mortality (14). Lack of clinical efficacy was attributed to inappropriate surfactant composition, inefficient aerosol delivery system, and patient selection. The surfactant used in the clinical trial (Exosurf) does not possess the apoproteins which are now believed essential for normal surfactant function (26). Also, the aerosol delivery system has previously been demonstrated to deposit only 4.3% of the administered surfactant in animal models of ARDS (27). A recurring criticism of sepsis trials is that the patient populations are so heterogenous that a subset of patients who may benefit from surfactant replacement are obscured (28). Indeed, the success of surfactant replacement in a recent Phase II clinical trial argues for continued investigation (16). With variable clinical results, the opportunity and potential for improvement exists. Utilization of severe injury animal models that effectively model ARDS will expose the benefits and limitations of surfactant replacement. We attribute the success of our study to the use of a surfactant with appropriate apoproteins. In previous studies we demonstrated the typical pulmonary recruitment of leukocytes in response to LPS infusion (18, 29). Additionally, both studies showed that instilled surfactant exacerbated the increase in leukocyte infiltration into the alveolus and interstitium. There is evidence that pulmonary lavage can increase leukocyte infiltration. Kawano and colleagues have found that lung lavage causes granulocyte margination with an associated increase in lung injury. In granulocytedepleted rabbits, lavage was associated with minimal lung injury as evidenced by only minor changes in oxygenation and lung morphometrics (30). The exacerbation of leukocyte infiltration by surfactant instillation is another reason why aerosolization is the superior delivery method. Moreover, in the present study aerosolized surfactant replacement actually decreased leukocyte infiltration in the alveolus and the interstitium. There is in vitro evidence that surfactant is anti-inflammatory and plays a role in leukocyte recruitment and infiltration. The location of the alveolar macrophage assures interaction with exogenous surfactant. Thomassen and coworkers have shown in a monocyte cell culture model that surfactant (Exosurf) decreases tumor necrosis factor (TNF) and interleukin-1 (IL-1) production in macrophages (31). The lipid portion of surfactant must be responsible for this decrease because Exosurf contains no apoproteins. There is evidence that the mechanism for this observation involves surfactant modulation of the transcription factor, nuclear factor kappa B (NF-kB) (32). Work from our laboratory demonstrated that the concentration of surfactant altered the release of chemotactic cytokines (chemokines) by LPS-stimulated mac-

Lutz, Carney, Finck, et al.: Aerosolized Surfactant Replacement in ARDS

rophages (33). Supernatant from LPS-stimulated macrophages caused PMN migration. If LPS-stimulated macrophages were incubated in surfactant, PMN migration was reduced. These data may represent a mechanism for decreased pulmonary leukocyte infiltration in the experimental groups treated with surfactant aerosol. Although we do not as yet have specific information on surfactant deposition with our nebulizer, preliminary deposition results are encouraging. In a porcine model, aerosol deposition of rhSOD was compared with the rhSOD deposited by instillation. The present rhSOD deposited was comparable with each technique but a more uniform distribution of rhSOD was measured with aerosolization. Instillation caused a disproportionate amount of rhSOD to be deposited in the left and right caudal lung lobe (unpublished observations). Also, approximately 55% of the aerosol generated was within the 1 to 5 mm range which improves lung deposition (unpublished measurements). In summary, we have demonstrated that exogenous, aerosolized surfactant replacement delivered therapeutically via an improved nebulizer decreases pulmonary leukocyte infiltration and improves oxygenation in our severe lung injury model of ARDS. References 1. Zeni, F., B. Freeman, and C. Natanson. 1997. Anti-inflammatory therapies to treat sepsis and septic shock: a reassessment. Crit. Care Med. 25:1095–1100. 2. Ashbaugh, D. G., D. B. Bigelow, T. L. Petty, and B. E. Levine. 1967. Acute respiratory distress in adults. Lancet 2:319–323. 3. Petty, T. L., G. W. Silvers, G. W. Paul, and R. E. Stanford. 1979. Abnormalities in lung elastic properties and surfactant function in adult respiratory distress syndrome. Chest 75:571–574. 4. Gregory, T., W. Longmore, M. Moxley, J. Whitsett, C. Reed, A. Fowler, L. Hudson, and R. Maunder, C. Crim, and T. Hyers. 1991. Surfactant chemical composition and biophysical activity in acute respiratory distress syndrome. J. Clin. Invest. 65:1976–1981. 5. Seeger, W., G. Stöhr, W. R. D. Wolf, and H. Neuhof. 1985. Alteration of surfactant function due to protein leakage: special interaction with fibrin monomer. J. Appl. Physiol. 56:326–338. 6. O’Brodovich, H., and V. Hannam. 1993. Exogenous surfactant rapidly increases PaO2 in mature rabbits with lungs that contain large amounts of saline. Am. Rev. Respir. Dis. 147:1087–1090. 7. Gommers, D., C. Vilstrup, J. A. H. Bos, A. Larsson, O. Werner, E. Hannappel, and B. Lachman. 1993. Exogenous surfactant therapy increases static lung compliance, and cannot be assessed by measurements of dynamic compliance alone. Crit. Care Med. 21:567–574. 8. Lewis, J. F., B. Tabor, M. Ikegami, A. H. Jobe, M. Joseph, and D. Absolom. 1993. Lung function and surfactant distribution in saline-lavaged sheep given instilled vs nebulized surfactant. J. Appl. Physiol. 74:1256– 1264. 9. Kobayashi, T., H. Kataoka, T. Ueda, S. Murakami, Y. Takada, and M. Kokubo. 1984. Effects of surfactant supplement and end-expiratory pressure in lung-lavaged rabbits. J. Appl. Physiol. 57:995–1001. 10. Huang, Y. T., S. P. Caminiti, T. A. Fawcett, R. E. Moon, P. J. Fracica, F. J. Miller, S. L. Young, and C. Piantadosi. 1994. Natural surfactant and hyperoxic lung injury in primates: I. Physiology and biochemistry. J. Appl. Physiol. 76:991–1001. 11. Lamm, W. J. E., and R. K. Albert. 1990. Surfactant replacement improves lung recoil in rabbit lungs after acid aspiration. Am. Rev. Respir. Dis. 142:1279–1283. 12. Zelter, M., B. J. Escudier, J. M. Hoeffel, and J. F. Murray. 1990. Effects of aerosolized artificial surfactant on repeated oleic acid injury in sheep. Am. Rev. Respir. Dis. 141:1014–1019. 13. Tashiro, K., K. Yamada, W. Z. Li, Y. Matsumoto, and T. Kobayashi. 1996. Aerosolized and instilled surfactant therapies for acute lung injury

845 caused by intratracheal endotoxin in rats. Crit. Care Med. 24:488–494. 14. Anzueto, A., R. P. Baughman, K. K. Guntupalli, J. G. Weg. H. P. Wiedemann, A. A. Raventos, F. Lemaire, W. Long, D. S. Zaccardelli, and E. N. Pattishall. 1996. Aerosolized surfactant in adults with sepsisinduced acute respiratory distress syndrome. N. Engl. J. Med. 334: 1417–1421. 15. Weg, J. G., R. A. Balk, S. Tharratt, S. G. Jenkinson, J. B. Shah, D. Zaccardelli, J. Horton, and E. N. Pattishall. 1994. Safety and potential efficacy of an aerosolized surfactant in human sepsis induced adult respiratory distress syndrome. J.A.M.A. 272:1433–1438. 16. Gregory, T. J., K. P. Steinberg, R. Spragg, J. E. Gadek, T. M. Hyers, W. J. Longmore, M. A. Moxley, G.-Z. Cai, P. Newton, B. R. Mitchell, and A. J. Gold. 1997. Bovine surfactant therapy for patients with acute respiratory distress syndrome. Am. J. Respir. Crit. Care Med. 155:1309– 1315. 17. Picone, A., L. A. Gatto, G. F. Nieman, A. M. Paskanik, and C. Lutz. 1996. Pulmonary surfactant function following endotoxin: effects of exogenous surfactant treatment. Shock 5:304–310. 18. Lutz, C., A. Picone, L. A. Gatto, A. Paskanik, S. Landas, and G. Nieman. 1998. Exogenous surfactant and positive end-expiratory pressure (PEEP) in the treatment of endotoxin induced lung injury. Crit. Care Med. (In press) 19. Fink, M. P., B. P. O’Sullivan, M. J. Menconi, P. S. Wollert, H. Wang, M. E. Youssrf, and J. H. Fleisch. 1993. A novel leukotriene B4-receptor antagonist in endotoxin shock: a prospective controlled trial in a porcine model. Crit. Care Med. 21:1825–1837. 20. Cummings, J. J., B. A. Holm, M. L. Hudak, B. B. Hudak, W. H. Ferguson, and E. A. Eagan. 1992. A controlled clinical comparison of four different surfactant preparations in surfactant-deficient preterm lambs. Am. Rev. Respir. Dis. 145:999–1004. 21. Barry, B. E., and J. D. Crapo. 1985. Application of morphometric methods to study diffuse and focal injury in the lung caused by toxic agents. Crit. Rev. Toxicol. 14:1–32. 22. Gundersen, H. G. J. 1997. Notes on the estimation of the numerical density of arbitrary profiles: the edge effect. J. Microsc. 111:219. 23. Zeger, S. L., and K. Y. Liang. 1992. An overview of methods for the analysis of longitudinal data. Stat. Med. 11:1825–1839. 24. Engstrom, P. C., B. A. Holm, and S. Matalon. 1989. Surfactant replacement attenuates the increase in alveolar permeability in hypoxia. J. Appl. Physiol. 67:688–693. 25. Harris, J. D., F. Jackson, M. A. Moxley, and W. J. Longmore. 1989. Effect of exogenous surfactant instillation on experimental acute lung injury. J. Appl. Physiol. 66:1846–1851. 26. Scarpelli, E. M., E. David, M. Cordova, and A. J. Mautone. 1993. Surface tension of therapeutic surfactant (exosurf neonatal, infasurf, and survanta) as evaluated by standard methods criteria. Am. J. Perinatol. 10:414–419. 27. MacIntyre, N. R., R. E. Coleman, F. S. Schuller, D. Zaccardelli, and E. Pattishall. 1994. Efficiency of the delivery of aerosolized artificial surfactant in intubated patients with the adult respiratory distress syndrome (abstract). Am. J. Respir. Crit. Care Med. 149:A125. 28. Vincent, J. L. 1997. Dear SIRS, I’m sorry to say that I don’t like you. . . . Crit. Care Med. 25:372–374. 29. Nieman, G. F., L. A. Gatto, A. M. Paskanik, B. Yang, R. Fluck, and A. Picone. 1996. Surfactant replacement in the treatment of sepsis induced adult respiratory distress syndrome (ARDS) in pigs. Crit. Care Med. 24:1025–1033. 30. Kawano, T., S. Mori, M. Cybulsky, R. Burger, A. Ballin, B. Cutz, and A. C. Bryan. 1987. Effect of granulocyte depletion in a ventilated surfactant-depleted lung. J. Appl. Physiol. 62:27–33. 31. Thomassen, M. J., J. M. Antal, M. J. Conners, D. P. Meeker, and H. P. Wiedemann. 1994. Characterization of Exosurf(surfactant)-mediated suppression of stimulated human alveolar macrophage cytokine responses. Am. J. Respir. Cell Mol. Biol. 10:399–404. 32. Antal, J. M., L. T. Davis, S. C. Erzurum, H. P. Wiedemann, and M. J. Thomassen. 1996. Surfactant suppresses NF-kappa B activation in human monocyte cells. Am. J. Respir. Cell Mol. Biol. 14:374–379. 33. Finck, C. M., M. G. Hodell, W. H. Marx, A. M. Paskanik, D. J. McGraw, C. J. Lutz, L. A. Gatto, A. L. Picone, and G. F. Nieman. 1998. Endotoxin stimulated alveolar macrophage recruitment of neutrophils and modulation with exogenous surfactant. Crit. Care Med. (In press)