European Journal of Clinical Investigation (1 992) 22, 7 12-7 I 8

Surfactant protein A (SP-A) is decreased in acute parenchymal lung injury associated with polytrauma" U. PISON, U. OBERTACKET, W. SEEGERS & S. HAWGOOD5 Department of Anesthesiology and Intensive Care Medicine, Universitatsklinikum Rudolf Virchow, Freie Universitat Berlin; ?Department of Trauma-Surgery Universitatsklinikum Essen; $Division of Clinical Pathophysiology/Departmentof Internal Medicine, Justus-Liebig Universitat Giessen and §Cardiovascular Research Institute, University of California at San Francisco Received 2 January 1992 and in revised form 15 June 1992; accepted 24 June 1992

Abstract. To further investigate if the pulmonary surfactant system is altered in acute parenchymal lung injury of adults following polytrauma we measured SP-A level and phospholipid composition in 150 sequentially obtained lung lavage samples from polytraumatized patients ( n = 19) beginning at the day of trauma and ending 18 days later or when the patient was extubated. Out of the 19 patients studied 10 had severe parenchymal lung injury (ARDS), nine had moderate lung injury. SP-A was measured using a twomonoclonal sandwich ELISA-assay. Phospholipids were separated using high-performance liquid chromatography and their composition was calculated by comparison with standard phospholipid mixtures. We found immunoreactive SP-A concentrations ranging from 0.1 pg ml-' to 8.5 pg ml-' lung lavage fluid obtained from all patients. The mean SP-A concentration in patients who had severe parenchymal lung injury (ARDS) was 1.06f0.16 pg ml-' lavage fluid, the mean concentration in patients who had only moderate parenchymal lung injury was 1.92 k 0.18 pg ml-' lavage fluid. Both concentrations were lower than in healthy controls (2.74k0.3 pg ml-' lavage fluid; n = 12). In patients who had moderate lung injury the SP-A level normalized, but in patients who had severe lung injury the SP-A level remained low during the timespan examined. SP-A alterations did not correlate to changes in phospholipid composition as determined in lung lavage samples of individual patients. We conclude that alveolar SP-A concentrations decrease in polytraumatized patients who have acute parenchymal lung injury soon after the trauma occurs. In patients who have lung injury of low severity the SP-A level normalizes with recovery, but with more severe parenchymal lung injury SP-A levels remain low. We speculate that the metabolic regulation of individual surfactant components might differ during lung injury and repair Correspondence: Ulrich Pison MD, Freie Universitat Berlin, Universitatsklinikum Rudolf Virchow, Department of Anesthesiology and Intensive Care Medicine, Augustenburger Platz I , lo00 Berlin 65, Germany. * Preliminary results of this study had been reported elsewhere [ I].

712

Keywords. Adult respiratory distress syndrome, parenchymal lung injury, pathomechanism, pulmonary surfactant, SP-A.

Introduction Acute parenchymal lung injury in adults, referred to as adult respiratory distress syndrome (ARDS) in its most severe form, is commonly associated with polytrauma and significantly contributes to patient mortality over and above that caused by the trauma itself [2]. Several animal and human studies have shown that a variety of inflammatory mediators react with blood and lung tissue cells, causing an increase in endothelial and epithelial permeability, resulting in 'high protein' pulmonary oedema, an early clinical sign of acute parenchymal lung injury [3]. The transition from this early stage to later stages of parenchymal lung injury is still poorly understood. Type I1 cells, the cells producing pulmonary surfactant [4], contribute to the repair of the alveolar epithelium after injury by division and differentiation into type I cells [5,6,7]. The susceptibility of epithelial type I1 cells to parenchymal lung injury after polytrauma, or their metabolic response during the repair process is poorly understood. One measure of type I1 cell metabolism is their ability to synthesize and secrete surfactant. Therefore, one way how the transition from early to later stages of parenchymal lung injury could be investigated is by measuring pulmonary surfactant components, the secretion products of type I1 cells, during the course of lung injury and repair. Biologically active pulmonary surfactant is a lipidprotein complex which lines the alveoli and makes breathing at normal transpulmonary pressures feasible [8,9]. Deficient pulmonary surfactant is the primary cause of the respiratory distress syndrome seen in premature infants [lo]. Alterations of pulmonary surfactant composition and function have been also reported to occur in adult patients who have acute parenchymal lung injury [ 1 I , 12,131. To date only changes in surfactant phospholipid composition have

SP-A IN ACUTE PARENCHYMAL LUNG INJURY

713

Table 1. Clinical data of the 10 healthy control subjects and the 19 polytrdumatized patients (who had parenchymal lung injury)*

Group

Number of Age (years) subjccts

10 Control group 19 All trauma patients Patients with moderate lung injury 9 10 Patients with severe lung injury

Survival (I%)

31k5.7 100 31k3.8 35 31k4.3 40 32 & 3.5 28

Lung injury Extravascular lung water score (ml kg-' body weight) 0.00 0.61 +0.01 0.40 k0.03 0.78 k 0.04t

ND 8.18k0.27 7.03 k 0. I6 9.34 0.4 It

* The table gives mean values for healthy control subjects and polytrauma patients averaged over the entire observation period of 18 days; t Significantly different from values formoderdtehng injury ( P< 0.05); ND: Not determined. been reported. However, optimal activity of lung surfactant depends not only on the phospholipid composition, but also on the presence of the specific surfactant proteins SP-A, SP-B, and SP-C [ 141. We have previously reported changes in bronchoalveolar lavage phospholipid content and composition in patients with lung injury following polytrauma [12,13]. We now report changes in lavage SP-A concentration in a subgroup of these patients after an assay to measure SP-A became available. We wanted to know, whether and when SP-A is altered in these patients. To answer these questions, we analyzed SP-A concentration and re-evaluated phospholipid composition in lung lavage samples obtained sequentially from polytrauma patients during the transition from early to later stages of parenchymal lung injury. Patients and methods Patient populurion The 19 patients in this study had polytrauma of class five according to the abbreviated injury scale. The severity of trauma estimated by the injury severity score [ 151 was above 30 points in all patients. None of the patients had direct chest trauma, a history of chronic respiratory or heart failure, or pulmonary capillary wedge pressures above 18 mmHg as a possible sign of acute left heart failure. All patients had acute parenchymal lung injury resulting from polytrauma and required ventilator treatment (Table 1). Positive end-expiratory pressure (PEEP) was applied according to the severity of pulmonary dysfunction. The inspired oxygen fraction (FlOz) was adjusted to maintain arterial oxygen tension above 80 mmHg. All patients received a similar therapeutic regimen. Specifically, crystalloid solutions were infused to maintain urine output above 70 ml h - ' , red blood cells were transfused to a final haemoglobin concentration of approximately 10 g dl-', and inotropic drugs (dopamine, dobutamine, or both) were given in response to changes in haemodynamic variables [ 161. Control group The control group consisted of 10 patients undergoing elective surgical intervention (peripheral plate extrac-

tion about 2 years after osteosynthesis) under general anesthesia. These patients had no history or acute clinical signs of lung or other organ failure. Study protocol

In all study subjects, lung lavage samples for analysis of pulmonary surfactant were obtained using a bronchoalveolar lavage technique. In the polytrauma patients bronchoalveolar lavage was usually carried out every day, beginning 3-8 h after trauma and ending on the 18th day after trauma or when the patient was extubated. Bronchoalveolar lavage was performed in lung areas that had neither visible signs of contusion by inspection during bronchoscopy, nor major infiltrates on chest radiographs. The sequential lavages of at least 2 days in a row were not obtained from the same segment of the lungs. In the control group lung lavage was carried out immediately after endotracheal intubation. To answer the question, whether SP-A is altered in polytrauma patients, we compared the data obtained from patients with data from the control group. To answer the question, when is SP-A altered in polytrauma patients, we looked at changes in SP-A concentration over time in groups of patients. We calculated mean values for each day after trauma and for the early stage of lung injury (0-2 days after trauma), the intermediate stage (3-6 days), and the late stage (7-18 days). Correlations between SP-A and the following variables were tested: extravascular lung water, severity of lung injury, severity of trauma, survival, total lung lavage protein and phospholipid content, and relative amounts of the phospholipids phosphatidylcholine and phosphatidylglycerol. The study design was approved by the local ethics committee. Written informed consent was obtained from patients, or nearest relatives of the patients gave permission. Bronchoalveolar lavage

For bronchoalveolar lavage all subjects received intravenous anesthesia without administration of muscle relaxants or anesthetic gases, as previously described [17]. Ventilator settings of PEEP and F102were not

714

U. PISON et al.

changed for at least 2 h before and during bronchoscopy. Minute ventilation was kept constant during bronchoscopy. A fibreoptic bronchoscope (Pentax, FB- 15A) was directed through the endotracheal tube and was gently wedged in a segment bronchus (e.g. lingula or lateral segment of the right middle lobe). Five consecutive 20-ml volumes of 0.15 M sodium chloride (Fresenius, Bad Homburg, Germany) were instilled and withdrawn under negative pressure. Recovery was 40-70°/0 by volume and volume recovery was without any correlation to severity of trauma or parenchymal lung injury. The lavage fluid was centrifuged (180xg,,,, 10 min, 4°C) to sediment cellular material. Cell-free supernatants were stored at - 7 0 T until further analysis. We did not detect a neutrophilic response by cell counts or an increase in lung permeability by extravascular lung water measurements due to repeat lavages (data not shown). Phospholipid determinations

Phospholipids were extracted from lung lavage samples according to the method of Folch et al. [18]. Total phospholipid content was quantified by the determination of organic phosphorus [19]. Phospholipids were separated using a high-performance liquid chromatography method [20]. Phospholipid composition was calculated by comparison with standard phospholipid mixtures obtained from Sigma, Miinchen, Germany. SP-A and protein determinations

SP-A was measured in lung lavage samples using a two-monoclonal sandwich ELISA-assay. Both monoclonal antibodies were human SP-A specific and directed to different antigenic determinants of SP-A. The monoclonal antibodies and the SP-A standard for this assay were kindly provided by California Biotechnology Inc., Mountain View, California. The first monoclonal antibody (10 pg ml-' diluted in 0.1 M NaHC03, ph 8.3) was placed into microtiter plates (Linbro/Titertek 96-multiwell plates; 100 pl well-'), and incubated overnight at 4°C. Wells were washed three times with phosphate-buffered saline containing O.2OA1 bovine serum albumin and 0.05% Tween-20 (PBS/BSA/Tween), letting the final wash sit for 30 min at room temperature as blocking step. 100 pI of SP-A standard (3.125-100 ng ml-' glycosylated human recombinant SP-A) or lung lavage samples were placed into the wells and incubated for 2 h at 37°C. Plates were washed three times with PBS/BSA/Tween and 100 p1 well - I of the biotinylated second monoclonal antibody, the concentration of which was determined in a Checkerboard assay, was added and incubated overnight at 4°C. Plates were washed with PBS/BSA/ Tween and 100 p1 well-' of streptavidin-horseradish peroxidase (Zymed) was added and incubated for 30 min at 37°C. Wells were washed five times with PBS/ BSA/Tween and 200 pl well-' of freshly prepared colour reagent (25 ml of 0.1 M citric phosphate, pH 5;

10 mg of o-phenylenediamine; 10 pl of 30% hydrogen peroxide) was added. Developing was stopped with 50 p1 of 2 M sulfuric acid and absorbance was measured with an automatic microplate reader (Biorad) at 490 nm. Samples were run in triplicates or duplicates depending on the amount of sample available. Human control serum ( 1 : 100) did not react in the ELISA assay, nor did addition of serum to lung lavage samples alter the assay. Total protein content was determined in lung lavage samples by the method of Lowry [2 I]. To reduce interference by lipids, the protein content was measured in the presence of 0.1 Yo sodium dodecyl sulfate.

Assessment of lung injury

The assessment of the severity of lung injury was based on a composite scoring system that combined separate evaluation of the chest radiograph, the oxygenation index, pulmonary artery pressure (PAP), and respiratory system compliance [13]. In brief, the composite lung injury score was calculated according to the following equation, where different components of the score were weighted by using different multipliers: lung injury score = 0.108 +(chest radiographic score index x 0.6208) + (mean x 0.15 12)+ (oxygenation PAP x 0.00725)-(respiratory system compliance x 0.00416). Patients were scored for lung injury every 6 h within the first 48 h after trauma and then every 24 h. Based on these sequential lung injury scores, a mean score was calculated for the whole observation period. A computerized regression model was used to divide the whole group of patients into two subgroups having either moderate or severe lung injury. Averaged over the whole observation period, the mean score of the first group was 0.40f0.03, and of the latter 0.78 f0.04. Measurement of clinical variables

Chest radiographs were obtained using routine mobile unit radiography. Radiographs were taken from patients in the semi-erect position during inspiration. Gas exchange abnormalities were quantified using the arterial-to-alveolar partial pressure of oxygen ratio (oxygenation index). Alveolar oxygen tension (PA02) was calculated from standard alveolar air equation [22]. Arterial oxygen tension (Pa02) was measured with a blood gas electrode (Corning). Pulmonary artery pressure was measured with a Swan-Ganz balloon catheter [23]. Dynamic respiratory system compliance was estimated with a Driiger instrument (Drager, Lubeck, Germany). Extravascular lung water content was determined with the thermal green-dye double-indicator dilution technique as described [24]. Data analysis

Results are expressed as mean & SEM. The Pearson product moment was calculated to determine the

SP-A IN ACUTE PARENCHYMAL LUNG INJURY

715

Table 2. Surfactant components in bronchoalveolar lung lavage (BAL) samples obtained from the 10 healthy control subjects and the 19 polytraumatized patients (who had parenchymal lung injury),

Group

Total Number of protein samples (pg m1-l)

12 Control group I50 All trauma patients Patients with moderate lung injury 75 75 Patients with severe lung injury

I 17.9f8.4 220.8 f 15.81 143.4&12.6 320,8+27.3$5

?4 of total

Total phospholipid SP-A (pg m I - ' ) (pgml-I)

PCt

84. I k 5.9 65.4 f4.8 65.7f7.4 65.1 f 5 . 8

62.8 f I .2 10.0 f0.7 52.3 0.9$ I .7 f0.2: 1.6+0.2$ 56.3k 1.01 48.1 f 1.415 1.9+0.3$

2.74 f0.30 I .49 & 0.26: 1.92+0.18$ 1.06+0.16$(j

PGt

* The tablc gives mean values for polytrauma patients averaged over the entire observation period of 18 days and mean values for healthy control subjects averaged over 12 single bronchoalveolar lavages; t PC: phosphatidylcholine. PG: phosphatidylglycerol; $ Significantly different from control ( P i 0.05); 0 Significantly different from values for moderate lung injury (P