Inhaled Nitric Oxide for Adult Respiratory Distress Syndrome After Pulmonary Resection

Inhaled Nitric Oxide for Adult Respiratory Distress Syndrome After Pulmonary Resection Douglas J. Mathisen, MD, Elbert Y. Kuo, MMS, Chiwon Hahn, MD, Ashby C. Moncure, MD, John C. Wain, MD, Hermes C. Grillo, MD, William E. Hurford, MD, and Cameron D. Wright, MD Division of General Thoracic Surgery and Department of Anesthesia, Massachusetts General Hospital, Boston, Massachusetts Background. The adult respiratory distress syndrome (ARDS) developing after pulmonary resection is usually a lethal complication. The etiology of this serious complication remains unknown despite many theories. Intubation, aspiration bronchoscopy, antibiotics, and diuresis have been the mainstays of treatment. Mortality rates from ARDS after have been reported as high as 90% to 100%. Methods. In 1991, nitric oxide became clinically available. We instituted an aggressive program to treat patients with ARDS after pulmonary resection. Patients were intubated and treated with standard supportive measures plus inhaled nitric oxide at 10 to 20 parts/ million. While being ventilated, all patients had postural changes to improve ventilation/perfusion matching and management of secretions. Systemic steroids were given to half of the patients. Results. Ten consecutive patients after pulmonary resection with severe ARDS (ARDS score 3.1 0.04) were treated. The mean ratio of partial pressure of arterial oxygen to the fraction of inspired oxygen at initiation of treatment was 95 13 mm Hg (mean SEM) and improved immediately to 128 24 mm Hg, a 31% 8% improvement (p < 0.05). The ratio improved steadily over the ensuing 96 hours. Chest x-rays improved in all patients and normalized in 8. No adverse reactions to nitric oxide were observed. Conclusions. We recommend the following treatment regimen for this lethal complication: intubation at the first radiographic sign of ARDS; immediate institution of inhaled nitric oxide (10 to 20 parts per million); aspiration bronchoscopy and postural changes to improve management of secretions and ventilation/perfusion matching; diuresis and antibiotics; and consideration of the addition of intravenous steroid therapy. (Ann Thorac Surg 1998;65:1894 902) 1998 by The Society of Thoracic Surgeons The Adult Respiratory Distress Syndrome (ARDS) is seen in a variety of clinical settings. The occurrence of ARDS after pulmonary resection is one such clinical setting. The syndrome has remarkably consistent findings, which include onset 24 to 72 hours postsurgery, interstitial infiltrate on chest x-ray, hypoxia, and rapid progression. This syndrome refers specifically to patients without signs of pneumonia, heart failure, pulmonary emboli, fluid overload, or bronchopleural fistula. It occurs more commonly after (2% to 5%; right more common than left) than lobectomy ( 1%) [1, 2]. No clear-cut etiology has been identified, but many have postulated excessive fluid, lymphatic interruption, barotrauma, cytokine release, or activation of the complement system as explanations for ARDS. Development of ARDS after is often fatal, with mortality rates reported between 30% and 100% [2 5]. Treatment has consisted mainly of aggressive supportive measures: mechanical ventilation, diuresis, broad-spectrum antibiotics, and pulmonary toilet. All too often these measures fail to reverse severe hypoxia and death. Specific therapy has been unavailable because of our lack of understanding of what triggers the problem and what mediates the injury to the lung that leads to the clinical syndrome. Nitric oxide (NO) was discovered in 1987 to be the vasodilator responsible for the biologic activity of endothelium-derived relaxing factor [6]. When inhaled at low levels, NO selectively dilates the pulmonary circulation. Significant systemic vasodilation does not occur because NO is inactivated by rapid binding to hemoglobin. Zapol and coworkers [7, 8] postulated that nitric oxide, when inhaled as a gas at low concentrations, should diffuse into the pulmonary vasculature of ventilated lung regions and cause relaxation of pulmonary vascular smooth muscle, thereby decreasing the pulmonary hypertension often seen in ARDS. Because NO is inhaled, the gas should be distributed predominantly to well-ventilated alveoli and not to collapsed or fluid-filled areas of the lung. Local vasodilation of well-ventilated lung regions should cause Presented at the Thirty-fourth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA Jan 26 28, 1998. Address reprint requests to Dr Mathisen, Massachusetts General Hospital, 55 Fruit Street, Blake 1570, Boston, MA 02114 (e-mail: mathisen. douglas@mgh.harvard.edu). This article has been selected for the open discussion forum on the STS Web site: http://www.sts.org/annals 1998 by The Society of Thoracic Surgeons 0003-4975/98/$19.00 Published by Elsevier Science Inc PII S0003-4975(98)01167-9

Ann Thorac Surg MATHISEN ET AL 1998;66:1894 902 NITRIC OXIDE FOR ARDS 1895 a diversion of pulmonary artery blood flow toward wellventilated alveoli, improving the matching of ventilation to perfusion and augmenting arterial oxygenation during ARDS. Nitric oxide became clinically available in 1991 and has been evaluated as a therapeutic option for ARDS from a variety of etiologies [7 12]. Nitric oxide was found to be a valuable therapeutic option, raising oxygenation, lowering requirements of mechanical ventilation, lowering pulmonary artery pressures, and decreasing the levels of inspired oxygen concentrations in many patients. We added inhaled NO as a therapeutic modality to the treatment of postpulmonary-resection ARDS in 1993. This report records our results of treatment in 10 consecutive patients with postpulmonary-resection ARDS treated by the standard supportive measures and inhaled nitric oxide. Material and Methods We identified all patients who developed ARDS after pulmonary resection from 1993 to 1997. All patients met the criteria for ARDS developed by the American- European Consensus Conference for ARDS: acute respiratory insufficiency requiring mechanical ventilation; ratio of partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO 2 /FIO 2 ratio) less than 200 mm Hg; bilateral pulmonary consolidations on chest x-ray; and absence of clinical signs of left atrial hypertension [13]. Patients with other explanations for ARDS such as sepsis, heart failure, aspiration, or bronchial fistula were excluded from consideration. Ten patients were identified from January 1993 to July 1997, who fulfilled these criteria and were treated with inhaled NO. The severity of ARDS just before administration of NO was calculated by the lung injury scoring system described by Murray and coworkers [15]. This method grades alveolar consolidation on chest radiograph, gas exchange abnormalities expressed as PaO 2 /FIO 2 ratio, the peak end-expiratory pressure score when ventilated, and respiratory system compliance. Scores greater than 2.5 are indicative of severe ARDS. Although this was not a prospective randomized study, we sought a comparable group of patients who were treated for ARDS with the methods described previously, before NO was clinically available. We reviewed charts from between 1986 and 1993 and identified 7 patients who developed ARDS after. Identical criteria were used for defining ARDS in these patients. Similar demographic and clinical data were obtained. Standard Therapy There is a great deal of uniformity in the intraoperative and postoperative management of patients undergoing pulmonary resection in our institution. Airways are managed with double-lumen endotracheal tubes. An inhalational anesthetic agent is used most often. Intraoperative fluids are scrupulously restricted because of concern regarding development of ARDS. All patients with signs of ARDS postoperatively are managed by diuresis, early intubation, mechanical ventilation, and aspiration bronchoscopies. Swan Ganz catheters are not routinely used because of concerns about pulmonary artery stumps and the need to place the catheters under fluoroscopy to ensure placement in the nonoperated lung. Ventilated patients in the intensive care units in our hospital have frequent changes in their position while being ventilated. Patients are placed on continuous-rotational beds for lateral positioning and are in the prone position for 6 hours once or twice daily. The attention to postural changes is felt to be important to enhance mobilization of secretions and redistribution of ventilation to the nondependent lung. Steroids, when administered, are given at the initiation of NO therapy. Solumedrol 50 to 100 mg intravenously every 6 hours for 2 to 3 days is used. Administration and Monitoring of Nitric Oxide Inhaled nitric oxide was administered immediately after the diagnosis of ARDS. The concentration of NO varied between 5 and 20 parts per million. Nitric oxide was administered to all patients using the methods described previously [15]. Nitric oxide source gas (780 to 880 parts per million; Airco, Riverton, NJ) was stored in nitrogen (N 2 ). Nitric oxide in nitrogen was then mixed with N 2 or air (Bird Blender, Palm Springs, CA) and delivered to the air intake of a Puritan-Bennett 7200 ventilator (Carlsbad, CA). In the ventilator, the NO/N 2 mixture was blended with oxygen and administered to the patient during inspiration. Inspired NO and NO 2 concentrations were measured with a chemoluminescence analyzer (model 14A, Thermo Environmental Instruments, Franklin, MA; or model CLD700AL, Eco Physics, Postfach, Switzerland) or with a NO electrochemical analyzer (Bedfont Scientific, Kent, UK). The inspired concentration of NO 2 was monitored intermittently with an electrochemical NO 2 analyzer (Bedfont Scientific). Blood methemoglobin levels were measured daily in every patient. Breathing circuits capable of delivering inhaled NO were available to allow manual ventilation of the patient during tracheal suctioning or transport. Measurements Clinical and radiographic data were collected each day while patients were receiving NO. Data collected included blood gas measurements, white blood cell count, maximum daily temperature, ventilator settings (positive end-expiratory pressure, peak inspiratory pressure), FIO 2, hemodynamics (heart rate, central venous pressure, and, when available, pulmonary artery pressure, pulmonary capillary wedge pressure, cardiac output, cardiac index, stroke volume), cardiac isoenzymes, culture data, electrocardiogram, bronchoscopy results, and chest roentgenogram. Statistical Analysis All statistical data are presented as mean values standard deviation. The effects of treatments are expressed as a comparison between the mean baseline value before treatment, the mean value immediately after treatment, and the mean value at the end of treatment.

1896 MATHISEN ET AL Ann Thorac Surg NITRIC OXIDE FOR ARDS 1998;66:1894 902 Two-tailed paired t tests with multiple comparison adjustments were used to compare values recorded during treatment with those recorded at baseline. For data that were analyzed using multiple-comparison adjustments p values less than 0.0083 were considered significant. Both two-tailed paired t test with multiple comparison adjustments and Wilcoxon signed rank test were performed on the steroids versus nonsteroids data. Fisher s exact test was used to compare mortality between the two groups of ARDS patients (with and without NO treatment). p values less than 0.05 were considered significant. Results The average age of the patients treated from 1993 to 1997 was 58.4 4.0 years, ranging from 39 to 73 years. There were six men and four women. Five of 10 patients received preoperative radiation (4200 cgy). Two patients underwent lobectomy (one of whom had a previous lobectomy on the contralateral side) and 8 underwent (left, 5 patients; right, 3 patients) (Table 1). Eighty-three pneumonectomies and 329 lobectomies were performed at the Massachusetts General Hospital during the period of this study. The incidence of ARDS after lung resection was 7.5% for pneumonectomies and 0.9% for lobectomies. Pathology revealed 6 patients with adenocarcinoma, 2 with large cell carcinomas, 1 with squamous cell carcinoma, and 1 with metastatic non small cell carcinoma. The mean preoperative vital capacity was 2.6 0.2 L (69.0 5.7% of predicted). The mean forced expiratory volume in 1 second was 2.1 0.2 L (65.4 6.1% of predicted). The predicted postoperative forced expiratory volume in 1 second for the 6 patients who had preoperative split-function ventilationperfusion scans was 1.4 0.2 L. Eight of 10 patients had diffusion capacity measured preoperatively, with a mean of 15.4 1.2 (57.6 4.6% of predicted). There were no intraoperative complications. There were no intraoperative blood transfusions. The average ARDS score just before induction of NO therapy was 3.1 0.1. The median number of days from operation to diagnosis of ARDS was 2.5. The median length of stay in the hospital was 31 days (range, 20 to 141 days). The median length of stay in the intensive care unit was 26 days (range, 12 to 90 days). The median time intubated was 24 days (range, 7 to 101 days). Effect on Oxygenation Inhalation of NO produced an immediate increase in the PaO 2 /FIO 2 ratio for each of the 10 patients in this study (Table 2). A significant increase in this ratio is considered to be greater than 20%. The mean PaO 2 /FIO 2 ratio for all patients increased significantly, from a baseline of 95 14 mm Hg before NO to 128 24 mm Hg ( p 0.004) immediately after NO, a 31 8% increase ( p 0.0003). This increase ranged from 10.0 mm Hg to 147 mm Hg. The percent increase ranged from 9% (patient 4) to 85%. The mean PaO 2 /FIO 2 ratio increased significantly, to 178 22, 197 23, 202 21, and 213 28 at 24, 48, 72, and 96 hours, respectively ( p 0.005) (Figs 1, 2). The median number of days receiving NO was 8, ranging from 5 to 39 days. The peak benefit in the PaO 2 /FI0 2 ratio for all patients averaged 285.0 19.5 mm Hg, with an average percent peak increase of 240.7 43.2% ( p 0.005). The median number of days between the initiation of NO treatment and the peak PaO 2 /FIO 2 ratio was 6.5, ranging from 3 to 39 days. In the 5 patients who were administered intravenous steroids, there appeared to be a benefit immediately after instituting therapy, compared with those patients who did not receive steroids (Table 3). The percent improvement was 34% for those receiving steroids, compared with 28% for those who did not receive steroids. This was not statistically significant, however. There was no difference in percent improvement at any other interval. Effect on Mechanical Ventilatory Requirements All patients showed a decrease in peak end-expiratory pressure and peak inspiratory pressure values during NO administration, compared with their baseline before treatment. Mean peak end-expiratory pressure decreased from 9.8 0.6 mm Hg to 6.4 0.7 mm Hg ( p 0.05). Mean peak inspiratory pressure decreased from 35 2mmHgto27 3mmHg(p 0.05). Effect of Nitric Oxide on Radiologic Appearance All patients showed an improvement in their chest roentgenogram within 48 hours of NO administration. Using the scoring system for ARDS described earlier, the chest roentgenograms improved from a mean chest radiograph score of 3.8 0.1 to 1.1 0.2 within 48 hours after initiation of NO to 0.2 0.1 by the time NO was discontinued ( p 0.005). Eight of the 10 patients had normal lung fields after treatment with NO. Hemodynamic Measurements There was no significant change in heart rate, systemic arterial pressure, or central venous pressure with NO administration. Only 2 of the 10 patients had pulmonary artery catheters placed. No conclusions can be drawn with respect to the effect that NO had on pulmonary artery pressures in the patients in this study, as 8 patients had no pulmonary artery catheter. Morbidity and Mortality Seven of 10 patients survived, yielding an overall inhospital mortality rate of 30.0%. No toxicity of NO was identified in any patient. Complications included cholecystitis in 1 patient, inferior wall myocardial infarction in 1 patient, Clostridium difficile colitis in 1 patient, and fungal sepsis in 1 patient. None of the three deaths was directly related to ARDS. One patient died of unexplained sepsis 20 days postoperatively. He had recovered from ARDS and had been maintained on NO (10 parts per million) for 17 days while weaning from mechanical ventilation. Two days before his death, he developed fevers to 104 F. No source of sepsis was identified. Two other patients died after recovering from their initial ARDS. Administration of NO

Table 1. Demographics and Clinical Summary of Patients Treated With Nitric Oxide Chemoand NO Patient No. Age (years) Sex Operation 1 58 F 4/7/94: right upper lobectomy 2 65 M 7/19/93: left intrapericardial 3 70 M 9/21/94: left 4 46 M 1/5/96: right 5 61 M 4/17/96: left lower lobectomy 6 72 M 8/14/95: right completion 7 73 F 2/20/96: right 8 41 F 7/19/96: left 9 39 F 9/11/96: left 10 59 M 6/12/97: left completion ARDS Dose Days on Days Days in Score (ppm) NO Intubated ICU Days in Radiation Hospital Therapy Steroids Complications Status 3.25 16 39 101 90 141 N N Acalculous cholecystitis A 3.25 20 8 47 32 64 Y N Inferior wall A myocardial infarction; Pseudomonas pneumonia 3.00 20 7 11 14 21 N N Positive A clostridium difficile 3.00 20 5 10 13 31 Y Y Fungemia A 3.00 10 17 17 20 20 N Y Infection of D unknown etiology 3.25 20 8 29 31 31 Y Y Staphylococcus D aureus and Pseudomonas aeruginosa pneumonia 3.00 20 8 19 19 23 N Y Bronchopleural D fistula 3.25 10 19 30 32 42 N N None A 3.00 10 26 34 34 43 N N None A 3.00 5 5 7 12 17 Y Y None A Mean SE 58.4 4.0 3.1 0.1 14.2 3.5 29.5 9.1 29.7 7.2 43.3 11.7 Mortality Rate 30.0% ARDS adult respiratory distress syndrome; ICU intensive care unit; NO nitric oxide; ppm parts per million; SE standard error. Ann Thorac Surg MATHISEN ET AL 1998;66:1894 902 NITRIC OXIDE FOR ARDS 1897

1898 MATHISEN ET AL Ann Thorac Surg NITRIC OXIDE FOR ARDS 1998;66:1894 902 Table 2. PaO 2 /FIO 2 Ratio, an Index of the Efficiency of Arterial Oxygenation in the First 4 Days and at the Time of Discontinuation of Nitric Oxide Treatment Day NO Discontinued 96 Hours on NO 72 Hours on NO 48 Hours on NO 24 Hours on NO Immediately After NO Before NO (baseline) Patient No. 1 76.7 88.8 15.7 105.7 37.8 98.8 28.8 72.5 5.5 76.7 0.0 165.5 115.8 2 112.9 137.1 21.5 262.0 132.1 287.5 154.7 314.0 178.1 325.0 187.9 350.0 210.0 3 63.0 73.8 17.1 132.5 110.3 138.2 119.4 171.1 171.6 325.0 415.9 297.5 372.2 4 169.0 184.4 9.1 272.0 60.9 307.5 82.0 250.0 47.9 257.5 52.4 257.5 52.4 5 91.1 111.1 22.0 92.5 1.5 122.9 23.9 155.7 71.9 207.3 127.6 68.4 24.9 6 64.0 82.1 28.3 123.8 93.4 150.0 134.4 190.0 196.9 164.4 156.9 350.0 446.9 7 64.0 82.1 26.3 136.7 113.6 156.0 143.8 242.0 278.1 162.0 153.1 212.0 231.2 8 69.0 117.9 70.9 210.0 204.3 202.2 193.0 188.9 173.8 143.6 108.1 302.5 338.4 9 68.9 78.9 14.5 172.9 150.9 240.0 248.3 190.0 175.8 154.5 124.1 282.0 309.3 10 173.3 320.0 84.7 267.5 54.4 265.0 52.9 250.0 44.3 318.3 83.7 318.3 83 Mean SE 95.2 13.6 127.6 24.0 31.0 8.1 177.6 22.2 95.9 18.8 196.8 23.4 118.1 22.9 202.4 20.8 133.3 27.9 213.4 27.9 141.0 35.1 260.4 28.1 213.4 48.8 NO nitric oxide; SE standard error. had been discontinued for 23 days in 1 patient and 11 days in the other. Both patients died of multisystem organ failure predominantly from pulmonary sepsis. Historical Group (Before Nitric Oxide) Between 1986 and 1993, 7 patients (four men, three women) developed ARDS after (4.4% of pneumonectomies during this time period). Five patients had undergone right and 2 had undergone left. The mean ARDS score was 2.5 0.2. The median length of stay in the intensive care unit was 20 8 days. The median number of days intubated was 17 7 days. Only 1 patient survived, for a mortality of 86%. All patients were treated in a similar fashion to the recent 10 patients, with the exception of inhaled NO and intravenous steroids. Comment The development of ARDS after pulmonary resection has been a difficult challenge for thoracic surgeons. The insidious onset of interstitial changes on a chest radiograph 1 to 2 days after pulmonary resection forewarning of impending respiratory distress is familiar to all thoracic surgeons. Despite aggressive supportive measures, a progressive downhill course often follows. One of the greatest sources of frustration has been the inability to identify what is responsible for the occurrence of this devastating clinical problem. The development of ARDS is completely unpredictable, with no apparent correlation to age, gender, lung function, comorbid conditions, blood loss, or anesthetic techniques. This was certainly true of our patients. All patients had good preoperative lung function, had no significant comorbid conditions, and more patients in this series had left pneumonectomies than right pneumonectomies. A disproportionate number of patients in this series had preoperative irradiation (5 of 10 patients). Further work needs to be done to quantitate the risk preoperative irradiation poses to the development of ARDS. Because of the relative infrequency ( 5% of pneumonectomies), it has been difficult to study the problem prospectively. Early speculation about the cause of this problem focused on fluid overload, interruption of lymphatic drainage, or barotrauma during surgery [1 3]. Fluid overload as an explanation seemed appealing, but has not been proved to be the case clinically or experimentally [16 18]. This is not the case in our own experience either. Our anesthesiologists pay strict attention to intraoperative fluid balance because of our concern over the role of excessive fluid administration and ARDS. None of the patients in this series had excessive fluid administration (intraoperatively or in the first 48 hours after surgery). The occurrence of alterations in lymphatic drainage secondary to surgery in the opposite hemithorax seems unlikely to account for this syndrome. Intraoperative barotrauma is difficult to evaluate and monitor minute to minute during an operation. Barotrauma and the damage it creates could explain the findings seen in postpulmonary-resection ARDS and deserves further investigation.

Ann Thorac Surg MATHISEN ET AL 1998;66:1894 902 NITRIC OXIDE FOR ARDS 1899 Fig 1. Mean PaO 2 /FIO 2 for all patients. PaO 2 /FIO 2 ratio of partial presence of arterial oxygen to the fraction of inspired oxygen. Interest in the pathogenic mechanisms of acute lung injury has focused recently on the cellular and biochemical mediators of ARDS. The most fundamental early physiologic characteristic of acute lung injury is an increase in protein permeability across the endothelial and epithelial barriers of the lung. The pulmonary endothelium is actively involved in the development of acute lung injury by mediating cell-cell adhesion, which is the initial step in leukocyte migration; by changing barrier permeability, which allows movement of protein-rich fluid into the interstitium of the lung; and by its ability to release and metabolize vasoactive and inflammatory substances (serotonin, bradykinin, endothelin, NO, and cytokines) [19]. It is quite possible that injury to the endothelium during pulmonary resection may be the inciting event leading to the development of ARDS. As our ability to understand the events of endothelial injury increases, it may be possible to identify specifically which of the possible mechanisms is responsible and to prevent the problem or to intervene therapeutically in a more precise manner. The lung epithelium may also play an important role in the pathogenesis of ARDS [19, 20]. Less is currently known about the epithelium because of the lack of specific markers of lung epithelial injury. Alterations in surfactant have been demonstrated in acute lung injury [21]. Further investigation in this area needs to be done. At the cellular level, the lung is known to be a reservoir for neutrophils, as well as alveolar and interstitial macrophages. These are known to be important sources of cytokines that may play an important role in acute lung inflammation and injury [22]. The activation of neutrophils and macrophages with the subsequent release of Fig 2. Mean percent increase in PaO 2 /FIO 2 ratio from baseline. PaO 2 /FIO 2 ratio of partial presence of arterial oxygen to the fraction of inspired oxygen.

1900 MATHISEN ET AL Ann Thorac Surg NITRIC OXIDE FOR ARDS 1998;66:1894 902 Table 3. Result of steroid administration on PaO 2 /FIO 2 Parameter Steroids No Steroids 112 78 Immediately after NO 160 99 change 34% 28% NO nitric oxide; PaO 2 /FIO 2 ratio of partial presence of arterial oxygen to the fraction of inspired oxygen. cytokines and activation of the complement system could explain many of the findings of ARDS in postsurgical patients. Further work is needed to clarify the role of these components in the process. Nitric oxide was first identified as a vascular smooth muscle relaxing factor. This is undoubtedly an important property in the response of patients with ARDS to the administration of inhaled NO. By reducing pulmonary hypertension and improving ventilation and perfusion matching, oxygenation should be improved. This was certainly the case in our patients. All showed an improvement in oxygenation immediately after institution of inhaled NO. This improvement in oxygenation, as measured by the ratio of PaO 2 /FIO 2, continued to improve throughout the first 96 hours after initiation of therapy. For this reason, inhaled NO should be continued until requirements for mechanical ventilation are minimal. Although only 2 of our patients had pulmonary artery catheters, the pulmonary vasodilating capacity of NO was undoubtedly an important factor in the response seen in our patients. Just as importantly, no patients had the systemic vasodilation so common with other dilating agents. As more becomes known about the mechanism of action of NO, it is apparent that it may be beneficial in many other ways in patients with ARDS [12]. It has been shown that NO inhibits neutrophil and platelet adhesion. As stated earlier, the release of cytokines from these cells may be involved in the acute lung injury of ARDS. By limiting adhesion of these cells to the endothelium, NO could ameliorate the effect of the cytokines released by these cells. Nitric oxide has also been shown to be a free radical scavenger. Free radicals have been implicated in acute lung injury as well. Nitric oxide also has been shown to relax bronchial smooth muscle. As more is known about NO, other mechanisms of action might be identified to explain the beneficial effects seen in ARDS. Many studies have been published evaluating the efficacy of inhaled NO in patients with ARDS from a variety of causes [4, 5]. Most of the reports have been retrospective. Manktelow and colleagues [10] reported a 58% response rate in patients with ARDS, as measured by a 20% increase in the PaO 2 /FIO 2 ratio or a 20% decrease in pulmonary vascular resistance. In this study of ARDS from a variety of causes, only 33% of patients with septic shock responded favorably. Preliminary results of a randomized double-blinded multicenter trial demonstrated that, compared with placebo (nitrogen), inhaled NO improved PaO 2, decreased PA pressure, and decreased intensity of mechanical ventilation [11]. Overall, mortality rate and duration of mechanical ventilation were unaffected by breathing NO. It is difficult to explain the 100% response rate seen in our small series, compared with the response rates reported for ARDS from other causes. The development of ARDS after pulmonary resection is not complicated by the effects of sepsis, transfusions, aspiration, and burns seen in most series of ARDS patients. Regardless of the exact mechanism of action, NO does seem to have a position impact on patients who develop ARDS after pulmonary resection. Before the availability of NO, aggressive supportive measures were all that was available. As evidenced by the 7 patients we treated from 1987 to 1993 without NO, the course of ARDS was progressive and often fatal (6 of 7 patients died). A report from the Mayo Clinic of 22 patients who developed ARDS after pulmonary resection had a 100% mortality rate [4]. All of the patients treated with aggressive supportive measures and inhaled NO in our series showed dramatic improvement and recovery from ARDS. Chest radiographs improved dramatically and normalized in 8 of 10 patients. The improved oxygenation allowed reduction in FIO 2, thereby avoiding the toxicity associated with prolonged administration of high concentrations of oxygen. The pressures required for mechanical ventilation (peak end-expiratory and peak inspiratory pressures) also decreased, thereby reducing these effects on the lung as well. None of the three deaths in this series were related directly to ARDS. All three patients ultimately died of sepsis (one from an unknown source and two late pneumonias). The 2 patients who died as a result of pneumonia had had NO discontinued 12 and 22 days earlier. The pneumonia was a complication of the requirement for prolonged mechanical ventilation in 1 patient and as a complication of a late bronchopleural fistula in another patient. These 3 patients all received intravenous steroids for 48 hours at the initiation of NO therapy. It is doubtful that the steroids played any role in the sepsis that developed much later in these 3 patients. Steroids were administered to half of the patients in this series. They were always administered with the initiation of NO therapy and discontinued after 2 to 3 days. There was an initial benefit of improved oxygenation, compared with those who didn t receive steroids, but it was not statistically significant. Steroid administration could reduce the inflammatory component of acute lung injury and deserves further investigation as an adjunct to NO administration. Inhaled nitric oxide has proved to be an effective and safe therapy for patients with ARDS. At the prescribed doses of 5 to 20 parts per million, few serious side effects have been identified. Precautions mentioned previously should be observed, however. Based on our experience, the following recommendations for the management of ARDS after pulmonary resection can be made: 1. early intubation and mechanical ventilation 2. diuresis to improve fluid balance

Ann Thorac Surg MATHISEN ET AL 1998;66:1894 902 NITRIC OXIDE FOR ARDS 1901 3. broad-spectrum antibiotic coverage 4. aggressive pulmonary toilet, including aspiration bronchoscopy and frequent position changes of the patient (to include prone and lateral positioning during mechanical ventilation) 5. early institution of inhaled NO 10 to 20 parts per million and continued until ventilation requirements are minimal (FIO 2 40%, peak endexpiratory pressure 10, NO requirement for ionotropic support) 6. consideration should be given for the addition of a short course of intravenous steroid therapy (48 hours). Further work needs to be done to determine what triggers the development of ARDS after pulmonary resection, what mediates the lung injury seen in ARDS, and to determine the mechanism of action responsible for the beneficial effects of NO seen in patients with ARDS after pulmonary resection. References 1. Slinger PD. Perioperative fluid management for thoracic surgery: the puzzle of post pulmonary edema. J Cardiothorac Vasc Anesth 1995;9:442 51. 2. Zeldin RA, Normadin D, Landtwing BS, et al. Post pulmonary edema. J Thorac Cardiovasc Surg 1984; 87:359 65. 3. Verheijen-Breemharr L, Bogaard JM, Van Den Berg B, et al. Post pulmonary edema. Thorax 1988;43: 323 6. 4. Turnage WS, Lunn JJ. Post pulmonary edema: a retrospective analysis of associated variables. Chest 1993;103:1646 50. 5. Waller DA, Gebitekin C, Saunders NR, Walker DR. Noncardiogenic pulmonary edema complicating lung resection. Ann Thorac Surg 1993;55:140 3. 6. Ignarro LJ, Buga GM, Woods KS, Byrns RE, Chaudhuri G. Endothelium-derived relaxing factor produced and released from artery and vein is nitric oxide. Proc Natl Acad Sci USA 1987;84:9265 9. 7. Frostell C, Fratacci MD, Wain JC, Jones R, Zapol WM. Inhaled nitric oxide: a selective pulmonary vasodilator for reversing hypoxic pulmonary vasoconstriction. Circulation 1991;83:2038 47. 8. Zapol WM, Hurford WE. Inhaled nitric oxide in adult respiratory distress syndrome and other lung diseases. Adv Pharmacol 1994;31:513 30. 9. Zapol WM, Hurford WE. Inhaled nitric oxide in the adult respiratory distress syndrome and other lung diseases. New Horizons 1993;1:638 50. 10. Manktelow C, Bigatello LM, Hess D, Hurford WE. Physiologic determinants of the response to inhaled nitric oxide in patients with acute respiratory distress syndrome. Anesthesiology 1997;87:297 307. 11. Dellinger RP, Zimmerman JL, Hyers TM, et al. Inhaled nitric oxide in ARDS: preliminary results of a multicenter clinical trial. Crit Care Med 1996;24:A29. 12. Dellinger RP, Zimmerman JL, Taylor RW, et al. Effects of inhaled nitric oxide in patients with acute respiratory distress syndrome: results of a randomized phase II trial. Crit Care Med 1998;26:15 23. 13. Hurford WE, Hess D, Zapol WM. Inhaled nitric oxide therapy. In: Yaksh TL, Lynch C, Zapol WM, Mase M, Biebuych JF, Saidman LJ, eds. Anesthesia Biologic Foundations. Philadelphia: Lippincott, 1997:1431 53. 14. Bernard GR, Luce GR, Sprung CL, et al. High dose corticosteroids in patients with the adult respiratory distress syndrome. N Engl J Med 1987;317:1565 70. 15. Murray JF, Matthay MA, Luce JM, Flick MR. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988;138:720 3. 16. Bigatello LM, Hurford WE, Kacmarek RM, Roberts JD, Zapol WM. Prolonged inhalation of low concentrations of nitric oxide in patients with severe adult respiratory distress syndrome. Anesthesiology 1994;80:761 70. 17. Lee E, Little AG, Wen-Hu H, et al. Effect of on extravascular lung water in dogs. J Surg Res 1985;38:566 73. 18. Mathru M, Blakeman B, Dires DJ, et al. Permeability pulmonary edema following lung resection. Chest 1990;98: 1216 8. 19. Waller DA, Gebitekin C, Saunders NR, Walker DR. Noncardiogenic pulmonary edema complicating lung resection. Ann Thorac Surg 1993;55:140 3. 20. Pittet JF, Mackersie RC, Martin TR, Matthay MA. Biological markers of acute lung injury: prognostic and pathogenetic significance. Am J Respir Crit Care Med 1997;155: 1187 204. 21. Matthay MA, Wiener-Kronish. Intact epithelial barrier function is critical for the resolution of alveolar edema in humans. Am Rev Respir Dis 1990;142:1250 7. 22. Lewis JF, Jobe A. Surfactant and the adult respiratory distress syndrome. Am Rev Respir Dis 1993;147:218 33. DISCUSSION DR THOMAS R. J. TODD (Toronto, Ont, Canada): Dr Anderson, Dr Pairolero, Doug, members, and guests. It is always a pleasure to discuss a paper from the MGH thoracic group, and I very much appreciate the opportunity of doing so and I thank Doug for allowing me to review the manuscript ahead of time. (Slide) Over the last several years, since the initial publication by Zapol et al, in 1993, there have been a number of anecdotal reports and, indeed, phase II trials demonstrating that inhaled nitric oxide will work in respiratory failure by lowering pulmonary artery pressures and increasing PaO 2. But it does not do it always. And in this particular trial, the study that you have just seen, there is no new information in terms of this ability. But there is, I believe, some very striking new information in the material that Doug has given you. And I think it is twofold. First, they noted a 100% improvement in all of their patients. My review of all this literature suggests that at best you may see improvement in 20% to 60% of your patients. The second thing that they showed was an apparent decrease in mortality in this condition. And I emphasize the word apparent, because this is a retrospective series with comparison of mortality to historical controls. Dellinger et al, in the journal Critical Care Medicine, and Doug quotes this paper in his manuscript, recently did a phase II study showing again that in respiratory failure inhaled nitric oxide improves PaO 2. But there was no improvement in mortality in that series of ICU patients. But if you think about it, that is not surprising, because most ICU studies involve a heterogeneous group of patients who have multisystem organ failure and sepsis. And it is usually the multisystem organ failure that leads to the mortality. This study