Inhaled Prostacyclin and Iloprost in Severe Pulmonary Hypertension Secondary to Lung Fibrosis

Transcription

1 Inhaled Prostacyclin and Iloprost in Severe Pulmonary Hypertension Secondary to Lung Fibrosis HORST OLSCHEWSKI, H. ARDESCHIR GHOFRANI, DIETER WALMRATH, RALPH SCHERMULY, BETTINA TEMMESFELD-WOLLBRÜCK, FRIEDRICH GRIMMINGER, and WERNER SEEGER Depatment of Internal Medicine II, Justus-Liebig-University, Giessen, Germany Pulmonary hypertension is a life-threatening complication of lung fibrosis. Vasodilator therapy is difficult owing to systemic side effects and pulmonary ventilation perfusion mismatch. We compared the effects of intravenous prostacyclin and inhaled NO and aerosolized prostacyclin in randomized order and, in addition, tested for effects of oxygen and systemic calcium antagonists (CAAs) in eight patients with lung fibrosis and pulmonary hypertension. Aerosolized prostaglandin (PG)I 2 caused preferential pulmonary vasodilatation with a decrease in mean pulmonary arterial pressure from to mm Hg, and pulmonary vascular resistance (RL) from to dyn s cm 5 (p 0.05, respectively). Systemic arterial pressure, arterial oxygen saturation, and pulmonary right-to-left shunt flow, measured by multiple inert gas analysis, were not significantly changed. Inhaled NO similarly resulted in selective pulmonary vasodilatation, with RL decreasing from to dyn s cm 5. In contrast, both intravenous PGI 2 and CAAs were not pulmonary selective, resulting in a significant drop in arterial pressure. In addition, PGI 2 infusion caused a marked increase in shunt flow. Long-term therapy with aerosolized iloprost (long-acting PGI 2 analog) resulted in unequivocal clinical improvement from a state of immobilization and severe resting dyspnea in a patient with decompensated right heart failure. We concluded that, in pulmonary hypertension secondary to lung fibrosis, aerosolization of PGI 2 or iloprost causes marked pulmonary vasodilatation with maintenance of gas exchange and systemic arterial pressure. Long-term therapy with inhaled iloprost may be life saving in decompensated right heart failure from pulmonary hypertension secondary to lung fibrosis. Olschewski H, Ghofrani HA, Walmrath D, Schermuly R, Temmesfeld-Wollbrück B, Grimminger F, Seeger W. Inhaled prostacyclin and iloprost in severe pulmonary hypertension secondary to lung fibrosis. AM J RESPIR CRIT CARE MED 1999;160: Lung fibrosis of various etiologies is often associated with pulmonary hypertension, which may become a major contributor to morbidity and mortality, or may even represent the major cause of death as in systemic sclerosis (1). In patients suffering from primary pulmonary hypertension (PPH), intravenous prostacyclin has been demonstrated to be a potent pulmonary vasodilator, and long-term infusion of the prostanoid was found to improve exercise tolerance and survival in these patients (2 4). However, in the presence of lung fibrosis, any systemic vasodilator therapy may be hampered by two drawbacks. (1) The percentage of increased pulmonary vascular resistance that is due to fibrotic and thus fixed remodeling processes, compared with the percentage caused by vasoconstriction, is unknown. Systemic vasodilatation due to medications in the absence of pulmonary vasodilatation may provoke (Received in original form October 2, 1998 and in revised form March 15, 1999) This study was supported by the Deutsche Forschungsgemeinschaft, Klinische Forschergruppe Respiratorische Insuffizienz. Correspondence and requests for reprints should be addressed to Dr. Horst Olschewski, Department of Internal Medicine II, Justus-Liebig-University, Klinikstr. 36, D Giessen, Germany. horst.olschewski@innere.med. unigiessen.de Am J Respir Crit Care Med Vol 160. pp , 1999 Internet address: severe systemic hypotension in these patients if cardiac output is not adequately increased. (2) Any systemic administration of vasodilators may increase the blood flow to low or nonventilated lung areas by interfering with the physiological hypoxic vasoconstrictor mechanism, thereby worsening preexistent ventilation ( V )/perfusion ( Q ) mismatch and shunt (5), resulting in arterial hypoxia and wasting of the small ventilatory reserve of these patients. Selective pulmonary vasodilatation by inhalation of the vasorelaxant agent is an appealing concept to circumvent some of the hazards inherent in systemic vasodilator therapy in pulmonary hypertension. The feasibility of this concept was demonstrated for inhalation of nitric oxide (NO) by children with persistent pulmonary hypertension of the newborn (6 9), in the adult respiratory distress syndrome (ARDS) (10), and also in scleroderma patients with isolated pulmonary hypertension that is characterized by severe pulmonary hypertension without interstitial lung disease (11). By employing an aerosol technique, our group demonstrated that the concept of pulmonary selectivity may be extended to inhalation of aerosolized prostacylin (prostaglandin I 2, PGI 2 ) (12, 13): in patients with ARDS both intravenous and aerosolized PGI 2 caused pulmonary vasodilatation; however, the former deteriorated whereas the latter improved shunt flow and gas exchange. This corresponds to studies of experimental pulmonary hypertension,

2 Olschewski, Ghofrani, Walmrath, et al.: Inhaled Prostanoids in Secondary Pulmonary Hypertension 601 wherein selective vasodilatation in the pulmonary vascular bed was achieved by aerosolized PGI 2 (14). Subsequent studies in patients with ARDS and severe pneumonia demonstrated that inhalation of NO and aerosolized PGI 2 resulted in comparable hemodynamic and gas exchange effects (15, 16). By employing this approach in patients with PPH and excessive pulmonary hypertension, we found that aerosolization of PGI 2 or its stable analog, iloprost, effected equipotent pulmonary vasodilatation to the maximum tolerable dose of intravenous prostacyclin, but induced fewer systemic side effects (17). In these patients, prostacyclin appeared to be more potent than NO in decreasing pulmonary resistance, corresponding to other data (18, 19). On the other hand, inhaled NO, owing to its immediate inactivation by hemoglobin binding on entering the intravascular space, is strictly selective to the pulmonary vessels, whereas inhaled prostacyclin will have some systemic effects due to overspill into the systemic circulation. In this study, we examined the effects of the most important systemic and inhalative vasodilators, NO and intravenous and inhaled prostacyclin, in comparison with O 2 and systemic calcium antagonists in patients with severe pulmonary hypertension associated with interstitial lung diseases of differing etiologies. METHODS Patients Eight patients with lung fibrosis and pulmonary hypertension, who were referred to the Division of Pulmonary and Critical Care (Department of Internal Medicine, Justus-Liebig-University, Giessen, Germany) between June 1995 and October 1996, were included in the study. Criteria for entry included a peak systolic pulmonary pressure 50 mm Hg, as suggested by echocardiography or a resting pulmonary mean pressure 30 mm Hg as measured by catheter investigation, and the diagnosis of chronic fibrotic lung disease. The underlying diseases included extrinsic allergic alveolitis (two patients), systemic sclerosis, CREST syndrome (calcinosis, the Raynaud phenomenon, esophageal hypomotility, sclerodactyly, and telangiectasia), collagen vascular overlap syndrome, bronchopulmonary dysplasia, postradiation lung fibrosis, and idiopathic pulmonary fibrosis (Table 1). The diagnosis of these diseases was based on history, lung function testing, chest X-ray, and high-resolution computed tomography, which demonstrated at least medium-grade bilateral interstitial fibrosis in all patients. Flexible bronchoscopy with biopsy and bronchoalveolar lavage was performed in five patients. All suffered from significant lung restriction or severe restriction in gas exchange capacity, with a mean vital capacity (VC) of 48% and CO-diffusing capacity (DCO) of 26% relative to unaffected control patients. Significant arterial hypoxia under resting conditions was noted in seven patients, and these received long-term oxygen therapy. Seven patients were treated with steroids (mean dose, 21 mg of methylprednisolone per day). Two patients were treated with low-dose calcium antagonists (CAAs): patient B had received felodipine because of a right heart catheterization 7 mo before this study, and had subsequently recovered from a period of right heart decompensation. The other patient (patient H) had received low-dose diltiazem for several years, but validation of this approach had never been performed. None of the patients were anticoagulated or treated with other vasodilatory agents. Underlying causes for pulmonary hypertension other than parenchymal lung diseases were excluded by transthoracic and transesophageal echocardiography (n 8), ventilation and perfusion pulmonary nuclear scintigraphy (n 8), and pulmonary angiogram (patient F). Seven patients were classified as stage III according to the New York Heart Association (NYHA) system of classification. Patient F, who later received long-term iloprost inhalation, presented with decompensating right heart failure, unable to walk despite continuous nasal oxygen. This patient was a 27-yr-old woman with a history of dry cough and intermittent pleurisy for 4 yr. High-titer anti-nuclear, anti- DNA, and anti-skin-sensitizing antibodies, suggested collagen vascular disease resulting in progressive lung fibrosis with continuous reduction of VC and DCO. Exercise tolerance had been declining more rapidly in the previous 12 mo. She had been treated with high-dose corticosteroids and azathioprine, and subsequently with bolus application of cyclophosphamide (two boluses with 1 g/m 2, 4 wk apart) without clinical benefit. Current therapy included high-dose corticosteroids, continuous nasal oxygen, and diuretics. Study Protocol All patients gave informed written consent to participate in the test protocol, which was approved by the institutional ethics committee of the Justus-Liebig-University. A fiberoptic thermodilution pulmonary artery catheter (Edwards Swan-Ganz, 93A-754H-7.5F; Baxter Healthcare, Irvine, CA) was inserted to measure central venous pressure (Pcv), pulmonary artery pressure (Ppa), pulmonary artery wedge pressure (Ppa,we), cardiac output ( Q. ; thermodilution technique), right ventricular ejection fraction (RVEF), and central venous oxygen saturation (Sv O2 ). A femoral artery catheter was used to assess mean arterial pressure ( Pa) and to draw arterial blood samples. The pulmonary shunt flow was measured by determining the retention and excretion values of sulfur hexafluoride (20) in all patients. In addition, in patient F, the retention and excretion values for ethane, cyclopropane, halothane, diethyl ether, and acetone were used to determine the pattern of ventilation and perfusion distributions (multiple inert gas elimination technique, MIGET) as described in detail by Wagner and co- TABLE 1 BASELINE CHARACTERISTICS AND LUNG FUNCTION Patient Underlying Disease Sex Age (yr) Height (cm) Weight (kg) Liters VC % pred FEV 1 (L) DCOc PO 2 PCO 2 A EAA F B SS M C CREST M D EAA M E BPD F ND F CVOL F G PRT F H IPF M Mean Definition of abbreviations: BPD bronchopulmonary dysplasia after premature birth and long-term ventilation; CREST calcinosis, the Raynaud phenomenon, esophageal dysfunction, sclerodactyly, and telangiectasia; CVOL collagen vascular overlap syndrome (see text); DCOc, diffusion capacity for CO gas, corrected for the actual hemoglobin concentration, given as a percentage of predicted values, corrected for age, sex, and body surface area (single-breath method in patients with VC 1.5 L and steady state in patients with VC 1.5 L); EAA chronic extrinsic allergic alveolitis; F female; FEV 1 forced expiratory volume within 1 s; IPF idiopathic pulmonary fibrosis; M male; ND not determined; PO 2 and PCO 2 resting arterial oxygen and carbon dioxide partial pressure during oxygen pause; PRT postradiation therapy presenting with lung fibrosis and kyphoscoliosis occurring after thoracic radiation during the first year of life and due to malignant neuroblastoma; SS systemic sclerosis; VC inspiratory vital capacity in liters and as a percentage of predicted values.

3 602 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL TABLE 2 BASELINE HEMODYNAMICS Q (L/min) Ppa Ppa,we RL [(dyn s)/cm 5 ] RVEF Pcv HR Pa (min 1 ) RL/Rva Shunt A B C D E ND F , G ND H Mean Definition of abbreviations: HR heart rate; Pa mean systemic arterial pressure; Pcv central venous pressure; Ppa mean pulmonary artery pressure; Ppa,we pulmonary artery wedge pressure; Q cardiac output; RL pulmonary vascular resistance; RL/Rva RL/Rva (systemic vascular resistance) ratio; RVEF right ventricular ejection fraction, assessed by thermodilution; Shunt right-to-left shunt blood flow as a percentage of total pulmonary blood flow, assessed by MIGET analysis. workers (20). Concerning the patients with long-term CAA therapy, the agent was discontinued 48 h before the test in patient H and continued in patient B, who had suffered from right ventricular decompensation before therapy with felodipine. Test Procedure In patients receiving long-term oxygen therapy (seven patients), catheterization was performed with ongoing nasal oxygen. Oxygen delivery was then stopped for 20 min, and baseline values were determined. Next, high-dose oxygen was administered (4 8 L/min) in order to increase arterial oxygen saturation above 95%, and measurements were performed. Subsequently, oxygen supply was titrated to maintain baseline arterial oxygen saturation above 85% (0 4 L/min). This oxygen flow was kept constant throughout the following tests, which were performed in randomized order. Inhaled nitric oxide (15 to 80 ppm; mean, 40 ppm) was titrated to achieve a maximum response of pulmonary artery pressure decrease without decline of arterial oxygen saturation, as assessed by fingertip oximetry. After 5 min on a constant dose, a complete set of hemodynamic measurements was performed. Intravenous prostacyclin (epoprostenol sodium; Wellcome Research Laboratories, Beckenham, Kent, UK) was increased in increments of 2 (ng/kg)/min until patients experienced discomfort (thoracic oppression, heat, headache) or until mean arterial pressure decreased to less than 70 mm Hg. The highest tolerated doses ranged from 5 to 16 (ng/kg)/min, with a mean of 8.0 (ng/kg)/min. Fifteen minutes after finding the highest tolerated dose, during continuous PGI 2 infusion, a complete set of hemodynamic measurements was performed. Aerosolized prostacyclin, diluted in glycine buffer (50 g/ml), was jet nebulized (Puritan-Bennett raindrop medication nebulizer) with room air at a pressure of 80 kpa (compressor from Pari Boy, Pari, Germany) (fluid flux, 0.09 ml/min; mass median aerodynamic diameter of particles, 3.5 m; geometric SD of 2.6, ascertained by impactor technique) and delivered to a spacer connected to the afferent limb of a Y-valve mouthpiece. The total inhalation time was 12 to 15 min (total nebulized dose, 54 to 68 g), depending on systemic pressure and fingertip oximetry. Hemodynamic measurements were performed every 3 min and arterial and central venous blood samples were drawn before and during the last minute of inhalation. After each test, 1 h was allowed to pass to achieve a new baseline. After termination of the randomized trial period, calcium antagonists were given to six patients. Nifedipine, 10 to 20 mg, was administered sublingually (patients A, C, D, E, and F) and hemodynamic measurements were then performed 30 min after ingestion of this dose. In one patient (patient H), instead of nifedipine, diltiazem was applied corresponding to the preceding therapy. Diltiazem, 40 mg, was applied intravenously dur- TABLE 3 ACUTE RESPONSE OF A PATIENT WITH DECOMPENSATED RIGHT HEART FAILURE TO VASODILATORY THERAPY* Q (L/min) Ppa RL [(dyn s)/cm 5 ] RVEF Pcv HR (min 1 ) Pa Pa O2 Shunt Before NO , During NO Before PGI 2, intravenous , During PGI 2, intravenous Before PGI 2, aerosolized , During PGI 2, aerosolized Before nifedipine , After nifedipine After 5 mo Before iloprost, aerosolized , During iloprost, aerosolized After 12 mo Before iloprost, aerosolized , During iloprost, aerosolized Definition of abbreviations: see Tables 1 and 2. * Response to NO, intravenous PGI 2, aerosolized PGI 2, and nifedipine during the first test trial and after 5 and 12 mo of iloprost inhalation in patient F. Measurements on the first day were performed during continuous oxygen flow (4 L/min); during subsequent trials, measurements were performed without nasal oxygen supply.

4 Olschewski, Ghofrani, Walmrath, et al.: Inhaled Prostanoids in Secondary Pulmonary Hypertension 603 Figure 1. Acute responses to oxygen, NO, intravenously administered and inhaled prostacyclin (PGI i.v. and PGI aero, respectively) in eight patients, and to calcium antagonists (CAAs) in six patients. Dark columns and light columns give mean values SE before and after drug administration, respectively, for arterial oxygen saturation (Sa O 2 ), right-to-left shunt flow (as a percentage of pulmonary blood flow; SHUNT), right ventricular ejection fraction (RVEF), and central venous pressure (CVP). p significance level for differences in the responses to the various agents (ANOVA for the intrapair differences); * significant difference pre- and postapplication, p 0.05; significant linear contrast between responses to different agents (Scheffé test, p 0.05). ing a 30-min infusion period. Measurements were performed 15 min after the end of infusion. Patient G refused to take calcium antagonists owing to intolerance during preceding episodes of treatment with these drugs, and patient B was not included owing to continuous felodipine therapy. Statistics One-way analysis of variance (ANOVA) was employed to evaluate changes in parameters during exposure to the different agents (preand postexposure percent differences for Q., pulmonary vascular resistance [RL], and systemic vascular resistance [Rva] and numerical differences for other parameters) and the Scheffé test was used as an a posteriori test for linear contrasts between these differences. The response to an agent was considered significant if the 95% confidence interval (p 0.05) or the 99% confidence interval (p 0.01) of the pre- and postexposure difference did not overlap with zero. The significance level for the Scheffé test was set at p RESULTS Baseline Data Values of Ppa and RL for all patients were significantly elevated and scattered over a wide range (Table 2). Ppa,we was in the normal range, excluding left heart failure as the underlying cause of the pulmonary hypertension. Q values were in the normal or lower normal range in all patients, except in patient F, who presented with low-output syndrome and decompensated right heart failure, and had elevated Pcv (Table 3). The RL/Rva ratio was increased and RVEF was decreased in all. Shunt flow was increased with great differences between individuals. Seven patients presented with significant arterial hypoxemia, whereas PCO 2 values were slightly elevated in only one patient. Oxygen During high-dose oxygen inhalation, Pa O2 increased to an average of 125 mm Hg (p 0.01), with a corresponding increase in Sa O2 (Figure 1). Heart rate decreased significantly from to min 1 (p 0.05) and Ppa decreased from to mm Hg (p 0.05; Figure 2). All other hemodynamic and gas exchange variables were not significantly altered. Continuous low-flow nasal oxygen was then administered to maintain baseline Sa O2 values above 85% (see increased baseline Sa O2 data in the following tests). NO Inhaled NO significantly decreased Ppa from to mm Hg (p 0.01), RL from to dyn s cm 5 (p 0.05), and the RL/Rva ratio from to (p 0.05). RVEF was increased from to % (p 0.05). All other hemodynamic and gas exchange variables were not significantly altered (see Figures 1 and 2). Intravenous PGI 2 Prostacyclin infusion increased Q from to L/min (p 0.01) and significantly decreased both Ppa from to and Pa from to mm Hg (p 0.05, respectively). Correspondingly, both RL and Rva were decreased by an average of 40%, resulting in an unchanged RL/Rva ratio. RVEF increased significantly from to %. A more than 2.5-fold increase in pulmonary shunt flow occurred, from an average of to % (see Figures 1 and 2). Owing to the increased central venous oxygen saturation (data not given), which reflected Q increase, this marked augmentation of shunt flow resulted in only a moderate decrease in Sa O2.

5 604 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL Figure 2. Acute responses to oxygen, NO, intravenously administered and inhaled prostacyclin (PGI i.v. and PGI aero, respectively), and calcium antagonists (CAAs). Dark columns and light columns give mean values SE before and after drug administration, respectively, for mean pulmonary artery pressure ( Ppa), pulmonary vascular resistance (RL), ratio of pulmonary to systemic vascular resistance (RL/Rva), mean systemic arterial pressure ( Pa), cardiac output ( Q ), and heart rate (HR). For statistics see Figure 1. Inhaled PGI 2 Aerosolized PGI 2 markedly decreased the pulmonary artery pressure from to mm Hg. This response was significantly greater than the response to intravenous PGI 2 (Scheffé test). Most of the difference in these responses, however, resulted from increased preinhalation rather than decreased postinhalation pressures. In contrast, Pa values were not significantly reduced. Q increased from to L/min. Correspondingly, the RL values were reduced by about 50% (from to ) and the RL/Rva ratio was significantly decreased (from to ). RVEF was increased from to %, and Pcv was significantly lowered from to mm Hg. Shunt flow and heart rate did not change significantly. Calcium Antagonists Calcium antagonists significantly decreased Ppa and, more impressively, Pa values. Q increased from to L/min and RVEF was increased from to % on average (p 0.05, respectively). Correspondingly, both RL and Rva were significantly reduced, with unchanged RL/Rva ratio. Shunt flow and arterial oxygenation were not significantly altered. Long-term Therapy On the basis of the results of the test trial, in three patients (A, C, and E), in whom nifedipine resulted in a substantial decrease in Ppa and RL without systemic side effects, treatment with calcium antagonists was started and in patient H the CAA was withdrawn owing to decreased arterial oxygen pressure after the test dose. In the patient with immobilization due to decompensated right heart failure (patient F), calcium antagonists were not tolerated. This may be related to a negative inotropic effect of the calcium antagonist suggested by the moderate Pcv decrease from 19 to 16 mm Hg despite an afterload reduction 50% (see Table 3) in this patient; long-term therapy with repetitive aerosolization of iloprost, the stable analog of prostacyclin, was started. This decision was based on the fact that both PGI 2 infusion and PGI 2 inhalation markedly decreased the excessively high Ppa and RL values, with approximate doubling of Q, but intravenous PGI 2 drastically in- creased the shunt flow in this patient (Figure 3), with concomitant drop in arterial oxygenation, and a sensation of dyspnea

6 Olschewski, Ghofrani, Walmrath, et al.: Inhaled Prostanoids in Secondary Pulmonary Hypertension 605 Figure 3. Dose response curve of prostacyclin inhalation. PGI 2 was continuously aerosolized at a rate of 4.5 g/min. Hemodynamic measurements were performed before onset of nebulization and after 3, 6, and 9 min, i.e., at cumulative nebulized doses of 13.5, 27, and 40.5 g of PGI 2. Aerosolization was terminated after min, depending on individual side effects (mean arterial pressure and arterial oxygen saturation [fingertip oximetry] starting to decrease) and the final measurement was then immediately performed. Mean values SEM are given for eight patients for Q, RL, and Ppa. and anginal oppression (Table 3). In contrast, aerosolized PGI 2, similar to NO, did not significantly increase shunt flow and increased Pa O2 values. NO, however, was not employed owing to possible rebound pulmonary hypertensive crisis on abrupt cessation of the gas flow (21). Aerosolization of iloprost (13.5 g within 15 min) exerted acute benefits on hemodynamics and gas exchange, comparable to those of inhaled PGI 2 (Figure 4), but the effect lasted longer ( 90 min). A daily total dose of 135 g of iloprost, divided in nine aerosol applications (15 g each) with two 4-h nocturnal intervals, was then continuously administered for 21 mo. The patient was discharged home after 4 wk, and is now performing light housework. Recatheterization after 5 and 12 mo showed a tendency toward improvement of baseline hemodynamics (assessed in the morning after a 6-h interval without nebulization), as well as a maintained acute response to the aerosolization of iloprost without any increase in dosage (Table 3). Exercise tolerance increased, and after 12 mo she performed a 6-min walk of 240 m just before and 314 m just after inhalation of iloprost. Interestingly, lung fibrosis did not progress further as assessed by VC, DCO, and chest X-ray. DISCUSSION In this study, systemic and inhaled vasodilators were applied to patients with significantly increased pulmonary vascular resistance and reduced right ventricular function, partly resulting in reduced cardiac output. In contrast to previous investigations addressing pulmonary vasodilatation in pulmonary hypertension (2, 3, 11, 17, 18), all of our patients suffered from severe interstitial lung disease with severe gas exchange abnormalities (mean CO diffusion capacity, 26%). Seven of the eight patients were receiving long-term oxygen therapy (mostly 24 h/d). To the best of our knowledge, this is the first investigation to compare the effects of different vasodilators in a population of patients with interstitial fibrosis accompanied by pulmonary hypertension. The contribution of increased right ventricular afterload to the reduction of exercise tolerance in the individual patient is difficult to assess, although pulmonary hypertension was found to be the most common cause of death in systemic sclerosis (1). In our patient with decompensated right ventricular failure due to a collagen vascular disease, it was evident that restriction of circulation was predominant over the restriction of ventilation. On persistent reduction of the pulmonary vascular resistance with concomitant increase in cardiac output, this patient improved from a state of being confined to bed to a 6-min walking distance of 314 m, although lung restriction (reduced vital capacity, FEV 1 ) and gas exchange abnormalities (CO diffusion) were entirely unchanged. This finding impressively supports the concept that lowering of the pulmonary resistance is a worthwhile goal in secondary pulmonary hypertension as long as arterial oxygenation and systemic pressure are not compromised. As anticipated for pulmonary hypertension secondary to lung parenchymal disease, nasal oxygen supply effected a moderate yet significant decrease in Ppa. A detailed analysis of the hemodynamic data did, however, disclose that this effect was only partially due to a relaxation of the pulmonary vasculature, with RL values being only minimally lowered. More importantly, the relief of systemic hypoxia caused a significant decrease in heart rate, most probably reflecting reduced sympathetic tone, with slightly decreased cardiac output. An oxygen challenge is thus clearly insufficient to discriminate between the reversible (vasoconstriction related) and the fixed component of the secondary pulmonary hypertension in lung fibrosis. Inhaled NO significantly reduced RL by 25%, with a concomitant Ppa decrease and increase in RVEF as a consequence of reduced right ventricular load. Q was slightly in- creased and Pa values were not significantly altered. This corresponds to a profile of selective pulmonary vasodilatation, as anticipated for NO, and this was also reflected by the significant decrease in the RL/Rva ratio. The response to NO in these patients with secondary pulmonary hypertension is thus similar to that in PPH patients (17, 18, 22) and in scleroderma patients suffering from isolated pulmonary hypertension (11). As reported from these previous investigations, the NOinduced changes in hemodynamics returned to baseline values within 2 5 min after cessation of inhalation. In contrast to NO, infused prostacyclin drastically reduced both RL and Rva, resulting in an unchanged RL/Rva ratio. Pa was significantly decreased, resulting in arterial baroreflex activation with heart rate elevation and overshooting Q in- crease, which partially antagonized the Ppa decrease. Nevertheless, improved right ventricular function was demonstrated by increased RVEF. Most disadvantageously, however, systemic prostacyclin administration caused a substantial increase in shunt flow in these patients with fibrosis-related secondary pulmonary hypertension, with concomitant drop in Pa O2 despite enhanced central venous oxygen saturation. As repetitive measurements to establish dose effect curves for in-

7 606 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL Figure 4. Ventilation perfusion ( V / Q ) distribution, estimated by the multiple inert gas elimination technique (MIGET), in patient F. (Upper left) V / Q distribution under baseline conditions. (Upper right) During inhalation of 10 ppm NO. (Lower left) During infusion of PGI 2, 5 (ng/kg)/min. (Lower right) At the end of inhalation of 68 g of PGI 2. Shunt and deadspace values represent percentages of total pulmonary blood flow and total ventilation, respectively. travenous PGI 2 were not undertaken, it may not be excluded that the ratio of beneficial to adverse effects of this approach might have been more favorable at a lower dose of infused PGI 2. At the highest tolerated dose of intravenously administered PGI 2, effecting an RL decrease equivalent to that in response to inhaled prostacyclin, the more disadvantageous effect of the intravenous application on gas exchange was obvious, accompanied by a sensation of dyspnea and oppression in some patients. This does not exclude the possibility that long-term effects of prostacyclin, particularly at lower doses, may be beneficial to such patients. Nevertheless, in our patient with decompensated right heart failure, we were not able to establish a long-term treatment with PGI 2 owing to severe worsening of gas exchange. Similar to intravenous PGI 2, CAAs effected a nonselective decrease in both RL and Rva, with an unchanged RL/Rva ratio. As cardiac output was not as markedly increased as noted in response to prostacyclin infusion, the drop in systemic arterial pressure was more prominent for the CAAs as compared with all other vasodilatory regimens. Interestingly, however, there was only a minor, insignificant increase in shunt flow. The reason for this better conservation of the ventilation perfusion matching in response to systemically applied CAAs as compared with intravenous prostacyclin in this group of patients with fibrosis is currently not known. One hypothesis is that shunt vessels are more sensitive to PGI 2 than to calcium antagonists; an alternative explanation could be that with calcium antagonists the highest tolerated dose was not titrated in this study. Aerosol application of PGI 2 caused the most pronounced decrease in Ppa among all vasodilator challenges investigated. Similar to NO inhalation, a marked drop in the RL/ Rva ratio was noted, which indicated predominant, although not exclusive, selectivity of the vasodilator effect for the pulmonary circulation. In marked contrast to infused prostacyclin, shunt flow was not significantly increased. These findings are consistent with our previous suggestion that aerosolization of PGI 2 allows targeting of the vasorelaxant properties to lung vessels in well-ventilated areas (12, 13, 15, 17). The increase in cardiac output, which surpasses that in response to NO, may be explained by a more marked relief of the right ventricular afterload, but also by a minor reduction in Rva, most probably owing to some spillover of the prostaglandin into the systemic circulation. Nevertheless, the marked pulmonary vasodilatation during PGI 2 inhalation concomitant with significant hemodynamic differences between aerosolized and infused PGI 2 suggests that an alveolar deposition of the drug along with marked regional vasodilatation was achieved, with only minor systemic drug effects. The notion that preferential pulmonary vasodilatation is achieved by the aerosolization technique is also supported by the absence of typical side effects attributed to infused prostacyclin, i.e., flushing, headache, jaw pain, and/ or diarrhea. In the patient presenting with decompensated right heart failure, the difference between the efficacy profiles of infused and inhaled PGI 2 was most prominent. Prostacylin infusion caused a drastic increase in shunt flow, which at the given ventilatory restriction resulted in dyspnea and thoracic oppression not tolerated by the patient, whereas PGI 2 nebulization resulted in a reduction in RL to about one-third of the baseline value and in near doubling of cardiac output while conserving ventilation perfusion matching. The same hemodynamic and gas exchange profile was achieved by aerosolization of iloprost, however, with longer lasting effect. After termination of aerosol inhalation, PGI 2 -induced changes returned to baseline within min and those in response to iloprost within min. In view of the poor clinical state of the patient and the favorable response to inhaled iloprost, long-term treatment with daily repetitive iloprost inhalation was started. The drug was well tolerated, and an impressive improvement in the clinical state over the subsequent 18-mo observation pe-

8 Olschewski, Ghofrani, Walmrath, et al.: Inhaled Prostanoids in Secondary Pulmonary Hypertension 607 riod was noted. Interestingly, in addition to some moderate baseline hemodynamics (preinhalation) the acute reactivity to the prostaglandin aerosolization was maintained, obviating any need to increase the dosage. Overshooting rebound pulmonary hypertension, as observed on cessation of continuous NO inhalation (21), was not observed during the nocturnal interruption of iloprost inhalation. However, more data on the chronic effect of iloprost are mandatory with regard to the fact that acute vasodilatory efficacy and benefit on chronic use may be correlated. We conclude that in severe pulmonary hypertension secondary to lung fibrosis, a considerable percentage of the lung vascular resistance increase is apparently caused by persistent vasoconstriction, which may be acutely antagonized by CAAs, intravenous PGI 2, inhaled NO, and inhaled PGI 2. The most profitable profile of changes in terms of hemodynamics and gas exchange was achieved by aerosolization of PGI 2 or its stable analog, iloprost. Daily repetitive iloprost administration may be considered as a new approach for long-term therapy in severe secondary pulmonary hypertension, particularly in those patients not tolerating systemic vasodilator therapy because of enhanced shunting and drop in systemic arterial pressure. Long-term studies addressing chronic effects of inhaled iloprost in those patients are requested. Acknowledgment : The authors thank Jens Bier for excellent technical assistance in the MIGET laboratory, Wolfgang Pabst (Institute for Medical Statistics, Justus-Liebig-University, Giessen) for the statistics, and Mary Kay Steen- Mueller, M.D., for carefully reviewing the manuscript. References 1. Lee, P., P. Langevitz, C. A. Alderdice, M. Aubrey, P. A. Baer, M. Baron, D. Buskila, J. P. Dutz, I. Khonstanteen, S. Piper, M. Ramsden, T. O. Rosenbach, S. Sukenik, S. Wilkinson, and E. C. Keystone Mortality in systemic sclerosis (scleroderma). Q. J. Med. 82: Barst, R. J., L. J. Rubin, M. D. McGoon, E. J. Caldwell, W. A. Long, and P. S. Levy Survival in primary pulmonary hypertension with long-term continuous intravenous prostacyclin. Ann. Intern. Med. 121: Barst, R. J., L. J. Rubin, W. A. Long, M. D. McGoon, S. Rich, D. B. Badesch, B. M. Groves, V. F. Tapson, R. C. Bourge, B. H. Brundage, S. K. Koerner, D. Langleben, C. A. Keller, S. Murali, B. F. Uretsky, L. M. Clayton, M. M. Jöbsis, S. D. Blackburn, D. Shortino, and J. W. Crow A comparison of continuous intravenous epoprostenol (prostacyclin) with conventional therapy for primary pulmonary hypertension. N. Engl. J. Med. 334: Rubin, L. J Primary pulmonary hypertension. N. Engl. J. Med. 336: Agusti, A. G. N., and R. Rodriguez-Roisin Effect of pulmonary hypertension on gas exchange. Eur. Respir. J. 6: Kinsella, J. P., S. R. Neish, E. Shaffer, and S. H. Abman Low-dose inhalation nitric oxide in persistent pulmonary hypertension of the newborn. Lancet 340: Roberts, J. D., D. M. Polaner, P. Lang, and W. M. Zapol Inhaled nitric oxide in persistent pulmonary hypertension of the newborn. Lancet 340: Neonatal Inhaled Nitric Oxide Study Group Inhaled nitric oxide in full-term and nearly full-term infants with hypoxic respiratory failure. N. Engl. J. Med. 336: Roberts, J. D., J. R. Fineman, W. M. Zapol, F. C. Morin, III, P. W. Shaul, S. Rimar, M. D. Schreiber, R. A. Polin, M. S. Zwass, M. M. Zayek, I. Gross, M. A. Heymann, J. Briegel, M. Welte, and K. Peter Inhaled nitric oxide and persistent pulmonary hypertension of the newborn. N. Engl. J. Med. 336: Rossaint, R., K. J. Falke, F. López, K. Slama, U. Pison, and W. M. Zapol Inhaled nitric oxide for the adult respiratory distress syndrome. N. Engl. J. Med. 328: Williamson, D. J., C. Hayward, P. Rogers, L. L. Wallman, A. D. Sturgess, R. Penny, and P. S. Macdonald Acute hemodynamic responses to inhaled nitric oxide in patients with limited scleroderma and isolated pulmonary hypertension. Circulation 94: Walmrath, D., T. Schneider, J. Pilch, F. Grimminger, and W. Seeger Aerosolised prostacyclin in adult respiratory distress syndrome. Lancet 342: Walmrath, D., T. Schneider, R. Schermuly, H. Olschewski, F. Grimminger, and W. Seeger Direct comparison of inhaled nitric oxide and aerosolized prostacyclin in acute respiratory distress syndrome. Am. J. Respir. Crit. Care Med. 153: Welte, M., B. Zwissler, H. Habazettl, and K. Messmer PGI 2 aerosol versus nitric oxide for selective pulmonary vasodilation in hypoxic pulmonary vasoconstriction. Eur. Surg. Res. 25: Walmrath, D., T. Schneider, J. Pilch, R. Schermuly, F. Grimminger, and W. Seeger Effects of aerosolized prostacyclin in severe pneumonia. Impact of fibrosis. Am. J. Respir. Crit. Care Med. 151: Zwissler, B., G. Kemming, O. Habler, M. Kleen, M. Merkel, M. Haller, J. Briegel, M. Welte, and K. Peter Inhaled prostacyclin versus inhaled nitric oxide in adult respiratory distress syndrome. Am. J. Respir. Crit. Care Med. 154: Olschewski, H., D. Walmrath, R. Schermuly, H. A. Ghofrani, F. Grimminger, and W. Seeger Aerosolized prostacyclin and iloprost in severe pulmonary hypertension. Ann. Intern. Med. 124: Sitbon, O., F. Brenot, A. Denjean, A. Bergeron, F. Parent, and R. Azarian Inhaled nitric oxide as a screening vasodilator agent in primary pulmonary hypertension. Am. J. Respir. Crit. Care Med. 151: Warren, J. B., and T. Higenbottam Caution with the use of inhaled NO. Lancet 348: Wagner, P. D., H. A. Saltzman, and J. B. West Measurement of continuous distributions of ventilation perfusion ratios: theory. J. Appl. Physiol. 36: Miller, O., S. Tang, A. Keech, and D. Celermajer Rebound pulmonary hypertension on withdrawal from inhaled nitric oxide. Lancet 346: Pepke-Zaba, J., T. W. Higenbottam, A. T. Ding-Xuan, D. Stone, and J. Wallwork Inhaled nitric oxide as a cause of selective pulmonary vasodilatation in pulmonary hypertension. Lancet 338: