Double-blind, Placebo-controlled Trial of Pirfenidone in Patients with Idiopathic Pulmonary Fibrosis

Transcription

1 Double-blind, Placebo-controlled Trial of Pirfenidone in Patients with Idiopathic Pulmonary Fibrosis Arata Azuma, Toshihiro Nukiwa, Eiyasu Tsuboi, Moritaka Suga, Shosaku Abe, Koichiro Nakata, Yoshio Taguchi, Sonoko Nagai, Harumi Itoh, Motoharu Ohi, Atsuhiko Sato, and Shoji Kudoh for the members of the Research Group for Diffuse Lung Diseases in Japan; and Ganesh Raghu Fourth Department of Internal Medicine, Nippon Medical School; Division of Respiratory Disease, Toranomon Hospital, Tokyo; Department of Respiratory Oncology and Molecular Medicine Division of Cancer Control Institute of Development, Aging, and Cancer, Tohoku University, Sendai; Department of Respiratory Medicine, Graduate School of Medical Sciences, Kumamoto University, Kumamoto; Third Department of Internal Medicine and Department of Biochemistry, Sapporo Medical University School of Medicine, Sapporo; Department of Respiratory Medicine, Tenri Hospital, Tenri; Department of Respiratory Medicine, Graduate School of Medicine, Kyoto University; Kyoto Preventive Medical Center, Kyoto; Department of Radiology, Fukui Medical School, Fukui; Sleep Medical Center, Osaka Kaisei Hospital, Osaka, Japan; and University of Washington Medical Center, Seattle, Washington Idiopathic pulmonary fibrosis (IPF) is a fatal disorder without an effective therapy to date. In a double-blind, randomized, placebocontrolled trial, 107 patients were prospectively evaluated for efficacy of a novel compound, pirfenidone. The difference in the change in the lowest oxygen saturation by pulse oximetry (Sp O2 ) during a 6-minute exercise test, the primary endpoint, from baseline to 6 months was not significant between the two groups (p ). In a prespecified subset of patients who maintained a Sp O2 greater than 80% during a 6-minute exercise test at baseline, the lowest Sp O2 improved during a 6-minute exercise test in the pirfenidone group at 6 and 9 months (p and , respectively). Positive treatment effect was demonstrated in secondary endpoints: (1) change in VC measurements at 9 months (p ) and (2) episodes of acute exacerbation of IPF occurring exclusively in the placebo group during the 9 months (p ). Significant adverse events were associated with pirfenidone; however, adherence to treatment regimen was similar between pirfenidone and placebo groups. In conclusion, treatment with pirfenidone improved VC and prevented acute exacerbation of IPF during the 9 months of follow-up. Future long-term studies are needed to clarify the overall safety and efficacy of pirfenidone in IPF. Keywords: acute exacerbation; antifibrotic agent; idiopathic pulmonary fibrosis; pirfenidone; 6-minute steady-state exercise test Idiopathic pulmonary fibrosis (IPF) is a relentlessly progressive and fatal disorder characterized by high-resolution computed tomography (HRCT) and histologic features of usual interstitial pneumonia (UIP) in adults over 50 years of age with exertional dyspnea, abnormal pulmonary function tests (PFTs), and ineffective therapy (1, 2). Resting PFTs are routinely used to assess the functional status and the rate of progression of IPF. Recently, serial changes in PFTs have been demonstrated to be predictive of survival in patients with IPF (3 5). The progression of IPF assessed by PFTs and survival were uninfluenced by IFN-, and the need for further trials has been emphasized (6 8). Endexercise hypoxemia during maximal and submaximal steadystate exercise is a useful measure of severity in IPF (9, 10). (Received in original form April 29, 2004; accepted in final form January 18, 2005) Supported by Shionogi & Co., Ltd., Osaka, Japan. Correspondence and requests for reprints should be addressed to Ganesh Raghu, M.D., Division of Pulmonary & Critical Care Medicine, Campus Box , University of Washington, Seattle, WA graghu@u.washington.edu This article has an online supplement, which is accessible from this issue s table of contents online at Am J Respir Crit Care Med Vol 171. pp , 2005 Originally Published in Press as DOI: /rccm OC on January 21, 2005 Internet address: In addition to the changes in resting PFTs, the international consensus statement on IPF (1) also suggested the use of measured oxygen saturation during exercise to monitor the functional status during follow-up as a measure of response to treatment. A recent study demonstrated that decreased oxygen saturation by pulse oximetry (Sp O2 ) during 6 minutes of walking predicted survival in IPF (11). In a modified version of the 6- minute-walk test, the decrease of Sp O2 to 80% during the walk, the walk velocity, and the distance covered by the patients to desaturate predicted survival in a recent study of patients with IPF followed for 5 years (12). We prospectively designed a novel study using the change in the lowest Sp O2 reached during the 6- minute-walk test as the primary endpoint. Pirfenidone (5-methyl-1-phenyl-2-[1H]-pyridone; Shionogi and Co., Ltd., Osaka, Japan; MARNAC Inc., Dallas, TX), a novel compound with combined anti-inflammatory, antioxidant, and antifibrotic effects in experimental models of pulmonary fibrosis (13 22), has therapeutic potential for IPF (23 25). Prompted by the encouraging results of an open label prospective phase II clinical study of pirfenidone in patients with advanced IPF (23), we undertook a prospective clinical trial to determine its effect on Sp O2 with exercise and its correlation with resting PFTs. The study was aborted in favor of pirfenidone treatment due to an increased number of acute exacerbations of IPF in the placebo group. The results of this trial were presented in part during the American Thoracic Society International Conference in 2002 (26). METHODS Consenting patients with well defined IPF participated in a doubleblind, placebo-controlled, randomized, multicenter, prospective clinical trial conducted at 25 sites in Japan. The protocol was approved by the institutional review board at each center, and ongoing efficacy and safety results were reviewed by an independent Data and Safety Monitoring Board (DSMB) at 6-month intervals. The diagnosis of IPF was in accordance with the international consensus statement (1). While histologic evidence of UIP was not mandatory, HRCT evidence of definite or probable UIP pattern in the appropriate clinical setting was required in all patients. Definite UIP pattern was defined by basal predominant, subpleural reticular abnormality with traction bronchiectasis and honeycomb cysts and without atypical features of UIP (2, 27, 28). Probable UIP pattern was defined as the same as definite UIP pattern, but without traction bronchiectasis. The presence of other typical clinical features, including bibasilar inspiratory crackles, abnormal PFTs (1), and increased serum levels of damaged-pneumocyte markers (KL-6 [29] and surfactant proteins A and D [30]), was required in addition. Eligible patients were 20 to 75 years of age with adequate oxygenation at rest (Pa O2 70 mm Hg) and

2 Azuma, Nukiwa, Tsuboi, et al.: Pirfenidone for IPF 1041 Figure 1. Sp O2 area: graphic display of Sp O2 measurements plotted during the 6-minute steady-state exercise/walk test on the treadmill (6MET). (a)sp O2 area in patients maintaining Sp O2 greater than 80% throughout the 6-minute duration of the 6MET. Sp O2 area (light-gray shaded area) between the black horizontal dashed line of 100% and the curve of Sp O2 during the 6MET (red arrowed line). Sp O2 area is the difference ( ) between the area at baseline (black-arrowed line) and that at 6 months. A blue arrow indicates the improvement of Sp O2 area. (b) Sp O2 area in patients desaturating to 80% prior to reaching the intended 6 minutes during the 6MET. When Sp O2 reached below 80% once or stayed between 80 to 85% for 30 seconds, the walk test was terminated to assure the patient s safety. In these cases, the Sp O2 area was calculated by drawing a horizontal line (red solid arrowed line) from the Sp O2 reached at the point (red arrow) of terminating the test to the intended 6-minute duration (light-gray shaded area). A blue arrow indicates the deterioration of the Sp O2 area. Sp O2 oxygen saturation by pulse oximetry; Sp O2 area graphic display of Sp O2 measurements plotted during the 6MET. demonstrated Sp O2 of 90% or less during exertion while breathing air, within 1 month before enrollment. Criteria for exclusion were a decrease in symptoms during the preceding 6 months, use of immunosuppressives and/or oral prednisone greater than 10 mg/day during the preceding 3 months, clinical suspicion of idiopathic interstitial pneumonia other than IPF (2), coexisting emphysema (HRCT images of low attenuated areas in upper lung fields), pulmonary hypertension, asthma, tuberculosis, sarcoidosis, bronchiectasis other than traction associated, aspergillosis or respiratory infection; uncontrolled diabetes, comorbid conditions including malignancy, severe hepatic, renal, or cardiac disease; pregnancy (or its pursuance), breastfeeding; previous use of pirfenidone, suspicion of poor compliance in adherence to protocol, or being unable to understand protocol/written informed consent. Whereas concomitant prednisone of 10 mg/day or less was allowed, the following immunosuppressants or other antiinflammatory/antifibrotic drugs were not allowed to be used concomitantly: cyclophosphamide, azathioprine, methotrexate, d-penicallimine, colchicine, erythromycin, IFNs, N-acetylcysteine, cyclosporine, tacrolimus, and other experimental agents under investigation for IPF. The primary endpoint was defined as the change in the lowest Sp O2 during a 6-minute steady-state exercise test (6MET). Patients were requested to walk on a treadmill at a constant speed while breathing air. The Sp O2 was continuously measured by a pulse oximeter (31, 32). The initial speed for each patient was started at 60 m/minute and the investigator, while monitoring the Sp O2 measurements, adjusted the treadmill speed to determine an appropriate speed (40 80 m/minute) Figure 2. Categorized analyses of the lowest Sp O2 during the 6MET and pulmonary functions in the full analysis set (all patients) at 9 months. The results are shown by the categories of improved (blue areas), stable (open areas), and deteriorated (red areas) as defined in METHODS. p Values are indicated at the right. Statistical analysis by Wilcoxon s test was used. on the basis of the patient s comfort to be able to perform the 6MET while the lowest Sp O2 reached 90% or below. This predetermined speed at screening was kept constant for the individual patient at baseline and follow-up visits every 3 months. Thus, every patient had their own determined speed tailored to their comfort and demonstrated O 2 desaturation to less than 90%. To assure patients safety, the test was stopped when Sp O2 reached 80%. A greater than 4% increase in the lowest Sp O2 value during the 6MET was classified as improved, a 4% decrease or less as deteriorated, and the other values as stable (1). Each patient s Sp O2 during the 6MET was quantified and assessed visually as Sp O2 area (Figure 1). In addition to the change in the lowest Sp O2 during the 6MET, the difference in the Sp O2 area between the baseline test and the follow-up test at 6 or 9 months was also determined. Secondary endpoints were changes in resting PFTs while breathing air (VC, TLC, Dl CO,Pa O2 ), disease progression by HRCT patterns, episodes of acute exacerbation of IPF, change in serum markers of pneumocyte damage, and changes in quality of life measurements (33 35). PFTs were measured at baseline and at 3-month intervals and classified into improved, stable and deteriorated (changes of 10% for VC and TLC, 15% for Dl CO,and4mmHgforPa O2 ) (1). All HRCT scans were performed in accordance with predetermined protocol at baseline and 6-month intervals. Three expert chest radiologists independently evaluated the pattern of lung fibrosis on HRCT scans. In cases of disagreement, the radiologists and Study Coordinating Committee reexamined the scans in question to reach a consensus. Disease worsening on HRCT was defined as progression in the extent of UIP pattern compared with baseline, and was based on the independent evaluation of the site investigator and one of the three radiologists. The radiologists and Study Coordinating Committee were blinded to patient identification, treatment assignment, and temporal sequence of the studies. The definition of acute exacerbation of IPF (36, 37) was prespecified as manifestation of all of the following: worsening, otherwise unexplained clinical features within 1 month: progression of dyspnea over a few days to less than 5 weeks, new radiographic/hrct parenchymal abnormalities without pneumothorax or pleural effusion (e.g., new, superimposed ground-glass opacities), a decrease in the Pa O2 by 10 mm Hg or more, and exclusion of apparent infection based on absence of Aspergillus and pneumococcus antibodies in blood, urine for Legionella pneumophila, and sputum cultures. Serum KL-6 (29) and surfactant protein-d (30) levels, Chronic Respiratory Disease Questionnaire Score (33, 34), and Hugh-Jones Classi-

3 1042 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL fication Score (35) were measured to assess the changes in blood levels and patient s perceived quality of life during the study. Patients were randomly assigned into pirfenidone or placebo (2:1) groups using a modified permuted-block randomization method with block sizes of six. A dose titration schedule was followed for all patients: patients received oral tablets (pirfenidone or placebo) at a dose of 200 mg three times a day for the first 2 days, 400 mg three times a day for the 2 following days, and 600 mg three times a day (maximum dose) for the last 3 days. The maximum dose was maintained in patients tolerating it throughout the study. This regimen was determined on the basis of pretrial studies in Japanese volunteers and was lower than that given to white patients in a previous study (23). Patients not tolerating the maximum dose of 1,800 mg/day received a reduced, prespecified regimen utilizing the Standards for Classification of Serious Adverse Drug Reactions (38). For an adverse event of Grade 2 or worse, the dosage was reduced in a stepwise manner: from 9 tablets per day to 6 tablets per day, 6 tablets per day were reduced to 3 tablets per day. After a period of 14 days of observation with reduced dosage, the dosage was reduced further by one more step if the adverse event had persisted or increased (from 6 tablets per day to 3 tablets per day), and patients were monitored for another 14 days while taking the subsequently reduced dose. When the adverse event of Grade 2 or worse persisted or increased despite reducing the dosage to 3 tablets per day, the study medication was discontinued and patients observed for another 14 days. If the adverse event had resolved or decreased with reduction in the dose, the investigator was allowed to increase the dose up to 9 tablets per day. For an adverse event of Grade 1, the dosage regimen/reduction was deferred to the investigator s clinical judgment. Statistical Analysis The prespecified sample size was 90 patients (pirfenidone: 60, placebo: 30) based on a simulation study with the use of the lowest Sp O2 achieved during a 6MET after 1-year duration of the study. This minimum number of patients provided statistical power greater than 0.8 to detect assumed efficacy at the significance level of Analysis of change from baseline was performed with the Welch s t test. Categorical variables were analyzed with the Wilcoxon s test. Analyses of incidences were performed with Fisher s exact test. For missing values, the principle of last observation carry forward was adopted. Immediately after initiating the trial, a decision to conduct a prespecified analysis (i.e., before breaking the code) in the subset of patients who were able to complete the 6MET without the Sp O2 reaching less than 80% at baseline was made (see Results). Based primarily on important 6-month trends in a secondary endpoint, the DSMB recommended early termination of the trial on ethical grounds. Due to the length of time needed to collect, analyze, report a minimum of 6 months of data for DSMB review, all patients in the trial completed a minimum of 9 months on their assigned treatment arm by the time the recommendation to break the code was made. While both 6- and 9-month results are presented in the Results section of this manuscript, results from the latest complete follow-up exam (i.e., 9 months) most closely match the planned length of follow-up and are considered of primary importance. The data were held and analyzed by the trial sponsor. All authors of this study actively participated in the study design and had full access to analyzed data. No restrictions were placed on the authors by the sponsor for data analyses or reporting. RESULTS One hundred nine patients (pirfenidone: 73, placebo: 36) were entered into the study from November 2000 to January 2001 for safety assessment. Of the 109 patients, 107 (pirfenidone: 72, placebo: 35) were evaluated for efficacy of pirfenidone as 2 patients were excluded per prespecifications because they had violated inclusion criteria. A total of 91 of the 107 patients had definite UIP by HRCT; the remainder (16/107) had the probable UIP pattern. All patients had other clinical features typical of IPF (1). The assessment by the three expert chest radiologists concurred with one another in 92% of cases. The few cases of disagreement were due to their assessment of ground-glass opacities and honeycombing versus cysts. The site investigator disagreed over the change in extent of disease with the radiologist in 36% of cases. In these circumstances, the radiologist s opinion was preferentially adopted. Contrary to our expectations, 27 patients were not able to complete the 6MET for the entire duration of 6 minutes as defined at baseline. Because we were concerned that this unexpected finding at baseline itself might adversely affect the prespecified statistical analysis, we amended our design at the beginning of the study and proactively decided to perform the analysis for the group of patients who completed the entire duration of the 6MET (n 80). Among 107 patients, 80 patients (pirfenidone: 55, placebo 25) were able to complete the entire duration of the 6MET at the baseline visit as their Sp O2 remained more than 80%; the other 27 patients were stopped per prespecified protocol when their Sp O2 dropped to 80% before the 6-minute duration. The patient characteristics and demographic parameters for both groups were similar (Table 1). The majority of patients were above 50 years of age with only four patients below the age of 50 years (three in the pirfenidone group and one in the placebo group), but none below 46 years of age. Each of the four patients had typical patterns of UIP in the HRCT. A total of 35 patients who had coexisting atypical features in HRCT underwent lung biopsy (23: surgical lung biopsy, 12: transbronchial lung biopsy). Of the 107 patients evaluated for efficacy of pirfenidone, 92 had not received prior treatment with corticosteroids, and thus, the majority of enrolled patients were corticosteroid naive. Primary Endpoint Statistical significance was not reached for change from baseline of the lowest Sp O2 during the 6MET in the full analysis set (Table 2). At 6 months, the pirfenidone group was observed to have a mean increase from baseline in lowest Sp O2 during the 6MET of 0.64% compared with a mean decrease of 0.55% in the placebo arm (p ). At 9 months, the pirfenidone group was observed to have a mean increase from baseline in lowest Sp O2 during the 6MET of 0.47% compared with a mean decrease of 0.94% in the placebo arm (p ). The difference between groups was, however, more pronounced at both 6 and 9 months in 80 patients who completed the baseline 6MET (p at 6 months, p at 9 months). Secondary Endpoints There was a marginal decline in VC from baseline at 6 months in the placebo group ( 0.08 L) compared with the pirfenidone group ( 0.01 L, p ). At 9 months, the difference in decline of VC between the placebo group ( 0.13 L) and the pirfenidone group ( 0.03 L) was statistically significant (p , Table 2). Changes in TLC, Dl CO, and resting Pa O2 between the two groups were not statistically significant at either 6 or 9 months. The VC and TLC were stable in more patients in the pirfenidone group compared with the placebo group (p and , respectively, at 9 months), but statistically significant differences in stability were not observed for Dl CO or resting Pa O2 (Figure 2). Changes in the lowest Sp O2 and Sp O2 area were weakly correlated with changes in VC, TLC, and Dl CO (Table 2). However, the change in the lowest Sp O2 was correlated with the change in Sp O2 area during 6MET at 9 months (Figure 3). The reduction of ground-glass and reticular opacities was recognized as improved patterns of the HRCT images. The proportion of patients who improved at 6 months was 15% (10/65) for the pirfenidone group and 7% (2/29) for the placebo group (p ). There were no differences observed in the extent or severity of the honeycomb pattern. Acute exacerbation of IPF was manifested in 14% of the

4 Azuma, Nukiwa, Tsuboi, et al.: Pirfenidone for IPF 1043 TABLE 1. CLINICAL CHARACTERISTICS AND DEMOGRAPHIC PARAMETERS OF PATIENTS WITH IDIOPATHIC PULMONARY FIBROSIS IN THE STUDY All Patients: Full Analysis Set (FAS*) Patients Completed 6MET Pirfenidone (n 72) Placebo (n 35) p Value Pirfenidone (n 55) Placebo (n 25) p Value Sex Male 62 (86%) 33 (94%) (89%) 24 (96%) Female 10 (14%) 2 (6%) 6 (11%) 1 (4%) Age, yr Smoking history Current 7 (10%) 3 (9%) (13%) 3 (12%) l Ex 57 (79%) 30 (86%) 43 (78%) 21 (84%) Never 8 (11%) 2 (6%) 5 (9%) 1 (4%) Years since IPF diagnosis 1 yr 20 (28%) 6 (17%) (24%) 4 (16%) yr 17 (24%) 10 (29%) 13 (24%) 7 (28%) 3 yr 35 (49%) 19 (54%) 29 (53%) 14 (56%) Surgical lung biopsy 15 (21%) 8 (23%) (18%) 5 (20%) Prior treatment with oral corticosteroids 62 (86%) 30 (86%) (87%) 21 (84%) (14%) 5 (14%) 7 (13%) 4 (16%) VC, % predicted TLC, % predicted DL CO, % predicted Pa O2 at rest, mm Hg Lowest Sp O2 during 6MET, % Log Sp O2 area KL-6, U/ml SP-D, ng/ml Definition of abbreviations: KL-6 serum levels of pneumocyte marker KL-6; N number of patients; 6MET 6-minute steady-state exercise test; SP-D serum level of surfactant protein D; Sp O2 oxygen saturation by pulse oximetry. Plus-minus values are mean SD. * FAS: analysis set for efficacy assessment; derived from the set of all randomized subjects by minimal and justified elimination of subjects (pirfenidone: 72, placebo: 35). Fisher s exact test. Welch s t test. Chi-square test. Wilcoxon s test. Patients completed 6MET: patients completed entire duration of six minutes walking test without desaturating below 80% at baseline (pirfenidone: 55, placebo: 25). placebo group (5/35) and in none of the pirfenidone group during the 9 months (p ). All five patients met the prespecified definition of acute exacerbation and required hospitalization for supportive care that included high-dose corticosteroid therapy and oxygen supplementation. Based on the interim acute exacerbation data at 6 months, the DSMB strongly advised that the study be aborted and pirfenidone be administered to patients in the placebo group. One of the five patients died after the onset of acute exacerbation in the placebo group, whereas there were no deaths in the pirfenidone group during the 9-month study period. At nine months, neither dyspnea nor the quality of life was affected by study medication (p and , respectively). There were no significant changes observed in serum SP-D or KL-6 measurements between groups. Safety Adverse events reported with the frequency of 10% or more at 6 months are listed in Table 3. The events associated with pirfenidone were similar to those of a previous report (23 25). In the pirfenidone group, 62 out of 73 patients (85%) adhered to the treatment protocol at 6 months and 57 (78%) patients adhered to the treatment regimen at 9 months compared with 29 out of 36 patients (81%) and 28 (78%) patients in the placebo group at 6 and 9 months, respectively. The number of patients discontinued is summarized by reason in Table 3. Adverse events causing discontinuation (n 11) in the active arm are photosensitivity (five patients) and, vomiting, fever, abnormality of hepatic function, dizziness, facial paralysis, and hepatoma (one patient for each cause), whereas in the placebo arm, adverse events causing discontinuation were headache and bradycardia (one patient for each). There was no significant difference in the discontinuation of treatment between the two groups at 9 months with the exception of the noted significant episodes of acute exacerbation in the placebo group. The major cause for discontinuation of treatment was adverse events in the pirfenidone group compared with acute exacerbation in the placebo group. Skin photosensitivity was the major adverse event for discontinuing or reducing pirfenidone dose. Most of the adverse events disappeared with decrease of the dose or temporarily holding the medication. These patients tolerated readministration of pirfenidone/placebo at a lower dose based on the prespecified protocol. DISCUSSION An efficacious treatment regimen for IPF is long overdue. While several new agents modulate a specific molecular and cytokine cascade of events implicated in the pathogenesis of IPF (8) and are the subject of ongoing studies, the results of this doubleblind, placebo-controlled, randomized, multicenter clinical trial provide important insights into management of IPF. First, the safety and efficacy of an experimental simple molecule, pirfenidone, is tested in a well defined population of IPF. Second, the clinical significance of acute exacerbation of IPF is highlighted by its manifestation exclusively in the placebo group. Third, change in VC measurement over time is observed to be a suitable

5 1044 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL TABLE 2. COMPARISON AND CORRELATION OF LOWEST OXYGEN SATURATION BY PULSE OXIMETRY AND OXYGEN SATURATION BY PULSE OXIMETRY AREA DURING THE 6-MINUTE WALK TEST AND CORRELATION WITH PULMONARY FUNCTION TESTS Comparison of Change from Baseline Lowest Sp O2 Value during the 6MET and Pulmonary Functions 6 Months 9 Months Pirfenidone Placebo Pirfenidone Placebo Analysis Group Mean SD Mean SD p Value Mean SD Mean SD p Value All patients: full analysis set (FAS*) the lowest Sp O2, % Patients completed 6MET the lowest Sp O2, % All patients: full analysis set (FAS*) VC, L TLC, L DL CO, ml/min/mm Hg Resting Pa O2,mmHg Correlation between the Change in the Lowest Sp O2 and Sp O2 Area During the 6MET and Pulmonary Function Parameters VC (L ) TLC (L ) DL CO (ml/min/mmhg) the lowest Sp O2 Coefficient of correlation p Value log Sp O2 area Coefficient of correlation p Value Definition of abbreviations: 6MET 6-minute steady-state exercise test; Sp O2 oxygen saturation by pulse oximetry; Sp O2 area graphic display of Sp O2 measurements plotted during the 6MET. * FAS: analysis set for efficacy assessment; derived from the set of all randomized subjects by minimal and justified elimination of subjects (pirfenidone: 72, placebo: 35). Patients completed 6MET: Patients completed entire duration of six minutes walking test without desaturating below 80% at baseline (pirfenidone: 55, placebo 25). t test. Welch s t test. outcome measure, and fourth, change in the lowest Sp O2 desaturation during exercise appeared to be a promising new functional measure for IPF. The study was aborted on the basis of the recommendations of the DSMB in favor of the pirfenidone treatment over the placebo group. The early termination before the intended 1-year period of follow-up and the yet-to-be validated primary endpoint are limitations of this study. Therefore, results of this study should be interpreted with caution. This study demonstrated no difference in primary endpoint but a significant difference in secondary endpoints of change in VC and acute exacerbation at 9 months. Whereas there was no significant difference in the primary endpoint between the two Figure 3. Correlation of the lowest Sp O2 with the Sp O2 area during the 6MET at 9 months. Scatter plots are shown for correlation of the lowest Sp O2 reached during the 6MET with the Sp O2 area. The bold black line is the regression of the values of the lowest Sp O2 with the Sp O2 area. r regression. groups when the data were analyzed for all patients (the full analysis set), a positive trend was noted at 9 months in the pirfenidone group. However, pirfenidone treatment suppressed the decrease in oxygenation during exercise in the subset of patients who did not demonstrate Sp O2 less than 80% during the 6MET at baseline. A decrease in Sp O2 of 4% or more during 6 minutes of walking has been recently demonstrated to predict survival in IPF (11). In a recent study, end-exercise Sp O2, change in Sp O2 with exercise, walk distance, and walk velocity were correlated with survival in a 5-year follow-up study in patients with IPF (12). Other studies have correlated exercise-induced hypoxia to the severity of IPF (9, 10). Our study was specifically designed to unmask the exercise-induced oxygen desaturation for each individual patient as the speed determined for each patient was tailored to the patient s ability to comfortably perform the 6MET while demonstrating a Sp O2 less than 90%. Whereas the change in Sp O2 measurements during the 6MET may be clinically relevant to the individual patient, the 6MET requires validation in future studies in IPF. In the categorized analysis of the lowest Sp O2 during the 6MET in the full analysis at 9 months, the number of patients who demonstrated improvement and stabilization was 16 (24%) and 38 (58%), respectively, in the pirfenidone group versus 2 (6%) and 20 (61%) in the placebo group (p ). The change in Sp O2 measurement occurred despite no detectable changes in resting Pa O2 and Dl CO measurements. The explanations for this discordance can only be speculative as the precise mechanism of the biological actions and clinical effects of pirfenidone are unknown. It is possible that the improved oxygenation during exercise is due to the improved ventilation and perfusion matching without reflecting Dl CO measurements, increased minute ventilation, and a decrease in physiologic dead space during exercise. In addition, the presence or absence of occult secondary pulmonary hypertension may have affected some of the clinical and physiologic findings observed in this study.

6 Azuma, Nukiwa, Tsuboi, et al.: Pirfenidone for IPF 1045 TABLE 3. ADVERSE EVENTS AND DISCONTINUATION OF THE STUDY MEDICATION Adverse Events Observed with the Frequency of 10% at 6 Months Pirfenidone Placebo Adverse Events Number of Patients (% ) p Value* Any adverse events 72 (98. 6) 32 (88.9) Photosensitivity 32 (43.8) 0 (0.0) Stomach discomfort 22 (30.1) 3 (8.3) Anorexia 23 (31.5) 2 (5.6) Nausea 16 (21.9) 2 (5.6) Heartburn 12 (16.4) 1 (2.8) Drowsiness 17 (23.3) 6 (16.7) Fatigue 16 (21.9) 1 (2.8) Upper respiratory tract infections 12 (16.4) 3 (8.3) Fever 6 (8.2) 4 (11.1) Elevation of GOT 4 (5.5) 6 (16.7) Elevation of -GTP 20 (27.4) 3 (8.3) Urinary occult blood positive 6 (8.2) 4 (11.1) Elevation of CRP 15 (20.5) 10 (27.8) Discontinuation of the Study Medication at 9 Months Pirfenidone Placebo Reason for Discontinuation Number of Patients (% ) p Value* Adverse events 11 (15.1) 2 (5.6) Photosensitivity 5 (6.8) 0 (0.0) Vomiting 1 (1.4) 0 (0.0) Fever 1 (1.4) 0 (0.0) Abnormality of hepatic function 1 (1.4) 0 (0.0) Dizziness 1 (1.4) 0 (0.0) Facial paralysis 1 (1.4) 0 (0.0) Hepatoma 1 (1.4) 0 (0.0) Headache 0 (0.0) 1 (2.8) Bradycardia 0 (0.0) 1 (2.8) Acute exacerbation 0 (0.0) 5 (13.9) Patient s request for discontinuation 3 (4.1) 0 (0.0) Progression of disease 1 (1.4) 1 (2.8) Protocol violation 1 (1.4) 0 (0.0) Definition of abbreviations: CRP C reactive protein; GOT glutamic oxaloacetic transaminase; -GTP -guanosine triphosphate; * p Value obtained using Fisher s exact test comparing variance between pirfenidone and placebo group. In our study, we displayed the continuous Sp O2 saturation monitored during the 6MET in a novel, visual, and graphic manner as Sp O2 area. There was an excellent correlation between the change in the lowest Sp O2 during exercise and the change in Sp O2 area (Figure 3). The Sp O2 area was significantly smaller in the pirfenidone group than in the placebo group at 6 months in patients who were able to complete the total duration of the 6MET at baseline (see Table E1 in the online supplement). The mean Sp O2 area of all patients increased slightly in the placebo group and decreased in the pirfenidone group during the study period of 9 months. The Sp O2 area may therefore be considered as a supplementary tool or an endpoint in future studies. There was a lesser decline in VC during the 9-month study period observed in the pirfenidone group. Correlation of changes in VC and TLC with the lowest Sp O2 during the 6MET was also significant in our study. Recently, serial changes of FVC at 6 and 12 months and a composite physiologic index have been demonstrated to predict survival in IPF (3 5). Thus, the improvement in VC measurements observed in the pirfenidone group over 9 months is encouraging. Long-term studies are needed to validate the primary endpoint chosen and determine the clinical relevance, significance, and potential benefits of using the Sp O2 during the 6MET as primary outcome measure over the conventional measurements of serial changes in FVC, VC, and other PFTs and/or using a composite parameter, such as the composite physiologic index and clinical radiographic physiologic score in IPF (39). Despite the relatively mild impairment based on VC and TLC, the patients enrolled in our study had clinically significant disease, as they all had significant reduction in Dl CO and Sp O2 on exertion. One reason for this relatively preserved VC could be that the disease may have been diagnosed early, as people get routine medical evaluations (including chest radiographs) in accordance with the national health insurance plan in Japan. Thus, it is apparent that several patients in this study were at an early stage of IPF based on VC, even 3 years after diagnosis and demonstrating Sp O2 less than 90% during the 6MET. Therefore, conventional measurements of VC and TLC may not necessarily reflect the severity of IPF based on oxygen desaturation during exertion in the patient population generally confronted by clinicians to intervene therapeutically in routine practice. Alternatively, patients may demonstrate Sp O2 less than 90% during the 6MET even in early stages of IPF. Future clinical trials will need to include patients with all degrees of impairment (resting PFTs and exercise-induced Sp O2 ) to clarify if there is a preferential effect on outcome measures in IPF. The significant difference in the incidence of acute exacerbation of IPF between the study groups is an important and clinically relevant measure of favorable outcome. The implication of the prevention of acute exacerbation by pirfenidone is potentially crucial for the prognosis of IPF, as these episodes often have a fatal outcome despite aggressive levels of supportive care in the intensive care unit (36, 37, 40). In our study, all the patients who manifested an acute exacerbation of IPF were exclusively in the placebo group. Despite the fact that this secondary endpoint was the main reason for the DSMB s recommendation to abort the study, we express caution in drawing conclusions based solely on this result for the following reasons: (1) limited follow-up data in the study, (2) a consensus among experts worldwide regarding definition is currently not available, and (3) the incidence of well defined acute exacerbation of IPF, its clinical significance, relevance, and relationship to outcome measures and mortality are unknown. Based on the prespecified definition of acute exacerbation of IPF in this study, at least 14% of patients in the placebo group manifested the acute syndrome during a short-term follow-up of 9 months. In a recent study, there appeared to be significant episodes of acute respiratory decompensations preceding death in patients with progressive IPF (6). Future studies in a well defined study population are needed to carefully define and prospectively monitor episodes of acute exacerbation during treatment and to investigate its associations with mortality in IPF. Whereas the precise molecular mechanisms of the effects of pirfenidone are unknown, one may speculate that its multiple biological functions, including scavenging hydroxyl radicals and antiinflammatory and antifibrotic properties (13 22), may in part protect the fibrotic lung from superimposed diffuse lung damage associated with acute exacerbation of IPF. A significant number of patients receiving pirfenidone manifested adverse effects. However, only 11/73 (15%) of the patients discontinued pirfenidone during the study. At 9 months, 54% of patients tolerated the maximum dose of 1,800 mg/day. Approximately half of the patients who did not tolerate the dose of 1,800 mg/day were able to tolerate the dose of 1,200 mg/day. Whereas the adverse events and positive clinical effects associated with pirfenidone were similar to those reported in 54 IPF patients treated with pirfenidone (23), there are differences in dosage between the two studies. The maximum dose used in the previous study done in the United States was 40 mg/kg/day and not more than 3,600 mg/day. While the dose range tolerated in those patients with IPF followed in the United States is unclear,

7 1046 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL Japanese patients tolerated at least 1,200 mg/day and the maximum dose was 1,800 mg/day in this study. Because the mean weight of patients enrolled in this study was 62 kg, it appears that Japanese patients tolerate a dose of mg/kg/day rather than the maximum dose of 40 mg/kg/day used in Caucasian patients. It must be noted that this study was not designed to determine the safety and efficacy in a dose-dependent manner. Phase III clinical trials are needed to clarify if the noted side effects will be tolerated or not by patients for it to be a realistic therapeutic option. In summary, this novel study provides encouraging evidence to pursue the potential of an efficacious treatment with pirfenidone for patients with IPF in a well designed phase III clinical trial. During the 9-month follow-up study, the most striking and clinically important findings noted were that the episodes of acute exacerbation of IPF manifested exclusively in the placebo group and there was a lesser decline in change in VC in patients receiving pirfenidone. Because lesser change in the lowest Sp O2 during the 6MET was noted in the pirfenidone group of patients who were able to complete the 6MET without desaturation to less than 80% at baseline, it can be hypothesized that pirfenidone may have a preferential therapeutic effect in patients with relatively milder disease. Acknowledging the limited follow-up data to date, this is the first prospectively conducted multicenter large clinical trial to demonstrate an improvement in prespecified clinical parameters of outcome in patients treated with oral pirfenidone for IPF. Significant adverse events were associated with pirfenidone; however, there were no differences in the discontinuation of study medications in the pirfenidone and placebo groups. The trial was aborted on the basis of episodes of acute exacerbation of IPF observed exclusively in the placebo group. Long-term studies are needed to confirm our encouraging results in IPF patients treated with pirfenidone. Conflict of Interest Statement : A.A. received from Shionogi & Co., Ltd., $1,984 in 2002 and $1,984 in 2003 for serving as a consultant, $2,480 in 2001, $992 in 2002, and $992 in 2004 for serving on an Advisory Board, and $992 in 2002 for lecture fees; T.N. received from Shionogi & Co., Ltd., $992 in 2002 and $1,984 in 2003 for serving as a consultant and $2,480 in 2001, $992 in 2002, and $992 in 2004 for serving on an Advisory Board; E.T. received from Shionogi & Co., Ltd., $992 in 2001 and $992 in 2002 for serving on an Advisory Board; M.S. received from Shionogi & Co., Ltd., $2,480 in 2001 and $992 in 2004 for serving on an Advisory Board; S.A. received from Shionogi & Co., Ltd., $2,480 in 2001 and $992 in 2002 for serving on an Advisory Board; K.N. received from Shionogi & Co., Ltd., $2,480 in 2001, $992 in 2002, and $992 in 2004 for serving on an Advisory Board; Y.T. received from Shionogi & Co., Ltd., $2,480 in 2001, $992 in 2002, and $992 in 2004 for serving on an Advisory Board; S.N. received from Shionogi & Co., Ltd., $12,276 in 2003 for serving as a consultant and $496 in 2001 and $992 in 2002 for serving on an Advisory Board; H.I. received from Shionogi & Co., Ltd., $3,472 in 2001 and $6,944 in 2002 for serving as an expert witness; M.O. received from Shionogi & Co., Ltd., $1,984 in 2002 for serving as a consultant; A.S. received from Shionogi & Co., Ltd., $4,687 in 2003 and $10,044 in 2004 for serving as a consultant and $992 in 2004 for serving on an Advisory Board; S.K. received from Shionogi & Co., Ltd., $1,984 in 2002 and $992 in 2003 for serving as a consultant, $2,480 in 2001 and $992 in 2004 for serving on an Advisory Board, and $992 in 2002 as lecture fees; G.R. is a consultant, advisor, and steering committee member for clinical trials using -IFN (unrelated to this study), and received the following from Intermune: $1,000 (lecture) and $500 (consultation) in 2001, $1,500 (advisor) and $3,125 (honoraria for lectures) in 2002, $2,500 (advisor) and $500 (consultation for pirfenidone studies) in 2003 and $2,000 for pirfenidone trial design (advisor) in 2004, and he received from Shionogi & Co., Ltd. $1,000 (consultant, advisor) in 2002, and he has received $20,000 in 2001, $17,750 in 2002, $2,000 in 2003 and $15,500 in 2004 as honoraria for CME lectures on management of IPF and related topics. Acknowledgment : The authors thank Dr. Talmadge King, UCSF, San Francisco, CA, for expert advice and Mr. Craig Johnson, M.S., University of Washington, Seattle, WA, Mr. Kiyoshi Shirai, Dr. Yasumasa Goh, Ms. Masayo Kawaoka, Mr. Masaaki Yamada, and Ms. Toshimi Kitai for their advice and for reviewing the manuscript. Members of Research Group for Diffuse Lung Diseases in Japan : Shigeru Tsukagoshi (Tokyo Cooperative Oncology Group); Keiichi Nagao (Chiba University); Ariyoshi Kondo (Niigata Tetsudo-Kenshin Center); Masato Takebe (Kitasato University Graduate School); Hiroki Takahashi (Sapporo Medical University School of Medicine); Masahito Ebina (Tohoku University); Eiichi Suzuki (Niigata University Medical); Yukihiko Sugiyama (Jichi Medical School); Yasuyuki Yoshizawa (Tokyo Medical and Dental University); Ken Ohta (Teikyo University); Yoshinosuke Fukuchi (Juntendo University); Atsushi Nagai (Tokyo Women s Medical University); Rokuro Matsuoka (Showa General Hospital); Masaru Oritsu (Japanese Red Cross Medical Center); Minoru Kanazawa (Saitama Cardiovascular and Respiratory Center); Hiroyuki Taniguchi (Tosei General Hospital); Kingo Chida and Hirotoshi Nakamura (Hamamatsu University School of Medicine); Michiaki Mishima (Kyoto University); Nobuyuki Katakami (Kobe City General Hospital); Yoshikazu Inoue (National Kinki- Chuo Hospital for Chest Diseases); Yoshiro Mochizuki (National Himeji Hospital); Takeshi Isobe and Nobuoki Kohno (Hiroshima University); Hironobu Hamada (Ehime University); Hiroshi Mukae and Shigeru Kohno (Nagasaki University School of Medicine); Tomiyasu Tsuda (Oita Medical University); Satoshi Noma (Tenri Hospital); and Kiyoshi Murata (Shiga University Hospital). References 1. American Thoracic Society/European Respiratory Society. Idiopathic pulmonary fibrosis: diagnosis and treatment, international consensus statement. Am J Respir Crit Care Med 2000;161: American Thoracic Society/European Respiratory Society. American Thoracic Society/European Respiratory Society International multidisciplinary consensus classification of the idiopathic interstitial pneumonias. Am J Respir Crit Care Med 2002;165: Collard HR, King TE Jr, Bartelson BB, Vourlekis JS, Schwarz MI, Brown KK. Changes in clinical and physiologic variables predict survival in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2003;168: Latsi PI, Du Bois RM, Nicholson AG, Colby TV, Bisirtzoglou D, Nikolakopoulou A, Veeraraghavan S, Hansell DM, Wells AU. Fibrotic idiopathic interstitial pneumonia: the prognostic value of longitudinal functional trends. Am J Respir Crit Care Med 2003;168: Flaherty KR, Mumford JA, Murray S, Kazerooni EA, Gross BH, Colby TV, Travis WD, Flint A, Toews GB, Lynch JP III, et al. Prognostic implications of physiologic and radiographic changes in idiopathic interstitial pneumonia. Am J Respir Crit Care Med 2003;168: Raghu G, Brown KK, Bradford WZ, Starko K, Noble PW, Schwartz DA, King TE. A placebo-controlled trial of interferon gamma-1b in patients with idiopathic pulmonary fibrosis. N Engl J Med 2004;350: Teirstein AS. The elusive goal of therapy for usual interstitial pneumonia. NEnglJMed2004;350: Mason RT, Schwarz MI, Hunninghake GW, Musson RA. Pharmacological therapy for idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1999;160: King TE, Schwarz MI, Brown K, Tooze JA, Colby TV, Waldron JA, Flint A, Thurlbeck W, Cherniack RM. Idiopathic pulmonary fibrosis: relationship between histopathologic features and mortality. Am J Respir Crit Care Med 2001;164: King TE, Tooze JA, Schwarz MI, Brown KR, Cherniack RM. Predicting survival in idiopathic pulmonary fibrosis: scoring system and survival model. Am J Respir Crit Care Med 2001;164: Lama VN, Flaherty KR, Toews GB, Colby TV, Travis WD, Long Q, Murray S, Kazerooni EA, Gross BH, Lynch JP III, et al. Prognostic value of desaturation during a 6-minute walk test in idiopathic interstitial pneumonia. Am J Respir Crit Care Med 2003;168: Hallstrand TS, Boitano LJ, Johnson WC, Spada CA, Hayes JG, Raghu G. The timed walk test as a measure of severity and survival in idiopathic pulmonary fibrosis. Eur Respir J 2005;25: Margolin SB, Lefkowitz S. Pirfenidone: a novel pharmacologic agent for prevention and resolution of lung fibrosis. FASEB J 1994;8:A Iyer SN, Gurujeyalakshmi G, Giri SN. Effects of pirfenidone on transforming growth factor- gene expression at the transcriptional level in bleomycin hamster model of lung fibrosis. J Pharmacol Exp Ther 1999;291: Gurujeyalakshmi G, Hollinger MA, Giri SN. Pirfenidone inhibits PDGF isoforms in bleomycin hamster model of lung fibrosis at the translational level. Am J Physiol 1999;276:L311 L Nakazato H, Oku H, Yamane S, Tsuruta Y, Suzuki R. A novel antifibrotic agent pirfenidone suppresses tumor necrosis factor- at the translational level. Eur J Pharmacol 2002;446: Oku H, Nakazato H, Horikawa T, Tsuruta Y, Suzuki R. Pirfenidone suppresses tumor necrosis factor-, enhances interleukin-10 and protects mice from endotoxic shock. Eur J Pharmacol 2002;446: Iyer SN, Gurujeyalakshmi G, Giri SN. Effects of pirfenidone on procollagen gene expression at the transcriptional level in bleomycin hamster model of lung fibrosis. J Pharmacol Exp Ther 1999;289: Misra HP, Rabideau C. Pirfenidone inhibits NADPH-dependent micro-

8 Azuma, Nukiwa, Tsuboi, et al.: Pirfenidone for IPF 1047 somal lipid peroxidation and scavenges hydroxyl radicals. Mol Cell Biochem 2000;204: Iyer SN, Wild JS, Schiedt MJ, Hyde DM, Margolin SB, Giri SN. Dietary intake of pirfenidone ameliorates bleomycin-induced lung fibrosis in hamsters. J Lab Clin Med 1995;125: Margolin S, Margolin B, Margolin D. Removal of interstitial pulmonary fibrosis (asbestos-induced) by oral chemotherapy with pirfenidone. Fed Proc 1982;41: Kehrer JP, Margolin SB. Pirfenidone diminishes cyclophosphamideinduced lung fibrosis in mice. Toxicol Lett 1997;90: Raghu G, Johnson WC, Lockhart D, Mageto Y. Treatment of idiopathic pulmonary fibrosis with a new antifibrotic agent, pirfenidone: results of a prospective, open-label Phase II study. Am J Respir Crit Care Med 1999;159: Nagai S, Hamada K, Shigematsu M, Taniyama M, Yamauchi S, Izumi T. Open-label compassionate use one year-treatment with pirfenidone to patients with chronic pulmonary fibrosis. Intern Med 2002;41: Gahl WA, Brantly M, Troendle J, Avila NA, Padua A, Montalvo C, Cardona H, Calis KA, Gochuico B. Effect of pirfenidone on the pulmonary fibrosis of Hermansky-Pudlak syndrome. Mol Genet Metab 2002;76: Azuma A, Tsuboi E, Abe S, Nukiwa T, Nakata K, Nagai S, Taguchi Y, Suga M, Itoh H, Ohi M, et al. A placebo control and double blind Phase II clinical study of pirfenidone in patients with idiopathic pulmonary fibrosis in Japan [abstract]. Am J Respir Crit Care Med 2002;165:A Raghu G, Mageto YN, Lockhart D, Schmidt RA, Wood DE, Godwin JD. The accuracy of the clinical diagnosis of new-onset idiopathic pulmonary fibrosis and other interstitial lung disease: a prospective study. Chest 1999;116: Nishimura K, Kitaichi M, Izumi T, Nagai S, Kanaoka M, Itoh H. Usual interstitial pneumonia: histologic correlation with high-resolution CT. Radiology 1992;182: Yokoyama A, Kohno N, Hamada H, Sakatani M, Ueda E, Kondo K, Hirasawa Y, Hiwada K. Circulating KL-6 predicts the outcome of rapidly progressive idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998;158: Takahashi H, Fujishima T, Koba H, Murakami S, Kurokawa K, Shibuya Y, Shiratori M, Kuroki Y, Abe S. Serum surfactant proteins A and D as prognostic factors in idiopathic pulmonary fibrosis and their relationship to disease extent. Am J Respir Crit Care Med 2000;162: Tsuboi E, Kawabata M, Kishi K, Narui K, Homma S, Nakatani T, Nakata K. Treadmill six-minute walk tests in patients with idiopathic pulmonary fibrosis. The Annual Report by Study Group of Ministry of Health and Welfare for Diffuse Lung Disease 2000; Damien S, Ellen E, Kailash S, Szidon P, Ankin M, Kesten S. Comparison of hallway and treadmill six-minute walk tests. Am J Respir Crit Care Med 1999;160: Wijkstra PJ, TenVergert EM, Van Altena R, Otten V, Postma DS, Kraan J, Koëter GH. Reliability and validity of the chronic respiratory questionnaire (CRQ). Thorax 1994;49: Chang JA, Curtis JR, Patrick DL, Raghu G. Assessment of healthrelated quality of life in patients with interstitial lung disease. Chest 1999;116: Fletcher CM. The clinical diagnosis of pulmonary emphysema; an experimental study. Proc R Soc Med 1952;45: Kondoh Y, Tanigushi H, Kawabata Y, Yokoi T, Suzuki K, Takagi K. Acute exacerbation in idiopathic pulmonary fibrosis: analysis of clinical and pathologic findings in three cases. Chest 1993;103: Akira M, Hamada H, Sakatani M, Kobayashi C, Nishioka M, Yamamoto S. CT findings during phase of accelerated deterioration in patients with idiopathic pulmonary fibrosis. Am J Roentgenol 1997;168: Safety Division of Ministry of Health and Welfare. Notification No. 80 of the Safety Division, Pharmaceutical Affairs Bureau dated June 29, Watters LC, Schwarz MI, Cherniack RM, Waldron JA, Dunn TL, Stanford RE, King TE. Idiopathic pulmonary fibrosis: pretreatment bronchoalveolar lavage cellular constituents and their relationships with lung histopathology and clinical response to therapy. Am Rev Respir Dis 1987;135: Gong MS, Mark EJ. Case records of the Massachusetts General Hospital: weekly clinicopathological exercises, Case , a 56-year-old man with rapidly worsening dyspnea. N Engl J Med 2002;347: