Colchicine versus Prednisone in the Treatment of Idiopathic Pulmonary Fibrosis A Randomized Prospective Study

Colchicine versus Prednisone in the Treatment of Idiopathic Pulmonary Fibrosis A Randomized Prospective Study WILLIAM W. DOUGLAS, JAY H. RYU, STEPHEN J. SWENSEN, KENNETH P. OFFORD, DARRELL R. SCHROEDER, GAIL M. CARON, RICHARD A. DEREMEE, and Members of the Lung Study Group* Division of Pulmonary and Critical Care Medicine, Department of Diagnostic Radiology, and Section of Biostatistics, Mayo Foundation, Rochester, Minnesota Twenty-six symptomatic subjects with clinical evidence plus either high-resolution computed tomography (HRCT, n 25) or open-lung biopsy (OLB, n 1) patterns typical for idiopathic usual interstitial pneumonia (idiopathic UIP) were entered into a randomized prospective treatment trial using high-dose prednisone (n 12) versus colchicine (n 14). The minimum dose of prednisone used was 60 mg/d for 1 mo, tapered to 40 mg/d over the second month, tapered to 40 mg every other day during the third month, with subsequent doses adjusted as clinically indicated. The dose of colchicine was 0.6 1.2 mg/d, as tolerated. The presence of a rim of subpleural honeycomb change was present in all of the 25 subjects who had HRCT. Subjects treated with high-dose prednisone alone experienced a higher incidence of serious side effects and also exhibited a trend (not statistically significant, p 0.391) to more rapid decline of pulmonary function and shortened survival than did those treated with colchicine alone. In most subjects with typical clinical and HRCT features of idiopathic UIP, neither prednisone nor colchicine resulted in objective improvement, and the disease continued to progress in the majority. Colchicine appears to be a safer alternative to a trial of high-dose prednisone but may be no different than no therapy. Douglas WW, Ryu JH, Swensen SJ, Offord KP, Schroeder DR, Caron GM, DeRemee RA, and members of the Lung Study Group. Colchicine versus prednisone in the treatment of idiopathic pulmonary fibrosis: a randomized prospective study. AM J RESPIR CRIT CARE MED 1998;158:220 225. Idiopathic pulmonary fibrosis (IPF), also known in the United Kingdom as cryptogenic fibrosing alveolitis, is a group of disorders of unknown cause, characterized by proliferation of excessive fibrous tissue in the lung. IPF has been estimated to compromise 46.2% of all interstitial lung diseases in males and 44.2% in females, and is the most common of the interstitial lung diseases (1). The prevalence of IPF has been estimated to be 20.2 per 100,000 for males and 13.2 per 100,000 for females; the incidence has been estimated to be 10.7 per 100,000/yr for males and 7.4 per 100,000/yr for females (1). For the IPF group as a whole, median survival averages 5 yr after diagnosis, with the majority of patients experiencing a progressive downhill course, usually characterized by dyspnea, nonproductive cough, arterial hypoxemia, restrictive impairment of pulmonary function, and diffuse pulmonary infiltrates on chest radiographs (2, 3). (Received in original form September 22, 1997 and in revised form January 13, 1998) * David M. Fisk, Geisinger, PA; Michael J. Krowka, Jacksonville, FL; Ashokakumar M. Patel, Vancouver, BC; and Oscar A. Schwartz, St. Louis, MO. Funding was provided by Mayo Institute funds. Correspondence and requests for reprints should be addressed to William W. Douglas, M.D., 200 First St. SW, Rochester, MN 55905. Am J Respir Crit Care Med Vol 158. pp 220 225, 1998 Internet address: www.atsjournals.org The most common clinicopathologic subgroup of IPF is usual interstitial pneumonia (UIP). Katzenstein and Myers have proposed a revision of the histologic classification of the idiopathic interstitial pneumonias (4), which includes UIP, desquamative interstitial pneumonia (DIP), nonspecific interstitial pneumonia (NSIP) (5), and acute interstitial pneumonia (AIP). In a recent study of biopsy-confirmed IPF using this system (6), UIP had a more rapidly progressive course and shortened mean survival that the IPF group as a whole, which included subjects with each of the above entities plus chronic hypersensitivity pneumonia (CHSP), bronchiolitis obliterans organizing pneumonia (BOOP), chronic eosinophilic pneumonia (CEP), and pulmonary eosinophilic granuloma (EG). The term IPF has been used inconsistently in previous reports, sometimes incorporating patients considered clinically to have connective-tissue diseases and sometimes including subjects who would now be judged histologically to have DIP and NSIP. Increasingly, the term IPF has been used to describe only subjects with idiopathic UIP. Standard therapy for idiopathic UIP has often included high-dose prednisone, with or without immunosuppressive drugs such as azathioprine or cyclophosphamide (7 10). Recent prospective studies of subjects with idiopathic UIP suggest that high-dose prednisone regimens are associated with a high incidence of serious and irreversible corticosteroid-related side effects (11). Even with high-dose prednisone therapy, there is

Douglas, Ryu, Swensen, et al.: Colchicine and Pulmonary Fibrosis 221 no prospective study that has shown a statistically significant decrease in mortality when compared with no therapy. When combined with corticosteroids, both cyclophosphamide (9) and azathioprine (10) have shown a trend toward improved survival compared with high-dose prednisone alone in subjects with idiopathic UIP. However, cyclophosphamide can produce an interstitial lung disease similar in appearance to that seen in early UIP before honeycomb change has occurred (12), and has caused bladder injury and bladder cancer. In addition, both cyclophosphamide and azathioprine may cause suppression of bone marrow function and may increase the risk of both opportunistic infection and lymphoreticular malignancy (13). Because of the problems associated with the currently recommended treatment of idiopathic UIP, alternative therapies have been sought. Over the past 10 yr at Mayo Clinic Rochester, subjects who refuse to take prednisone, have contraindications to prednisone, or have failed prednisone have been treated with colchicine (14). Colchicine initially was chosen because of its relative lack of side effects, its potential as an antifibrotic agent, and its low cost. Preliminary studies from this institution (15, 16) have demonstrated that treatment using colchicine is well tolerated by most subjects. When compared with high-dose prednisone, colchicine is as effective (as measured by the rate of decline of pulmonary function) and is associated with fewer serious and irreversible side effects. This prospective study was implemented because previous studies of colchicine therapy for idiopathic UIP are retrospective and have other limitations. METHODS The minimum dose of prednisone was 60 mg/d for 1 mo, tapered by 10 mg every 2 wk to 40 mg/d at the end of the second month, then tapered gradually to 40 mg every other day by the end of the third month, with subsequent doses determined by clinical response at the discretion of the treating clinician. This dose could be increased if the patient declined symptomatically during weaning. The duration of 3 mo was selected because it was at the upper range of the duration of unsuccessful trials in patients referred to us by others. The dose of prednisone used after 3 mo was not precisely defined, so as to accommodate the needs of both responders and nonresponders, with and without side effects. In the colchicine arm, the dose was 0.6 1.2 mg/d, at the highest dose tolerated by the subject. To permit flexibility in the treatment programs, the use of colchicine in the prednisone arm for less than 2 wk, and of prednisone at less than 20 mg/d for less than 2 wk in the colchicine arm was permitted for reasons other than as treatment of IPF. At entry, subjects were stratified into groups according to whether the subject had previous treatment for IPF with corticosteroids (yes, no), whether open-lung biopsy (OLB) had been done (yes, no) and by severity of impairment of pulmonary function (severe, less than severe) using American Thoracic Society (ATS) criteria. Within each stratum defined by the three stratification factors described previously, the randomization was done in blocks of four, ensuring that after every fourth subject was entered in a given stratum the number of subjects in the stratum assigned to prednisone was the same as the number of subjects in the strata assigned to colchicine. Exclusion Criteria Exclusion criteria were as follows: history of allergy, intolerance, or unwillingness to take either study drug; pregnancy, lactation, or women capable of becoming pregnant who were without adequate birth control; history of chronic asthma and/or treated for asthma within the previous year; diabetes treated (including dietary therapy) within the previous year; active tuberculosis treated within the previous year; use of either study drug within the previous 2 mo. Inclusion Criteria We used the following inclusion criteria: conforms to clinical plus either high-resolution computed tomography (HRCT) or histopathologic criteria for the diagnosis of idiopathic UIP; baseline tests performed, including pulmonary function, chest radiograph, serum creatinine, liver function tests, and complete blood count; willing to return for followup examination at 3-mo intervals for 1 yr; age 18 yr or older; written informed consent given. Definition of Clinical Criteria for a Diagnosis of Idiopathic UIP 1. Typical bibasilar end-inspiratory crackles (often described as velcro or cellophane in character) present on auscultation of the lungs. 2. Clinical evidence of UIP present for at least 3 mo with progression of symptoms, pulmonary function impairment, or radiographic abnormality. 3. No evidence of a specific cause of UIP, including concomitant clinical evidence of connective tissue disease, significant asbestosis exposure, previous radiation therapy to the chest, previous cancer chemotherapy, fibrogenic drug therapy at onset of pulmonary disease, history of acute lung injury at onset of pulmonary disease, evidence for recurrent aspiration, or evidence for hypersensitivity pneumonitis. Definition of HRCT Criteria for a Diagnosis of UIP Criteria for the HRCT diagnosis of UIP were modified from the work of several investigators (17 21): 1. A reasonably symmetrical bilateral interstitial pulmonary process must be demonstrated. 2. Pulmonary infiltrates must be predominantly irregular linear or reticular opacities with areas of associated architectural distortion including traction bronchiolectasis and/or bronchiectasis. 3. These infiltrates must be distributed preferentially to subpleural locations and must involve both extreme lung bases. 4. Some honeycomb change must be present, and must be distributed preferentially to subpleural regions of the lung. 5. Minimal areas of associated ground glass opacity may be present in addition to the above pattern, but must never be the dominant or exclusive pattern. 6. The presence of a centrilobular micronodular process disqualifies the subject. Definition of Histopathologic Criteria for a Diagnosis of UIP Criteria for the diagnosis of UIP are taken from the work of Katzenstein and Askin (22): 1. A patchy interstitial process with variable distribution characterized by inflammation and fibrosis must be present. 2. This process must demonstrate a predilection for peripheral and subpleural regions of the lung. 3. This process should include areas that appear temporally heterogeneous, with foci of scarring, active inflammation ( fibroblast foci ), and normal lung. 4. For this study, the presence of any granulomas excluded the patient. Study Endpoints The primary study failure endpoint was defined at the time of the first occurrence of any of the following: death, significant deterioration of pulmonary function, intolerance or adverse event due to the study drug requiring cessation of therapy, addition of a second drug for treatment of UIP, and study dropout for any reason. Failure was assigned to one cause only according to the above hierarchy. Significant pulmonary function deterioration was defined as a decline from baseline of FVC by 15% or more and at least 200 ml, and/or pulmonary diffusing capacity for carbon monoxide (DL CO ) by 20% or more and at least 3 ml/min/mm Hg (23, 24). Nonsteroidal anti-inflammatory drugs, aspirin, and antioxidants or multivitamins were not considered to be other specific drugs for treatment of UIP, and were not prohibited nor were a basis for declaring the subject as an endpoint failure, since none of these agents has yet been reported to alter outcome in UIP. Compliance with therapy was documented by review of a daily diary, in which missed doses were recorded, accompanied by reasons

222 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 158 1998 for the missed doses. These diaries were kept for the first 3 mo of the trial. Subjects were recruited by pulmonologists at Mayo Clinic Rochester, or from other sites by members of the Lung Study Group who had appropriate institutional review board (IRB) approval. A screening questionnaire inquiring about potential causes for pulmonary fibrosis was administered at entry. At entry, a modified dyspnea index questionnaire, and a functional status questionnaire (SF36) were used to assess pulmonary symptoms and impairment. HRCT scans were reviewed by a chest radiologist with a special interest in interstitial lung diseases (S.J.S.). The single OLB, done at another institution, was reviewed at Mayo Clinic Rochester by a pathologist with experience in evaluating interstitial lung diseases using the Katzenstein-Myers criteria. Randomization was accomplished by the Section of Biostatistics at Mayo Rochester. Subjects were evaluated at regular 3-mo intervals. This evaluation included limited physical examination, chest radiography, pulmonary function tests with diffusing capacity, complete blood count, and screening blood chemistries including fasting glucose, creatinine, alkaline phosphatase, bilirubin, and aspartate aminotransferase. The ATS, functional status, and modified dyspnea index questionnaires were again administered, and an additional questionnaire addressing drug side effects was completed. For subjects arriving at an endpoint, a form detailing the specific reason or reasons for failure was completed. Pulmonary Function Measurements Pulmonary function tests (PFTs) were measured using Medical Graphics (St. Paul, MN) 1070 and 1085 equipment, using a pneumotach spirometer, plethysmographic lung volumes, and single-breath diffusing capacity measured with the use of a chromatograph to analyze gases. Spirometers were calibrated daily and met all American Thoracic Society performance specifications. Predicted equations for FVC and DL CO were from Miller and associates (25, 26). Statistical Methods Patient demographics and baseline pulmonary function measurements were compared for prednisone-treated versus colchicine-treated patients using the rank sum test for continuous variables and the chi-square test for categorical variables. The percentage of subjects meeting failure criteria in the first 12 mo after randomization was compared for prednisone versus colchicine using the Fisher exact test. Cumulative survival probabilities were estimated using the Kaplan-Meier method and the log-rank test was used to compare survival for prednisone versus colchicine. The log-rank test was used to compare overall survival for subjects included in this trial versus those included in a recent retrospective review of patients diagnosed with idiopathic UIP by openlung biopsy at Mayo Clinic Rochester during the years 1976 through 1985 (6). RESULTS Enrollment Twenty-six subjects with idiopathic UIP were enrolled, 14 of whom were randomized to colchicine, and 12 to prednisone. None of these subjects has been included in previous reports from this institution. All but four of the subjects were enrolled at Mayo Clinic Rochester; one subject each was recruited and followed elsewhere by a member of the Lung Study Group. One of the subjects in this study had OLB and fulfilled the clinicopathologic criteria; the other 25 fulfilled the clinical and HRCT clinicoradiologic criteria for the diagnosis of idiopathic UIP without lung biopsy. Demographics and Baseline Pulmonary Function Patient demographics and baseline pulmonary function tests are presented in Tables 1 and 2. None of the baseline characteristics was found to be significantly different for prednisoneversus colchicine-treated subjects. Antibody to nuclear antigens (ANA) and rheumatoid factor was positive in low titer in many patients, but the incidence of a positive ANA and rheumatoid factor was not different between the two study groups. Characteristic TABLE 1 PATIENT CHARACTERISTICS Prednisone (n 12) Colchicine (n 14) Age, yr (mean SD) 69.9 4.0 65.9 12.3 Male, % 83 71 Digital clubbing, % 42 36 Family history of IPF, % 17 7 Previous corticosteroid therapy, %* 8 17 Duration of symptoms, yr 3.2 1.7 2.9 3.0 * Use of either colchicine or prednisone within the 2 mo prior to enrollment was an exclusion criterion. Data were missing on two prednisone subjects and one colchicine subject. Success and Failure Success, defined as completing 1 yr of therapy without failure from any cause, was not found to be significantly different between groups (p 0.391) and occurred in 2 of 12 (17%, 95% confidence interval 2 to 48%) prednisone-treated subjects and 5 of 14 (36%, 95% confidence interval 13 to 65%) colchicinetreated subjects. The 95% confidence interval for the difference in success rates is consistent with colchicine being better than prednisone by 52 percentage points and worse than prednisone by 14 percentage points. Treatment failures are summarized in Table 3 according to reason and time of failure. No subject in either treatment group demonstrated a significant improvement in pulmonary function compared with baseline using ATS criteria. Of the 10 prednisone-treated patients who failed, seven subjects experienced documented pulmonary function decline, one subject became profoundly weak and dyspneic and was admitted to a nursing home (6 wk after starting treatment when taking 50 mg/d) without any follow-up PFTs and was unable to return thereafter, and one subject withdrew consent. One subject, at 9 mo after starting treatment when taking 20 mg/d without documented PFT failure, died after being intubated and mechanically ventilated after developing accelerated decline with features of the adult respiratory distress syndrome superimposed on previously slowly progressive idiopathic UIP. Of the nine colchicine-treated subjects meeting failure criteria, five subjects experienced documented pulmonary function decline, one subject failed because he developed intolerable diarrhea on the 0.6 mg/d dose, and one subject withdrew consent. Two subjects failed as a result of a second drug being added, one because of lack of symptomatic improvement, the other because he exceeded the 14-d limit for prednisone which was substituted (unsuccessfully) for colchicine because of severe diarrhea which antedated colchicine therapy. After meet- TABLE 2 BASELINE PULMONARY FUNCTION Prednisone Colchicine Characteristic n Mean SD n Mean SD FVC, % predicted 12 66.8 14.9 14 62.4 14.8 DL CO, % predicted 11* 46.9 10.8 14 51.9 18.1 Sa O2 at rest, % 12 92.8 2.5 13 91.1 7.0 Sa O2 with exercise, % 10 80.8 7.9 10 86.3 9.3 Definition of abbreviations: FVC forced vital capacity; DL CO diffusing capacity for carbon monoxide; Sa O2 arterial oxygen saturation. * Because of a small vital capacity it was technically not possible to measure baseline DL CO for one patient randomized to prednisone. Sa O2 data are not available for one colchicine-treated patient. Two prednisone-treated and three colchicine-treated patients were unable to exercise.

Douglas, Ryu, Swensen, et al.: Colchicine and Pulmonary Fibrosis 223 TABLE 3 NUMBER OF FAILURES ACCORDING TO TREATMENT GROUP, REASON, AND TIME OF FAILURE* Prednisone (n 12) Colchicine (n 14) Reason 3 mo 6 mo 9 mo 12 mo 3 mo 6 mo 9 mo 12 mo PFT decline 4 2 1 0 1 3 0 1 Drug intolerance 0 0 0 0 0 1 0 0 Other drug added 0 0 0 0 0 2 0 0 Withdrew consent 0 1 0 0 1 0 0 0 Other 1 0 1 0 0 0 0 0 Cumulative total 5 8 10 10 2 8 8 9 42% 67% 83% 83% 14% 57% 57% 64% * There were 2 of 12 (17%) prednisone-treated patients and 5 of 14 (36%) colchicine-treated patients who did not meet any failure criteria during the 12-mo study period. Success, defined as completing 1 yr of therapy without failure for any reason, was not found to be significantly different between the two treatment groups (p 0.391). Note that subjects were considered to have failed for one failure criterion only, according to the hierarchy discussed in METHODS. After meeting any failure criteria patients were followed for survival information only (i.e., additional follow-up PFT information was not obtained). One prednisone-treated patient became profoundly weak and dyspneic, was admitted to a nursing home and was unable to return for any follow-up PFTs. Another prednisone-treated patient returned for 3- and 6-mo follow-up visits but died without documented PFT failure 9 mo following enrollment after developing accelerated decline with features of the adult respiratory distress syndrome superimposed on previously slowly progressive idiopathic UIP. ing failure criteria, subjects were followed for survival information only (i.e., subsequent follow-up pulmonary function tests were not obtained). For this reason, analyses of individual failure criteria were not performed. Pulmonary Function Because subjects were followed for only vital status information after meeting failure criteria, follow-up pulmonary function is not consistently available for all subjects. FVC was measured in 11 prednisone- and 13 colchicine-treated subjects at their first follow-up visit, whereas DL CO was measured in 10 and 12 subjects respectively. The time to the first follow-up PFT measurement was not significantly different for prednisone- versus colchicine-treated subjects (3.3 0.8 and 3.5 1.6 mo respectively). Since not all subjects returned at exactly 3 mo following their randomization date, FVC and DL CO at 3 mo was obtained via linear interpolation. At 3 mo, the change Figure 1. Survival after initiation of treatment of idiopathic usual interstitial pneumonitis using either colchicine alone (dashed line) or prednisone alone (solid line). There is a trend for more prolonged survival in the colchicine-treated group which is not statistically significant (p 0.108). Figure 2. Survival of subjects enrolled in this study with idiopathic UIP diagnosed by clinical and HRCT criteria (dashed line) compared with historical subjects with idiopathic UIP diagnosed by OLB at the Mayo Clinic between the years 1976 1985 (solid line). There is no significant difference in mortality between the two groups. from baseline for FVC (percent predicted) was not significantly different for prednisone- versus colchicine-treated subjects ( 6.9 6.8 and 5.1 7.7, respectively, p 0.385), although both treatment groups experienced a significant decline from baseline (p 0.012 and p 0.027, respectively). The change from baseline DL CO (ml/min/mm Hg) was not significantly different for prednisone- versus colchicine-treated subjects ( 2.0 2.1 and 1.1 1.3, p 0.529), however both treatment groups experienced a significant decrease from baseline (p 0.031 and p 0.017, respectively). Survival The survival curves for subjects treated with prednisone versus colchicine are presented in Figure 1. Median length of follow-up was 1.5 yr. There was a trend toward worse survival for subjects treated with prednisone versus colchicine. However, this difference was not statistically significant (p 0.108). The overall estimated 1-yr mortality for subjects included in this trial was 21% and the estimated 2-yr mortality was 58%. The 95% confidence interval for the difference in 1-yr survival is consistent with colchicine being better than prednisone by 47 percentage points and worse than prednisone by 19 percentage points. Survival for subjects included in this trial was not significantly different from that noted in subjects previously diagnosed with idiopathic UIP by OLB at Mayo Clinic Rochester between the years 1976 1985; see Figure 2. No significant difference in mortality was found for subjects 65 yr of age or older when compared with those 65 yr, either for both treatment groups together, or for either treatment group alone. Drug Side Effects Drug side effects are listed in Table 4. Death due to accelerated decline occurred in one prednisone-treated subject at 9 mo at a dose of 20 mg/d. One near-fatal episode of gram-negative bacteremic shock due to urosepsis superimposed on corticosteroid-induced hyperglycemia occurred after 4 wk of therapy at a dose of 60 mg prednisone/day. Cushingoid appearance, compression fracture of a spinal vertebra, diabetes mellitus, and myopathy (defined as objective quadriceps weakness) were seen only in prednisone-treated subjects, and insomnia was more common in prednisone-treated subjects. Nausea was

224 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 158 1998 Adverse Event reported only in colchicine-treated subjects, and diarrhea was more commonly seen in colchicine-treated subjects. DISCUSSION TABLE 4 ADVERSE EVENTS DURING TREATMENT, PERCENT* Prednisone, % (n 12) Colchicine, % (n 14) p Value Cushingoid 75 0 0.001 Depression 25 7 0.10 Diabetes 50 0 0.004 Diarrhea 17 57 0.051 Epigastric pain 25 0 0.10 Insomnia 50 7 0.026 Myopathy 42 0 0.012 Muscle cramps 42 14 0.10 Nausea 0 21 0.10 Proteinuria 8 21 0.10 * Adverse events that were experienced one or more times by 20% of subjects in either treatment group are summarized. No significant differences were found between treatment groups for adverse events experienced by 20% of subjects in each treatment group. Fisher exact test. The purpose of this randomized prospective trial was to evaluate and compare the impact of colchicine versus prednisone as single-drug therapy for subjects with IPF who have experienced symptomatic progression. In addition, we compared the effect of the two regimens on pulmonary function as well as subject compliance and adverse effects of therapy. We found that subjects treated with prednisone experienced a trend, though not statistically significant, to more rapid decline of pulmonary function and death, and also had a higher incidence of side effects than did subjects treated with colchicine. This is the third prospective study to compare an alternative treatment program to high-dose corticosteroid therapy alone, and all three studies are similar in showing a trend to better outcomes for the alternative treatment arm as compared with corticosteroids alone. In the Johnson study (9), prednisolone alone at a dose of 60 mg/d for 1 mo then slowly tapered to 20 mg every other day, was compared to oral cyclophosphamide 100 120 mg/d plus prednisolone 20 mg every other day. At 3 yr, 10 of 22 patients in the prednisolone only group had died; whereas only 3 of 21 of the cyclophosphamide prednisolone group had died; this difference was not statistically significant (p 0.1). In the Raghu study (10), prednisone at 1.5 mg/kg alone for 2 wk then tapered slowly to 20 mg/d plus placebo was compared with prednisone at the same dose and schedule plus azathioprine at 3 mg/kg/d. Mortality was not different between the two groups (p 0.16). No deaths were seen after 3 yr in the azathioprine group, but follow-up for 36 mo was incomplete. These and other previous studies suggested a 20 to 30% overall favorable response rate to corticosteroid therapy (8 10), which has lead to the recommendation by some physicians that all subjects with idiopathic UIP undergo a trial of high-dose corticosteroid therapy. In retrospect, it seems likely that many of those responding favorably may have done so because they had disorders such as DIP or NSIP rather than UIP. There is some evidence that this may have been the case in the Johnson and Raghu studies, in that their subjects were younger than the subjects in this study (40% and 44% less than 55 yr of age respectively as opposed to 8% less than 55 yr of age in this study). Because subjects with NSIP tend to be younger (two-thirds are younger than 55 yr) (27) than subjects with UIP (5% of our Mayo Clinic Rochester patients seen in 1996 were younger than 55 yr), we advise OLB and often a trial of corticosteroid therapy for subjects less than 55 yr of age. Whether both OLB and HRCT need to be done in all patients with idiopathic interstitial pneumonias is controversial (28), but increasingly, HRCT of the lungs has been recognized as an alternative to OLB when the clinical features and HRCT findings are typical for idiopathic UIP (29). During 1996 at Mayo Clinic Rochester, only 16% of patients diagnosed with idiopathic UIP had undergone OLB, whereas 91% had HRCT. We believe that it is important to define the outcome and response to therapy of the majority of patients with this disorder, most of whom are now diagnosed by HRCT rather than limiting study to the minority who have been diagnosed by OLB criteria. HRCT is associated with lower morbidity, mortality, and cost compared with OLB in patients with IPF. Not only does the initial computerized tomography (CT) pattern offer important diagnostic and prognostic information in IPF (30), but CT also allows for the possibility of follow-up CT with comparison to assess the progression of disease over time (31). Subjects with ground-glass patterns on their HRCT have a better prognosis than do those with irregular linear opacityreticular-honeycomb patterns; in one study (30), those with pure ground-glass patterns had benign outcomes, whereas those with reticular or mixed patterns had outcomes similar to the outcomes reported for idiopathic UIP. OLB in patients with predominantly ground-glass opacities may show a variety of disorders including DIP and NSIP, most of which are corticosteroid-responsive. We thus favor OLB and a trial of corticosteroid therapy for most subjects with predominantly groundglass patterns on HRCT. Those with HRCT patterns showing the characteristic UIP pattern of subpleural honeycombing involving both extreme lung bases usually have a UIP pattern of histopathology at OLB. However, chronic hypersensitivity pneumonia (CHSP) may stimulate the subpleural rim of honeycombing typical for UIP. In Lynch s study (21), the accuracy of HRCT to discriminate between UIP and CHSP was only 90% even when the radiologist made a CT diagnosis with a high degree of confidence. Similar HRCT patterns also can be seen infrequently late in the course of DIP and NSIP (27, 31, 32) and after acute lung injury. Our review of subjects with NSIP suggests that the HRCT pattern seems to predict outcome and corticosteroid responsiveness better than does the OLB pattern (27); those with OLB showing a NSIP pattern but HRCT showing a UIP pattern tended to respond poorly to prednisone, and to develop progressive disease. Mortality rates were similar for subjects in this study diagnosed using mostly clinical and HRCT criteria without OLB compared with an historical group of idiopathic UIP subjects diagnosed by OLB (6) without HRCT (Figure 2). This finding supports our contention that rigorous clinical and HRCT criteria can serve as a surrogate for OLB in distinguishing idiopathic UIP from other more corticosteroid-responsive idiopathic interstitial lung disorders. Subjectively, a few patients on prednisone therapy in this study noted symptomatic improvement in cough and dyspnea during the first few weeks, but then experienced a relapse of symptoms as prednisone was tapered. By contrast, colchicinetreated patients usually reported no change in symptoms other than incurring diarrhea. The frequency of objective significant improvement in the pulmonary function of patients with idiopathic UIP to corticosteroids is not known, but is probably less than 20%; in our retrospective study (16) it was 1 of 22 (5%); in the present study, it was 0/12. Our findings

Douglas, Ryu, Swensen, et al.: Colchicine and Pulmonary Fibrosis 225 suggest that UIP as we have defined it is usually a steroid refractory disorder for which a trial of high-dose corticosteroid therapy is unlikely to be beneficial and may do more harm than good. In conclusion, in most subjects with typical features of idiopathic UIP, neither prednisone nor colchicine resulted in objective improvement or stopped the progression of the disease. Prednisone therapy was associated with a higher incidence of serious side effects than was colchicine. Both drugs were relatively ineffective as antifibrotic agents and may not be different than no therapy. Colchicine should be considered to be an acceptable regimen against which to compare newer drugs with antifibrotic potential. Acknowledgment : The authors thank Dr. H. D. Tazelaar for reviewing the open-lung biopsy and for offering helpful critical comments. References 1. Coultas, D. B., R. E. Zumwalt, W. C. Black, and R. E. Sobonya. 1994. The epidemiology of interstitial lung diseases. Am. J. Respir. Crit. Care Med. 150:967 972. 2. Carrington, C. B., E. A. Gaensler, R. E. Coutu, M. X. FitzGerald, and R. G. Gupta. 1978. Natural history and treated course of usual and desquamative interstitial pneumonia. N. Engl. J. Med. 298:801 909. 3. Turner-Warwick, M., B. Burrows, and A. Johnson. 1980. Cryptogenic fibrosing alveolitis: clinical features and their influence on survival. Thorax 35:171 180. 4. Katzenstein, A. A., and J. L. Myers. 1998. State of the art: idiopathic pulmonary fibrosis: clinical relevance of pathologic classification. Am. J. Respir. Crit. Care Med. 157:1301 1315. 5. Katzenstein, A. A., and R. F. Fiorelli. 1994. Nonspecific interstitial pneumonia/fibrosis: histologic features and clinical significance. Am. J. Surg. Pathol. 18:136 147. 6. Bjoraker, J. A., J. H. Ryu, M. K. Edwin, J. L. Myers, H. D. Tazelaar, D. R. Schroeder, and K. P. Offord. 1998. Prognostic significance of histopathologic subsets in idiopathic pulmonary fibrosis. Am. J. Respir. Crit. Care Med. 157:199 203. 7. Panos, R. J., and T. E. King, Jr. 1991. Idiopathic pulmonary fibrosis. In J. P. Lynch and R. A. DeRemee, editors. Immunologically Mediated Pulmonary Diseases. J. P. Lippincott, Boston. 1 39. 8. Turner-Warwick, M., B. Burrows, and A. Johnson. 1980. Cryptogenic fibrosing alveolitis: response to corticosteroid treatment and its effect on survival. Thorax 37:593 599. 9. Johnson, M. A., S. Kwan, B. J. Snell, A. J. Nunn, J. H. Darbyshire, and M. Turner-Warwick. 1989. Randomized controlled trial comparing prednisolone alone with cyclophosphamide and low dose prednisolone in combination in cryptogenic fibrosing alveolitis. Thorax 44:280 288. 10. Raghu, G., W. J. Depaso, K. Cain, S. P. Hammar, C. E. Wetzel, D. F. Dreis, J. Hutchinson, N. E. Pardee, and R. H. Winterbauer. 1991. Azathioprine combined with prednisone in the treatment of idiopathic pulmonary fibrosis: a prospective double-blind, randomized, placebo-controlled clinical trial. Am. Rev. Respir. Dis. 144:291 296. 11. Hampton, J., F. Martinez, J. Orens, G. Toews, and J. P. Lynch, III. 1994. Corticosteroids in idiopathic pulmonary fibrosis (IPF): toxicity may outweigh benefits (abstract). Am. Rev. Respir. Dis. 149:A878. 12. Malik, S. W., J. L. Myers, R. A. DeRemee, and U. Specks. 1996. Lung toxicity associated with cyclophosphamide use; two distinct patterns. Am. J. Respir. Crit. Care Med. 154:1851 1856. 13. Lynch, J. P., III, and W. J. McCune. 1997. Immunosuppressive and cytotoxic pharmacotherapy for pulmonary disorders. Am. J. Respir. Crit. Care Med. 155:395 420. 14. DeRemee, R. A. 1990. Clinical Profiles of Diffuse Interstitial Pulmonary Disease. Futura, Mount Kisco, NY. 152. 15. Peters, S. G., J. C. McDougall, W. W. Douglas, D. T. Coles, and R. A. DeRemee. 1993. Colchicine in the treatment of pulmonary fibrosis. Chest 103:101 104. 16. Douglas, W. W., J. H. Ryu, J. A. Bjoraker, D. R. Schroeder, J. L. Myers, H. D. Tazelaar, S. J. Swensen, P. D. Scanlon, S. G. Peters, and R. A. DeRemee. 1997. Colchicine versus prednisone as treatment of usual interstitial pneumonia. Mayo Clinic Proc. 72:201 209. 17. Müller, N. L., R. R. Miller, W. R. Webb, K. G. Evans, and D. N. Ostrow. 1986. Fibrosing alveolitis: CT pathologic correlation. Radiology 160: 585 588. 18. Strickland, B., and N. H. Strickland. 1988. The value of high definition, narrow section computed tomography in fibrosing alveolitis. Clin. Radiol. 39:589 594. 19. Tung, K. T., A. U. Wells, M. B. Rubens, J. M. E. Kirk, R. M. dubois, and D. M. Hansell. 1993. Accuracy of the typical computed tomographic appearances of fibrosing alveolitis. Thorax 48:334 338. 20. Remy-Jardin, M., F. Giraud, J. Remy, M. C. Copin, B. Gosselin, and A. Duhamel. 1993. Importance of ground-glass attenuation in chronic diffuse infiltrative lung disease: pathologic CT correlation. Radiology 189:693 698. 21. Lynch, D. A., J. D. Newell, P. M. Logan, T. E. King, Jr., and N. L. Müller. 1995. Can CT distinguish hypersensitivity pneumonitis from idiopathic pulmonary fibrosis? A.J.R. 165:807 811. 22. Katzenstein, A. L., and F. B. Askin. 1982. Idiopathic interstitial pneumonia/idiopathic pulmonary fibrosis. In A. L. Katzenstein and F. B. Askin, editors. Surgical Pathology of Non-neoplastic Lung Disease. W. B. Saunders, Philadelphia. 58 96. 23. American Thoracic Society. 1991. Lung function testing: selection of reference values and interpretive strategies. Am. Rev. Respir. Dis. 144: 1202 1218. 24. American Thoracic Society. 1987. Single breath carbon monoxide diffusing capacity (transfer factor): recommendations for a standard technique. Am. Rev. Respir. Dis. 136:1299 1307. 25. Miller, A., J. C. Thornton, R. Warshaw, J. Anderson, A. S. Tierstein, and I. J. Selikoff. 1983. Single breath diffusing capacity in a representative sample of the population of Michigan, a large industrial state. Am. Rev. Respir. Dis. 127:270 277. 26. Miller, A., J. C. Thornton, R. Warshaw, J. Bernstein, I. J. Selikoff, and A. S. Tierstein. 1986. Mean and instantaneous expiratory flows, FVC and FEV 1 : prediction equations from a probability sample of Michigan, a large industrial state. Bull. Eur. Physiopathol. Respir. 22:589 597. 27. Midthun, D. E., J. H. Ryu, J. L. Myers, H. D. Tazelaar, T. E. Hartman, and W. W. Douglas. 1998. Nonspecific interstitial pneumonia: clinical, radiographic, and pathologic features (abstract). Am. J. Respir. Crit. Care Med. 157:A277. 28. Raghu, G. 1995. Interstitial lung disease: a diagnostic approach. Am. J. Respir. Crit. Care Med. 151:909 914. 29. Hansell, D. M., and A. U. Wells. 1996. State of the art: CT evaluation of fibrosing alveolitis applications and insights. J. Thorac. Imaging 11: 231 249. 30. Wells, A. U., D. M. Hansell, M. B. Rubens, P. Cullinan, C. M. Black, and R. M. dubois. 1993. The predictive value of appearances on thin-section computed tomography in fibrosing alveolitis. Am. Rev. Respir. Dis. 148:1076 1082. 31. Hartman, T. E., S. L. Primack, E. Y. Kang, S. J. Swensen, D. M. Hansell, G. McGuinness, and N. L. Müller. 1996. Disease progression in usual interstitial pneumonia compared with desquamative interstitial pneumonia: assessment with serial CT. Chest 110:378 382. 32. Hartman, T. E., S. L. Primack, S. J. Swensen, D. Hansell, G. McGuinness, and N. L. Müller. 1993. Desquamative interstitial pneumonia: thin Section CT findings in 22 patients. Radiology 187:787 790.