Published ahead of print on May 1, 2008, doi:10.1164/rccm.200709-1365OC
© 2008 American Thoracic Society doi: 10.1164/rccm.200709-1365OC
Computed Tomography Findings in Acute Exacerbation of Idiopathic Pulmonary FibrosisDepartments of 1 Radiology, 2 Pathology, and 3 Internal Medicine, National Hospital Organization, Kinki-Chuo Chest Medical Center, Osaka, Japan Correspondence and requests for reprints should be addressed to Masanori Akira, M.D., Department of Radiology, National Hospital Organization, Kinki-Chuo Chest Medical Center, 1180 Nagasone-cho, Kita-ku, Sakai City, Osaka 591-8555, Japan. E-mail: Akira{at}kch.hosp.go.jp
Rationale: The serial computed tomography findings and prognosis of the acute exacerbation of idiopathic pulmonary fibrosis (IPF) are not yet well defined in a larger number of cases. Objectives: To evaluate the parenchymal abnormalities and prognosis using high-resolution computed tomography (HRCT) in acute exacerbation of IPF. Methods: The study consisted of clinical, laboratory, and HRCT data before and at the time of acute exacerbation in 64 episodes of 58 patients with IPF. A semiquantitative analysis of overall extent of parenchymal abnormalities, extent of alveolar opacity (ground-glass attenuation and consolidation), and extent of fibrotic opacity (reticulation and honeycombing) on CT was performed by two chest radiologists. The newly appeared parenchymal abnormalities were also classified into three patterns: peripheral, multifocal, and diffuse. Measurements and Main Results: In all patients, HRCT scans taken at the exacerbation showed typical signs of IPF and newly developing alveolar opacity. They included 34 patients of peripheral pattern, 8 of multifocal pattern, and 16 of diffuse pattern. Twenty-five patients died and 33 survived after the initial exacerbation. Worse survival was associated with patients with diffuse type compared with patients with multifocal and peripheral type. The CT patterns and overall CT extent were associated with an increased hazard of death after adjusting for age, sex, smoking, baseline diffusion capacity for carbon monoxide, baseline FVC, and disease extent on CT. On multivariate analysis, the strongest correlations were observed between CT patterns (combined diffuse and multifocal versus peripheral) and survival (odds ratio, 4.629; 95% confidence interval, 1.900–11.278; P = 0.001). Conclusions: HRCT extent and patterns are predictive of survival in acute exacerbation of IPF.
Key Words: lung disease, interstitial • computed tomography, high-resolution • idiopathic pulmonary fibrosis • acute exacerbation • acute lung injury
Idiopathic pulmonary fibrosis (IPF) is a progressive and lethal fibrotic lung disease of unknown etiology. In a consensus statement of the American Thoracic Society (ATS) and the European Respiratory Society (ERS), IPF is defined as a distinctive type of chronic fibrosing interstitial pneumonia of unknown cause limited to the lungs and associated with a histologic pattern of usual interstitial pneumonia (UIP) (1). Typical high-resolution computed tomography (HRCT) signs of IPF are basal and peripheral reticular abnormalities, associated with minimal ground-glass opacities (1, 2). The natural history is invariably one of gradual and progressive deterioration, with median length of survival from the time of diagnosis ranging from 2 to 3 years (3). Although chronic in nature, an accelerated phase may occur at any stage in the history of the disease without identifiable cause, leading to death in a period of a few weeks to a few months (4–11). These episodes are called "acute exacerbations" of IPF. Acute exacerbations of IPF are increasingly recognized as common and highly morbid clinical events. We have previously described the HRCT features of acute exacerbation of IPF in 17 patients (6). In these 17 patients, all 5 patients with diffuse type had died, all 6 patients with peripheral type had survived, and half of 6 patients with multifocal type had survived. HRCT findings on 17 patients were descriptively reported. The clinical profile, laboratory data, computed tomography findings, and prognosis are not yet well defined in a larger number of cases. The purpose of this study was to confirm the predictive value of HRCT findings in patients with acute exacerbation of IPF.
Patients The present retrospective study included patients with pathologically proven idiopathic pulmonary fibrosis (IPF) who were admitted to our hospital for acute exacerbation on their follow-up between January 1994 and March 2006. Patients had no known causes of interstitial lung disease, such as collagen vascular disease, drug toxicity, and environmental exposure. None of the subjects were at risk of acute respiratory distress syndrome. Sixteen subjects had ever received corticosteroid treatment before acute exacerbation. Four of these 16 patients received dosage reduction of corticosteroid immediately before acute exacerbation. Six patients were on home oxygen (O2) therapy. Two patients underwent open lung biopsy and one underwent lung resections for lung cancer within 2 weeks before the exacerbation. Six of 58 patients had two episodes of acute exacerbation. Eleven of the 17 patients described in the previous article (6) were included in this study. Four patients were excluded from the current study because of cases without pathologically proven usual interstitial pneumonia, and two were excluded because of lost of clinical and radiologic data. The acute exacerbation of IPF was an increase in symptoms beyond normal day-to-day variation requiring a change in medication and hospital admission and was diagnosed based on the following criteria (4–6): previous or concurrent diagnosis of IPF; unexplained worsening or development of dyspnea within 30 days; newly developing bilateral ground-glass abnormality and/or consolidation superimposed on a background reticular or honeycomb pattern consistent with UIP pattern on chest HRCT; absence of apparent infectious agents and heart failure; and no cause other than IPF for the radiologic progression. Echocardiography demonstrated no evidence of heart failure in all patients. Cultures of sputum, blood, urine, and bronchoalveolar lavage fluid (BALF) examined for mycobacteria, fungi, and bacteria were negative in all patients. All serologic studies for viruses, Chlamydia, mycoplasma, and legionella were negative. In 29 patients, bronchoalveolar lavage (BAL) at acute exacerbation was performed. Cultures of BALF for various viruses (cytomegalovirus, herpes simplex, varicella-zoster virus, measles virus, adenovirus, influenza A, influenza B, parainfluenza 1–3, and respiratory syncytial virus) were performed. In BAL samples additional stains were used; Ziehl-Neelsen staining for mycobacteria; Gomori methenamine silver stain for Pneumocystis. In the remaining 29 patients, the biopsies were taken during acute exacerbation in 6 patients and autopsy was performed in 23 patients. Although there was no evidence of infection, 31 patients were treated with broad-spectrum antibiotics without any improvement. No antibiotic therapy was given before obtaining cultures. Pathological samples were investigated using monoclonal antibodies against herpes virus type 1, cytomegalovirus, and with specials stains for fungi, Pneumocystis (Gomori methenamine silver stain), and acid-fast bacilli (Ziehl Neelsen stain). All lung biopsy specimens and autopsy slides were reviewed by three pathologists, according to the 2000 American Thoracic Society/European Respiratory Society consensus classification (1). The pathologic diagnosis of the presence or absence of UIP was based on agreement of two or more pathologists. There was no histologic evidence of infection. All patients underwent HRCT scan of the chest within 6 months before and at the onset of acute exacerbation. Clinical data, laboratory findings, and pulmonary function test within 6 months before acute exacerbation and at the acute exacerbation were extracted from case records. The CT scans in acute exacerbation were obtained at the day of hospitalization. They were obtained within 1 day to 28 days after the onset of symptoms (mean ± SD, 9 ± 7 d). All patients were treated with high doses of methylprednisolone (1,000 mg · day–1 for 3 d) for acute exacerbation. Corticosteroid therapy was followed by a tapered dosage and was combined with intravenous cyclophosphamide, with the choice dependent on the patient's pulmonary physiologic response. In 44 of the 58 patients, multiple CT scans were obtained for each patient after the exacerbation. Second CT scans were obtained 3 to 18 days (median, 9 d) after initial CT of acute exacerbation. This retrospective study had Institutional Review Board approval; informed consent was not required.
CT Examination After reviewing the CT images independently in random order, all CT scans for each patient were also examined in sequence to assess sequential changes over time. The pattern, extent, and distribution of abnormal CT findings on follow-up CT studies were evaluated and compared with findings in the same region on initial CT scans. The newly developing pulmonary opacities were classified into peripheral, multifocal, and diffuse pattern according to the classification of Akira and coworkers (Figure 1) (6). In peripheral pattern, parenchymal opacification appeared in the inner peripheral zone adjacent to preexisting subpleural honeycombing or was increased in areas with preexisting peripheral interstitial opacity (Figure 2). In multifocal pattern, parenchymal opacification was apparent in central and peripheral regions; however, the multifocal abnormalities were at a few sites and limited in extent (Figure 3). Diffuse pattern had generalized pulmonary involvement but regional inhomogeneity (Figure 4).
Statistical Analysis Data are expressed as mean ± SD. Data analyses were performed using a commercially available software package (SPSS, version 12.0J; SPSS, Chicago, IL). Comparison between data before and those at the onset of the acute exacerbation were made using paired t tests. Comparison between survivors and nonsurvivors were made using Mann-Whitney tests or chi-squared statistics as appropriate. For all statistical analyses, P < 0.05 was considered to indicate a significant difference. To compare among CT patterns, survival curves were estimated using the Kaplan-Meier approach; survival curves were compared using the log-rank test. Cox proportional hazards regression was used to examine the relationship between CT patterns and mortality, adjusting for demographic characteristics (age, sex, and smoking status), physiologic data (FVC, DLCO), and disease extent on CT.
Clinical, pulmonary function, and HRCT scan data before acute exacerbation are shown in Table 1. Laboratory and HRCT scan data before and at the time of acute exacerbation are shown in Table 2. The time between the diagnosis of IPF and the onset of the exacerbation ranged from 1 to 65 months.
There was a significant statistical difference between values of white blood cells (WBC), C-reactive protein (CRP), lactate dehydrogenase (LDH), and serum albumin before the exacerbation and those at the exacerbation. Forty-four patients presented an increase of WBC on the exacerbation. Elevated values CRP and LDH were found in 55 and 43 cases, respectively. Hypoalbuminemia was found in 48 patients. In all cases typical signs of IPF/UIP (reticular pattern more prominent at the lungs bases, traction bronchiectasis, and areas of honeycombing) were found on the CT scans before the exacerbation. The disease extent of parenchymal abnormalities revealed by CT scans taken within 6 months of acute exacerbation ranged from 10 to 47% of the lung parenchyma (mean ± SD, 24 ± 9%). The disease extent of parenchymal abnormalities revealed by CT scans taken at the acute exacerbation ranged from 15 to 83% of the lung parenchyma (mean ± SD, 47 ± 16%). The parenchymal abnormalities could be classified into three patterns: peripheral (n = 34), multifocal (n = 8), and diffuse (n = 16). The overall extent of parenchymal abnormalities and the extent of ground-glass attenuation were significantly higher in diffuse type than in both multifocal and peripheral type. The CT scans in patients of multifocal pattern were obtained within 1 to 5 days (median, 2 d) after the onset of symptoms. The CT scans in patients with diffuse pattern were obtained within 3 to 24 days (median, 11 d) after the onset of symptoms. The CT scans of peripheral patterns were obtained within 2 to 28 days (median, 7 d) after the onset of symptoms. The CT scans of the multifocal patterns were obtained in the earlier phase of acute exacerbation than those of the diffuse patterns (P = 0.022). There was no significant difference between patients with peripheral pattern and those with multifocal pattern or diffuse pattern in the time obtained CT scans after the onset of symptoms.
Outcome Follow-up CT scans were available in 8 patients with diffuse type, 8 patients with multifocal type, and 28 patients with peripheral type. In survivors with peripheral type, much of the alveolar opacities (ground-glass opacity and consolidation) regressed to return the preexist parenchymal abnormality on the CT before acute exacerbation, although fibrotic opacities (reticulation and honeycombing) had no change. Consolidation had changed to ground-glass opacity and then disappeared in 1 to 3 months. In five survivors of multifocal type, areas of ground-glass attenuation and consolidation disappeared with corticosteroid therapy. On follow-up CT scans in three nonsurvivors with multifocal type, new areas of ground-glass attenuation and consolidation occurred in previously normal-appearing lung, and areas of ground-glass attenuation and consolidation showed marked extension for a short period and evolved into diffuse pattern (Figure 5). In diffuse type, follow-up CT scans showed marked extension and nearly complete opacification of both lungs. On follow-up CT scans, some areas of ground-glass attenuation had changed to consolidation. Some consolidation became more compact and were accompanied by traction bronchiectasis.
In 6 of 33 patients who recovered, acute exacerbation of diffuse pattern occurred at 6 to 20 months after recovery from the initial acute exacerbation of peripheral pattern. These six patients died of respiratory failure after the second exacerbation. In these six patients, new diffuse alveolar opacities were appeared after the peripheral alveolar opacities had resolved (Figure 6).
As shown in Table 3, age, sex, duration of disease, laboratory data at the exacerbation, baseline DLCO, and baseline FVC did not differ between survivors and nonsurvivors. There was no significant difference between extent of reticulation, honeycombing, and consolidation on CT at the acute exacerbation in survivors and those in nonsurvivors. The overall extent of parenchymal abnormalities and extent of ground-glass attenuation on CT at the acute exacerbation were significantly higher in nonsurvivors than in survivors.
The median survival was 16 days for patients with diffuse type, 240 days for patients with multifocal type, and 540 days for patients with peripheral type. The median follow-up time was 420 days for the patients with peripheral pattern. Kaplan-Meier curves were constructed to illustrate differences in survival based on three CT patterns (Figure 7). There was statistically significant difference in survival from the time of acute exacerbation among CT patterns. Patients with diffuse pattern on CT had a higher risk of mortality than those with multifocal or peripheral pattern.
Because multifocal type was considered to be early phase of diffuse type and progressed to diffuse type, patients with multifocal type and diffuse type have been combined. By univariate analysis, CT patterns, higher CT extent, higher CT alveolar opacity, and higher LDH were found to be significant predictors of mortality (Table 4). Importantly, CT patterns, overall extent on CT, and LDH remained a significant predictor of mortality after adjusting for age, sex, smoking history, baseline DLCOo percent predicted, and baseline FVC percent predicted (Table 5).
At autopsy of 23 patients, the histologic findings showed the aspects of UIP with superimposed features of diffuse alveolar damage. These 23 patients included 20 patients of diffuse pattern and 3 patients in multifocal pattern on HRCT. The biopsies taken during acute exacerbation in six patients showed the aspects of UIP with superimposed features of acute lung injury. The acute lung injury included diffuse alveolar damage in one patient, numerous fibroblastic foci in three, and organizing pneumonia in two patients. These six patients had peripheral pattern on HRCT.
IPF usually presents insidiously with symptoms and signs of progressive respiratory insufficiency that slowly progress over a period of months to years. In addition to a chronic course, an accelerated phase may occur at any stage in the history of the disease (4–11). Acute deterioration in IPF may occur secondary to infections, pulmonary embolism, pneumothorax, or heart failure. Often, however, there is no identifiable cause for the acute decline, and these episodes are called acute exacerbation of IPF (4–11). The histologic findings from lung biopsy specimens show variable aspects; the typical usual interstitial pneumonia pattern is associated with signs of acute lesions, such as diffuse alveolar damage (DAD) with or without hyaline membranes, numerous fibroblastic foci, organizing pneumonia (OP), and hemorrhage with capillaritis (5–8). Churg and colleagues (13) described that three microscopic patterns of acute lung injury were seen in acute exacerbation of UIP—DAD, OP, and a pattern of numerous very large fibroblastic foci superimposed on underlying fibrosis—and that patients with OP or extensive fibroblastic foci as the acute pattern seem to do better than those with DAD. Martinez and colleagues retrospectively analyzed the course of 168 patients with IPF (9). Over a median period of 76 weeks, 21% of these patients died, and 47% of these deaths followed an acute deterioration in the patient's clinical status. Kim and coworkers described that among 147 patients with biopsy-proven IPF the 1-year frequency of acute exacerbation was 8.5% and the 2-year frequency was 9.6% (11). Elevated values of white blood cells and C-reactive protein in serum were found in most patients at the onset of the acute exacerbation. Lactate dehydrogenase (LDH) values were also increased in most patients. These findings are consistent with those of other reports (5, 7). Similar findings were also found in acute interstitial pneumonia (AIP) (14, 15). In nonspecific blood tests, LDH values were predictive of survival time after adjustment for age, sex, smoking history, and baseline FVC. We have previously described the HRCT features of acute exacerbation of IPF in 17 patients (6). In these 17 patients, all patients with diffuse type had died, all patients with peripheral type had survived, and half of patients with multifocal type had survived. In the present study, we found that the disease extent on HRCT has considerable prognostic value in survivors and nonsurvivors. In acute respiratory distress syndrome, lower extension of diffuse alveolar damage areas is associated with a better prognosis (7). Patients presenting a focal lung involvement improve, whereas those with a massive lung injury die, with an overall survival rate of approximately 50% (7). We also found that the CT patterns were the most discriminative feature for separating nonsurvivors from survivors. Survival was worse for patients with diffuse pattern than for those with both peripheral and multifocal pattern. CT scans of multifocal pattern were obtained in the earlier phase of acute exacerbation compared with those of diffuse pattern and multifocal pattern evolved into diffuse pattern for a short period. Multifocal pattern is considered to be the early phase of diffuse pattern. Peripheral pattern does not evolve into diffuse pattern for a short period. In the acute exacerbation of IPF, two basic CT patterns may be classified: new areas of parenchymal opacification mainly within the peripheral region with relatively limited damage, and new areas of parenchymal opacification that spread rapidly throughout the lung with a fulminant clinical course. On the basis of CT-pathologic correlation in our cases, diffuse pattern corresponded to DAD, whereas peripheral pattern mainly correlated with OP or numerous fibroblastic foci. The CT extent and CT patterns had a higher predictive value regarding patient survival than did the clinical and laboratory data. When the CT extent is less severe in the earlier phase of acute exacerbation, CT patterns may be useful for predicting patient prognosis in acute exacerbation of IPF. The absence of significant ground-glass opacity is an important feature of definite UIP pattern (3, 16). The presence of extensive ground-glass opacities in patients with UIP or the development of rapidly progressive new ground-glass opacity away from areas of fibrosis should raise the possibility of acute exacerbation. Also, acute exacerbation of IPF and NSIP can look similar on HRCT in the early phase of IPF. All cases in our study developed acute on chronic respiratory failure in the absence of any precipitating factor. In our subjects, however, 3 of the 54 patients developed acute respiratory failure within 2 weeks of open lung biopsy (n = 2) and lung resection surgery (n = 1). Utz and colleagues (17) reported that 10 of 46 patients with IPF/UIP died within 30 days of surgical biopsy. Saydain and coworkers (18) described one patient who developed acute on chronic respiratory failure after surgery for mitral valve replacement. Kumar and colleagues (19) indicated that acute respiratory distress syndrome occurred in four of 22 patients with lung fibrosis undergoing resection for cancer. Acute exacerbation may repeat in some patients with UIP. In addition, acute exacerbation may be the presenting manifestation in some patients with UIP (8). An accelerated variant of UIP in two previously healthy patients with no known interstitial lung disease is reported (20). The accelerated variant of UIP is associated with evidence of peripheral ground-glass opacities and consolidation. Subpleural reticulation or honeycombing is not been demonstrable on the initial CT examination. Traction bronchiectasis and cysts were not seen on CT in the two patients until 37 days after presentation. Pathological findings in these two cases were consistent with a diagnosis of UIP (20). Accelerating variant of UIP had no a chronic course of follow-up and no reticular or honeycomb pattern consistent with IPF on the initial chest HRCT scans, but might have subclinical UIP. Accelerating variant of UIP may be an acute exacerbation in patients with subclinical UIP. Our study has several limitations. This study was performed retrospectively. Not all the patients with acute exacerbation were included in this study because the study is limited to patients with proven histologic diagnosis. Only half of our patients underwent BAL, and the remaining half had no histologic evidence of infection in lung biopsy or autopsy. Infection is the most important factor in differential diagnosis of acute exacerbation. We were not able to obtain BAL samples from patients with more severe disease. We had performed HRCT scans at the interval of 6 months in patients with IPF during a certain period. So we obtained HRCT scans within 6 months before the onset of acute exacerbation in all subjects, although they did not represent more rapidly progressive underlying disease.
Originally Published in Press as DOI: 10.1164/rccm.200709-1365OC on May 1, 2008 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form September 14, 2007; accepted in final form May 1, 2008
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
