Published ahead of print on February 15, 2007, doi:10.1164/rccm.200603-343OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200603-343OC
Computed Tomography Reflects Lower Airway Inflammation and Tracks Changes in Early Cystic Fibrosis1 Division of Pulmonology, Department of Pediatrics, and 2 Department of Radiology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina; 3 Department of Radiology, Cincinnati Children's Hospital, Cincinnati, Ohio; and 4 Department of Biostatistics, School of Public Health, University of North Carolina at Chapel Hill Correspondence and requests for reprints should be addressed to Stephanie D. Davis, M.D., Division of Pediatric Pulmonology, 130 Mason Farm Road, Campus Box 7220, Chapel Hill, NC 27599-7220. E-mail: sddavis{at}med.unc.edu
Rationale: Detecting and tracking early cystic fibrosis (CF) lung disease are difficult due to lack of sensitive markers of airway dysfunction. Objectives: The goals were to detect regional distribution of airway disease through high-resolution computed tomography, correlate abnormalities to lower airway inflammation/infection, and compare computed tomography findings before and after intravenous antibiotic therapy in children with CF younger than 4 years experiencing a pulmonary exacerbation. Methods: High-resolution computed tomography was performed in 17 children scheduled for bronchoscopy. The radiologist identified the lobes with the "greatest" and "least" disease based on computed tomography, and bronchoalveolar lavage was performed in these areas. In 13 subjects, imaging was repeated after antibiotic completion. Modified Brody scores were assigned by two radiologists. Measurements and Main Results: The lobe with greatest disease was predominantly localized to the right and had higher modified Brody scores, indicating more severe abnormalities (p < 0.01), compared with the lobe with least disease. The total modified Brody score (p < 0.01), hyperinflation subscore (p < 0.01), and bronchial dilatation/bronchiectasis subscore (p < 0.01) improved after antibiotics and intensified airway clearance. Interleukin-8 levels (p < 0.01) and % neutrophils (p = 0.04) were increased in the lobe with greatest disease compared with the lobe with least disease. Conclusions: These results indicate that, in young children with CF experiencing a pulmonary exacerbation, computed tomography detects regional differences in airway inflammation, may be a sensitive outcome to evaluate therapeutic interventions, and identifies early lung disease as being more prominent on the right.
Key Words: infant • child • computed tomography scanners, X-ray • bronchoalveolar lavage • bronchopneumonia
Lung disease in cystic fibrosis (CF) often begins in infancy, before significant respiratory symptoms occur (1, 2). Detection of early CF lung disease is essential for understanding disease progression and providing endpoints for clinical trials, but it is difficult due to lack of sensitive, noninvasive outcome measures. Chest X-rays (CXRs) typically show minimal, nonspecific abnormalities, are less sensitive than high-resolution computed tomography (HRCT) for detection of CF lung disease (3), and are relatively insensitive during pulmonary exacerbations (4). Bronchoalveolar lavage (BAL) studies and HRCT have the potential to detect early lung disease (5). Studies on BAL fluid (BALF) in infants with CF have identified the presence of infection and inflammation (1, 2, 6–10); however, BALF can only be obtained from a small portion of lung and is subject to sampling error. Although HRCT has been performed in older children and adults with CF (3, 11–24), only recently has this imaging technique in the young CF population (< 4 yr) been studied (25–28). A technique described by Long and colleagues (25–27) allows the young child's ventilation to be controlled noninvasively, minimizing motion artifact and improving details of airway and parenchymal structures. Using controlled ventilation HRCT, Long and colleagues (26) and Martinez and associates (28) demonstrated that infants with CF have thicker airway walls compared with control subjects. Air trapping was present in middle and lower lobes, but not in upper lobes (28). A challenge in evaluating HRCT has been developing a scoring system to quantify abnormalities in mild CF lung disease. Brody and colleagues (23) developed a scoring system applicable to older children with mild disease. Investigators have demonstrated HRCT score improvement after therapy for pulmonary exacerbation in subjects between the ages of 5 and 43 years (12, 23); however, no study has documented that these scoring systems are applicable to the young CF population. Furthermore, no published studies have performed HRCT in the young CF population as a means of identifying disease pattern or severity during pulmonary exacerbation or determined if these HRCT changes due to inflammation/infection reverse with therapy. Using HRCT to direct BAL sites has potential to detect early, localized lung pathology and to compare these findings with the presence of inflammation and infection associated with CF lung disease. To address some of these questions, we undertook a study of HRCT in young children with CF with a pulmonary exacerbation. Our primary hypotheses were as follows: (1) HRCT will identify regional distribution of disease during a pulmonary exacerbation, (2) BALF markers of inflammation and infection will be increased in areas with the "greatest" amount of disease identified by HRCT, and (3) HRCT will detect improvement of airway disease after intravenous antibiotic and intensified airway clearance therapy. Preliminary results of this study were presented as abstracts (29, 30).
Additional detail on the method for making these measurements is provided in an online supplement.
Subjects
Study Protocol
Controlled Ventilation HRCT Protocol
Image Analyses
Statistics
Subjects Seventeen subjects were enrolled into the study (5 females, 12 males). All 17 underwent initial HRCT followed immediately by bronchoscopy. Two subjects did not receive a second HRCT due to sedation difficulties and two others did not receive a follow-up HRCT because there was no intravenous antibiotic therapy after the bronchoscopy. Thus, 17 subjects contributed data to the interlobar comparisons between the lobes with the greatest and least amount of disease, whereas only 13 subjects contributed data to comparisons before and after treatment. The subjects enrolled were between the ages of 2 and 44 months (mean age, 17 mo) with a mean weight of 9 kg (range, 4–19 kg). Half of the subjects were homozygous for the delta F508 genetic mutation.
Radiographic Findings
On the basis of HRCT scores (Table 2), the area identified as having the greatest disease was predominantly on the right side and the least disease was predominantly on the left side as also indicated by the qualitative assessment. However, the specific lobes identified as having the greatest or least disease sometimes differed between qualitative and quantitative assessments. In 10 of 17 subjects (59%), the lobes qualitatively chosen as the lobe with the greatest disease were also scored as the lobe with greatest disease, and in 4 of 17 subjects (24%), the lobes qualitatively chosen as the lobe with the least disease were scored as the lobe with the least disease. A lobe qualitatively chosen as having the least amount of disease was later quantitatively scored as being the lobe with the greatest amount of disease. This score was primarily attributable to marked hyperinflation in this lobe; however, peribronchial thickening and bronchial dilatation were the main parameters used to qualitatively choose lobes (see the online supplement).
Table 3 shows the HRCT scores assessed by the two HRCT readers. Due to mild motion artifact on one HRCT, reader 1 scored 16 scans instead of 17. As noted, there was a significant difference in scores between the lobe initially chosen as having the least disease and the lobe chosen as having the greatest disease for both readers. Figure 1 shows an example of a scan showing the difference between lobes. There was also a significant difference between the total right-sided lung score and the left-sided lung score, with the right side scoring higher. The mean total HRCT score was 18.18 (SD, 8.92; n = 16) and 12.18 (SD, 7.54; n = 13) for visits 1 and 2, respectively. The interobserver correlation between the two readers for the total HRCT score was 0.45 (p = 0.08) and 0.61 (p = 0.03) for visits 1 and 2, respectively. The total HRCT score (right- plus left-sided scores) significantly improved between visit 1 and visit 2. The mean hyperinflation subscore and the bronchiectasis/bronchial dilatation subscore significantly improved, indicating that intravenous antibiotics and intensified airway clearance improved these changes noted on HRCT (Table 3). Although not statistically significant, the other subscores also tended to decrease, thus including these five subscores for the total HRCT score led to a highly statistically significant improvement in the total score after antibiotic and intensified airway clearance therapy. The average HRCT score of the lobe qualitatively chosen as having the greatest abnormality significantly decreased between visit 1 and visit 2 (p = 0.002). The average HRCT score did not significantly change in the lobe qualitatively chosen as having the least amount of disease between visit 1 and visit 2 (Figure 2). Figure 3 shows an example of resolution of a parenchymal opacity after treatment. Figure 4 shows an example of improvement in air trapping after intravenous antibiotic therapy, and Figure 5 demonstrates a decrease in size of the airways relative to the size of the corresponding arteries, suggesting resolution of early-stage bronchiectasis/bronchial dilatation.
The CXR Brasfield score was greater than or equal to 20 in all subjects except for two and only showed a tendency to improve between visits (p = 0.06). Reader 1 (L.A.F.) attempted to choose a lobe with the greatest disease and a lobe with the least amount of disease based on the CXRs. For 6 of 17 (35%) and 11 of 17 (65%) subjects, reader 1 was unable to identify a lobe as having the greatest abnormality or least abnormality, respectively. Reader 1 chose a lobe with greatest or least abnormality in only three subjects and one subject, respectively, that corresponded to HRCT quantitative findings.
BALF Findings
Our results suggest that HRCT at the time of pulmonary exacerbation in very young subjects with CF detects regional differences in airway inflammation as reflected in BALF IL-8 levels and % neutrophils. The significant improvement in HRCT scores after antibiotic and intensified airway clearance therapy indicates that this tool may also be a sensitive outcome measure in the evaluation of therapeutic interventions in the young CF population. Qualitative and quantitative assessments of the HRCT showed that early CF lung disease was more prominent on the right side. On the basis of these findings, the modified Brody score (23) used in this study is applicable to young children with CF with evidence of mild disease. To our knowledge, no other published studies have compared HRCT findings to BALF results in this population. A recent study in a small number of infants with CF showed that airway structural changes measured through HRCT correlated with airway function measured through the raised volume technique (28). Recent reports have also demonstrated that HRCT is a more sensitive marker of early airway disease than spirometry in young children older than 5 years (21, 24) and in older children and adults, peripheral bronchiectasis on CT declines more rapidly than lung function parameters (32). Studies in older children and adults conclude that HRCT is most useful in patients with mild disease (19, 20). One of these studies showed that some children had substantial structural lung damage despite normal pulmonary function values (24). Our study subjects had significantly worse disease on the right compared with the left side as reported in some prior studies (16, 33). Furthermore, 70% of subjects had the least amount of disease localized to the left upper lobe and lingula. The inability to perform a detailed score before bronchoscopy is clearly a weakness of this study; however, the predominance of right-sided disease was demonstrated both qualitatively and quantitatively. The qualitative evaluation before bronchoscopy uses the same criteria as the quantitative scoring system without detailed measurements. CXRs were insensitive in detecting this early regional lung disease compared with HRCT findings. Maffessanti and coworkers (33) reported in 36 patients (mean age, 13 yr) that the upper lobes had more significant disease, especially on the right side. Santis and colleagues (16) also reported that the earliest abnormality on HRCT in 38 adult subjects with mild CF disease was located in the right upper lobe. Martinez and associates (28) reported that air trapping was more prominent in the middle and lower lobes in infants with CF; however, we did not find a regional difference in our hyperinflation subscores (data not shown). Our findings may differ from the study by Martinez and colleagues (28) due to methodologic differences. Two possible etiologies of the significant right-sided disease may be gastroesophageal reflux with aspiration or relatively more difficulty in clearing secretions from the right upper lobe. Infants with CF do have an increased incidence of gastroesophageal reflux disease (34–36) and positioning for airway clearance could promote gastroesophageal reflux/aspiration (37). On the basis of these findings, further studies are needed to better define the role of gastroesophageal reflux disease in CF. BALF inflammation as indicated by % neutrophils and IL-8 levels was significantly higher in the area identified as having the greatest disease compared with the area with the least disease. Investigators have reported regional differences in BALF of subjects with CF (38, 39); however, in contrast to our study, imaging was not used to identify diseased areas and the subjects with CF were at their clinical baseline. Our findings of regional variation are consistent with those reported by Meyer and colleagues (38) in 12 adult subjects with CF. This group reported regional variation in inflammatory markers by demonstrating that absolute numbers of neutrophils and neutrophil elastase levels were significantly higher in the right upper lobe compared with the right lower lobe. Gutierrez and colleagues (39) recently reported higher bacterial counts in the right middle lobe compared with the lingula in six of nine young children with CF with 10,000 cfu/ml or more of bacteria, but inflammatory indices were similar between the lobes. Similar to our findings, only two of these nine children had discordance in bacterial species between the two lobes and P. aeruginosa, S. aurerus, and Moraxella catarrhalis were common pathogens. In contrast to our findings, Dakin and colleagues (40) reported that overall HRCT scores did not correlate with sputum inflammatory markers in children with CF between the ages of 6 and 21 years. Our study suggests that regional variability in inflammation and disease may account for this apparent discrepancy, because sputum presumably is composed of a mixture of secretions from multiple sites within the lung. The regional differences reported in our study highlight the potential importance of performing BAL in more than one area in the young CF population. This finding also has important implications for bronchoscopy/BAL protocols used in multicenter clinical trials. There was a significant decrease in the total modified Brody score before and after intravenous antibiotic and intensified airway clearance therapy in these young children with CF. This finding has been reported in older subjects treated for a CF exacerbation with intravenous antibiotics (12, 23), but has never been demonstrated in the young CF population (< 4 yr old). HRCT has also been used as an outcome measure when assessing the efficacy of DNase (41, 42) in the CF population, and one study (41) reported that HRCT scores significantly improved after treatment with DNase in children younger than 5 years. However, these investigators stated that motion artifact was present in five of six subjects younger than 2.2 years. In our study, the use of controlled breathing during the HRCT diminished motion artifact, allowing more accurate scoring and interpretation. In addition, we demonstrated that the score of the lobe chosen as having the greatest amount of disease significantly improved after therapy, but not the score of the lobe chosen as having the least amount of disease. These findings suggest that reversible changes did occur in the area chosen as having the greatest disease; thus, BAL was performed in an area corresponding to an exacerbation. The lack of improvement after intravenous antibiotic therapy in the lobe having the least amount of disease may be due to the lack of disease in this area, the insensitivity of the scoring system in detecting subtle changes in this lobe chosen as having the least amount of disease, or, alternatively, the presence of irreversibility. Due to the risk of radiation, we chose to perform only three expiratory images at specific levels. This limitation could lead to difficulties matching the expiratory area scanned before and after therapy; however, regional air trapping and the bronchiectasis/bronchial dilatation score significantly improved after intravenous antibiotic and airway clearance therapy. Castile and coworkers (43) have demonstrated a significant elevation of the residual volume to total lung capacity ratio, a measure of air trapping, in infants with CF compared with control subjects. Because small airway disease due to mucous plugging is believed to be the earliest sign of lung involvement, the reversibility of air trapping after antibiotic therapy indicates improvement of peripheral airway inflammation and obstruction secondary to these plugs. Volumetric scanning was not used in this study; therefore, evaluating the image at the exact location before and after therapeutic intervention was a limitation of this study. The reversibility of the bronchiectasis/bronchial dilatation score in this young age group is encouraging and suggests that aggressive therapy may delay or potentially reverse early evidence of airway damage before irreversible bronchiectasis is established. Brasfield scores did not demonstrate as significant of a change pre– and post–antibiotic and intensified airway clearance therapy compared with the total HRCT scores, emphasizing that plain CXRs are relatively insensitive outcome measures for assessing therapeutic interventions in this young population. Radiation dosage is a potential concern in the pediatric population. The dosage used in this study is similar to recent publications (5); however, since completing enrollment for this study, the radiation doses have been actively decreased at our institution, with quality imaging results. Improvement of CT scanners and technique has shown that excellent results can be obtained with much lower radiation exposure than in the present study. On the basis of current knowledge, a dose of 20 mAs and 80 kV(p) could have been used, thus decreasing radiation exposure by more than 50%. Using a computational model to estimate the risk of biennial CT scans in the CF population, de Jong and colleagues (44) recently concluded that routine scans have a low risk of leading to radiation-induced mortality. However, as the median age of survival improves in this population, the risk to benefit ratio in this population must be considered when performing HRCT scans as an outcome measure (5, 44). HRCT scoring using the modified Brody scale appeared to yield consistent results between different radiologists. A comparison of five different scoring systems has been performed (45) and the results were similar among these different systems, thus verifying that when used accurately, the scoring systems are able to identify and track disease. Because early, aggressive treatment in CF lung disease may improve prognosis, identifying the presence of early airway disease and sensitive outcome measures in the young CF population is critical. These outcome measures are needed for future trials that identify the most effective therapies in this population. The regional distribution of lower airway inflammation and the predominance of right-sided disease have important implications for both patient management and clinical trials. The significant improvement in the modified Brody score and the reversibility of air trapping and bronchiectasis/bronchial dilatation after intravenous antibiotic and intensified airway clearance therapy demonstrates both the importance of aggressive therapy in treating early peripheral airway disease and the potential use of this scoring system in the young population with CF. On the basis of our results, HRCT has the potential to be a sensitive outcome measure in this young population. Further research will need to address standardization of protocols in young children, radiation risks versus clinical benefit, and the applicability of this tool in the clinical setting before routine CT scanning can be recommended for management of CF lung disease.
The authors thank Kathy Abode, R.N., for her assistance in the bronchoscopy suite; the Pediatric Specialty Care Team for assisting in sedating the children; Paula Murphy for her assistance in processing the BALF samples; and Ned Beasley, R.T.R. (C.T.), along with the HRCT radiology technicians of North Carolina Children's Hospital, in assisting with scheduling and acquisition of the HRCT scans.
Supported by a Cystic Fibrosis Foundation Shwachman Clinical Investigator Award (DAVIS00Q0) and the Cystic Fibrosis Foundation Therapeutics, Inc. (CFFT DAVIS02A). This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200603-343OC on February 15, 2007 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 March 8, 2006; accepted in final form February 12, 2007
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
