| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Departments of Respiratory Medicine (Drs. Judge and Gallagher) and Radiology (Drs. Dodd and Masterson), St. Vincent’s University Hospital, Dublin, Ireland.
Correspondence to: Jonathan Dodd, MD, CIMIT Office, Suite 400, Massachusetts General Hospital, Charles River Plaza, Cambridge St, Boston, MA 02114-2969; e-mail: jddodd{at}partners.org
Abstract
Background: FEV1 may remain stable while high-resolution CT (HRCT) appearances deteriorate in children with cystic fibrosis (CF). However, spirometry results commonly decline in older age groups.
Objectives: To compare the rate of decline in HRCT abnormalities and spirometry results over time in an adult cohort with CF.
Methods: The HRCT scans of 39 consecutive patients (19 males and 20 females; mean age, 22 years; range, 16 to 48 years) with two HRCT scans > 18 months apart were randomly and blindly scored using a modified Bhalla scoring system by two independent chest radiologists. Age, body mass index, spirometry, and sputum cultures were recorded at the time of both HRCTs. Rates of change in clinical parameters and HRCT abnormalities were calculated and compared using repeated-measures analysis of variance.
Results: Mean FEV1 declined at a rate of – 2.3% per year, while mean HRCT total score declined at a rate of –2.7% per year. Several individual HRCT abnormalities as well as HRCT total scores declined significantly faster than FEV1 (p < 0.001). Six patients showed stable spirometry results but worsening HRCT scores. Mucus plugging and extent of bronchiectasis deteriorated at a more rapid rate in the group with mildly impaired lung function. Air trapping, collapse/consolidation, peribronchial thickening, severity of bronchiectasis, and generations of bronchial divisions involved deteriorated at a more rapid rate in the group with moderate-to-severely impaired lung function.
Conclusions: Adult CF patients have more rapid rates of decline in HRCT abnormalities than in spirometry results. Individual HRCT abnormalities decline at different rates depending on the degree of lung function impairment.
Key Words: cystic fibrosis • high-resolution CT • longitudinal study • respiratory function tests
Cystic fibrosis (CF) is the most common autosomal recessive inherited disorder in whites, with a prevalence of 1 in 1,461 births in the Irish population.1 By the later teenage years, significant irreversible lung damage has occurred in many patients. The majority die of recurrent respiratory sepsis and respiratory failure in adulthood.2
Spirometry, particularly FEV1, is used as a marker of lung disease because of its reproducibility and accuracy as an outcome surrogate.3456 It is well known that spirometry results decline as lung disease progresses.56 In a thought-provoking article, de Jong et al7 showed that the rate of decline in high-resolution CT (HRCT) appearances was more significant than the rate of decline in spirometry results in children, and this has been confirmed by others.8 More recently, further novel work has shown that several HRCT abnormalities decline at a faster rate than FEV1 in adults.9 As a result, proposals for the use of HRCT as an outcome surrogate have been suggested.1011 In this regard, a combined HRCT/spirometric score has been shown to be a sensitive marker in detecting treatment effects.1213 More recently, HRCT has been shown to correlate with the number of respiratory exacerbations in children over time.14
In adults with CF, the rate of change in spirometry is well documented, but the rate of change in HRCT abnormalities has only been described in a single study.9 Since adults generally have more progressive lung function impairment than children, we hypothesized that the rate of decline in lung spirometry in this age group would be similar or worse than the rate of decline in HRCT abnormalities. The aim of this study was to compare the rates of change in HRCT and clinical measurements over a significant time period in an adult cohort of patients with CF.
Materials and Methods
Patient Population
Since 1997, all adult patients undergo elective HRCT every 18 to 24 months as part of scheduled long-term patient assessment at the Irish National Referral Centre for Adult CF. Patients 
16 years old from the entire country are referred to the center for multidisciplinary care. The number of patients attending our center has steadily increased from 1997 (approximately 170 patients) to 2005 (approximately 250 patients) as treatment has become more centralized and life expectancy continues to improve. The CF radiology database was retrospectively searched from January 1997 to January 2005 by a single investigator (E.J.) for 39 consecutive patients with two HRCT scans (HRCT-1 and HRCT-2) > 18 months apart (19 males and 20 females; mean age, 22 years; range, 16 to 48 years). All patients had documented clinical, radiologic, or genotypic features of CF as well as abnormal sweat test results (sweat sodium and chloride > 60 mmol/L).
Clinical charts were reviewed and age, body mass index (BMI), spirometry, and sputum cultures were recorded at the time of the two HRCT scans. Only clinically stable patients at the time of a HRCT scan were included. Patients were excluded if they had the following: (1) only one HRCT (or two HRCTs < 18 months apart); (2) symptoms or signs of acute respiratory exacerbation at the time of HRCT or spirometry; (3) unstable spirometry results at the time of either HRCT (> 10% decrease in FEV1 compared with baseline values in the preceding 2 months); (4) required hospitalization for IV antibiotics in the 2 weeks prior to either HRCT for a respiratory tract infection; and (5) died before undergoing sequential HRCT. This resulted in approximately 210 patients being excluded from the study, leaving 39 in the final study group. The hospital ethics committee of out institution approved the study.
HRCT Protocol and Scoring
Images were obtained using spiral CT with patients in the supine position (Somatom Plus 4; Siemens; Erlangen, Germany). Inspiratory images were obtained in suspended deep inspiration with 1-mm slice thickness every 10 mm from the apices to the costophrenic angles. Expiratory images were obtained in full expiration at three levels: the top of the aortic arch, the carina, and 2 cm above the diaphragm. Scanning parameters were 140 kilovolt peak and 140 mA, with images reconstructed using a high spatial frequency bone algorithm and a 512 x 512 matrix. Lung windows with a width of 2,000 Hounsfield units and level of – 700 Hounsfield units were applied. Images were retrospectively, independently scored in random order by two chest radiologists (J.D., J.M.), who were blinded to patient identification, clinical severity of disease, spirometry, and date of all HRCT scans. Of the two observers, one observer (J.M.) was a senior chest radiologist with > 20 years experience. The other observer (J.D.) was a thoracic clinical fellow in chest radiology with 2 years experience scoring HRCT. A consensus meeting was held prior to formal scoring to standardize scoring between observers. HRCT scans were scored over a 2-week time period.
Images were scored using a modified Bhalla scoring system.15 Modifications included additional HRCT abnormalities described in CF lung disease16, 17 since the original publication by Bhalla et al.15 Abnormalities were defined according to recommendations of the nomenclature committee of the Fleischner society.18 Bronchiectasis was defined as a bronchus with an internal diameter larger then its accompanying pulmonary artery, lack of tapering of the bronchial lumen for > 2 cm, and visualization of a bronchus within 1 cm of the costal pleura. Peribronchial thickening was defined by a bronchial wall > 1 mm in thickness. Mucus plugging was identified by visualizing plugs in large airways or by the presence of tree-in-bud peripherally. Sacculations were defined when dilated bronchi had a cystic or saccular appearance. A bulla was defined as a round, focal airspace > 1 cm in diameter, demarcated by a thin wall. Emphysema was defined as areas of decreased attenuation with disruption of the underlying vascular pattern and absence of well-defined walls. Consolidation was identified when increased lung opacification was detected that obscured the underlying parenchyma. We modified the score by including an assessment of acinar nodules, thickening of interlobular septa, ground glass and mosaic perfusion on inspiratory images, and air trapping on expiratory images, as these changes are commonly seen and generally included in contemporary scoring systems for CF lung disease.19 A nodule was defined as a round opacity at least moderately well marginated and not more than 3 cm. A septal line was defined as a thin line corresponding to an interlobular septum. Ground-glass opacity was identified when increased lung opacification was detected but did not obscure the underlying parenchyma. Areas of hyperlucency with hypovascularity defined mosaic perfusion on inspiratory scans. Areas of hyperlucency on expiratory images defined air trapping. The total score was derived by adding the scores for each abnormality, and ranged from 0 to 37. For scoring purposes, the average of the two observers’ individual scores were calculated. Absolute total scores were converted to percentages of the potential total maximum score. The extent and severity of abnormalities were scored as in Table 1 .
|
1% over the time period for each patient (median ± SD time interval, 42 ± 7 months; range, 20 to 73 months). Patients were also subgrouped into mild, moderate, and severely impaired lung function groups based on FEV1 percentage of predicted values. Patients in the mild group had FEV1 > 80% of predicted, patients in the moderate group had FEV1 from 50 to 80% of predicted, and patients in the severe group had FEV1 < 50% of predicted.
Sputum
Sputum samples were collected following spontaneous expectoration at time 1 and time 2 and were processed by the microbiology laboratory at our CF center using standardized methodology.2122 Samples were plated onto selective media using standard techniques, and pathogen colony morphotypes were identified visually.
Longitudinal Measurements and Statistical Calculations
The initial HRCT and spirometry findings obtained at time 1 are reported as HRCT-1 and PFT-1, and follow-up HRCT and spirometry findings obtained at time 2 are reported as HRCT-2 and PFT-2. Median time interval was 42 ± 16 months (range, 19 to 73 months). The annual rate of change in HRCT score (
HRCT) was calculated for each patient as follows: (HRCT-2 – HRCT-1)/individual time interval for each patient. A positive HRCT value indicated a worsening in HRCT appearances. Similarly, the annual rate of change in PFT score (PFT) was calculated for each patient as follows: (PFT-2 – PFT-1)/individual time interval for each patient. A negative PFT value indicated a worsening in spirometry results.
The Wilcoxon signed-rank test was used to compare absolute numbers of patients for abnormalities between time 1 and time 2. Annual changes in composite CT, individual CT abnormalities, and PFTs were compared using repeated-measures analysis of variance, as described elsewhere.9 These comparisons were made for three separate groups: (1) all patients, (2) patients with mildly impaired lung function, and (3) patients with moderate-to-severely impaired lung function. Interobserver agreement for HRCT total scores was assessed using the intraclass correlation coefficient for total scores. A score > 0.8 is considered good agreement. Weighted 
analysis was used to assess interobserver agreement for individual abnormalities. When the value is 1.00, interobserver agreement is perfect; a value of zero indicates a level of agreement no better than chance. A value of 0.21 to 0.40 indicates fair agreement, 0.41 to 0.60 indicates moderate agreement, 0.61 to 0.80 indicates good agreement, and 0.81 to 1.00 indicates excellent agreement.23 Bland and Altman plots were utilized to evaluate for systematic bias between observers.24 Statistical significance was set at a p value < 0.05. Data are presented as mean ± SD.
Results
Patient Demographics, Spirometry, and Sputum Cultures
Mean age of the patients increased by 4 years (Table 2
). Mean BMI remained stable, but four patients had BMIs < 18.5 kg/m2 at time 1, indicating malnourishment. Genotype was available for 37 patients. Forty-four percent of patients were homozygous and 36% were heterozygous for the F508 mutation. All spirometry measurements except RV deteriorated significantly. Mean FEV1 was 81.8 ± 24.9% of predicted at time 1, indicating a group with overall mild lung impairment (Fig 1
, top, A). This decreased by 11.7% between PFT-1 and PFT-2 (42 ± 7 months). Sixteen patients had mild lung function abnormalities, with mean FEV1 decreasing from 102.1 ± 14.5 to 87.5 ± 18.2%. Nineteen patients had moderate lung function abnormalities, with mean FEV1 decreasing from 69.8 ± 6.0 to 59.5 ± 15.8%. Four patients had severe lung function abnormalities, with mean FEV1 decreasing from 34.0 ± 2.7 to 30.5 ± 2.6%. Sputum cultures were available for 28 patients and 34 patients at time 1 and time 2, respectively. The most commonly cultured organism was Pseudomonas aeruginosa (75% of patients at time 1, 74% of patients at time 2). Over the course of the study period, 72% of patients were prescribed pancreatic supplements; 64% were prescribed multivitamins; 23% were prescribed deoxyribonuclease; 64% were prescribed inhaled nebulized antibiotics (56% inhaled colomycin, 8% inhaled tobramycin); 64% were prescribed oral prophylactic antibiotics; 18% were prescribed nasal steroids for sinusitis; 59% were prescribed inhaled ß-agonists; and 33% were prescribed inhaled steroids.
|
|
|
|
|
> 0.8, p < 0.0001), generations of bronchial divisions ( = 0.50, p < 0.0001), air trapping ( = 0.33, p = 0.003), and severity of bronchiectasis ( = 0.25, p = 0.004). At time 2, the best interobserver agreement was seen for mosaic perfusion ( >0.8, p < 0.0001), generations of bronchial divisions ( 0.44, p < 0.0001), collapse/consolidation ( = 0.37, p = 0.001), bullae ( = 0.36, p = 0.02), and air trapping ( = 0.31, p = 0.009). Severity of bronchiectasis, extent of mucus plugging, sacculations/abscesses, generations of bronchial divisions involved, and air trapping demonstrated systematic differences between readers on Bland and Altman plots. Observer 1 scored extent of mucus plugging and air trapping worse on average than observer 2, and this was more pronounced for mild scores. Observer 2 scored severity of bronchiectasis, sacculations/abscesses, and generations of bronchial divisions involved worse on average than observer 1. For severity of bronchiectasis and generations of bronchial divisions involved, this was more pronounced for moderate scores; for sacculations/abscesses, this was more pronounced for severe scores.
Discussion
The major findings of this study are as follows: (1) 15% of adults in our cohort demonstrated stable spirometry results but deteriorating HRCT scores; (2) in the remaining adults with declining FEV1, HRCT scores for several abnormalities deteriorated at a more rapid rate; and (3) the rate of decline of individual HRCT abnormalities differed depending on the severity of lung function impairment. In recent innovative work, de Jong et al9 evaluated a Swedish group of children and adults with CF over time, and found several CT abnormalities as well as total HRCT score declined in the adult cohort while FEV1 remained stable. Six patients in our study demonstrated stable FEV1 but deterioration in many HRCT abnormalities, similar to findings in the Swedish cohort. In contrast, most patients in our adult CF population had moderate-to-severe lung impairment with chronic loss of lung function over time. Nevertheless, even in this group several CT abnormalities as well as total HRCT scores declined more rapidly than FEV1.
We subgrouped patients into mild and moderate-to-severe lung function impairment and found certain HRCT abnormalities worsened in patients with mild impairment, while others worsened in patients with moderate-to-severe impairment. In particular, based on the degree of lung impairment, the extent of bronchiectasis deteriorated more markedly in the mildly impaired group, whereas the severity of bronchiectasis deteriorated more markedly in the moderate-to-severely impaired group. de Jong et al9 found that peripheral bronchiectasis showed the most significant deterioration on HRCT over time in their adult Swedish cohort. Although our scoring system did not assess abnormalities in terms of central and peripheral locations, both extent and severity of bronchiectasis deteriorated significantly compared with FEV1. Whichever scoring system is utilized, both studies would appear to suggest that bronchiectasis is one of the most important factors in the longitudinal deterioration in CF lung disease.
Clearly there were differences between the Irish and Swedish cohorts in longitudinal progression, with the annual change in CT score equal to + 1.6% and FEV1 equal to – 0.4%, compared to – 2.7% and – 2.3% in the Irish cohort, respectively. For individual HRCT abnormalities, de Jong et al9 found no change in mucus plugging over time in their adult patients in contrast to our results in which mucus plugging worsened predominantly in the group with mildly impaired lung function. Such discrepancies may be related to differences in treatment or in patient populations. Some European centers have followed a more aggressive policy of early antibiotic intervention from time of diagnosis, which may influence decline in spirometry results over time.25 Alternatively, the genetic makeup of our patients may differ from other populations, and it is clear that genotype strongly influences the phenotypic expression of CF lung disease.26 It would be extremely interesting to perform a multicenter study evaluating such differences between populations and their influence on CT and spirometric changes over time. Despite these individual abnormality differences, our overall findings corroborate those of de Jong et al7 and suggest that mucus plugging occurs earlier in CF lung disease and at a milder stage of lung function impairment, and that bronchiectasis may be an end sequel of such antecedent abnormalities.
Reports of the prevalence of air trapping in adults with CF are limited.2728 The prevalence increases with progressively worsening radiologic appearances. We found air trapping deteriorated significantly faster than FEV1 over time, and this decline was most marked in the moderate-severely impaired lung function group. This finding in conjunction with the reduction in FEF25–75% suggests ongoing peripheral airways disease. CT pathology studies29 evaluating the association between bronchiectasis and air trapping have shown that hyperlucency on expiratory CT adjacent to bronchiectasis represents constrictive bronchiolitis histologically. Because it showed more rapid decline than FEV1 over time, air trapping may serve as a useful additional outcome measure in patients with CF. A potential obstacle is the poor-to-moderate interobserver agreement between readers.19 However, quantitative methods using spirometrically triggered HRCT may improve this variability, as they are more sensitive to regional changes in air trapping than spirometry alone.30 Additionally, in a recent, double-blind, randomized controlled trial31 of dornase alfa in children with CF, quantitative air trapping measurements were more sensitive to treatment-related changes in regional air trapping than spirometry or total HRCT scores alone.
Our findings would support the concept that while longitudinal structural abnormalities are clearly seen with HRCT, their longitudinal functional significance is less clear-cut.1132 Other imaging techniques such as hyperpolarized 3He MRI may prove a useful alternative in obtaining longitudinal functional information. We recently evaluated 3He MRI in CF adults and found it to be a strong correlate of structural HRCT abnormalities and a stronger correlate of spirometry than HRCT.33 We envisage, although currently hypothetical, that spirometry, 3He MRI, and HRCT may work synergistically to provide longitudinal functional and structural information in CF.
Our study has several limitations. It is retrospective, and spirometry was not performed on the same day as HRCT in some patients. However, we sought to be meticulous in excluding patients with clinical symptoms/signs of a respiratory exacerbation in the immediate period prior to HRCT. We included all consecutive patients with HRCT scans > 18 months apart, rather than just patients with stable spirometry results over a significant time period. We believe this reflects the more typical clinical spectrum of adults with CF. Since all patients in our center electively undergo HRCT, we hoped this would minimize potential patient selection bias. Nevertheless, it is possible that healthier patients with HRCT scans at clinically stable periods were selected over patients with more severe disease. However, our group had a wide spectrum of lung function impairment, and we believe selection bias to be minimal. Not all patients underwent comprehensive spirometric evaluation, although FEV1 and FVC, which are the major spirometric measures used in CF, were available in all patients.
Radiation dose to patients remains an inherent limitation of HRCT. Recent estimates of lifelong HRCT surveillance in patients with CF show a low risk for cancer-induced mortality.34 Technological advances such as automatic tube modulation and use of low-dose HRCT protocols may further reduce radiation exposure.3536 Such findings have implications for routine HRCT surveillance and use as an outcome measure in CF.101114 Our results would support such an application in adults, but further work demonstrating that independent changes in HRCT consistently correlate with changes in clinical outcome measures across several studies is required to substantiate this proposal.
In conclusion, in CF adults with stable FEV1 several HRCT abnormalities deteriorate significantly. Even in adults with declining FEV1, HRCT abnormalities show more rapid deterioration. Different HRCT abnormalities worsen at different rates in groups with mild and moderate-to-severe lung function impairment.
Footnotes
Abbreviations: BMI = body mass index; CF = cystic fibrosis; DLCO = diffusion capacity of the lung for carbon monoxide; FEF25–75 = forced expiratory flow between 25% and 75% of expiratory vital capacity; HRCT = high-resolution CT; HRCT = annual rate of change in high-resolution CT score; PFT = pulmonary function testing; PFT = annual rate of change in pulmonary function testing score; RV = residual volume; TLC = total lung capacity
Received for publication October 15, 2005. Accepted for publication May 23, 2006.
References
This article has been cited by other articles:
|
|
 |
|
  J. Donadieu, C. Roudier, M. Saguintaah, C. Maccia, and R. Chiron Estimation of the Radiation Dose From Thoracic CT Scans in a Cystic Fibrosis Population Chest, October 1, 2007; 132(4): 1233 - 1238. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
