| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From Depts of Diagnostic Radiology and Organ Imaging (K.T.W., G.E.A., A.T.A.), Medicine and Therapeutics (D.S.C.H., N.L., A.W., V.W., W.L., J.C.W., L.S.T., J.J.Y.S.), Anesthesia and Intensive Care (G.M.J.), and Surgery (S.S.C.C.), and Centre for Clinical Trials and Epidemiological Research (L.M.Y.), The Chinese Univ of Hong Kong, Prince of Wales Hosp, 30–32 Ngan Shing St, Shatin, Hong Kong SAR. Received Oct 10, 2003; revision requested Jan 5, 2004; revision received Jan 27; accepted Mar 2. Supported by the Research Fund for the Control of Infectious Diseases of the Health, Welfare and Food Bureau, Hong Kong. Address correspondence to K.T.W. (e-mail: wongkatakjeffrey@hotmail.com).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: The clinical, laboratory, and radiologic features of 138 patients with SARS were analyzed. Three radiologists in consensus retrospectively assessed the frontal chest radiographs obtained at presentation and during treatment (n = 2045) for the distribution (each lung was divided into upper, middle, and lower zones) and extent of lung parenchymal abnormality. Clinical end points included intensive care unit (ICU) admission and death.
RESULTS: Thirty-six (26.1%) patients required ICU care, and eight (5.8%) died. The patients who required ICU care and/or died had more extensive consolidation on chest radiographs obtained initially (median percentage of consolidation, 3.30%, with interquartile range [IR] of 1.70%–8.78% vs 1.70% [IR, 0%–3.30%]; P < .001) and on day 7 after fever onset (median percentage of consolidation, 15.00% [IR, 6.48%–28.73%] vs 5.00% [IR, 2.50%–7.50%]; P < .001) than did surviving patients who did not require ICU care. Patients with involvement of more than one lung zone on initial and day 7 chest radiographs were more likely to require ICU care and/or die than were those with involvement of one or fewer zones (P < .001). Patients with bilateral pneumonic changes at presentation were more likely to have an adverse outcome than were those with unilateral pneumonia (P < .001). Involvement of more than one lung zone at baseline chest radiography was an independent predictor of ICU admission and/or death (odds ratio, 3.16; 95% confidence interval: 1.07, 9.32; P = .037) after adjustments for other significant factors (ie, patient age, and baseline neutrophil count and lactate dehydrogenase level).
CONCLUSION: More extensive airspace disease at presentation is an independent predictor of adverse outcome in patients with SARS.
© Research Fund for the Control of Infectious Diseases, 2004
Index terms: Lung, consolidation • Lung, radiography, 68.11 • Pneumonia, 68.21 • Severe acute respiratory syndrome (SARS), 68.21
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Imaging has an important role in the diagnosis and daily treatment of patients with SARS. Chest radiography remains the first-line radiologic examination performed for suspected cases and is helpful for monitoring patient progress during treatment. Airspace opacity distributed peripherally in the lower lung zone is the most commonly encountered radiographic appearance in patients with SARS at presentation (2). Radiographic progression to multifocal or bilateral lung involvement followed by radiographic improvement occurs during treatment in about 70% of patients with SARS (2).
There is emerging evidence of a positive correlation between severity of lung abnormalities at chest radiography and clinical and laboratory parameters, such as arterial oxygen saturation and levels of liver enzymes, including alanine aminotransferase and aspartate aminotransferase (3). There are significant relationships among radiographic parameters, oxygen supplementation, and treatment response (4). The purpose of our study was to evaluate whether there is a correlation between the clinical outcomes and radiologic features of SARS.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Altogether, 138 patients were identified as either secondary cases—that is, those of patients who had had contact with the index case (n = 112)—or tertiary cases—that is, those of patients who had had contact with the secondary cases (n = 26). The 138 patients were 66 male and 72 female patients with a mean age of 39.3 years ± 16.8 (standard deviation) (age range, 20–83 years). There was no significant difference in age or sex between the secondary- and tertiary-case patients. The demographic data on these patients have been previously described in detail (1). All patients were Chinese. Clinical observations and laboratory data were documented. Analyses were based on data collected until April 5, 2003. The study was approved by our institutional review board; patient informed consent was not required.
Diagnosis and Monitoring of Patient Progress
The diagnosis of SARS was based on Centers for Disease Control and Prevention criteria (5). The criteria met in the study cases were (a) fever, as indicated by a body temperature higher than 38°C; (b) evidence of lung parenchymal consolidation at chest radiography and/or thoracic computed tomography (CT); and (c) history of exposure to the index case or direct contact with a patient or patients who became sick after being exposed to the index case. All 138 patients included in this study met all three of these criteria, and all were initially admitted to medical wards of Prince of Wales Hospital with isolation facilities.
The initial examinations performed in these patients included complete blood cell count (including differential count), clotting profile (prothrombin time, activated partial thromboplastin time, international normalized ratio, and d-dimer measurements), and serum biochemical (including electrolyte, renal function, liver function, creatine kinase, and lactate dehydrogenase [LDH] measurements) tests. These laboratory tests and chest radiography were performed daily. The requirements for supplemental oxygen and pulse methylprednisolone therapy during the study period also were recorded. A team of clinicians (N.L., A.W., V.W., W.L., J.C.W., L.S.T.) was responsible for clinical data collection.
Treatment
Initial treatment included the administration of cefotaxime (Claforan; Patheon, Swindon, United Kingdom) and either clarithromycin (Klacid; Abbott Laboratories, Berkshire, England) or levofloxacin (Cravit; Daiichi Pharmaceutical, Tokyo, Japan) to address the common pathogens that cause community-acquired pneumonia, according to current recommendations (6,7). When fever persisted and the complete blood cell count indicated leukopenia and/or thrombocytopenia, ribavirin (Rebetol; Schering-Plough, Las Piedras, Puerto Rico) and prednisolone (Sigma, Victoria, Australia) (0.5–1.0 mg/kg/d) were administered orally as a combination regimen. One hundred seven of the 138 patients with persistent fever and progressing lung opacity were given intravenous ribavirin and pulse methylprednisolone (Solumedrol; Pharmacia and Upjoin, Belgium) (0.5 g daily for 3–6 days).
Patients who developed hypoxia were treated with oxygen therapy through nasal cannulae. Patients were admitted to the ICU when respiratory failure, as indicated by the inability to maintain an arterial oxygen saturation of at least 90% while receiving supplemental oxygen of 50% and/or a respiratory rate greater than 35 breaths per minute, developed. Nineteen (13.8%) patients required invasive mechanical ventilation.
Radiologic Assessment
Frontal chest radiography was performed in the 138 patients at initial clinical presentation and during treatment. Only frontal chest radiographs—posteroanterior views for patients who could stand or anteroposterior views for those who could not stand—were obtained. All radiographic examinations were performed by using portable computed radiographic equipment (Mobilett Plus; Siemens, Erlangen, Germany) and a standardized technique: 75 kV, 4 mAs, and a 180-cm film-focus distance for the posteroanterior views and 70 kV, 4 mAs, and a 100-cm film-focus distance for the anteroposterior views; a broad tube focus was used to obtain both views (2). The images were assessed by using a picture archiving and communication system, or PACS, viewer workstation with a 2048 x 2048-pixel monitor (Magicview, version VA22E; Siemens). All 2045 frontal chest radiographs obtained (mean of 14.8 radiographs per patient; range, three to 26 radiographs) were assessed.
The frontal chest radiographs obtained at clinical presentation and at follow-up during treatment were retrospectively reviewed in consensus by three radiologists (including K.T.W. and G.E.A), who were unaware of the patients’ clinical progress. Two of these radiologists (one of whom was G.E.A.) had practiced as radiologists for more than 10 years, whereas the other (K.T.W.) had 7 years of experience in chest radiograph interpretation.
Follow-up radiographs were obtained daily during the patients’ hospital stay. Each lung was divided into upper, middle, and lower zones. Each of the three zones encompassed one-third of the craniocaudal distance of the lung on a frontal radiograph and was evaluated separately. The observers assessed and recorded the presence, appearance (airspace opacity, reticular opacity, mass, nodule, pleural effusion, and/or lymphadenopathy), distribution (upper, middle, or lower lung zone involvement and unilateral or bilateral involvement), and size (percentage area of involved lung parenchyma and number of involved zones) of the lung parenchymal abnormality seen on each chest radiograph obtained in all 138 patients. The percentage area of parenchyma involved in each zone on each lung was assessed by means of visual estimation, with the maximum percentage of each zone being 100%. The overall mean percentage of lung parenchymal involvement was calculated by averaging the percentage involvement of the six lung zones and ranged from 0% to 100% (2).
In addition, the disease progression pattern of findings seen on serial chest radiographs obtained in each patient during the treatment period was categorized into one of four pattern groups, according to our previous study protocol (2): The type 1 disease progression pattern was that of initial radiographic deterioration to the peak level followed by radiographic improvement, with peak level defined as overall mean lung parenchymal involvement of greater than 25% of the initial extent of involvement. The type 2 pattern was that of fluctuating radiographic changes, with at least two radiographic peaks and an intervening trough, with trough defined as overall mean lung parenchymal involvement that differed from the peak level by more than 25%. The type 3 pattern was that of static radiographic changes, with no discernible radiographic peak—that is, a change in overall mean lung parenchymal involvement of less than 25% of the initial extent of involvement—for more than 10 days. The type 4 pattern was that of progressive radiographic deterioration with no improvement.
Statistical Analyses
Our study cohort included all secondary and tertiary cases of SARS. The patients’ demographic, clinical, laboratory, and radiologic features were reported and analyzed. Requirement of ICU care and/or death was chosen as the clinical composite end point. The extent of consolidation was compared between the two outcome groups (patients who required ICU care and/or died and surviving patients who did not require ICU care) by using the Mann-Whitney test. Chest radiographic zonal involvement, unilateral or bilateral consolidation, and disease progression pattern were analyzed by using the 
2 test. Associations between rate of change in biochemical measurements and chest radiograph finding trend were assessed by using Spearman correlation. The extents of pneumonia (ie, percentages of lung parenchymal involvement) on chest radiographs obtained in the patients who did not require supplementary oxygen were illustrated with Kaplan-Meier survival curves.
The relationship between initial chest radiographic zonal involvement and clinical outcome was further explored by performing multivariable logistic regression analysis, with adjustments for patient age and LDH level and neutrophil count at presentation. All analyses were performed by using computer software (SAS, version 7.0; SPSS, Chicago, Ill). P < .05 was considered to indicate statistical significance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The percentages of consolidation and the numbers of involved lung zones at initial chest radiography are summarized in Tables 1 and 2, respectively. With regard to radiographic disease progression pattern, type 1, which was seen in 97 (70.3%) of the 138 patients, was the most common. Type 2, seen in 24 (17.4%); type 3, seen in 10 (7.2%); and type 4, seen in seven (5.1%) patients, were the next most common patterns (2).
|
|
|
Biochemical measurements revealed elevated serum alanine aminotransaminase levels in 32 (23.1%) patients (mean level, 60.4 IU/L ± 150.4 [standard deviation]; normal range, <55 IU/L). Creatine kinase levels were elevated in 44 (31.9%) patients (median level, 126 U/L; range, 29–4644 U/L; normal range, 42–218 U/L), whereas LDH levels were elevated in 98 (71.0%) patients. At presentation, the mean LDH level was 287.7 U/L ± 143.3 (normal range, 87–213 U/L) in the patients who did not require ICU care and 558.0 U/L ± 258.0 in those who required ICU care and/or died (P < .001). The mean peak LDH level was 310.0 U/L ± 153.8 in the patients who did not require ICU care and 629.7 U/L ± 283.5 in the patients who required ICU care and/or died (P < .001) (1).
Clinical Outcomes
Thirty-six (26.1%) of the 138 patients were admitted to the ICU, and all of these cases were due to respiratory failure. In the first 4 weeks of the SARS outbreak, there were eight deaths (crude mortality rate, 5.8%): Six of these patients died while they were in the ICU, whereas two died while they were in medical wards. A total of 38 patients reached the clinical end points for poor outcome—that is, they required ICU admission and/or they died. All eight patients who died were originally admitted to the hospital owing to major medical conditions. Two of these patients had myelodysplastic syndrome, four had cardiac disease (one with congestive heart failure, two with ischemic heart disease, and one with rheumatic heart disease), one had alcohol-related liver cirrhosis, and one had hepatitis B reactivation. All eight deaths were related to severe respiratory failure.
Correlation between Clinical Outcomes and Radiographic Findings
Extent of consolidation.—Clinical outcomes and radiographic findings are summarized in Table 3 and Figures 2 and 3. The patients who required ICU care and/or died had more extensive evidence of pneumonia on chest radiographs obtained initially (median percentage of consolidation, 3.30% [interquartile range, 1.70%–8.78%] vs 1.70% [interquartile range, 0%–3.30%]; P < .001) and on day 7 after fever onset (median percentage of consolidation, 15.00% [interquartile range, 6.48%–28.73%] vs 5.00% [interquartile range, 2.50%–7.50%]; P < .001) compared with the patients who survived and did not require ICU care.
|
|
|
Unilateral versus bilateral disease.—Patients with bilateral pneumonic changes (22 of 41 patients) at initial chest radiography were more likely to require ICU care or have a fatal outcome than were patients with unilateral pneumonia (13 of 67 patients) at initial radiography (P < .001).
Disease progression pattern on chest radiographs.—Of the 97 patients with the type 1 disease progression pattern on serial chest radiographs, only 17 (18%) were admitted to the ICU and/or died, whereas 21 of the 41 patients (51%) with a pattern other than type 1 were admitted to the ICU and/or died (P < .001). Fourteen of these 21 patients had the type 2 pattern, and seven had the type 4 pattern. The remaining 20 patients (10 with the type 2 pattern, 10 with the type 3 pattern) neither required ICU care nor had a fatal outcome.
Chest radiograph involvement and oxygen requirement.—The cumulative percentage of patients with SARS who did not require supplementary oxygen is correlated with the extent of consolidation in the Kaplan-Meier curve in Figure 4. Approximately 50% of the patients needed supplemental oxygen to maintain oxygen saturation above 90%, even when the area of consolidation on the chest radiograph was 10% at any time during treatment.
|
= 0.40, P = .014). However, the rates of change in neutrophil count (Spearman = 0.13, P = .2), creatine kinase level (Spearman = –0.16, P = .2), platelet count (Spearman = 0.17, P = .07), lymphocyte count (Spearman = –0.103, P = .3), activated partial thromboplastin time (Spearman = 0.11, P = .32), and d-dimer level (Spearman = 0.08, P = .7) did not correlate with radiographic changes. Independent predictive factors of ICU admission and/or death.—Our previous univariate analysis of data from the same patient cohort revealed that advanced age, male sex, peak creatine kinase value, peak LDH level at presentation, higher absolute neutrophil count at presentation, and low serum sodium levels were significant predictive factors of ICU admission and/or death (1). Further multivariable analyses, including assessment of the number of involved lung zones at initial chest radiography, revealed that advanced age (odds ratio [per year of age], 1.06; 95% confidence interval: 1.03, 1.10; P < .001), high LDH level at presentation (odds ratio [per unit per liter], 1.01; 95% confidence interval: 1.002, 1.008; P = .001), higher absolute neutrophil count at presentation (odds ratio, 1.37; 95% confidence interval: 1.04, 1.80; P = .025), and involvement of more than one lung zone at initial chest radiography (odds ratio, 3.16; 95% confidence interval: 1.07, 9.32; P = .037) were independent predictive factors of adverse outcome in our study cohort.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
At presentation, the majority (78.3%) of our study patients with SARS had evidence of airspace consolidation on their chest radiographs. The initial radiographic appearance was normal in more than 20% of cases, but serial chest radiography performed a median of 3 days later depicted airspace disease (2). Thin-section thoracic CT has had an important diagnostic role for patients who were clinically suspected of having SARS but had normal initial chest radiographs. Multifocal peripheral subpleural ground-glass opacification or consolidation has been the most commonly observed CT feature in patients with SARS at the time of diagnosis (9).
The substantial morbidity associated with SARS was reflected by the wide spectrum of clinical symptoms followed by the progression of consolidation that led to respiratory failure, the need for intensive care, and death in some cases. Radiographically depicted progression of consolidation occurred at a peak level a mean of 8.6 days ± 3.1 (standard deviation) after the onset of fever. There appears to be a strong correlation between the extent of radiographic abnormality and the degree of respiratory failure, as indicated by the Kaplan-Meier curve data (Fig 4), which show that even when only a small percentage (10%) of the lung was consolidated, approximately 50% of the patients required supplementary oxygen.
The thin-section CT features of SARS closely resemble those of bronchiolitis obliterans organizing pneumonia (10,11), which is generally a steroid-responsive condition. On the basis of the evidence of disease progression seen on serial chest radiographs, we initiated treatment with intravenous pulse methylprednisolone to reduce immunopathologic lung damage from an overexuberant host response (12,13), with a median time of commencement of this treatment of 8 days after the onset of fever. Although the treatment regimen yielded a favorable clinical response in the majority of the patients, 36 (26.1%) patients required ICU care and eight died.
In view of the substantial number of patients with SARS who required ICU care, it would be desirable to have some parameters (clinical, laboratory, or radiologic) early in the disease that would help predict the clinical outcome. The extent of pneumonia at presentation radiologically appears to correlate with an adverse clinical outcome of SARS. More extensive airspace disease (as reflected by the higher percentage of consolidation and involvement of more than one lung zone on chest radiographs obtained both at admission and on day 7 after fever onset and by bilateral disease at admission) was associated with subsequent need for ICU care and/or death. The finding of more extensive airspace disease is similar to other forms of community-acquired pneumonia in general and in ICU-treated populations, with abnormal airspace opacities in both lungs, involvement of more than one lobe, and rapid radiographic progression found to be independent prognostic factors for poor outcome (14–17).
In addition, we found that the pattern of radiographic disease progression correlated well with clinical outcome. Patients with the type 1 progression pattern (initial radiographic progression followed by improvement) on serial chest radiographs seemed to have a more favorable outcome, whereas all seven patients with the type 4 pattern (progressive deterioration) had an adverse clinical outcome: Six of them died, and the seventh patient was still critically ill in the ICU at the time of this writing.
There was a positive correlation between chest radiograph finding trend and rate of change in LDH level, a marker of tissue damage. We believe that the LDH level can reflect the extent of lung injury in this setting and that both serial chest radiograph findings and LDH levels are important in the management of SARS. Our multivariable analysis revealed that involvement of more than one lung zone at chest radiography performed at presentation is an independent predictor of adverse outcome after adjustments for high baseline LDH level, advanced age, and high baseline neutrophil count.
There were several limitations to the study. First, the radiographic data were collected retrospectively, and this might have led to observer bias. It was virtually impossible to review a substantial number of radiographs prospectively during the major outbreak of this highly infectious disease, which affected daily operations and services. This potential pitfall was minimized by having the same group of observers blinded to the clinical progress and outcomes of the patients. Another limitation of this study was the visual assessment of radiographic opacities without proper evaluations of intraobserver and interobserver variability. Given that the study was performed during the period of a highly infectious and haphazard epidemic, our method of radiographic assessment was deemed appropriate at that time. The number and extent of errors were minimized by having each radiograph read in consensus by three independent and experienced radiologists.
Another potential limitation was related to the practical difficulty in obtaining chest radiographs of consistently good quality with a portable radiographic unit; differences in radiation exposure would have seriously affected the perception of lung parenchymal opacity. This potential technical pitfall was minimized by using computed radiography and a standardized exposure technique.
In conclusion, the results of this study show that chest radiography is useful in the diagnosis and management of SARS. More extensive radiographically depicted lung involvement at presentation and during treatment is associated with adverse clinical outcome. In addition, extensive radiographic involvement at presentation is an independent predictor of ICU admission and/or death.
|
ACKNOWLEDGMENTS |
|---|
Guarantors of integrity of entire study, D.S.C.H., A.T.A., J.J.Y.S.; study concepts, D.S.C.H., A.T.A., K.T.W., J.J.Y.S., S.S.C.C.; study design, D.S.C.H., K.T.W.; literature research, K.T.W., G.E.A., N.L., A.W.; clinical studies, D.S.C.H., N.L., A.W., V.W., W.L., J.C.W., L.S.T., G.M.J., J.J.Y.S.; data acquisition, N.L., A.W., K.T.W., G.E.A.; data analysis/interpretation, N.L., A.W., K.T.W., L.M.Y.; statistical analysis, N.L., A.W., K.T.W., L.M.Y.; manuscript preparation, D.S.C.H., K.T.W., N.L., A.W., A.T.A., J.J.Y.S.; manuscript definition of intellectual content and revision/review, D.S.C.H., K.T.W., A.T.A.; manuscript editing and final version approval, D.S.C.H., K.T.W., A.T.A., J.J.Y.S.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, D.S.C.H., A.T.A., J.J.Y.S. The complete list of author contributions is cited at the end of this article.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  V. C. C. Cheng, S. K. P. Lau, P. C. Y. Woo, and K. Y. Yuen Severe Acute Respiratory Syndrome Coronavirus as an Agent of Emerging and Reemerging Infection Clin. Microbiol. Rev., October 1, 2007; 20(4): 660 - 694. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. S. Hui, S. D. Hall, M. T.V. Chan, B. K. Chow, S. S. Ng, T. Gin, and J. J.Y. Sung Exhaled Air Dispersion During Oxygen Delivery Via a Simple Oxygen Mask Chest, August 1, 2007; 132(2): 540 - 546. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. J. Cowling, M. P. Muller, I. O. L. Wong, L.-M. Ho, S.-V. Lo, T. Tsang, T. H. Lam, M. Louie, and G. M. Leung Clinical prognostic rules for severe acute respiratory syndrome in low- and high-resource settings. Arch Intern Med, July 24, 2006; 166(14): 1505 - 1511. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Hui, K. T. Wong, F. W. Ko, L. S. Tam, D. P. Chan, J. Woo, and J. J.Y. Sung The 1-Year Impact of Severe Acute Respiratory Syndrome on Pulmonary Function, Exercise Capacity, and Quality of Life in a Cohort of Survivors Chest, October 1, 2005; 128(4): 2247 - 2261. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
