Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2373041919
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Antonio, G. E.
Right arrow Articles by Ahuja, A. T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Antonio, G. E.
Right arrow Articles by Ahuja, A. T.
(Radiology 2005;237:1081-1090.)
© RSNA, 2005

Thoracic Imaging

Radiographic-Clinical Correlation in Severe Acute Respiratory Syndrome: Study of 1373 Patients in Hong Kong1

Gregory E. Antonio, MD, Clara G. C. Ooi, MD, Ka Tak Wong, MD, Eva L. H. Tsui, MSocSc, Joyce S. W. Wong, MD, Alan N. L. Sy, MD, Joyce Y. H. Hui, MD, Cho Yin Chan, MD, Henry Y. H. Huang, MD, Yip Fai Chan, MD, Tai Pang Wong, MD, Lilian L. Y. Leong, MD, Jane C. K. Chan, MD and Anil T. Ahuja, MD

1 From the Dept of Diagnostic Radiology and Organ Imaging, Chinese Univ of Hong Kong, Prince of Wales Hosp, Shatin, Hong Kong (G.E.A., A.T.A.); Dept of Diagnostic Radiology, Univ of Hong Kong, Hong Kong (C.G.C.O.); Dept of Diagnostic Radiology and Organ Imaging, Prince of Wales Hosp, Hong Kong (K.T.W.); Statistics and Research Unit (E.L.H.T.), Professional Services and Medical Development Div (J.C.K.C.), Hosp Authority Head Office, Hong Kong; Dept of Radiology, Queen Mary Hosp, Hong Kong (J.S.W.W., A.N.L.S., L.L.Y.L.); Dept of Radiology and Organ Imaging, United Christian Hosp, Hong Kong (J.Y.H.H.); Dept of Diagnostic Radiology and Nuclear Medicine, Tuen Mun Hosp, Hong Kong (C.Y.C.); Dept of Radiology, Tseung Kwan O Hosp, Hong Kong (H.Y.H.H.); and Dept of Radiology, Pamela Youde Nethersole Eastern Hosp, Hong Kong (Y.F.C., T.P.W.). Received Nov 26, 2004; revision requested Jan 31, 2005; revision received Apr 29; accepted May 31. Address correspondence to G.E.A. (e-mail: gregantonio{at}cuhk.edu.hk).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To retrospectively analyze serial chest radiographs in all patients with severe acute respiratory syndrome (SARS) in Hong Kong for temporal changes and differences between patients who died and those who were discharged from the hospital and to compare radiographic and clinical parameters.

MATERIALS AND METHODS: This retrospective study had ethics review board endorsement, and the need for informed consent was waived. Selected serial chest radiographs obtained from the time of presentation until discharge or death in 1373 patients with laboratory-confirmed SARS were scored. Scoring was based on the area and location of lung opacification on radiographs obtained at each of five milestones (presentation, beginning of ribavirin therapy, beginning of corticosteroid therapy, time of most severe radiographic appearance of disease, and before discharge or death). Extents of lung opacification at these five milestones were compared between patients who died and those who survived (by using a repeated-measures analysis of variance model), and the temporal trend of the radiographic-clinical parameters was analyzed (by using Cochran-Armitage trend testing, Kendall {tau} correlation coefficients, and descriptive graphic analysis).

RESULTS: The final cohort consisted of 1373 patients (1212 of whom [485 male and 727 female patients; mean age, 38.4 years] survived and 161 of whom [84 male and 77 female patients; mean age, 63.0 years] died). Among survivors, older patients had more extensive radiographic changes than younger ones. However, among patients who died, older patients had less extensive radiographic opacification at the worst stage of disease and just before death than did younger patients. Despite a higher mortality risk for male patients, both sexes in the same outcome group had similar radiographic findings. For both outcome groups, the rate of radiographic progression was similar for the first 11 days but diverged afterwards. The extent of opacification increased by approximately one zone every 4–5 days for the initial 11 days. Radiographic scores correlated with the ratio of PaO2 to the fraction of inspired oxygen, lymphocyte count, lactate dehydrogenase level, and neutrophil count at each milestone and in terms of changes between milestones (P < .01 for all correlation coefficients, except for radiographic score and neutrophil count between the first two milestones).

CONCLUSION: The initial extent of radiographic opacification may be useful for prognostic prediction. Radiographic progression correlates well with that of important clinical and laboratory parameters and may be used as an objective prognostic indicator early in SARS.

© RSNA, 2005


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
There was a 4-month outbreak of severe acute respiratory syndrome (SARS) in Hong Kong beginning in March 2003. By the end of the outbreak, there were 1755 probable cases of infection and 299 deaths (1).

Previous efforts by investigators at various institutions in Hong Kong to analyze radiographic findings in SARS have continued so that the radiographic changes in all patients with laboratory-confirmed SARS (2) in Hong Kong during the acute phase of the illness could be documented. Such analyses have been based on serial changes in the extent of air-space opacification—the major radiographic abnormality seen in SARS (3–12)—during the patient's hospital stay. The purpose of our study was to retrospectively analyze serial chest radiographs obtained in all patients with SARS in Hong Kong for temporal changes and differences between patients who died and those who were discharged from the hospital and to compare radiographic and clinical parameters.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
This retrospective study and the use of the SARS clinical database were commissioned by the Hong Kong Hospital Authority SARS Collaborative Group and endorsed by the ethics review board of the Hospital Authority. Informed consent for this retrospective study was waived. The authors acknowledge the permission granted by the Hospital Authority SARS Collaborative Group for the use of the SARS clinical database. Data collection for and management of the SARS clinical database were supported by funding from the Health, Welfare and Food Bureau and the Research Fund for the Control of Infectious Diseases of the Hong Kong government. The authors of this manuscript had control of the data and the information submitted for publication.

Patient Cohort
Between March and July 2003, a total of 1755 patients in 14 hospital centers in Hong Kong were identified as probably having SARS. The diagnosis of probable SARS was rendered by using a slightly modified version of the prevailing World Health Organization clinical case definition for SARS (2). Of these 1755 patients, 1462 (602 males, 860 females; mean age, 41.1 years; age range, 1–98 years; mortality rate, 11.8% [172 of 1462 patients]) were identified as having laboratory-confirmed SARS on the basis of the World Health Organization laboratory case definition for SARS issued in August 2003 (13).

Treatment Protocol
To enable a better understanding of the relevance of the milestones chosen for radiographic scoring (see below), we will briefly describe the recommended therapeutic protocol at that time. Clinicians were at liberty to adjust the therapy to meet individual patient needs. Patients were initially treated with broad-spectrum antibiotics and an antiviral agent for 2 days. If fever persisted for more than 48 hours despite treatment, the antiviral agent ribavirin (Rebetol; Schering-Plough, Las Piedras, Puerto Rico) and "low-dose" corticosteroids (0.5–1.0 mg of prednisolone [Sigma, Victoria, Australia] per kilogram of body weight per day) were administered. If there was persistence or recurrence of fever and radiographic progression of lung opacification or hypoxemia despite this initial combination therapy, pulses of high-dose methylprednisolone (Solumedrol; Pharmacia and Upjohn, Brussels, Belgium) were given (0.5 g intravenously for 3 consecutive days).

Milestone Selection
Five milestones during the patient's hospital stay were used for the purpose of radiographic scoring and comparison. Two of the milestones were determined by using clinical data retrieved from the Hong Kong Hospital Authority SARS clinical database: the date of starting ribavirin therapy and the date of starting pulse corticosteroid therapy (if administered). The radiographs obtained on these two dates were then scored by using a standard scoring system (see below). If no radiograph was obtained on the specific date, the radiograph obtained the day before the specific date was used for scoring, and if such a radiograph was not available, the radiograph obtained the day after the specific date was used. The dates of symptom onset and treatment were retrieved from the Hong Kong Hospital Authority SARS clinical database.

Three other milestones were determined at the time of radiographic review: the time of initial presentation or diagnosis, the time during the hospital stay at which the radiographic appearance of opacification was the worst (the "worst radiograph," the determination of which required a review, but not scoring, of all radiographs), and the time at which the last radiograph was obtained before death or discharge.

Imaging Technique
Frontal chest radiographs were obtained by using standard techniques in each of the 14 hospitals. These techniques included, for example, the use of 75 kV, 4 mAs, and a 180-cm film-to-focus distance for posteroanterior projection and 70 kV, 4 mAs, and a 100-cm film-to-focus distance for anteroposterior projection and the use of a broad tube focus for both projections. Owing to the highly infectious nature of the disease, cross-infection was an important consideration, and inpatient chest radiographs were preferentially acquired with mobile units in the wards (4655 radiographs) rather than in the radiology department (1325 radiographs). Digital images were reviewed by using proprietary viewing stations with 2K monitors.

Image Evaluation
Most of the serial radiographs acquired were single frontal chest radiographs; hence, only these were assessed. The radiographs were reviewed once by experienced radiologists (G.E.A., C.G.C.O., K.T.W., J.S.W.W., A.N.L.S., J.Y.H.H., C.Y.C., H.Y.H., Y.F.C., T.P.W., L.L.Y.L., and A.T.A.; range of experience in interpreting chest radiographs, 5–18 years) working in pairs, and scores were determined by consensus. The radiologists were blinded to the clinical progress or final outcome of the patients.

Results of previous studies have shown that the main radiographic abnormality observed in SARS is lung parenchymal opacification (3–12); this finding was therefore chosen for scoring in this study (Fig 1 ). The method for radiographic scoring used in this study was a simplified version of that used previously in an earlier study (3). On the frontal radiograph, each lung was divided craniocaudally into three equidistant zones (upper, middle, and lower), resulting in a total of six zones to be scored. The extent of opacification was assigned a score on the basis of the visually estimated area of opacification. In the five-point scoring system we used, a score of 0 denotes no opacification; a score of 1, an area of opacification constituting 1%–25% of the area of the lung; a score of 2, an area of opacification constituting 26%–50% of the area of the lung; a score of 3, an area of opacification constituting 51%–75% of the area of the lung; and a score of 4, an area of opacification constituting more than 75% of the area of the lung. Each zone was separately scored for lung parenchymal opacification. The overall score was obtained by summing the scores of the six lung zones. The number of zones involved was obtained by counting the zones with a degree of involvement greater than zero. Two sets of radiographic data were obtained from each radiograph: first, a score based on the percentage area of lung opacification, and second, the number of zones with opacification.



View larger version (155K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a. Selected serial frontal chest radiographs in 44-year-old man with laboratory-confirmed SARS at each of five milestones. (a) On radiograph acquired at admission (first milestone, 3 days after symptom onset), there was an area of opacification in the right lower zone. (b) On radiograph acquired at the start of ribavirin (second milestone, 5 days after symptom onset), a new area of opacification is seen to have developed in the left lower zone. (c) On radiograph acquired at first pulse of corticosteroid (third milestone, 8 days after symptom onset), new involvement of the right upper zone is seen. The other areas of opacification have enlarged and become less well defined. (d) Worst radiograph (fourth milestone, 11 days after symptom onset) shows widespread lung opacification affecting middle and lower zones bilaterally that is worse on the right. (e) Radiograph obtained before hospital discharge (fifth milestone, 24 days after symptom onset) shows that opacification has largely resolved. Reticulations were present in both the middle and lower zones.

 


View larger version (145K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b. Selected serial frontal chest radiographs in 44-year-old man with laboratory-confirmed SARS at each of five milestones. (a) On radiograph acquired at admission (first milestone, 3 days after symptom onset), there was an area of opacification in the right lower zone. (b) On radiograph acquired at the start of ribavirin (second milestone, 5 days after symptom onset), a new area of opacification is seen to have developed in the left lower zone. (c) On radiograph acquired at first pulse of corticosteroid (third milestone, 8 days after symptom onset), new involvement of the right upper zone is seen. The other areas of opacification have enlarged and become less well defined. (d) Worst radiograph (fourth milestone, 11 days after symptom onset) shows widespread lung opacification affecting middle and lower zones bilaterally that is worse on the right. (e) Radiograph obtained before hospital discharge (fifth milestone, 24 days after symptom onset) shows that opacification has largely resolved. Reticulations were present in both the middle and lower zones.

 


View larger version (134K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1c. Selected serial frontal chest radiographs in 44-year-old man with laboratory-confirmed SARS at each of five milestones. (a) On radiograph acquired at admission (first milestone, 3 days after symptom onset), there was an area of opacification in the right lower zone. (b) On radiograph acquired at the start of ribavirin (second milestone, 5 days after symptom onset), a new area of opacification is seen to have developed in the left lower zone. (c) On radiograph acquired at first pulse of corticosteroid (third milestone, 8 days after symptom onset), new involvement of the right upper zone is seen. The other areas of opacification have enlarged and become less well defined. (d) Worst radiograph (fourth milestone, 11 days after symptom onset) shows widespread lung opacification affecting middle and lower zones bilaterally that is worse on the right. (e) Radiograph obtained before hospital discharge (fifth milestone, 24 days after symptom onset) shows that opacification has largely resolved. Reticulations were present in both the middle and lower zones.

 


View larger version (134K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1d. Selected serial frontal chest radiographs in 44-year-old man with laboratory-confirmed SARS at each of five milestones. (a) On radiograph acquired at admission (first milestone, 3 days after symptom onset), there was an area of opacification in the right lower zone. (b) On radiograph acquired at the start of ribavirin (second milestone, 5 days after symptom onset), a new area of opacification is seen to have developed in the left lower zone. (c) On radiograph acquired at first pulse of corticosteroid (third milestone, 8 days after symptom onset), new involvement of the right upper zone is seen. The other areas of opacification have enlarged and become less well defined. (d) Worst radiograph (fourth milestone, 11 days after symptom onset) shows widespread lung opacification affecting middle and lower zones bilaterally that is worse on the right. (e) Radiograph obtained before hospital discharge (fifth milestone, 24 days after symptom onset) shows that opacification has largely resolved. Reticulations were present in both the middle and lower zones.

 


View larger version (151K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1e. Selected serial frontal chest radiographs in 44-year-old man with laboratory-confirmed SARS at each of five milestones. (a) On radiograph acquired at admission (first milestone, 3 days after symptom onset), there was an area of opacification in the right lower zone. (b) On radiograph acquired at the start of ribavirin (second milestone, 5 days after symptom onset), a new area of opacification is seen to have developed in the left lower zone. (c) On radiograph acquired at first pulse of corticosteroid (third milestone, 8 days after symptom onset), new involvement of the right upper zone is seen. The other areas of opacification have enlarged and become less well defined. (d) Worst radiograph (fourth milestone, 11 days after symptom onset) shows widespread lung opacification affecting middle and lower zones bilaterally that is worse on the right. (e) Radiograph obtained before hospital discharge (fifth milestone, 24 days after symptom onset) shows that opacification has largely resolved. Reticulations were present in both the middle and lower zones.

 
Clinical Parameters
The clinical and laboratory parameters used for analysis were retrieved from the Hong Kong Hospital Authority SARS clinical database. These included the serum lactate dehydrogenase (LDH) level, the neutrophil count, and the lymphocyte count. In addition, a calculated ratio of the PaO2 to the fraction of inspired oxygen (FIO2) was included as an indicator of the patients' blood oxygen saturation level. These parameters were chosen because they have been reported to correlate well with the clinical condition of patients with SARS and to be important as prognostic indicators (4,6,14–19). Corresponding radiographic and clinical observations were compiled for the five milestones mentioned above. For each radiographic score for a patient, the corresponding clinical parameters on the same day, if any, were retrieved from the clinical database.

Statistical Analysis
The two sets of radiographic data (radiographic score and number of opacified zones) for all 1373 patients (569 [41.4%] male patients and 804 [58.6%] female patients with a mean age of 41.2 years [range, younger than 1 year to 98 years]) were merged with the patients' demographic, clinical, and outcome data for statistical analysis. The patients were sorted into two clinical outcome groups (patients in whom SARS was fatal [the "fatal" group] and patients who were discharged alive from the hospital [the "discharged" group]), two sex groups, and three age groups (<35 years, 35–64 years, and ≥65 years of age). The following statistical analyses were performed:

1. Profile analysis (by means of a repeated-measures analysis of variance [ANOVA] model) was used to assess differences between the fatal and discharged groups in terms of their radiographic progression profile over the five milestones, with adjustment for sex and age effects and between- and within-subject interactions. Least-squares means adjusted for other effects in the ANOVA model were compared.

2. The Cochran-Armitage trend test was used to assess whether mortality rate increased with radiographic score.

3. Descriptive graphic analysis was used to examine the temporal progression of radiographic scores by plotting the mean day from symptom onset for each milestone against the corresponding mean radiographic score. The slope between milestones was estimated on the basis of the average rate of increase in radiographic score or number of opacified zones per day.

4. Descriptive graphic analysis of the temporal trend of the radiographic score against that of the corresponding clinical parameters was used to assess the radiographic-clinical correlation longitudinally. The Kendall {tau} coefficient was computed to measure the strength of the correlations between the radiographic and clinical observations at each milestone and between the difference in radiographic scores over two consecutive milestones and the corresponding difference in clinical observations. The Kendall {tau} equals 1.0 if all pairs of radiographic-clinical observations are concordant (in the same direction) and –1.0 if all pairs are discordant (in the opposite direction).

Statistical analysis was performed by using statistical software (SAS, version 8.2; SAS Institute, Cary, NC). P < .05 was considered to indicate a statistically significant difference.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Demographic Data
Among the 1462 patients with laboratory-confirmed SARS, sets of serial radiographs were available for scoring for 1373 (93.9%). Not all centers were equipped with electronic film archives, and there were 89 patients (6.1%) whose film records were not retrievable.

The cohort consisted of 569 male patients (41.4%) and 804 female patients (58.6%) with a mean age of 41.2 years (range, younger than 1 year to 98 years). Of these patients, 374 (27.2%) were health care workers or medical students (Table 1). A total of 5980 radiographs were scored in this study (average number of radiographs scored per patient, 4.4).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Demographic Characteristics and Health Care Worker Status of Patients in Whom SARS Was Fatal versus Those Who Were Discharged

 
There were 1212 survivors, including 485 male patients (40.0%) and 727 female patients (60.0%). One hundred sixty-one patients, including 84 male (52.2%) and 77 female (47.8%) patients, died, for a mortality rate of 11.7%. The mean age was 38.4 years (range, 1–91 years) for the survivors and 63.0 years (range, 28–98 years) for patients who died. The mortality rate for males was 14.8% (84 of 569 patients), compared with 9.6% (77 of 804 patients) for females. The mortality rate was 47.5% (84 of 177 patients) for elderly (≥65 years) patients, compared with 11.2% (70 of 625 patients) for the 35–64-year-old age group and 1.2% (seven of 571 patients) for the younger-than-35-year-old age group.

Radiographic Scores at Milestones
On the basis of the dates on which their initial radiographs were acquired, the patients in whom SARS was fatal presented a mean of 3.8 days after symptom onset, marginally but significantly (P = .01) earlier than the survivors (who presented a mean of 4.5 days after symptom onset). One thousand and ninety-two of 1325 patients had an abnormal initial chest radiograph; hence, radiography had a sensitivity of 82.4% in the detection of SARS.

The number of days since symptom onset for the second and third milestones were similar (P = .849 and P = .492, respectively) between the two outcome groups; hence, so were the number of days since symptom onset on which the radiographs were scored (Table 2). Ribavirin therapy was begun around day 6, and therapy with pulsed steroids was begun around day 9. The radiographic appearance of SARS in the patients in whom it was fatal continued to deteriorate until the point at which the worst radiograph was acquired (the fourth milestone) at the end of the 3rd week (mean number of days since symptom onset, 20.3). The survivors reached the point at which the worst radiograph was acquired sooner—around 11.4 days after symptom onset.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Mean Chest Radiographic Scores and Number of Opacified Zones across the Five Milestones in Patients in Whom SARS Was Fatal Versus Those Who Were Discharged

 
According to results of a two-way repeated-measures ANOVA model, the sex effect, the interactions between outcome and sex and age and sex, and the associated between- and within-variable interactions were not statistically significant (P > .05 for all). After we observed these results, we created a reduced model in which the nonsignificant parameters were removed (Table 3). Significant between-subject effects were found for the three age groups (P = .027), the two outcome groups (P < .001), and for the interaction between age and outcome (P < .001). The radiographic scores were also significantly different between milestones (P < .001). Given the results of the ANOVA model in Table 3, the differences between the least-squares mean radiographic scores for the two outcome groups at each milestone, with adjustment for the other model parameters, were statistically significant (P < .05 for all) (Table 2).


View this table:
[in this window]
[in a new window]

  		TABLE 3. Results of Repeated-Measures ANOVA of Radiographic Scores Across Five Milestones

 
Similarly, the mean chest radiograph scores for each age group within each outcome group were compared for each milestone (Table 4). Within the discharged group, the radiographic scores generally increased with age for all five milestones. Using pairwise comparisons, we found that at each milestone, the younger group scored significantly (P < .001) lower than did the middle and elderly age groups but that no difference existed between the middle and elderly age groups except at the start of ribavirin treatment. No such age-related worsening of chest radiographic findings was present in the fatal group. On the contrary, radiographs in elderly patients showed less opacification than did radiographs in patients in the younger groups when the radiographic findings were at their worst and before death (P < .001).


View this table:
[in this window]
[in a new window]

  		TABLE 4. Mean Chest Radiograph Scores for the Three Age Groups across the Five Milestones according to Outcome

 
Radiographic Progression Rate
There appeared to be a linearly increasing trend over the first four milestones for both outcome groups (Fig 2). The two curves appeared to be approximately parallel between the two groups at least up to day 11, which was the point of peak lung opacification in the discharged group. A rough estimate of the rate of radiographic progression (using the gradient between the first and fourth milestones [Fig 2, Table 3]) shows that the radiographic scores increased at a rate of 0.75 and 0.85 per day for the fatal and discharged groups, respectively. In the survivors, chest radiographic findings improved from day 11 onward, as opposed to the continued deterioration seen in the patients in whom SARS was fatal.



View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Graph shows mean chest radiographic (CXR) scores plotted against mean time for each milestone for patients in whom SARS was fatal and those who survived to discharge. The rate of radiographic progression (slope of each curve) for the two outcome groups is roughly parallel over the first 11 days. After this, radiographic scores in the group in whom SARS was fatal continued to deteriorate, whereas the scores in the patients who were discharged improved. The five data points on each curve refer to the five milestones in sequence. The approximate rate of change in radiographic score between the first and fourth milestones (gradient of slope) was calculated as (17.8 – 4.5)/(20.3 – 3.8 days) = 0.75 per day (where 17.8 is the mean total chest radiographic score and 20.3 is the mean number of days since symptom onset) in the patients in whom SARS was fatal and as (7.6 – 2.6)/(11.4 – 4.5 days) = 0.85 per day in the patients who were discharged.

 
Performing the same estimation by using the number of opacified zones instead of the radiographic scores yielded similar results. The rate of increase in the number of zones opacified between milestones was similar—0.19 zone per day for the fatal group and 0.25 zone per day for the discharged group (approximately one extra zone opacified every 4–5 days) for the first four milestones.

Radiographic and Clinical Comparison
The PaO2/FIO2 ratio progression was almost a mirror image (in terms of negative correlation) of the chest radiographic score progression (Fig 3). Like the radiographic score curves, the two PaO2/FIO2 ratio curves showed a parallel continual decrease for the first three milestones followed by a divergence between survivors and patients in whom SARS was fatal. In the survivors, the PaO2/FIO2 ratio improved, while it continued to deteriorate in patients in whom SARS was fatal, until their demise. Similarly, the two lymphocyte count progression curves were approximate mirror images of the respective radiographic score curves (Fig 4).



View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Graph shows mean chest radiographic (CXR) scores and PaO2/FIO2 ratios plotted for five milestones for patients in whom SARS was fatal and those who survived to discharge. The curves for the radiographic scores were roughly mirror images of the respective ones for PaO2/FIO2 ratios. A divergent point between the outcome groups was present at the third milestone. The five data points on each curve refer to the five milestones in sequence. Error bars indicate 95% confidence intervals.

 


View larger version (26K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Graph shows mean chest radiographic (CXR) scores and lymphocyte count plotted for five milestones for patients in whom SARS was fatal and those who survived to discharge. The curves for the radiographic scores were roughly mirror images of the respective ones for lymphocyte count. A divergent point between the outcome groups was present at the third milestone. The five data points on each curve refer to the five milestones in sequence. Error bars indicate 95% confidence intervals.

 
Both the serum LDH level and the neutrophil count had progression curves that approximated those for the radiographic scores (positive correlation). For the two outcome-based curves for each parameter, there was an approximately parallel increasing trend for the first three milestones, followed by a divergence between survivors and patients in whom SARS was fatal (Figs 5, 6).



View larger version (26K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5. Graph shows mean chest radiographic (CXR) score and LDH level (as the ratio of LDH to its upper reference limit) plotted for five milestones for patients in whom SARS was fatal and those who survived to discharge. The curves for the radiographic scores showed temporal trends that were similar to the ones shown by the respective curves for LDH level. A divergent point between the outcome groups was present at the third milestone. The five data points on each curve refer to the five milestones in sequence. Error bars indicate 95% confidence intervals.

 


View larger version (26K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6. Graph shows mean chest radiographic (CXR) scores and neutrophil count plotted for five milestones for patients in whom SARS was fatal and those who survived to discharge. The curves for the radiographic scores showed temporal trends that were similar to the ones shown by the respective curves for neutrophil count. A divergent point between the outcome groups was present at the third milestone. The five data points on each curve refer to the five milestones in sequence. Error bars indicate 95% confidence intervals.

 
On the basis of results of Kendall {tau} correlation coefficient analysis (Table 5), the PaO2/FIO2 ratio and LDH level had a good correlation with the radiographic scores (overall coefficient, –0.44 [P < .001] and +0.41 [P < .001], respectively). The neutrophil and lymphocyte counts correlated moderately with the radiographic scores (overall coefficient, +0.22 [P < .001] and –0.24 [P < .001], respectively).


View this table:
[in this window]
[in a new window]

  		TABLE 5. Correlations between Radiographic and Clinical Data at Each Milestone

 
Kendall {tau} correlation coefficients were also computed to assess the strength of correlation of the between-milestone differences for radiographic and clinical parameters (Table 6). All correlation coefficients (–0.11 to –0.32 for lymphocyte count and PaO2/FIO2 ratio; +0.17 to +0.38 for neutrophil count and LDH level) were significant at P < .001, except that for radiographic score and neutrophil count between the first two milestones. The moderate correlation again reflects the concordance in progression between the radiographic score and the clinical parameters during the course of the disease.


View this table:
[in this window]
[in a new window]

  		TABLE 6. Correlations between the Difference in Radiographic Data between Two Milestones versus the Corresponding Difference in Clinical Data

 
Risk of Death
The mortality rate increased with radiographic score for the first four milestones in a monotonic increasing trend (Fig 7). Results of the Cochran-Armitage test for trend revealed significance for each milestone (P < .001 for all). By using cutoffs in mortality rate at 10% increments, mortality zones were seen to shift in a time-related fashion. For the radiographic scores, the same extent of radiographic opacification conferred a different mortality risk at different times. Except in the extremely severe cases (ie, those in which the overall chest radiograph score was greater than 20–24), the extent of opacification was associated with a mortality risk that decreased with time.



View larger version (84K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7. Chart shows that an increasing and linear trend for the mortality rate was seen for the radiographic scores. When 10% mortality rate increments are arbitrarily used as cutoffs for mortality rate zones, the "safety margin" is seen to shift with time. More extensive lung opacification indicated a worse prognosis earlier in the disease than later. N = number of patients. Green indicates mortality rates between 0% and 10%; yellow, mortality rates between 10.1% and 20%; pink, mortality rates between 20.1% and 30%; and blue, mortality rates between 30.1% and 100%.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
According to the results of this study, patients in Hong Kong underwent their first radiographic examinations (at the time of presentation to the hospital) around the 4th day after the onset of symptoms. It appears that the sensitivity of the radiography performed at this time was around 82.4%. This sensitivity is in line with the results of investigators in other geographic regions (60%–100%) (6,8,12,20,21). At the time of the outbreak, it was important that a more sensitive imaging modality, computed tomography (CT), be used in cases with a high index of suspicion but unremarkable radiographs (22). The Centers for Disease Control and Prevention of the United States currently recommends that in these situations patients undergo a CT examination on day 6 after symptom onset for community-acquired disease (23).

Age and Sex Differences
Previous reports by investigators from different countries indicated that older patients were sicker and had a higher mortality rate (6,15,16,24); similar results were seen in Hong Kong. For our patients, the mortality rate increased from 1.2% (for patients younger than 35 years) to 11.2% (for patients between 35 and 64 years of age) to 47.5% (for patients 65 years or older).

By classifying our patients into three age groups, we found an age-related trend for radiographic scores among the patients who were discharged. Among patients who were discharged, those in the middle and older age groups had worse findings on radiographs than did those younger than 35 years. This age-related worsening of findings on chest radiographs was not present in the group of patients in whom SARS was fatal. On the contrary, the elderly patients in the latter group had better-appearing radiographs (compared with the radiographs in the younger age groups) at the point of acquisition of the worst radiograph and just before death. This suggests that the respiratory failure may not have been the only cause of death in these patients and that comorbidity may have played a part (4,25).

Most countries had a higher proportion of female patients with SARS (1), but male patients have been shown to have a higher mortality risk (26). Results of the present study indicate that there was no difference in radiographic progression between the sexes overall and within each outcome group. This similarity in radiographic scores between the two sexes suggests that the lungs' reaction to this infection, as reflected by radiographic findings, was similar for male and female patients within each outcome group. Patients of both sexes in the same outcome group fared similarly in terms of the radiographic appearance, despite the different mortality risks.

Radiographic Progression
The mean number of days after symptom onset for the first three milestones were similar between the two outcome groups. These similarities suggest that the timing of presentation and the treatment regimens (the days of commencing ribavirin and pulsed corticosteroid therapy) for both outcome groups were comparable.

For both outcome groups, the rate of radiographic progression was similar for the first 11 days. The extent of opacification increased by approximately one zone every 4–5 days. Therefore, the rate of radiographic deterioration for the first 11 days could not be used to estimate outcome. It was only after these 11 days when the patients who survived deviated from the patients in whom SARS was fatal by diverging away from the track of radiographic deterioration.

On the other hand, from the day of acquisition of the first radiograph to day 11, there was a consistent margin of difference between the patients who survived and those who eventually died. Results of another study involving the analysis of daily radiographic scores in 313 patients with SARS (27) indicated that findings on the radiograph acquired on day 7 were the earliest and most accurate in predicting prognosis.

The turning point on day 11 observed in the present study correlates well with the peak SARS viral load and the beginning of seroconversion, both of which occur around day 10 (9). From then on, lung damage may be due to immunopathologic damage as a result of an overexuberant host response rather than to viral replication (9). It is believed that this immunopathologic response is one of the main causes of death in SARS (9,28).

Radiographic and Clinical Comparison
The initially parallel and subsequently divergent progression curves between patients who survived and those in whom SARS was fatal that were seen in terms of radiographic scores were similarly seen when the corresponding curves were plotted for clinical and laboratory parameters, such as serum LDH level and neutrophil count, PaO2/FIO2 ratio, and lymphocyte count. The former two parameters correlated positively with the radiographic scores and showed a temporal trend similar to that of the radiographic course. The latter two parameters correlated negatively and approximately mirrored the radiographic course. Radiographic scores could therefore be used as an objective parameter for the patient's clinical condition.

The characteristic early laboratory findings of elevated serum LDH and lymphopenia in SARS have previously been described (6,7,12,13,24,29). A high peak LDH level has been identified as a predictor for intensive care unit admission, mechanical ventilation, and mortality (4,14–17). LDH is considered to be an indicator of pulmonary tissue destruction, and a high level reflects severe disease during acute SARS. The LDH level also correlates with lung changes suggestive of fibrosis at 6-month follow-up after SARS (18).

Results of previous studies have also indicated that a high neutrophil count (6,14,17,30), a low PaO2/FIO2 ratio or other parameters of hypoxemia (15,16,19), and a low lymphocyte count (16,30) are associated with poor prognosis. In this study, the progression of all of these clinical and laboratory parameters correlated well with the radiographic course, adding further weight to the radiograph's utility in prognostic indication.

Risk of Death
There was a monotonic increasing trend in the mortality rate with an increase in the radiographic score and the number of zones opacified. This suggests a "dose-response" relationship between the extent of lung opacification and the mortality rate. This predictive ability of chest imaging results is in line with the results of previous studies (3,5,16,27,31). In the present study, by using this intramilestone increasing trend together with the progressive radiographic deterioration through the first four milestones, a time-dependent shifting safety margin could be demonstrated. Such safety margins could potentially be adapted for monitoring patient progress and response to therapy at different points in time during the initial course of the disease. This may be particularly useful when new treatment regimens are used in the future.

There were limitations to this study. Owing to limited resources, it was not possible to score every serial radiograph for all 1373 patients. Representative milestones during the patient's acute disease were therefore chosen. The fact that data were missing for radiographic and clinical parameters was unavoidable in this retrospective analysis because such investigations had not been conducted for each patient at the evaluated time points. In this retrospective study, the actual predictive ability of the radiographic findings could not be tested. In addition, by dividing the lung into six zones for scoring, region-region interaction may have affected the total radiographic score. We believe this to be a trade-off for using a simplified scoring system.

Conclusion
Chest radiographic opacification scores correlated well with clinical parameters of disease progression. For the first 11 days, patients who survived and those who eventually died had similar rates of radiographic deterioration. Patients in whom SARS was fatal had a consistent margin of more extensive lung opacification than did survivors. Therefore, it appears that the final outcome of SARS may be predicted by the initial extent of radiographic lung opacification.


   ACKNOWLEDGMENTS
 
The authors thank members of the Hospital Authority SARS Collaborative Group for their advice and support on the use of the SARS clinical database. The authors also thank the Statistics and Research Unit of the Hospital Authority Head Office for their help in preparing this article.


   FOOTNOTES
 

Abbreviations: ANOVA = analysis of variance • FIO2 = fraction of inspired oxygen • LDH = lactate dehydrogenase • SARS = severe acute respiratory syndrome

See Materials and Methods for pertinent disclosures.

Author contributions: Guarantors of integrity of entire study, G.E.A., K.T.W., E.L.H.T., J.C.K.C., A.T.A.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, G.E.A., J.C.K.C.; clinical studies, G.E.A., C.G.C.O., K.T.W., J.S.W.W., J.Y.H.H., C.Y.C., H.Y.H.H., Y.F.C., T.P.W., L.L.Y.L., J.C.K.C., A.T.A.; statistical analysis, G.E.A., E.L.H.T., A.N.L.S., J.C.K.C.; and manuscript editing, G.E.A., C.G.C.O., K.T.W., E.L.H.T., A.N.L.S., Y.F.C., J.C.K.C., A.T.A.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. World Health Organization for Severe Acute Respiratory Syndrome, Communicable Disease Surveillance & Response. Summary of probable SARS cases with onset of illness from 1 November 2002 to 31 July 2003. (Based on data as of December 31, 2003.) http://www.who.int/csr/sars/country/table2004_04_21/en/index.html. Accessed August 8, 2004.
  2. World Health Organization case definition for severe acute respiratory syndrome. http://www.who.int/csr/sars/casedefinition/en/. Revised May 1, 2003. Accessed July 30, 2004.
  3. Wong KT, Antonio GE, Hui DS, et al. Severe acute respiratory syndrome: radiographic appearances and pattern of progression in 138 patients. Radiology 2003;228:401–406.[Abstract/Free Full Text]
  4. Choi KW, Chau TN, Tsang O, et al. Outcomes and prognostic factors in 267 patients with severe acute respiratory syndrome in Hong Kong. Ann Intern Med 2003;139:715–723.[Abstract/Free Full Text]
  5. Ooi CG, Khong PL, Lam B, et al. Severe acute respiratory syndrome: relationship between radiologic and clinical parameters. Radiology 2003;229:492–499.[Abstract/Free Full Text]
  6. Booth CM, Matukas LM, Tomlinson GA, et al. Clinical features and short-term outcomes of 144 patients with SARS in the greater Toronto area. JAMA 2003;289:2801–2809.[Abstract/Free Full Text]
  7. Poutanen SM, Low DE, Henry B, et al. Identification of severe acute respiratory syndrome in Canada. N Engl J Med 2003;348:1995–2005.[Abstract/Free Full Text]
  8. Grinblat L, Shulman H, Glickman A, et al. Severe acute respiratory syndrome: radiographic review of 40 probable cases in Toronto, Canada. Radiology 2003;228:802–809.[Abstract/Free Full Text]
  9. Peiris JS, Chu CM, Cheng VC, et al. Clinical progression and viral load in a community outbreak of coronavirus-associated SARS pneumonia: a prospective study. Lancet 2003;361:1767–1172.[CrossRef][Medline]
  10. Bitar R, Weiser WJ, Avendano M, et al. Chest radiographic manifestations of severe acute respiratory syndrome in health care workers: the Toronto experience. AJR Am J Roentgenol 2004;182:45–48.[Abstract/Free Full Text]
  11. Muller NL, Ooi GC, Khong PL, et al. Severe acute respiratory syndrome: radiographic and CT findings. AJR Am J Roentgenol 2003;181:3–8.[Abstract/Free Full Text]
  12. Vu HT, Leitmeyer KC, Le DH, et al. Clinical description of a completed outbreak of SARS in Vietnam, February–May 2003. Emerg Infect Dis 2004;10:334–338.[Medline]
  13. World Health Organization. Use of laboratory methods for SARS diagnosis. http://www.who.int/csr/sars/labmethods/en/#confirm. Accessed March 2, 2004.
  14. Lee N, Hui DS, Wu A, et al. A major outbreak of severe acute respiratory syndrome in Hong Kong. N Engl J Med 2003;348:1986–1994.[Abstract/Free Full Text]
  15. Lew TW, Kwek TK, Tai D, et al. Acute respiratory distress syndrome in critically ill patients with severe acute respiratory syndrome. JAMA 2003;290:374–380.[Abstract/Free Full Text]
  16. Zou Z, Yang Y, Chen J, et al. Prognostic factors for severe acute respiratory syndrome: a clinical analysis of 165 cases. Clin Infect Dis 2004;38:483–489.[CrossRef][Medline]
  17. Tsui PT, Kwok ML, Yuen H, Lai ST. Severe acute respiratory syndrome: clinical outcome and prognostic correlates. Emerg Infect Dis 2003;9:1064–1069.[Medline]
  18. Wong KT, Antonio GE, Hui DS, et al. Severe acute respiratory syndrome: thin-section computed tomography features, temporal changes, and clinicoradiologic correlation during the convalescent period. J Comput Assist Tomogr 2004;28:790–795.[CrossRef][Medline]
  19. Ooi CG, Khong PL, Ho JC, et al. Severe acute respiratory syndrome: radiographic evaluation and clinical outcome measures. Radiology 2003;229:500–506.[Abstract/Free Full Text]
  20. Wang TL, Jang TN, Huang CH, et al. Establishing a clinical decision rule of severe acute respiratory syndrome at the emergency department. Ann Emerg Med 2004;43:17–22.[CrossRef][Medline]
  21. Puthucheary J, Lim D, Chan I, Chay OM, Choo P. Severe acute respiratory syndrome in Singapore. Arch Dis Child 2004;89:551–556.[Abstract/Free Full Text]
  22. Wong KT, Antonio GE, Hui DS, et al. Thin-section CT of severe acute respiratory syndrome: evaluation of 73 patients exposed to or with the disease. Radiology 2003;228:395–400.[Abstract/Free Full Text]
  23. Centers for Disease Control and Prevention. Updated interim U.S. case definition for Severe Acute Respiratory Syndrome (SARS). http://www.cdc.gov/ncidod/sars/casedefinition.htm. Accessed January 8, 2004.
  24. Lim HK, Liu CP, Huang FY, et al. Severe acute respiratory syndrome in a medical center in Taipei. J Microbiol Immunol Infect 2003;36:161–168.[Medline]
  25. Gomersall CD, Joynt GM, Lam P, et al. Short-term outcome of critically ill patients with severe acute respiratory syndrome. Intensive Care Med 2004;30:381–387.[CrossRef][Medline]
  26. Karlberg J, Chong DS, Lai WY. Do men have a higher case fatality rate of severe acute respiratory syndrome than women do? Am J Epidemiol 2004;159:229–231.[Abstract/Free Full Text]
  27. Antonio GE, Wong KT, Tsui EL, et al. Chest radiograph scores as potential prognostic indicators in severe acute respiratory syndrome (SARS). AJR Am J Roentgenol 2005;184:734–741.[Abstract/Free Full Text]
  28. Nicholls JM, Poon LL, Lee KC, et al. Lung pathology of fatal severe acute respiratory syndrome. Lancet 2003;361:1773–1778.[CrossRef][Medline]
  29. Tsang KW, Ho PL, Ooi GC, et al. A cluster of cases of severe acute respiratory syndrome in Hong Kong. N Engl J Med 2003;348:1977–1985.[Abstract/Free Full Text]
  30. Wong RS, Wu A, To KF, et al. Haematological manifestations in patients with severe acute respiratory syndrome: retrospective analysis. BMJ 2003;326:1358–1362.[Abstract/Free Full Text]
  31. Hui JY, Cho DH, Yang MK, et al. Severe acute respiratory syndrome: spectrum of high-resolution CT findings and temporal progression of the disease. AJR Am J Roentgenol 2003;181:1525–1538.[Free Full Text]



This article has been cited by other articles:


Home page
 Clin. Microbiol. Rev.Home page
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]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Antonio, G. E.
Right arrow Articles by Ahuja, A. T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Antonio, G. E.
Right arrow Articles by Ahuja, A. T.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE