Cardiopulmonar y Imaging Original Research

Wong et al.
Radiography of Respiratory Syncytial Virus Infection

Cardiopulmonary Imaging
Original Research

Initial Radiographic Features

Downloaded from www.ajronline.org by 180.245.204.167 on 11/28/17 from IP address 180.245.204.167. Copyright ARRS. For personal use only; all rights reserved

as Outcome Predictor of Adult

Respiratory Syncytial Virus
Respiratory Tract Infection
Simon S. M. Wong1 OBJECTIVE. The purposes of this study were to evaluate the chest radiographic features
Janice W. L. Yu1 of adult patients hospitalized for respiratory syncytial virus respiratory tract infections and to
Ka Tak Wong1 assess whether initial chest radiographic findings help predict clinical outcome.
Nelson Lee2 MATERIALS AND METHODS. All adult patients hospitalized from January 2009 to
Grace C. Y. Lui2 December 2011 with laboratory-confirmed respiratory syncytial virus infection were includ-
ed in the study. Patient clinical data and admission chest radiographs were retrospectively re-
Paul K. S. Chan 3
viewed. Adverse outcomes included need for supplemental oxygen, need for assisted ventila-
Anil T. Ahuja1 tion, and death.
Wong SSM, Yu JWL, Wong KT, et al. RESULTS. Of 285 patients (mean age, 74 16 years) included, 199 (69.8%) had abnor-
mal chest radiographic findings: 49.5% (141/285) had acute changes, and 47.7% (136/285) had
chronic changes. Consolidation (68/141 [48.2%]) and ground-glass opacity (57/141 [40.4%])
were the predominant types of acute changes and were most common in unilateral single-low-
er-zone involvement. Consolidation, ground-glass opacity, and chronic changes occurred with
significantly higher frequency in patients with adverse outcomes. The presence of acute (odds
ratio, 3.6) and chronic (odds ratio, 2.2) changes were independent risk factors for mortality.
CONCLUSIONS. A large proportion of adult patients hospitalized with respiratory
syncytial virus respiratory tract infection had changes on initial chest radiographs. Consoli-
dation or ground-glass opacity in a unilateral single-lower-zone distribution were the most
common findings. The presence of acute and chronic radiographic lung changes was associ-
Keywords: adult respiratory tract infection, human
respiratory syncytial virus, radiography, thoracic, ated with adverse outcomes.
treatment outcome

DOI:10.2214/AJR.13.11356
Received June 2, 2013; accepted after revision
January1,2014.
Presented in part at the 2013 European Congress of
Radiology, Vienna, Austria.
1
Department of Imaging and Interventional Radiology,
The Chinese University of Hong Kong, Prince of Wales
Hospital, 30-32 Ngan Shing St, Hong Kong, SAR. Address
correspondence to S. S. M. Wong (wsm652@ha.org.hk).
2
Department of Medicine and Therapeutics, Prince of
Wales Hospital, The Chinese University of Hong Kong,
Hong Kong, SAR.
3
Department of Microbiology, Prince of Wales Hospital,
The Chinese University of Hong Kong, Hong Kong, SAR.
This article is available for credit.
AJR 2014; 203:280286
0361803X/14/2032280
American Roentgen Ray Society
R
espiratory syncytial virus (RSV)
is a common etiologic agent in
childhood respiratory tract in-
fection, causing hospitalizations
and deaths. In adults, RSV infection is asso-
ciated with a wide range of clinical syn-
dromes, including upper respiratory illness,
acute bronchitis, exacerbation of asthma, and
chronic obstructive pulmonary disease
(COPD) and pneumonia. In a prospective
study by Falsey et al. [1], RSV infection was
an important illness in adults who were el-
derly and at high risk. It accounted for 10.6%
of hospitalizations for pneumonia, 11.4% for
COPD, 5.4% for congestive heart failure, and
7.2% for asthma. Evidence of the importance
of RSV infection in community-acquired
pneumonia and acute exacerbation of COPD
has been reported worldwide [2, 3].
Most studies of the radiographic appear-
ances of RSV-related infection of the lower
respiratory tract have been focused on pedi-
atric patients [4, 5]. Despite the recognized
importance of RSV-related respiratory tract
infection in adults, the radiographic features
and the potential prognostic value of changes
on chest radiographs have not been well doc-
umented in the literature. In addition, RSV
vaccine and therapeutics are currently under-
going active development [68]. Therefore
knowledge of the chest radiographic patterns
of RSV infection is essential not only to as-
sist diagnosis and prognostication but also to
conduct future clinical trials. The aims of our
study were to evaluate the chest radiograph-
ic features of adult patients hospitalized for
RSV respiratory tract infections and to assess
whether admission chest radiographic find-
ings are predictive of clinical outcome.
Materials and Methods
Patient Collection
This study was approved by the institutional
review board. The requirement for informed con-
sent was waived owing to the retrospective and
observational nature of the study. All subjects

280 AJR:203, August 2014

 Radiography of Respiratory Syncytial Virus Infection
were consecutively registered adult patients ( 18
years) hospitalized for laboratory-confirmed RSV
infection during the period January 2009 to December
2011 who underwent frontal chest radiography at
admission [9]. This study was conducted at a sin-
gle university-affiliated acute care hospital man-
Downloaded from www.ajronline.org by 180.245.204.167 on 11/28/17 from IP address 180.245.204.167. Copyright ARRS. For personal use only; all rights reserved
aged by a hospital authority.
Patients presenting with symptoms of acute fe-
brile respiratory illnesses (e.g., fever, cough, spu-
tum production, shortness of breath, wheezing)
were admitted if potentially serious medical con-
ditions developed, if the chronic illness became
exacerbated, or the symptoms were considered
impossible to manage at home. All these hospital-
ized patients underwent routine testing for RSV
infection immediately after admission. RSV in-
fection was diagnosed when a positive result of an
immunofluorescence assay of nasopharyngeal as-
pirate was obtained. If nasopharyngeal aspiration
could not be performed, a nasopharyngeal flocked
swab was collected for testing. Sputum samples
were routinely collected for bacterial culture.
Blood culture was also performed if signs of sep-
sis were present. All patients routinely underwent
frontal chest radiography at admission. Support-
ive treatments, including oxygen therapy and as-
sisted ventilation, were administered if indicated
(indications described in Study Design). Patients
whose condition improved with supportive treat-
ment were discharged to home or to a convales-
cent hospital. Additional investigations and treat-
ment were considered individually for patients
whose condition did not improve. No routine clin-
ic or ward follow-up was conducted for patients
whose condition was deemed fit for discharge. Pa-
tients with chronic illnesses underwent follow-up
in the respective specialty departments.
Study Design
Clinical notes and charts of all patients included
in the study during the hospital stay were reviewed
for demographic characteristics (including age, sex,
smoking status); background diseases and condi-
tions (including COPD, asthma, congestive heart
failure, ischemic heart disease, immunocompro-
A B
AJR:203, August 2014 281
Fig. 158-year-old man with respiratory syncytial
virus infection. Initial posteroanterior chest
radiograph shows focal consolidation with air
bronchogram in lower zone of right lung. Patient
received supplemental oxygen during hospital stay,
and his condition gradually improved with treatment.
mised state); time from symptom onset to hospi-
tal admission; bacterial culture result (community-
acquired bacterial coinfection defined as positive
culture result of a bacterial pathogen from sputum
or blood samples obtained at admission or within
2 days of admission); treatment received during
hospitalization (including need for supplemental
oxygen and assisted ventilation); presence of re-
spiratory complications (e.g., pneumonia, acute
bronchitis, COPD exacerbation, asthma attack);
and outcome (discharge or death).
The main adverse outcome measures were need
for supplemental oxygen therapy, need for assist-
ed ventilation, and death. Patients were treated ac-
cording to a protocol. Antiviral treatment was not
given because there is no effective drug for RSV
infection, unlike influenza viral infection. Supple-
mental oxygen therapy was given if oxygen sat-
uration was less than 92% on room air. Assisted
ventilation (invasive or noninvasive) was required
if patients had persistent respiratory failure de-
spite supplemental oxygen therapy.
Chest Radiograph Evaluation
Chest radiographs were obtained for clinical
suspicion of pneumonia (e.g., fever, cough, dys-
Fig. 249-year-old man with respiratory syncytial
virus infection. Magnified image of right lower zone
of posteroanterior chest radiograph shows multiple
small (5 mm) uniformly sized soft-tissue nodules.
pnea, desaturation, positive auscultatory find-
ings). All chest radiographs were obtained with-
in 24 hours of admission. Posteroanterior chest
radiographs were acquired with a digital radiog-
raphy system (Definium 8000, GE Healthcare),
and portable anteroposterior chest radiographs
were obtained with a computed radiography sys-
tem (CR85-X, Agfa HealthCare; or FCR XG5000,
FujiFilm). The chest radiographs were reviewed
in consensus by two independent radiologists (5
years experience in interpretation of chest ra-
diographs) blinded to all clinical data and pa-
tient outcome. Whenever a consensus could not
be reached, a third radiologist (15 years of ex-
perience) was consulted. The review of chest ra-
diographs was performed at a dedicated PACS
workstation (Carestream PACS version 11.0,
Carestream Health). Previous chest radiographs
were used for comparison whenever available.
The chest radiographic findings were first clas-
sified as normal or abnormal. Abnormalities were
further classified as acute or chronic lung chang-
es. Abnormal findings detected on previous radio-
graphs, if available for comparison, were classified
as chronic lung changes, and any new abnormal-
ities were considered acute lung changes. When
Fig. 3Focal versus multifocal distribution of lung
parenchymal abnormalities.
A, 83-year-old woman with focal respiratory
syncytial virus infection. Posteroanterior chest
radiograph shows single zonal involvement. Ground-
glass opacity is evident in right upper zone.
B, 84-year-old man with multifocal respiratory
syncytial virus infection. Posteroanterior chest
radiograph shows multifocal consolidations in right
middle and lower and left lower zones.
 Wong et al.

Fig. 4Central versus peripheral distribution of lung

parenchymal abnormalities.
A, 69-year-old woman with central involvement of
respiratory syncytial virus infection. Posteroanterior
chest radiograph shows centrally located airspace
opacity in right lower zone silhouetting right
heart border in keeping with right middle lobe
Downloaded from www.ajronline.org by 180.245.204.167 on 11/28/17 from IP address 180.245.204.167. Copyright ARRS. For personal use only; all rights reserved

consolidation.
B, 97-year-old man with peripheral involvement of
respiratory syncytial virus infection. Posteroanterior
chest radiograph shows small focal areas of
peripherally located consolidation in bilateral lower
zones. Multiple small nodules also are evident in right
upper zone.

A B

previous radiographs were unavailable for com-
parison, the reviewers classified the radiographic
findings as acute or chronic lung changes on the
basis of morphologic features and published de-
scriptions of viral pneumonia in adults [10]. Com-
mon conditions considered chronic lung changes
included pulmonary fibrotic changes secondary to
previous tuberculosis, emphysema, or bronchiec-
tasis. Radiographic abnormalities determined to
be acute changes were then localized to the lung
parenchyma, pleura, hila, or mediastinum and
were further characterized [11, 12].
Evaluation of lung parenchymaThe lung pa-
renchyma and airways were evaluated for consoli-
dation, ground-glass opacity, increased peribron-
chial markings, nodular opacity, reticular opacity,
and hyperinflation [13]. Consolidation was de-
fined as an area of homogeneous increase in pul-
monary parenchymal attenuation that obscured
the margins of vessels and airways with or with-
out air bronchogram (Fig. 1). Ground-glass opac-
ity was defined as an area of increased hazy opac-
ity within which definition of lung structures was
preserved. A diagnosis of nodular opacity was
made when a focal round opacity was present (Fig.
2). Reticular opacity was defined as linear opaci-
ties forming a meshlike pattern. Increased peri-
bronchial marking was defined as coarse linear
markings radiating from the hila into the lungs.
Hyperinflation was defined as the presence of a
depression of the diaphragm on frontal or lateral
chest radiographs. The diaphragm was considered
depressed if it was projected at a level lower than
the sixth anterior rib on frontal chest radiographs
and had a straight (not domed) slope [11, 14].
The anatomic distributions of the lung paren-
chymal abnormalities were characterized as unilat-
eral or bilateral involvement; focal (a single focus
of abnormality) (Fig. 3A), multifocal (more than
one focus of abnormality) (Fig. 3B), or diffuse; in-
volvement of the upper, middle, or lower lung zones
(division made with imaginary horizontal lines tra-
versing the lungs at levels one-third and two-thirds
the vertical distance between the lung apices and
diaphragmatic domes); central ( 4 cm from the
hila) (Fig. 4A), peripheral (Fig. 4B), or mixed dis-
tribution; and extent of parenchymal involvement
estimated by evaluation of the percentage of lung
involved by the radiographic abnormalities.
Evaluation of pleural abnormalityThe pleura
was evaluated for the presence of pleural effusion.
Evaluation of mediastinum and hilaThe hila
and mediastinum were evaluated for lymphadenop-
athy. The presence of hilar lymphadenopathy was
defined as identification of increased size and opac-
ity of the hilum and a lobulated hilar contour and
obscuration of the interlobar artery. The presence
of mediastinal lymphadenopathy was inferred by
identification of widening of one or more medias-
tinal contours at known anatomic lymph node sites.

TABLE 1: Demographic Characteristics and Background Conditions of 285 Patients With Respiratory Syncytial Virus Infection
Requirement for Supplemental Requirement for Assisted
Oxygen Therapy (n = 282)a Ventilation (n = 282)a Death (n = 285)
Characteristic No (n = 86) Yes (n = 196) p No (n = 247) Yes (n = 35) p No (n = 250) Yes (n = 35) p
Age (y) 68 19 77 13 < 0.001 74 16 74 15 NS 73.7 16 76.5 18 NS
Sex (no.) NS NS NS
Men 41 (48) 104 (53) 125 (51) 20 (57) 131 (52) 17 (49)
Women 45 (52) 92 (47) 122 (49) 15 (43) 119 (48) 18 (51)
Congestive heart failure 9 (10) 41 (21) 0.04 45 (18) 5 (14) NS 45 (18) 5 (14) NS
Chronic obstructive 5 (9) 66 (34) < 0.001 56 (23) 15 (43) 0.02 60 (24) 13 (37) NS
pulmonary disease
Asthma 3 (3) 20 (10) NS 21 (9) 2 (6) NS 22 (9) 1 (3) NS
Diabetes mellitus 26 (30) 43 (22) NS 60 (24) 9 (26) NS 60 (24) 9 (26) NS
Ischemic heart disease 3 (3) 37 (19) < 0.001 35 (14) 5 (14) NS 32 (13) 8 (23) NS
Immunocompromised 16 (19) 27 (14) NS 34 (14) 9 (26) NS 35 (14) 8 (23) NS
Smokerb 4 (5) 18 (9) NS 16 (7) 6 (17) 0.04 19 (8) 3 (9) NS
NoteExcept for age (mean SD), data are raw numbers with percentages in parentheses. NS = not significant.
aData on three patients were missing.
bSmoking status was missing for two patients.

282 AJR:203, August 2014

 Radiography of Respiratory Syncytial Virus Infection

Statistical Analysis
All metric data were summarized and dis-
played as mean SD. Comparison of parametric
variables between groups was performed by Stu-
dent t test. Frequencies of categoric variables were
compared by use of the Fisher exact test (two cat-
Downloaded from www.ajronline.org by 180.245.204.167 on 11/28/17 from IP address 180.245.204.167. Copyright ARRS. For personal use only; all rights reserved
49.5% (141/285) had acute lung changes;
and 47.7% (136/285) had chronic chang-
es, including emphysema, fibrosis related
to previous pulmonary tuberculosis, and
bronchiectasis. Table 2 presents the pat-
tern of radiologic findings on chest radio-
Relation Between Radiologic Abnormalities and
Bacterial Coinfection and Congestive Heart Failure
Thirty-six of 285 patients (12.6%) had
culture results positive for community-ac-
quired bacterial coinfection (35 specimens
from sputum, one from blood). The percent-

egories) and chi-square test (multiple categories). graphs obtained at admission for patients ages of bacterial coinfection were not statis-
Correlation was tested with the Spearman correla- with detectable acute lung changes. The tically different (p= 0.22) between patients
tion test. Analysis of contribution of independent most commonly encountered lung paren- with consolidation (9/68 [13.2%]) and those
risk factors to the outcome was assessed by binary chymal abnormalities were consolidation without (27/217 [12.4%]). This was also true
logistic regression. In all analyses, statistical sig- (68/141 [48.2%]) (Fig. 1) and ground-glass for patients with ground-glass opacity, 11 of
nificance was considered p< 0.05. All statistical opacity (57/141 [40.4%]). Other abnormali- 57 of whom (19.3%) had coinfection, com-
analysis was performed with SPSS for Microsoft ties (such as reticular opacity, nodular opac- pared with 25 of 228 patients (11.0%) without
Windows software (version 19.0, IBM SPSS). ity, and increased peribronchial markings) ground-glass opacity (p= 0.229). Spearman
were present in smaller percentages (Table correlation analysis also showed no signifi-
Results 2). One third of patients with acute lung cant correlation between coinfection and pres-
Patients changes had pleural effusion. No patient ence of consolidation or ground-glass opacity.
During the study period, 288 adult patients had intrathoracic lymphadenopathy.
TABLE 3: Distribution of Lung
were admitted for RSV infection; 286 patients Among the 141 patients with acute lung
Parenchymal Radiographic
underwent frontal chest radiography at ad- changes, 31 patients had changes that were Changes in 110 Adults With
mission. One patient was excluded because not zonally distributed, including peri- Respiratory Syncytial Virus
a lung mass was found on the admission ra- bronchial thickening, pleural effusion, and Infection and Acute Lung
diograph and persisted after resolution of the hyperinflation; 110 patients had changes Parenchymal Changes
chest infection and was therefore considered that could be zonally classified. Table 3
Distribution No. of Patients
an incidental finding. Therefore, the records shows the distribution of acute lung paren-
of 285 patients (148 men [51.9%], 137 wom- chymal abnormality. Most of the patients Side of Involvement
en [48.1%]; mean age, 74 16 years; range, (90/110 [81.8%]) had unilateral lung opac- Unilateral 90 (81.8)
18103 years) were analyzed. The mean du- ity, most frequently affecting a single zone
Bilateral 20 (18.2)
ration between symptom onset and admission (60/110 [54.5%]) with lower zone predom-
was 2.9 3.0 days. The mean length of hospi- inance. The lung periphery was more com- Focality of involvement
talization was 14 16 days. monly involved (44/110 [40.0%]), a mixed Focal 71 (64.5)
Lower respiratory complications developed central and peripheral location was also Multifocal 36 (32.7)
in 221 patients (77.5%). The complications in- frequently observed (43/110 [39.1%]). Ap-
Diffuse 3 (2.7)
cluded pneumonia, acute bronchitis, COPD proximately two thirds of patients (66.4%)
exacerbation, and asthma attack. Thirty-five had less than 10% parenchymal involve- Lung zone involved
patients (12.3%) died of RSV infection. Three ment, and approximately one third (31.8%) Right upper 17 (15.5)
patients had missing data in the analysis of had 1050% involvement. Right middle 31 (28.2)
oxygen and ventilation requirements; 196 of
Right lower 67 (60.9)
the other 282 patients (69.5%) needed supple- TABLE 2: Radiographic Changes in
mental oxygen, and 35 patients (12.4%) need- 141 Adults With Respiratory Left upper 9 (8.2)
ed assisted ventilation. Two patients had no Syncytial Virus Infection Left middle 21 (19.1)
available data on smoking status. and Acute Changes on the
Left lower 38 (34.5)
Table 1 shows the clinical variables among Admission Chest Radiograph
Number of zones involved
patients with different adverse outcomes. Pa- Acute Radiographic Change No. of Patients
tients who needed supplemental oxygen were Single 60 (54.5)
Consolidation 68 (48.2)
significantly older (mean age, 77 13 vs 68 Multiple 50 (45.5)
19 years; p< 0.001). More patients who needed Ground-glass opacity 57 (40.4)
Percentage of involvement
oxygen had underlying chronic diseases than Reticular opacity 12 (8.5) (both lungs, 100%)
those who did not, including congestive heart Nodular opacity 13 (9.2) < 10% 73 (66.4)
failure (p= 0.04), ischemic heart disease (p<
Increased peribronchial marking 19 (13.5) 1050% 35 (31.8)
0.001), and COPD (p< 0.001). Significantly
more smokers (p= 0.04) and patients with un- Pleural effusion 47 (33.3) > 50% 2 (1.8)
derlying COPD (p= 0.02) needed assisted ven- Left 20 (14.2) Distribution of involvement
tilation during their hospital course. Right 18 (12.8) Central 23 (20.9)
Bilateral 9 (6.4) Peripheral 44 (40.0)
Radiologic Changes
At admission 199 of 285 patients (69.8%) Hyperinflation 2 (1.4) Both 43 (39.1)
had abnormal chest radiographic findings; NoteValues in parentheses are percentages. NoteValues in parentheses are percentages.

AJR:203, August 2014 283


TABLE 4: Relation Between Adverse Outcome and Radiographic Findings in 285 Adults with Respiratory Syncytial
Virus Infection
Requirement for Supplemental
Oxygen Therapy (n = 282)a
Finding No (n = 86)
Downloaded from www.ajronline.org by 180.245.204.167 on 11/28/17 from IP address 180.245.204.167. Copyright ARRS. For personal use only; all rights reserved
Acute changes 31 (36)
Consolidation 17 (20)
Ground-glass opacity 12 (14)
Reticular shadow 1 (1)
Nodules 3 (3)
Chronic lung changes 31 (36)
NoteData are numbers of patients with percentages in parentheses. NS = not significant.
aData on three patients were missing.
There was no statistical difference in coin-
fection rate between patients with pleural ef-
fusion (7/47 [14.9%]) and those without effu-
sion (29/238 [12.2%]) (p= 0.752). However,
there was a higher proportion of conges-
tive heart failure (12/47 [25.5%] vs 38/238
[16.0%]), although this difference did not
reach statistical significance (p= 0.141).
With logistic regression, the odds ratio of
having pleural effusion in the presence of
bacterial coinfection (1.305, p= 0.354) and
congestive heart failure (1.830, p= 0.119) did
not reach statistical significance.
Relation Between Radiologic Abnormalities and
Adverse Outcome
Table 4 shows the frequency of radiologic
findings on the admission chest radiographs
of the study cohort in relation to adverse out-
come, that is, need for supplemental oxygen,
need for assisted ventilation, and death. The
radiographic finding of acute lung chang-
es was significantly more common in pa-
tients who needed supplemental oxygen (p=
0.004) and those who died (27/35 [77%]
vs 114/250 [46%]; p= 0.01). Among those
with the acute radiologic pattern, consolida-
tion and ground-glass opacity were more of-
ten observed in patients who needed assisted
ventilation (consolidation, p= 0.03; ground-
glass opacity, p= 0.04) and those who died
(consolidation, p= 0.01; ground-glass opac-
ity, p= 0.01). The presence of chronic lung
changes was also more frequently associat-
ed with the need for supplemental oxygen
(p= 0.01), need for assisted ventilation (p=
0.003), and death (p= 0.01). No association
was found between distribution or extent of
lung change and adverse outcome.
There was also difference in outcome for
patients with and without acute chest ra-
diographic changes. Patients without acute
284 AJR:203, August 2014
Wong et al.
Requirement for Assisted
Ventilation (n = 282)a
Yes (n = 196) p No (n = 247) Yes (n = 35)
108 (55) 0.004 117 (47) 22 (63)
51 (26) NS 54 (22) 14 (40)
44 (22) NS 44 (18) 12 (34)
11 (6) NS 11 (4) 1 (3)
10 (5) NS 12 (5) 1 (3)
103 (53) 0.01 109 (44) 25 (71)
chest radiographic changes had statistically
significantly lower mortality (8/144 [5.6%]
vs 27/141 [19.1%]) (p= 0.001) and a lower
percentage of oxygen therapy requirement
(88/143 [61.5%] vs 108/139 [77.7%]) (p=
0.04). They also had a lower percentage of
ventilator requirement (13/143 [9.1%] vs
22/139 [15.8%]), although this difference did
not reach statistical significance.
Logistic regression evaluation of the con-
tribution of age, underlying chronic condi-
tions, smoking status, and presence of acute
and chronic radiologic changes in predicting
adverse outcome showed that the presence of
acute and chronic lung changes on admission
chest radiographs were independent risk fac-
tors in predicting mortality (acute changes
odds ratio, 3.6 [95% CI, 1.538.25]; chronic
changes odds ratio, 2.2 [95% CI, 1.014.86]).
Acute lung changes was found to be an inde-
pendent risk factor for need for supplemental
oxygen (odds ratio, 2.4 [95% CI, 1.324.29])
and chronic lung changes to be an indepen-
dent predictor of need for assisted ventilation
(odds ratio, 3.1 [95% CI, 1.366.85]).
Discussion
Most studies of the radiographic appear-
ance of RSV-related infection of the low-
er respiratory tract have focused on pediat-
ric patients [4, 5]. Kern et al. [5] reported on
a series of 108 children with RSV infection
of the lower respiratory tract, and the most
common radiologic findings were central
pneumonia (32%), peribronchitis (26%), and
normal chest radiographic findings (30%).
Despite the magnitude of disease burden
of RSV infection in adults, the radiograph-
ic features are not well described. In a se-
ries described by Walsh et al. [15], 42% of
chest radiographs of 118 pure RSV infections
showed no acute disease; 20% of the radio-
Death (n = 285)
p No (n = 250) Yes (n = 35) p
NS 114 (46) 27 (77) 0.01
0.03 53 (21) 15 (43) 0.01
0.04 44 (18) 13 (37) 0.01
NS 10 (4) 2 (6) NS
NS 11 (4) 2 (6) NS
0.003 112 (45) 24 (69) 0.01
graphs showed new opacities consistent with
pneumonia; 8% showed atelectasis; and 13%
showed infiltrates compatible with either at-
electasis or pneumonia. The radiographic
changes were generally unilateral and basal
in location, relatively small, and rarely dense.
Ariza-Heredia et al. [16] described the most
common radiologic findings of RSV infec-
tion in 24 immunocompromised adults as
pulmonary nodules, ground-glass opacities,
small cavities, and pleural effusions. Lrida
et al. [2] found that patients with acute RSV
infection more frequently had bilateral in-
volvement on chest radiographs.
To our knowledge, our study is the largest
series investigating the radiographic features
of adult RSV respiratory tract infection. Ap-
proximately one half of the 285 patients (141
patients) had acute lung changes related to
laboratory-confirmed RSV infection, a high-
er proportion than in other reported series [17,
18]. This finding may be explained by the fact
that a large proportion of patients in our co-
hort had underlying chronic diseases, which
are known predisposing factors for severe
RSV infection in adults [19]. The predomi-
nant patterns of acute lung changes were con-
solidation and ground-glass opacity, believed
to be related to underlying pathologic mecha-
nisms that include bronchiolitis, alveolar dam-
age with hyaline membranes formation, and
organizing pneumonia [20]. The chest radio-
graphic findings were in accordance with rec-
ognized CT appearances of RSV pneumonia,
which are a combination of small centrilobular
nodules, tree-in-bud opacities, air-space con-
solidation, ground-glass opacities, and bron-
chial wall thickening [2126]. However, ap-
proximately one third of patients in our cohort
with acute lung changes had pleural effusion
detectable on chest radiographs. This is con-
trary to the traditional belief that viral chest in-
 Radiography of Respiratory Syncytial Virus Infection
fections tend to cause predominantly lung pa-
renchymal changes without pleural effusion.
The cause of the high proportion of pleural ef-
fusion at presentation in RSV-infected adults
in our cohort remains unexplained. Bacterial
coinfection did not appear to play a role, as ev-
Downloaded from www.ajronline.org by 180.245.204.167 on 11/28/17 from IP address 180.245.204.167. Copyright ARRS. For personal use only; all rights reserved
idenced by the similar coinfection rates in pa-
tients with and those without effusion. On the
other hand, in view of a statistically insigni-
ficantly higher proportion of congestive heart
failure among patients with pleural effusion,
one possible explanation is coexistence or pre-
cipitation of underlying congestive heart fail-
ure by RSV infection. Nevertheless, that most
of the patients had unilateral (27.0%) rather
than bilateral (6.4%) pleural effusion may ne-
gate this possible explanation, because pleu-
ral effusion related to congestive heart failure
should have bilateral involvement. Whether
the finding is directly related to RSV infection
or another lung parenchymal or systemic reac-
tion warrants further investigation.
A minority of our patient cohort (12.6%)
had bacterial coinfection, which is not surpris-
ing in patients with RSV infection, who are
predisposed to secondary bacterial infection.
This is similar to the rate reported in the litera-
ture [9, 27]. This small percentage of bacteri-
al coinfection does not appear to confound our
analysis of the radiographic features of RSV
chest infection, evidenced by the lack of statis-
tical difference in coinfection rate between pa-
tients with and those without consolidation or
ground-glass opacity. We believe that most of
the active radiographic changes on the admis-
sion chest radiographs were related to RSV in-
fection, given the short time between onset of
symptoms and admission radiography.
There is lack of evidence in the literature
about the prognostic value of initial chest ra-
diography in RSV-related respiratory tract in-
fections. A study by Prodhan et al. [28] showed
that among children younger than 1 year with
infection, those who needed prolonged me-
chanical ventilation were identified on the ba-
sis of chest radiographic patterns around the
time of intubation [28]. Our results suggest that
the presence of both acute and chronic radio-
graphic changes on the admission chest radio-
graph should help predict which patients need
supplemental oxygen therapy or assisted venti-
lation and those who will die. Therefore admis-
sion chest radiography may play an important
role in triage of patients with RSV infection
at high risk of having an adverse clinical out-
come necessitating close monitoring and more
aggressive treatment. The role of initial chest
radiographic findings in predicting patient out-
AJR:203, August 2014 285
come of H1N1 influenza infection was stud-
ied by Aviram et al. [12]. Those authors found
that extensive involvement of both lungs (i.e.,
presence of multizonal and bilateral peripheral
opacities) on initial chest radiographs was as-
sociated with need for mechanical ventilation
and with death. Similar prognostic significance
of initial radiographic opacities was found for
severe acute respiratory syndrome [29]. We did
not find any significant relation between extent
of lung parenchymal involvement and adverse
clinical outcome.
Concerning the specimen used for vi-
ral testing, we used mainly nasopharyngeal
aspiration because it is a simple, quick and
relatively noninvasive method of specimen
collection. Other specimens, such as spu-
tum and bronchoalveolar lavage, can also be
used for diagnosis. Bronchoscopy with bron-
choalveolar lavage is particularly diagnostic
in patients with immunosuppression or unex-
plained ground-glass attenuation on CT im-
ages [30]. However, bronchoalveolar lavage
is invasive and is usually reserved for in-
stances in which other noninvasive tests fail
to reveal the causative pathogen.
There were several limitations to our
study. First, it was a retrospective study that
included only hospitalized patients. There-
fore, the full spectrum of radiographic
changes of RSV respiratory tract infection
in adults was not truly represented. Second,
we used consensus reading without perform-
ing intraobserver and interobserver variation
in interpretation of radiographic changes.
However, it is recognized that interobserv-
er agreement for assessment of communi-
ty-acquired pneumonia is poor with regard
to opacity type and fair to good in relation
to the actual presence of abnormal opacities
and their localization [31].
Third, RSV infection was identified with
an immunofluorescence assay rather than re-
verse transcriptase polymerase chain reaction
(RT-PCR). The specificity and sensitivity of
immunofluorescence assay is high, ranging
from 95% to 99% and 80% to 94% [9]. The
assay performs particularly well with naso-
pharyngeal aspirate samples and in children,
likely because of more cellular samples and
higher viral titer. Its performance in adults is
generally poorer because of lower viral titers
and short viral shedding time. RT-PCR has
higher sensitivity in the diagnosis of RSV in-
fection, even in situations in which the viral
titer is low and the viral shedding period is
short [32]. Unfortunately, at our center only
immunofluorescence assay was available as a
routine service at the time this study was per-
formed. RT-PCR is being developed at our
hospital and will be used as the test of choice
in future studies.
Lack of CT correlation was another limita-
tion. In our series, 86 of 285 patients (30.2%)
had normal chest radiographic findings. CT
will certainly show abnormalities in a propor-
tion of this group of patients. However, the CT
findings of RSV infection are often nonspe-
cific and overlap considerably with those of
other viral and bacterial infections found in
aspirate [21, 24]. Therefore, there are limited
literature findings and guidelines supporting
CT as a tool in diagnosis and management.
When our center balanced the possible addi-
tional benefit and the risk of radiation and IV
contrast administration, CT was not routinely
performed for patients with normal chest ra-
diographic findings. The indications for CT
were similar to those for other chest infec-
tions, such as those in patients in whom com-
plications (e.g., abscess) were suspected. De-
spite this, CT can be incorporated in future
prospective studies of its role in the diagnosis
and management of RSV infection in patients
with normal chest radiographic findings.
Another limitation was that there was no
routine follow-up chest radiography; therefore,
there were no data on the percentage of patients
with normal initial chest radiographic findings
later found to have radiographic abnormalities.
Nevertheless, the outcome measures (mortal-
ity, oxygen requirement, ventilatory require-
ment) would not be affected by whether these
patients would subsequently have chest radio-
graphic abnormalities. In other words, admis-
sion chest radiography still has value for pre-
dicting patient outcome. A future prospective
study with follow-up chest radiography will be
helpful for investigating this issue.
Conclusion
One half of adult patients hospitalized for
RSV respiratory tract infection had acute
changes on initial chest radiographs. The pre-
dominant pattern was consolidation or ground-
glass opacity in a single unilateral lower zone.
The presence of acute and chronic lung chang-
es on admission chest radiographs was associ-
ated with an adverse outcome.
Acknowledgments
We thank Timothy C. M. Li and Ernie C.
M. Tse, Department of Medicine and Thera-
peutics, Prince of Wales Hospital, The Chi-
nese University of Hong Kong, Hong Kong,
SAR, for their contribution to this study.

References
1. Falsey AR, Hennessey PA, Formica MA, Cox C,
Walsh EE. Respiratory syncytial virus infection
in elderly and high-risk adults. N Engl J Med
2005; 352:17491759
2. Lrida A, Marrn A, Casanova A, et al. Respira-
Downloaded from www.ajronline.org by 180.245.204.167 on 11/28/17 from IP address 180.245.204.167. Copyright ARRS. For personal use only; all rights reserved
tory syncytial virus infection in adult patients
hospitalized with community-acquired pneumo-
nia [in Spanish]. Enferm Infecc Microbiol Clin
2000; 18:177181
3. Ko FWS, Ip M, Chan PKS, et al. Viral etiology of
acute exacerbations of COPD in Hong Kong.
Chest 2007; 132:900908
4. Guo W, Wang J, Sheng M, Zhou M, Fang L. Radi-
ological findings in 210 paediatric patients with
viral pneumonia: a retrospective case study. Br J
Radiol 2012; 85:13851389
5. Kern S, Uhl M, Berner R, Schwoerer T, Langer
M. Respiratory syncytial virus infection of the
lower respiratory tract: radiological findings in
108 children. Eur Radiol 2001; 11:25812584
6. Hurwitz JL. Respiratory syncytial virus vaccine
development. Expert Rev Vaccines 2011; 10:1415
1433
7. Teng MN. The non-structural proteins of RSV:
targeting interferon antagonists for vaccine devel-
opment. Infect Disord Drug Targets 2012;
12:129137
8. Zeng R, Cui Y, Hai Y, Liu Y. Pattern recognition re-
ceptors for respiratory syncytial virus infection and
design of vaccines. Virus Res 2012; 167:138145
9. Lee N, Lui GCY, Wong KT, et al. High morbidity
and mortality in adults hospitalized for respirato-
ry syncytial virus infections. Clin Infect Di 2013;
57:10691077
10. Kim EA, Lee KS, Primack SL, et al. Viral pneumo-
nias in adults: radiologic and pathologic findings.
RadioGraphics 2002; 22(spec no):S137S149
11. Lee EY, McAdam AJ, Chaudry G, et al. Swine-
origin influenza a (H1N1) viral infection in chil-
dren: initial chest radiographic findings. Radiolo-
gy 2010; 254:934941
12. Aviram G, Bar-Shai A, Sosna J, et al. H1N1 influ-
F O R YO U R I N F O R M AT I O N
This article is available for CME and Self-Assessment (SA-CME) credit that satisfies Part II requirements for
maintenance of certification (MOC). To access the examination for this article, follow the prompts associated with
the online version of the article.
286 AJR:203, August 2014
Wong et al.
enza: initial chest radiographic findings in helping
predict patient outcome. Radiology 2010;
255:252259
13. Hansell DM, Bankier AA, MacMahon H, et al.
Fleischner Society: glossary of terms for thoracic
imaging. Radiology 2008; 246:697722
14. Bramson RT, Griscom NT, Cleveland RH. Inter-
pretation of chest radiographs in infants with
cough and fever. Radiology 2005; 236:2229
15. Walsh EE, Peterson DR, Falsey AR. Is clinical
recognition of respiratory syncytial virus infec-
tion in hospitalized elderly and high-risk adults
possible? J Infect Dis 2007; 195:10461051
16. Ariza-Heredia EJ, Fishman JE, Cleary T, et al.
Clinical and radiological features of respiratory
syncytial virus in solid organ transplant recipi-
ents: a single-center experience. Transpl Infect
Dis 2012; 14:6471
17. Dowell SF, Anderson LJ, Gary HE Jr, et al. Respi-
ratory syncytial virus is an important cause of
community-acquired lower respiratory infection
among hospitalized adults. J Infect Dis 1996;
174:456462
18. Falsey AR, Cunningham CK, Barker WH, et al.
Respiratory syncytial virus and influenza A infec-
tions in the hospitalized elderly. J Infect Dis 1995;
172:389394
19. Falsey AR, Walsh EE. Respiratory syncytial virus
infection in adults. Clin Microbiol Rev 2000;
13:371384
20. Ebbert JO, Limper AH. Respiratory syncytial vi-
rus pneumonitis in immunocompromised adults:
clinical features and outcome. Respiration 2005;
72:263269
21. Shiley KT, Van Deerlin VM, Miller WT Jr. Chest
CT features of community-acquired respiratory
viral infections in adult inpatients with lower re-
spiratory tract infections. J Thorac Imaging 2010;
25:6875
22. Gasparetto EL, Escuissato DL, Marchiori E, et al.
High-resolution CT findings of respiratory syncy-
tial virus pneumonia after bone marrow trans-
plantation. AJR 2004; 182:11331137
23. Ko JP, Shepard JA, Sproule MW, et al. CT mani-
festations of respiratory syncytial virus infection
in lung transplant recipients. J Comput Assist To-
mogr 2000; 24:235241
24. Miller WT Jr, Mickus TJ, Barbosa E Jr, et al. CT
of viral lower respiratory tract infections in
adults: comparison among viral organisms and
between viral and bacterial infections. AJR 2011;
197:10881095
25. Miller WT Jr, Barbosa E Jr, Mickus TJ, et al.
Chest computed tomographic imaging character-
istics of viral acute lower respiratory tract illness-
es: a case-control study. J Comput Assist Tomogr
2011; 35:524530
26. Syha R, Beck R, Hetzel J, et al. Human meta-
pneumovirus (HMPV) associated pulmonary
infections in immunocompromised adults: ini-
tial CT findings, disease course and comparison
to respiratory-syncytial-virus (RSV) induced
pulmonary infections. Eur J Radiol 2012;
81:41734178
27. Lee N, Choi KW, Chan PK, et al. Outcomes of
adults hospitalised with severe influenza. Thorax
2010; 65:510515
28. Prodhan P, Westra SJ, Lin J, et al. Chest radiologi-
cal patterns predict the duration of mechanical
ventilation in children with RSV infection. Pedi-
atr Radiol 2009; 39:117123
29. Antonio GE, Ooi CG, Wong KT, et al. Radio-
graphic-clinical correlation in severe acute respi-
ratory syndrome: study of 1373 patients in Hong
Kong. Radiology 2005; 237:10811090
30. Jouneau S, Poineuf JS, Minjolle S, et al. Which
patients should be tested for viruses on bronchoal-
veolar lavage fluid? Eur J Clin Microbiol Infect
Dis 2013; 32:671677
31. Albaum MN, Hill LC, Murphy M, et al. Interob-
server reliability of the chest radiograph in com-
munity-acquired pneumonia. PORT Investiga-
tors. Chest 1996; 110:343350
32. Popow-Kraupp T, Aberle JH. Diagnosis of respi-
ratory syncytial virus infection. Open Microbiol J
2011; 5:128134