The WHO guidelines(1) for detection of
plasma leakage for diagnosis of Dengue Hemorrhagic Fever (DHF) by
various indirect evidences have revolutionized the concept of managing
cases of DHF. The need to have area specific hematocrit cut off values
to identify hemoconcentration as evidence of plasma leakage has been
suggested(2) and studies from Delhi and Chennai have recommended such
values(3,4). There has been paucity of data regarding the usefulness
of ultrasonography and radiography in the diagnosis of DHF. Hence this
study was carried out to study the radiological and ultrasonographic
features in DHF and their role in the detection of plasma leakage.
This study also compared the various parameters indicating plasma
leakage including area specific hematocrit values with the
radiological and ultrasonographic evidence of the same.
Subjects and Methods
A prospective study was undertaken at Kanchi Kamakoti
CHILDS Trust Hospital, a tertiary care referral Pediatric hospital, in
Chennai from December 2003 to May 2004. All children (>1 month upto 18
years of age) who were diagnosed to have dengue infection and
hospitalized for the same were included in the study. Apart from a
meticulous clinical examination carried out daily, all the children
underwent the following investigations as per the requirement of the
study–complete blood count, liver function tests, renal function tests,
chest X-ray and ultrasonography. Repeat hematocrit and platelet
count were done at recovery or as and when required depending on the
clinical situation. All children were confirmed to have diagnosis of
dengue virus infection by doing the dengue ELISA IgG and IgM test (Pathozyme
dengue IgG and Pathozyme dengue IgM, Omega Diagnostics Ltd., Alloa,
Scotland, UK) on their serum samples. All clinical and investigation
parameters were recorded in a well-structured format from the time of
admission to the time of discharge. Clinically, presence of tachypnea,
chest retractions, decreased breath sounds and decreased vocal resonance
were considered signs of pleural effusion. Presence of abdominal
distension with fullness of the flanks and presence of shifting
dullness/fluid thrill was taken as evidence of ascites. The extent of
hemoconcentration in our study was quantitated by taking a difference
between the maximum hematocrit at admission or anytime during the
hospital stay and the minimum hematocrit recording at convalescence or
discharge. For the purpose of the study DHF was diagnosed as per the WHO
guidelines(1) as a probable case of dengue fever with hemorrhagic
tendencies and thrombocytopenia along with the presence of evidence of
plasma leakage manifested by any one or more of the following i.e.,
a rise in the average hematocrit for the age and sex by >20%; a >20%
drop in the hematocrit following volume replacement compared to the
baseline; signs of plasma leakage i.e., pleural effusion, ascites,
hypoproteinemia.
Children who were found to be sero-negative for
dengue infection, children with overlap infection with other organisms
in addition to dengue infection, children receiving treatment for dengue
infection on out patient basis and children who received blood
transfusion during the present admission were excluded.
The area specific hematocrit cut off values for
hemoconcentration was defined by the study done in Chennai as >34.8% in
less than 5 years age group and >37.5% in the age group of more than 5
years(4). Our study also attempted to test the utility of this
hematocrit cut off value in recognizing hemoconcentration.
The radiological and ultrasonographic evidence of
plasma leakage was compared with the other parameters of plasma leakage.
Percentage analysis was carried out on the data of
various evidences of plasma leakage as well as area specific hematocrit
cut off values. Sensitivity, specificity, positive predictive value and
negative predictive values were derived from the available data
pertaining to the imaging modalities and were compared with the other
parameters of plasma leakage.
Results
In the present study of a total number of 65 cases,
30 cases (46.15%) were Dengue fever (DF) and the remaining 35 cases
(53.84%) were Dengue Hemorrhagic Fever (DHF) based on the WHO case
definition. Only 6 (9.23%) of the 65 cases had Dengue Shock Syndrome (DSS).
The various USG findings in cases of Dengue virus
infection in our study are depicted in Table I and the comparison
of the parameters giving evidence of plasma leakage in children with DHF
depicted in Table II.
TABLE I
Ultrasonographic Findings in DF* and DHF*.
|
Ultrasonographic findings |
Total (n = 65) |
DF (n = 30) |
DHF (n = 35) |
|
Ascites |
46(70.76%) |
14(46.66%) |
32(91.42%) |
|
Pleural fluid |
44(67.69%) |
13(43.33%) |
31(88.57%) |
|
Gallbladder Changes |
42(64.61%) |
10(33.33%) |
32(91.42%) |
|
Hepatomegaly |
56(86.1%) |
22(73.33%) |
34(97.14%) |
|
Splenomegaly |
18(27.6%) |
7(23.33%) |
11(31.42%) |
|
Perinephric fluid |
10(15.3%) |
0 |
10(28.57%) |
*DF and DHF based on the WHO case definition.
TABLE II
Evidence for Plasma Leakage in Children with DHF*.
|
Evidence for plasma leakage |
DHF(n = 35) |
|
Hemoconcentration >20% |
20(57.14%) |
|
Hypoproteinemia |
11(31.42%) |
|
Clinical evidence of plasma leakage(pleural effusion and or
ascites) |
25(71.42%) |
|
Hemoconcentration as per Area specific hematocrit cut off levels |
32(91.42%) |
|
Evidence of plasma leakage on Radiography(pleural effusion) |
23(65.71%) |
|
Evidence of plasma leakage on USG( pleural effusion and or ascites) |
32(91.42%) |
*DHF as per the WHO case definition
The comparison of various parameters that give
evidence of plasma leakage is as follows.
Hemoconcentration (>20%) was detected in 20 (57.14%)
out of the 35 cases of DHF. Hemoconcentration based on the area specific
hematocrit cut off values was observed in 32 cases (91.42%) and it had a
better sensitivity and negative predictive value as an indicator of
plasma leakage. (Table IV)
TABLE III
Comparative Analysis of Parameters for Evidence of Plasma Leakage.
| |
Clinical
Evidence |
Hemoconcentration
>20% |
Hypoprote-inemia |
Ultrasound |
X-ray |
|
Sensitivity |
56.00% |
48.27% |
25.00% |
91.42% |
65.71% |
|
Specificity |
72.50% |
83.33% |
90.90% |
53.33% |
80.00% |
|
Positive predictive value |
56.00% |
70.00% |
72.72% |
69.56% |
79.31% |
|
Negative predictive value |
72.50% |
66.66% |
46.87% |
84.21% |
66.66% |
Table IV
Comparison between WHO Criteria for Hemoconcentration and Area
Specific Hematocrit Cut off Values as Evidence of Plasma Leakage.
| |
Hematocrit >20% |
Area specific hematocrit cut off values |
|
Sensitivity |
48.27% |
91.42% |
|
Specificity |
83.33% |
46.66% |
|
Positive predictive value |
70.00% |
66.66% |
|
Negative predictive value |
66.66% |
82.35% |
Hypoproteinemia was seen only in 11 cases (31.42%) of
DHF in our study. Most of the cases of DHF (24 cases, 68.57%) did not
have hypoproteinemia. Hypoproteinemia had a sensitivity of 25% and
negative predictive value of 46.87% as evidence for plasma leakage (Table
III).
Clinical evidence of plasma leakage in the form of
pleural effusion and or ascites was present in 25 out of 35 cases
(71.42%) of DHF. However, in comparison to the laboratory parameters of
plasma leakage (hypoproteinemia and hemoconcentration), clinical
evidence of plasma leakage (signs of pleural effusion and or ascites)
was observed more frequently. This parameter had a sensitivity of 56%
and negative predictive value of 72.50%.
Radiographic evidence of pleural effusion was present
in 23 (65.71%) out of the 35 cases of DHF. Among those who had
radiographic evidence of pleural effusion only 9 cases (39.13%) had
clinical evidence of the same. Out of 30 cases labelled as DF as per the
WHO recommended case definition, 6 cases (20%) had pleural effusion on
radiography. These 6 cases did not have 20% hemoconcentration, clinical
evidence of plasma leakage or hypoproteinemia to be labelled as DHF.
Chest radiography had a sensitivity of 65.71%, specificity of 80%,
positive predictive value of 79.31% and negative predictive value of
66.66% in identifying plasma leakage.
Out of the 35 cases of DHF, pleural effusion was
detected by ultrasonography in 31 cases (88.57%). Among them only 11
cases (35.48%) had pleural effusion on clinical examination and 23 cases
(74.19%) had radiographic evidence of the same. 20 cases (64.51%) did
not have any clinical evidence of pleural effusion and 8 cases (25.80%)
did not have radiographic evidence of the same. With the assistance of
ultrasonography, 13 cases labelled as DF as per the WHO criteria
(43.33%) were found to have pleural effusion and therefore would qualify
to be labelled DHF. There was no case in our study group that had
pleural effusion in chest radiography that was not evident in
ultrasonography.
Out of the 35 DHF cases ascites on ultrasonography
was present in 32 cases (91.42%) while clinical evidence of the same was
seen in only 20 cases (57.14%). When ultrasonography was included as a
tool in the detection of ascites 14 out of the 30 cases (46.66%)
labelled as DF as per the WHO parameters were found to have ascites and
therefore would qualify to be labelled as DHF instead. Ultrasonography
had the highest sensitivity (91.42%) and negative predictive value
(84.21%) when compared with all the other parameters in detecting plasma
leakage (Table III).
Discussion
There are difficulties in following the WHO criteria
for recognizing plasma leakage, for diagnosis of DHF(1), since hemoconcentration
(>20%) is usually diagnosed retrospectively, hypoproteinemia is
infrequent and clinical recognition of plasma leakage is difficult in a
sick child. Ultrasonography and radiography of the chest can reliably
detect presence of pleural effusion and ascites in children with DHF.
In our study, ultrasonography was found to be
superior when compared with radiography in detecting plasma leakage. The
low sensitivity of radiography could be attributed to the following
reasons.
Radiographic films are not ideal for detecting small
amounts of effusion, while sonography is highly useful(5-9). Lateral
decubitus chest radiography and chest sonography have been proven to be
highly efficient methods compared to conventional radiography to detect
small amounts of pleural effusions(5-9). In our study, lateral decubitus
films were not carried out. Also, the practical difficulty in getting
these two investigations done at the same time could have contributed to
these results. An earlier study from Indonesia also reported such
discrepancies in the findings related to the two investiga-tions(10).
The chest radiographs in our study were carried out at the time of
admission or at the time when there was a dilemma in diagnosing DHF. A
repeat chest radiograph at an appropriate time, which could have
resulted in a higher yield, was not carried out on our children for the
concern of radiation exposure. Six cases that were defined as DF as per
the WHO case definition did not have hemoconcentration, hypoproteinemia
or clinical signs of pleural effusion but had radiographic evidence of
pleural effusion confirming plasma leakage. Thus radiography was better
in comparison to the WHO recommended parameters in detecting plasma
leakage.
Out of the 65 cases screened by ultrasonography,
evidence of plasma leakage in the form of pleural effusion and or
ascites was seen in 46 cases (70.76%) and, therefore, would qualify to
be DHF. However we had only 35 cases of DHF based on the WHO parameters,
i.e., not using USG to detect pleural effusion or ascites.
Therefore when USG was also used apart from the clinical and the
biochemical criteria to screen these cases the total no of cases of DHF
would be 49 and DF cases would be 16. The reason for this discrepancy
could be that at the time ultrasonography was done, the amount of
capillary leak was probably minimal to be detected clinically. The
sensitivity of detecting ascitic fluid might have improved by clinical
methods other than shifting dullness like elicitation of puddle’s sign,
which was not carried out in the present study since it was not feasible
practically in these sick children. When the various parameters
suggested by WHO as indirect evidences of capillary leak were compared
with ultrasonography and radiography, ultrasonography was found to have
the highest sensitivity and negative predictive value in detecting
plasma leakage (Table III). Ultrasonography would be ideal owing
to its safety in that it is non-ionizing and would assist detecting
plasma leakage even before it clinically manifests. Similar findings
have been reported earlier from Indonesia(11).
Hemoconcentration of more than 20% was found to have
a lower sensitivity and negative predictive value compared to hemoconcentration
based on the area specific hematocrit cut off levels (Table IV),
thus supporting the use of area specific hematocrit cut off values as
recommended in earlier studies(2-4).
Clinical evidence of pleural effusion and or ascites
was present more often among cases of DHF than hemoconcentration (>20%)
and hypoproteinemia. This could be attributed to the fact that clinical
examination of our cases was being carried out more often and
periodically while blood investigations were not so frequently done.
This highlights the importance of a meticulous clinical examination
repeatedly in children with dengue virus infection in the recognition of
plasma leakage early.
The low sensitivity and low negative predictive value
of hypoproteinemia indicated that this tool was not sufficiently
sensitive to pick up plasma leakage early.
Acknowledgements
The authors wish to acknowledge Dr. C. Ravichandran
Assistant Professor of Pediatrics at Institute of Child Health and
Hospital for Children, Chennai, for his guidance in the statistical
analysis and modification of the manuscript.
Contributors: SB conceived the idea, supervised
data collection and clinical examination, revised the manuscript and
acts as guarantor for the study; LJ supervised data collection and
clinical examination; SS reviewed the subject, analysed the data and
drafted the paper; SM carried out the imaging study; SS collected the
clinical data and analyzed the same.
Funding: None.
Competing interests: None stated.
|
Key Messages |
|
• Repeated clinical assessment helps
recognize plasma leakage in DHF.
• Ultrasonography is superior to clinical and
laboratory parameters for diagnosing plasma leakage.
• Area specific hematocrit cutoff values are
useful in the diagnosis of DHF.
|
