Escolar Documentos
Profissional Documentos
Cultura Documentos
Background. Serologic testing in malaria has traditionally been done by immunouorescence antibody testing (IFA),
but the use of commercially available enzyme immunoassays (EIAs) has become more widespread.
Methods. We compared IFA with two commercial EIA kits, the Cellabs Pan Malaria CELISA and the Newmarket
Malaria EIA. Seventy-ve samples from 74 patients with clinically suspected malaria were examined by both EIA kits.
The samples were also examined by IFA (n = 48) and/or Giemsa-stained blood smear (n = 48).
Results. Using a consensus result as a gold standard, the agreement, sensitivity, and specicity were, respectively, as follows: Cellabs EIA 93.2, 95.5, and 92.2%; Newmarket EIA 87.7, 68.2, and 96.1%; and IFA 89.1, 86.4, and 91.7%. Compared to positive Giemsa-stained smears, the sensitivities were as follows: Cellabs EIA 90.9% (10/11), Newmarket EIA
54.5% (6/11), and IFA 100% (11/11). Antinuclear antibody (ANA)positive sera (n = 11) and rheumatoid factor (RF)
positive sera (n = 11) showed no cross-reactivity with the Newmarket EIA, while the Cellabs EIA yielded positive results
in one ANA-positive and two RF-positive sera. Among healthy blood donors (n = 50), the Newmarket EIA showed
100% specicity (50/50) and the Cellabs EIA showed a specicity of 92% (46/50).
Conclusions. While the Newmarket EIA was a generally more specic assay, it was insufciently sensitive relative to the
IFA and the Cellabs EIA for screening purposes for malaria antibodies. The Cellabs EIA demonstrated the best overall
sensitivity and is a reasonable choice as a serodiagnostic tool for malaria. It may also be useful as an adjunct to Giemsastained smear examination, to aid in cases of low parasitemia in previously nonimmune individuals.
106
donations with EIA rather than with IFA. Several
studies have validated the use of EIA in blood bank
screening in Europe and Australia.69
The Cellabs EIA detects IgG antibody captured
by recombinant antigens from P falciparum, P vivax,
P ovale, and P malariae. Further specications on
the nature of the antigens used are not available.
There are no published studies evaluating its performance to date. The Newmarket Malaria EIA
detects IgG, IgM, and IgA antibodies using recombinant merozoite-specic proteins of P falciparum
and P vivax. It has been evaluated with satisfactory
results for the purpose of blood bank screening.6,8 In
this study, we compared the performance characteristics of the IFA and two commercial EIA kits,
the Cellabs Pan Malaria Antibody CELISA and the
Newmarket Malaria EIA.
Materials and Methods
Human Plasma and Sera
The study was approved by the Institutional Review
Board of the University of Utah, IRB 7275. Between
January 2002 and June 2005, serum or plasma from
clinical samples tested for malaria by either Giemsastained blood smear or malaria IFA were collected
at our reference laboratory in Salt Lake City, Utah.
Seventy-four samples from individual patients were
evaluated. Of these, 46 were evaluated by Giemsa
stain, 27 by IFA, and two by both Giemsa stain and
IFA. A second plasma sample from patient 72 drawn
4 days after the original sample was included in this
study, for a total of 75 specimens. To test for specicity of the malaria EIA in healthy adults, 50 serum
samples were collected from random healthy blood
donors in the Salt Lake City, Utah area (all collected
in 2003 and stored at 20C).
Cellabs Pan Malaria Antibody CELISA
All specimens were tested according to the recommended protocol of the manufacturer (Cellabs,
Sydney, New South Wales, Australia). Briey, samples and controls were diluted 1 to 100 in phosphate-buffered saline (PBS)/Tween. One hundred
microliters of each sample was added to the coated
wells and incubated for 1 hour at room temperature.
The wells were washed four times with PBS/Tween
using an automatic washer (Columbus Pro, Tecan
Trading AG, Switzerland). One hundred microliters of conjugate (enzyme-labeled anti-human
immunoglobulin) was added to each well and allowed
to incubate for 1 hour at room temperature. The
wells were washed four times with PBS/Tween.
One hundred microliters of chromogen substrate
J Travel Med 2007; 14: 105111
She et al.
was added to each well and incubated for 15 minutes
at room temperature in the dark. Fifty microliters
of stop solution (6.5% hydrochloric acid) was then
added to each well. Optical densities (ODs) were
immediately read by a microplate reader (Spectramax Plus; Molecular Devices Corp., Sunnyvale,
CA, USA) at 450 nm with a reference of 620 nm.
The cutoff value for each run was determined to be
the mean OD of two negative controls plus 0.100.
Values were then expressed as an index value of OD
sample/OD cutoff. Any specimen with a result different from that obtained by the other test methods
used was repeated in duplicate.
Newmarket Malaria EIA
Tests were carried out on all samples according to
specications of the manufacturer (Newmarket
Laboratories Ltd, Newmarket, UK). Briey, 50 L
of undiluted sample or control was added to the
wells. The wells were placed on a plate shaker for 30
seconds and incubated for 30 minutes at 37C. The
wells were washed ve times with 30-second soaks
between each wash using the automated plate
washer. Fifty microliters of recombinant antigen
conjugated to horseradish peroxidase was added to
each well and allowed to incubate for 30 minutes at
37C. The wells were again washed ve times with
30-second soaks between each wash. Fifty microliters of substrate solution (urea peroxide and tetramethylbenzidine) was added to each well and
incubated for 30 minutes in the dark at room temperature. Fifty microliters of stop solution (0.5 M
sulfuric acid) was added to each well. ODs were immediately read at 450 nm using a reference of 650 nm.
The cutoff value for each run was determined to be
the mean OD of three negative controls plus 0.100.
Index values were calculated by dividing results by
the cutoff OD value. Any specimen with discrepant
results versus the other test methods used was repeated in duplicate.
Immunouorescence Antibody Testing
Of the patient samples collected, 29 had IFA testing performed prior to the study. An additional 19
IFA tests were performed on the specimens with
discrepant results between EIA testing and/or the
Giemsa-stained blood smear. All IFA testing was
performed by Parasitic Disease Consultants
(Tucker, GA, USA) based on a method previously
described using thick smears of washed, heparanized blood containing Plasmodium sp. schizonts
from infected primates as antigen.5 Serum samples
were tested for P falciparum, P vivax, P ovale, and
P malariae separately by twofold dilutions between
107
as positive, which included various combinations
of all four malarial species of Plasmodium. There
were an additional 14 positive samples by IFA
among the other 19 samples tested because of discrepant EIA and/or Giemsa stain results. Results of
all IFA-positive specimens are included in Table 1.
Of the 21 total positive IFA results, 13 were positive for two or more species of Plasmodium.
Enzyme Immunoassays
All 75 specimens were tested by the two EIA kits. By
the Cellabs EIA, 27 tested positive and 48 tested
negative. By the Newmarket Malaria EIA, 19 tested
positive and 56 negative. Comparison of the two
EIA kits is shown (Figures 1 and 2). Figure 1B uses
the logarithm of the index values to compare the
two kits. The left lower quadrant represents samples negative by both EIA kits, the left upper quadrant represents samples positive by the Cellabs EIA
and negative by the Newmarket EIA, the right upper quadrant represents samples positive by both
EIA kits, and the right lower quadrant represents
samples positive by the Newmarket EIA and negative by the Cellabs EIA. Agreement between the
two kits was 84.0%.
Comparison of Serologic Tests to Consensus Results
By consensus results, the 75 total samples consisted
of 22 positive, 2 equivocal, and 51 negative results.
Agreement, sensitivity, and specicity for each test
method are shown in Table 2. As demonstrated by
95% CIs, the Cellabs EIA was signicantly more
sensitive than the Newmarket EIA. The 95% CIs
otherwise showed no other statistically signicant
differences between the serologic tests. Samples
positive by one or more test method are summarized in Table 1. Among these samples, the most
frequently observed pattern, seen in ve cases, was
one in which the IFA, the two EIA kits, and Giemsastained smear were positive for a given sample. In
four cases, the Newmarket EIA tested negative
when the Cellabs EIA, IFA, and Giemsa-stained
smear were interpreted as positive.
Comparison of EIA and IFA to Giemsa-Stained
Blood Smear
The IFA detected 11 of 11 positive Giemsa-stained
smears, followed by the Cellabs EIA (10/11) and the
Newmarket EIA (6/11). The one case missed by the
Cellabs EIA was patient 72, who had a second sample drawn 4 days later that tested positive by the
Cellabs kit. Patients 34 and 49 were positive for malarial antibodies by all three serologic tests but negative by the Giemsa smear examination. Of the 36
J Travel Med 2007; 14: 105111
108
Table 1
She et al.
Results of cases positive by at least one test method
Case no.
Newmarket EIA
Cellabs EIA
IFA*
Giemsa stain
Consensus result
3
5
6
12
Pos
Pos
Pos
Pos
Pos
Neg
Pos
Pos
Not done
Not done
Not done
Not done
Pos
Neg
Pos
Pos
13
14
16
19
20
22
24
26
30
31
33
Pos
Neg
Pos
Neg
Pos
Neg
Neg
Pos
Pos
Neg
Pos
Pos
Neg
Pos
Pos
Neg
Pos
Neg
Pos
Pos
Pos
Pos
Pos
Neg
Pos
Pos
Neg
Pos
Pos
Pos
Pos
Pos
Pos
Pos
Not done
Not done
Not done
Not done
Not done
Not done
Not done
Not done
P falciparum
P vivax
P vivax versus
P ovale
Neg
Neg
Neg
Neg
P vivax
P vivax
Pos
Neg
Pos
Neg
Neg
Neg
Neg
Pos
Pos
Pos
Pos
34
47
49
50
51
53
55
62
64
68
Neg
Pos
Pos
Pos
Pos
Pos
Pos
Pos
Neg
Neg
P falciparum
P falciparum
Neg
Equ
Pos
Pos
69
Neg
Pos
P falciparum
Pos
70
71
Neg
Pos
Pos
Pos
P falciparum
P vivax
Pos
Pos
72
72-2
73
Neg
Neg
Pos
Neg
Pos
Pos
P ovale
Not done
Not done
Pos
Pos
Pos
74
Neg
Pos
Neg
Neg
Neg
Plasmodium malariae,
P vivax, P ovale
P falciparum, P vivax
P falciparum
P falciparum, P malariae
Neg
Neg
Neg
P malariae
Neg
P vivax
P vivax, P malariae
P falciparum, P vivax,
P ovale, P malariae
P falciparum
Neg
P falciparum
Neg
P ovale
P falciparum, P ovale,
P malariae
Neg
Neg
P falciparum
P falciparum, P vivax,
P ovale, P malariae
P falciparum, P vivax,
P ovale
P falciparum
P falciparum, P vivax,
P ovale
P ovale, P vivax
P ovale, P malariae
P falciparum, P vivax,
P malariae
P vivax, P ovale
Not done
Pos
Pos
Neg
Pos
Equ
Pos
Pos
EIA = enzyme immunoassay; IFA = immunouorescence antibody testing; pos = positive; neg = negative; equ = equivocal.
*Positive titers for IFA tests are greater than or equal to 1:64.
Consensus results: positive = 2/3, 3/3, 3/4, or 4/4 test methods are positive for a given sample, or 2/4 positive including positive Giemsa stain; negative =
0/3, 1/3, 0/4, or 1/4 positive; equivocal = 2/4 positive and Giemsa stain is negative.
Cross-Reactivity Studies
Using the Newmarket EIA, 0 of 11 ANA-positive
serum samples and 11 RF-positive serum samples
gave positive results. Using the Cellabs EIA, 1 of 11
ANA-positive serum samples and 2 of 11 RF-positive samples gave positive results. Both Newmarket
and Cellabs EIA kits gave negative results with the
single Babesia antibodypositive sample.
Discussion
The serologic tests examined yielded disparate results in many cases. This may be explained by the fact
that each method uses different antigens to capture
109
Table 2 Agreement, sensitivity, and specicity of the Cellabs EIA, Newmarket EIA, and IFA test using
consensus results as the gold standard
Cellabs EIA (n = 75)
Newmarket EIA (n = 75)
IFA (n = 48)
% Agreement
% Sensitivity*
% Specicity*
93.2
87.7
89.1
95.5 (82.799.2)
68.2 (54.174.6)
86.4 (73.792.4)
92.2 (86.793.8)
96.1 (90.098.8)
91.7 (80.197.2)
110
clinical history of the patients in this study was not
known; therefore, demographic information was not
available to aid the interpretation of results that were
discrepant between the various methods tested.
Recent larger studies have yielded results different from those of this study in regard to the performance of the Newmarket EIA. Kitchen and
colleagues7 demonstrated that it had comparable if
not superior sensitivity to a traditional IFA method
in acute, Giemsa stainproven malaria. In this study,
the Newmarket EIA detected 82.6% (114/138) of
acute P falciparum cases, 84.6% (11/13) of acute
P vivax cases, and 70% (7/10) of acute P ovale cases.
In comparison, the IFA, employing P falciparum
schizont as antigen, detected 74.3% (103/138) of
acute P falciparum cases, 61.5% (8/13) of acute
P vivax cases, and 80% (8/10) of acute P ovale cases.
Seed and colleagues8 demonstrated the Newmarket
EIA to have a sensitivity of 98% (106/108) in acute
P falciparum cases and 100% (12/12) in acute P vivax
cases. Both studies reported specicities of near
100% for this assay among healthy blood donors,
similar to ndings in the present study. There have
been few published studies on the performance of
the Cellabs Pan Malaria IgG CELISA to date.
Ozbilge and colleagues11 found that when compared
to light microscopy, the kit had a sensitivity and
specicity of 83 and 85%, respectively, similar to
the results found here.
There were two cases in this study that were negative by Giemsa-stained smear but positive by all
three serologic tests studied. While this may be due
to past exposure, these smears were evaluated under the clinical suspicion of malaria and thus there
is a concerning possibility of low parasitemia not
detectable by smear examination. Reexamination
of a negative smear may be done in future cases
should a concomitant serologic test be positive in a
previously unexposed individual. Although this
would not be a useful approach in malaria-endemic
areas, in which antimalarial antibodies are prevalent in the general population, this may be benecial at institutions in nonendemic countries that
often have limited experience examining blood
smears for malaria. The Cellabs EIA kit has in this
study shown to be reasonably sensitive in the acute
setting, detecting 10 of 11 cases of malaria. The
single case that was missed was detected in plasma
collected from the same patient 4 days later. This
corresponds to the development of antimalarial antibodies, which occur within 7 days of onset of parasitemia.2,3
Therefore, one additional use for the malaria
EIA could be as an adjunctive diagnostic tool to the
J Travel Med 2007; 14: 105111
She et al.
Giemsa-stained peripheral blood smear examination in the acute setting of suspected malaria. The
methods for EIA used in this study yielded results in
less than 3 hours. IFA, while correlating well in this
study with Giemsa smear evaluation, is time consuming and more labor intensive than the EIA. It
has some ability to differentiate antibodies from different Plasmodium species, but cross-reactivity is a
well-documented occurrence.5,6,9 The fact that 13
of the 21 positive IFA samples in this study had positive titers for two or more Plasmodium species suggests a signicant degree of cross-reactivity of
antimalarial antibodies. Patient 72, interpreted as
P ovale by Giemsa-stained smear evaluation, was
positive by IFA for P ovale and P vivax on the rst
sample submitted. The sample submitted 4 days
later was positive for P ovale and P malariae. In addition to cross-reactivity, subjectivity likely affects the
reproducibility of the IFA.
There are diagnostic methods available that are
more rapid, such as antigen detection (eg, immunochromatography) and nucleic acid detection. One
disadvantage of immunochromatography is its decreased sensitivity as parasitemia falls below a certain threshold, such as 100 parasites/L, and these
are the instances in which an adjunctive test to the
Giemsa smear would be most useful. Polymerase
chain reaction has shown to be an excellent diagnostic tool in the acute setting and may eventually
become more widely available.12
The specicity of the Cellabs EIA kit among
healthy blood donors was only 92% using the manufacturers recommended method for determining
the cutoff OD value. On the other hand, the manufacturer has previously published a study using an
OD of 3 standard deviations above the mean of negative controls as a method of determining the cutoff
value.13 If this method were used, the Cellabs EIA
would have yielded a higher specicity without
compromise of the sensitivity (data not shown).
Studies of the Newmarket Malaria EIA have followed the manufacturers protocol for determining
cutoff OD values for each run.6,8 Standardization of
these assays by testing additional random healthy
donors will yield a more accurate method of determining positive cutoff OD value for optimal sensitivity, specicity, and normalization of between-run
variations.
Acknowledgments
This work was supported by the ARUP Institute for
Clinical and Experimental Pathology. A special
thanks to Cellabs Pty Limited and Newmarket
111
7.
References
1. Eliades MJ, Shah S, Nguyen-Dinh P, et al. Malaria
surveillanceUnited States 2003. MMWR 2005;
54:2539.
2. Draper CC, Sirr SS. Serological investigations in
retrospective diagnosis in malaria. Br Med J 1980;
280:15751576.
3. Collins WE, Skinner JC, Jeffery GM. Studies on the
persistence of malarial antibody response. Am J Epidemiol 1968; 87:592598.
4. Tobie JE, Abele DC, Hill GJ, et al. Fluorescent antibody studies on the immune response in sporozoiteinduced and blood-induced vivax malaria and the
relationship of antibody production to parasitemia.
Am J Trop Med Hyg 1966; 15:676683.
5. Sulzer AJ, Wilson M, Hall EC. Indirect uorescentantibody tests for parasitic diseases. Am J Trop Med
Hyg 1969; 18:199205.
6. Chiodini PL, Hartley S, Hewitt PE, et al. Evaluation
of a malaria antibody ELISA and its value in reducing
9.
10.
11.
12.
13.