The efficacy of a malarial antibody enzyme immunoassay for establishing the reinstatement status of blood donors potentially exposed to malaria.

Abstract

The two key objectives of the study were, first, to evaluate the sensitivity and specificity of a recombinant antigen-based malarial enzyme-linked immunoassay (EIA) and, second, to estimate the risk associated with implementing this test with a shortened cellular component restriction period (6 months rather than the standard 12-36 months) for blood donors with a malarial risk exposure.
Blood donors were recruited into four distinct groups [non-exposed (control), malarial area 'visitors', 'residents' and 'previous infection') and screened by using the Newmarket malarial antibody EIA. Assay specificity was evaluated in unexposed blood donors, and sensitivity was determined in acute clinical samples.
No parasitaemic donors were detected amongst 337 malarial 'visitors' who had returned from a malaria-endemic area less than 6 months previously, or for 402 'visitors' or 'residents' who had returned from a malaria-endemic area more than 6 months previously. The incidence of malarial antibodies within the exposed blood donor groups was 1.33% (10/751). In acute clinical non-donor samples, the Newmarket EIA detected 106/108 (98.1; 93.5-99.5%) 'film' positive Plasmodium falciparum infections and 12/12 (100, 75.7-100.0%) P. vivax infections. The estimated additional risk exposure of the proposed new strategy was one infectious P. falciparum donation per 175 years or 1 per 4.2 years for P. vivax.
The study findings support the efficacy and safety of a targeted screening strategy combining antibody screening with a 6-month cellular component restriction period for donors with a declared malarial risk.

Full text

Page 1

98
Vox Sanguinis
(2005)
88
, 98–106
ORIGINAL PAPER
©
2005 Blackwell Publishing
Blackwell Publishing, Ltd.
The efficacy of a malarial antibody enzyme immunoassay for
establishing the reinstatement status of blood donors
potentially exposed to malaria
C. R. Seed,
1
Australian Red Cross Blood Service, Perth, Australia
2
School of Medicine and Pharmacology, University of Western Australia, Fremantle Hospital, Fremantle, Australia
3
Australian Red Cross Blood Service, Sydney, Australia
4
National Transfusion Microbiology Reference Laboratory, English National Blood Service, London, UK
1
A. Cheng,
1
T. M. E. Davis,
2
W. V. Bolton,
3
A. J. Keller,
1
A. Kitchen
4
& T. J. Cobain
1
Background and Objectives
the sensitivity and specificity of a recombinant antigen-based malarial enzyme-linked
immunoassay (EIA) and, second, to estimate the risk associated with implementing
this test with a shortened cellular component restriction period (6 months rather than
the standard 12–36 months) for blood donors with a malarial risk exposure.
Materials and Methods
Blood donors were recruited into four distinct groups
[non-exposed (control), malarial area ‘visitors’, ‘residents’ and ‘previous infection’)
and screened by using the Newmarket malarial antibody EIA. Assay specificity
was evaluated in unexposed blood donors, and sensitivity was determined in acute
clinical samples.
Results
No parasitaemic donors were detected amongst 337 malarial ‘visitors’ who
had returned from a malaria-endemic area less than 6 months previously, or for
402 ‘visitors’ or ‘residents’ who had returned from a malaria-endemic area more
than 6 months previously. The incidence of malarial antibodies within the exposed
blood donor groups was 1·33% (10/751). In acute clinical non-donor samples,
the Newmarket EIA detected 106/108 (98·1; 93·5–99·5%) ‘film’ positive
falciparum
infections and 12/12 (100, 75·7–100·0%)
ated additional risk exposure of the proposed new strategy was one infectious
P. falciparum
donation per 175 years or 1 per 4·2 years for
Conclusions
The study findings support the efficacy and safety of a targeted screen-
ing strategy combining antibody screening with a 6-month cellular component
restriction period for donors with a declared malarial risk.
Key words:
blood donor, falciparum, malaria, screening, transfusion.
The two key objectives of the study were, first, to evaluate
Plasmodium
P. vivax
infections. The estim-
P. vivax
.
Received: 5 July 2004,
revised 17 November 2004,
accepted 4 December 2004
Introduction
Transfusion-transmitted malaria, although rare, continues
to pose a risk to blood services worldwide [1,2]. Of the four
species of
Plasmodium
that cause human malaria,
is the most serious transfusion risk because it is fatal in
P. falciparum
approximately 10% of patients infected in this way [2]. The
last documented Australian incidence of transfusion-
transmitted malaria, a fatal infection with
occurred in 1991. This was the first reported transfusion
transmission of malaria since 1960 [3]. As approximately
1
×
10
donations are collected annually in Australia, the
incidence of reported transfusion-transmitted malaria since
1991 is probably < 1 in 1
×
10
Owing to the lack of a suitable high-throughput laboratory
test, screening for malaria in Australian blood donors
currently relies on collecting a comprehensive medical and
P. falciparum
,
6
7
.
Correspondence
6000, Australia
E-mail: cseed@arcbs.redcross.org.au
: Clive R. Seed, 290 Wellington St., Perth, Western Australia

Page 2

©
2005 Blackwell Publishing Ltd.
Vox Sanguinis
(2005)
88
, 98–106
Malarial antibody screening of blood donors
99
travel history as part of the donor assessment process and
exclusion of cellular blood components from those with
potential malaria exposure. The Australian Red Cross Blood
Service (ARCBS)
Guidelines for the Selection of Blood Donors
specifies that donors who have visited a country where
malaria is endemic (termed ‘visitors’) are restricted to donat-
ing only plasma for fractionation for a period of 12 months
after their return. In addition, donors who have spent a total
of 6 months or more within the last 3 years in an endemic
area (termed ‘residents’), or those with a previous history of
malaria, are restricted to fractionated plasma donations for
3 years after their return. It is noteworthy that the ARCBS
definition of a ‘resident’ differs from that used by the American
Association of Blood Banks (AABB) and Council of Europe,
which both identify residents based on birth or residence
in a malarial country in the first 5 years of life. Despite
this divergence, the ARCBS policy is effective in preventing
transfusion transmission, given the observed low incidence.
However, it results in the ARCBS discarding
cells per annum (ARCBS unpublished data), most of which
are non-infectious. This level of wastage, in the light of
mounting pressure to maintain sufficiency as a result of
continually more restrictive donor selection criteria, has
intensified the need for an efficient and reliable laboratory
donor-screening test for malaria.
Potential candidates for such a test include traditional light
microscopy, plasmodial antigen assays, and plasmodial
RNA/DNA and serological assays. Of these methods, sero-
logical assays based on the reference indirect fluorescence
antibody test (IFAT) or enzyme immunoassay (EIA) have been
most widely applied to screen blood donors identified at risk
by questioning [1,4–6].
The high-throughput EIA serological screening assays are
attractive candidates for donor screening. Malarial antibod-
ies to all four species are produced in virtually all individuals
1–2 weeks after initial infection and persist for 3–6 months
after parasite clearance [7–9]. As antibodies are maintained
at high titre whilst parasites are present, infectious blood
donations should be effectively excluded by using a sensitive
antibody assay. However, antibodies persist after parasite
clearance, resulting in immune, previously infected but now
non-infectious blood donors also being excluded. The impact
of this loss on the overall efficacy of any antibody-based
screening programme will be dependent on the antibody
prevalence in the donor population in addition to the assay
specificity.
Previous studies on antibody EIAs have demonstrated
sensitivity for
P. falciparum
in the range of 76–100% in acute
samples, while sensitivity for
species in Australia, was limited to 50–56% because of the
lack of
P. vivax
-specific antigens [1,5]. Recently, improved
EIAs incorporating specific recombinant antigens for
have become commercially available. These assays potentially
≈
35 000 red
P. vivax
, the most prevalent
P. vivax
address the limitation in sensitivity, particularly for
expressed by some in respect of single antigen tests [6,10,11].
In this report, a multiantigen malaria antibody EIA was
evaluated to determine its suitability as a candidate test for
use within a proposed early reinstatement protocol for malarial
risk donors. In addition, a mathematical model was developed
and applied to estimate the risk associated with implement-
ing the new protocol in comparison with the existing policy.
P. vivax
,
Materials and methods
Study samples
We recruited four groups of blood donors based on a history
of malaria and/or potential exposure (Fig. 1), and also had
access to serum samples from patients recruited to clinical
studies of malaria pathophysiology and treatment.
Non-exposed blood donors (Group 1)
Five-hundred and three random ARCBS blood donors from
regional and metropolitan collection sites within the state of
Western Australia were tested to determine assay specificity.
These donors had no previous history of malaria and had not
travelled to a malaria-endemic area in the previous 3 years,
making it highly unlikely that they had been exposed to
malaria during this time-period. Each donor was invited to
participate in the study either at the time of donation, or by
telerecruitment, using an informed consent protocol approved
by the ARCBS institutional ethics committee. Briefly, donors
were requested to read an information statement and to sign
a separate consent form after which an additional 6 ml of
blood (EDTA) was collected for study purposes. Donors
initially reactive (IR) on the screening EIA were retested in
duplicate. Any repeatedly reactive (RR) donors (i.e. reactive
on one or both retests) were subjected to confirmatory test-
ing, counselled and referred for specialist follow-up if found
to be positive in any malarial confirmatory assays used
within the study protocol.
Potentially exposed malarial donors (Groups 2, 3
and 4)
A total of 751 blood donors restricted from cellular compon-
ent donation from regional and metropolitan collection
sites within the state of Western Australia were tested in four
different subgroups. All donors were recruited with use of the
same study protocol and consent process described for
non-exposed (Group 1) donors.
Visitors (Group 2).
all of whom had visited a malaria-endemic area within the
preceding 12 months. Under current ARCBS guidelines, such
donors are restricted to donating plasma only for further
fractionation for a period of 12 months after leaving the affected
This comprised a total of 701 donors,

Page 3

100
C. R. Seed
et al.
©
2005 Blackwell Publishing Ltd.
Vox Sanguinis
(2005)
88
, 98–106
area. Visitors were further subdivided into Group 2A (those
who had returned from the risk area less than 6 months
previously;
n
= 337) and Group 2B (those who had returned
between 6 and 12 months previously;
n
= 364).
Residents (Group 3).
all of whom had resided in a malaria-endemic area for a
cumulative time-period of
≥
6 months during the preceding
3 years and had subsequently resided in Australia for more
than 6 months. Current ARCBS guidelines restrict such
donors to plasma donation (fractionated) only for a period of
3 years after leaving the affected area.
The residents group comprised 38 donors,
Previously infected donors (Group 4).
12 blood donors with a previous history of clinical malaria:
six having had malaria within the past 3 years, and six
having had malaria more than 3 years previously.
This group comprised
Clinical samples
One-hundred and twenty samples from patients with acute
infection, all with microscopy-confirmed parasitaemia, were
sourced from the School of Medicine and Pharmacology,
University of Western Australia. One-hundred and eight
P. falciparum
samples were collected between August 1991
and August 2002 from China, Vietnam, Malaysia and Cam-
bodia. The age of the patients ranged from 3 to 62 years, and
comprised 30 females and 78 males. Twelve
were collected in August 1991 and August 2002 from China
and Cambodia. The age of the patients ranged from 16 years
of age to 48 years, and comprised four females and eight males.
P. vivax
samples
Malarial antibody screening EIAs
All donor samples were tested in accordance with the manu-
facturer’s recommended protocol by using the Newmarket
malaria EIA (Newmarket Laboratories Ltd, Newmarket, UK;
distributed in Australia by PANBIO Ltd, Brisbane, Australia).
This assay is a sandwich EIA incorporating four recombinant
antigens for
P. falciparum
and
immunogobulin G (IgG), immunogobulin M (IgM) and
immunogobulin A (IgA).
P. vivax
, detecting specific
Supplementary malarial testing
All Newmarket RR donors, in addition to all samples from
Groups 2B, 3 and 4, were further tested in the following assays
in accordance with the manufacturer’s instructions:
(1) NOW® ICT malaria Pf/Pv test (Binax Inc., Portland,
ME): a rapid immunochromatographic test for the detection
of circulating
P. falciparum
antigen and a separate antigen
common to
P. vivax
,
P. malariae
(2) OptiMAL® rapid malaria test (DiaMed AG, 1785 Cressier,
Switzerland): a rapid malarial test based on the detection of
parasite lactate dehydrogenase (pLDH), an enzyme produced
by all four
Plasmodium
spp.
(3) Malarial polymerase chain reaction (PCR) assay (
Malaria PCR kit; Artus GmbH, Hamburg, Germany): a malarial
DNA assay based on the detection of a 140-bp plasmodial
DNA sequence by using Real-Time PCR in a Lightcycler
(Roche Molecular Diagnostics, Pleasanton, CA) instrument.
Plasmodial DNA was extracted from 200
by using the QIAmp DNA blood Mini-Kit (Qiagen GmbH,
and
P. ovale
in whole blood.
RealArt
®
µ
l of whole blood
Fig. 1 Donor study population.

Page 4

©
2005 Blackwell Publishing Ltd.
Vox Sanguinis
(2005)
88
, 98–106
Malarial antibody screening of blood donors
101
Hilden, Germany), according to the manufacturer’s recom-
mended protocol. As defined in the Lightcycler (LC) PCR assay
protocol, 2
µ
l of eluate from each donor sample was used as
the starting volume for each PCR reaction. The product insert
claims a sensitivity of 10–100 genomes/
using a dilution series of cloned
evaluation of acute samples from Thailand, the manufacturer
claims a detection limit of 1 parasite/
data).
µ
l, determined by
DNA. In a recent
P. falciparum
µ
l (Artus-Biotech internal
Donor testing algorithm
All donors in Groups 1–4 were tested by the Newmarket EIA
singly and, if reactive, the tests were repeated in duplicate.
RR samples were subject to all three supplementary tests to
determine their final infection status. In addition, all donors
in Groups 2B, 3 and 4 were subject to supplemental testing
to determine if any had detectable parasitaemia. In order to
clarify the antibody status of RR EIA samples, each was further
tested by IFAT.
Risk analysis model
In order to estimate the risk associated with implementing
our proposed reinstatement protocol (Fig. 2), combining
antibody screening with a reduction of the current cellular
product restriction to 6 months, we developed a mathematical
model. The model assumes that the additional risk of imple-
menting the proposed protocol over the existing policy can be
considered as the probability that an asymptomatic parasitaemic
blood donor with undetectable antibody attends to donate more
than 6 months after their declared risk exposure. This
probability can be subdivided into three components:
(1) Incidence (I): the probability of an asymptomatic par-
asitaemic donor attending more than 6 months after leaving
a risk (malarious) area.
(2) Test failure rate (S): the probability that the implemented
test will fail to detect an asymptomatic parasitaemic donor.
(3) Exposure risk (E): the probability that any asymptomatic
parasitaemic donor is infected with a particular
species.
The additional risk exposure then is the product of these
three probabilities:
Plasmodium
Risk = (I
×
S
×
E)(1)
Incidence
The most accurate estimate of the incidence of asymptomatic
parasitaemia would be derived directly from that in Groups
2B and 3, who represent a sample of donors from our popu-
lation in whom the elapsed period since the last exposure
ranges from 6 to 36 months. This was not possible owing to
the lack of a parasitaemic patient in these two groups. There-
fore, by necessity, we indirectly estimated the incidence by
using the following method.
Data from three studies investigating the interval between
infection and diagnosis were analysed to define the incidence
of malaria diagnosed within 6 months of exposure in com-
parison with that after 6 months of exposure. In the first
study of imported malaria in the USA in 1995, Williams
[12] found that 364 of 367 (99·2%)
and 325 of 432 (75·2%)
P. vivax
within 6 months of arrival in the USA. Second, in a similar
UK study of imported malaria, 1427 of 1434 (99·5%)
ciparum
and 543 of 829 (65·5%)
within 6 months of arrival in the UK [13]. Available data on
diagnoses from Western Australia for the period 1990–2001
are consistent with these reports in that for
all 85 (100%) infections with falciparum malaria occurred
within 6 months of arrival. The data for
are markedly different, with 180 of 184 (97·8%) being detected
within 6 months of arrival in Australia [14]. Because of the
close agreement of the
P. falciparum
sources, they were pooled to increase the confidence of the
estimate. This resulted in a combined total of 1886 infections,
of which 1876 (99·5%) were diagnosed within 6 months. In
et al
.
P. falciparum
infections, were diagnosed
infections,
P. fal-
P. vivax
infections occurred
P. falciparum
,
P. vivax,
however,
data from the three


Fig. 2 Proposed reinstatement protocol. EIA, enzyme immunoassay.

Page 5

102
C. R. Seed
et al.
©
2005 Blackwell Publishing Ltd.
Vox Sanguinis
(2005)
88
, 98–106
the case of
as it showed little agreement, ranging from 65·5 to 97·8% of
infections diagnosed within 6 months. Given that the donor
population assessed was from Western Australia, it was
decided that the best estimate would be the corresponding
population of patients. Therefore, the subsequent
risk calculations are based on an estimate of 97·8% infections
being detected within 6 months. It is possible to use these
data, in addition to the observed asymptomatic incidence in
visitors who had returned from malaria-endemic areas for less
than 6 months (Group 2A), to estimate the incidence of an
asymptomatic parasitaemic donor in Group 2B/3. First, based
on the above diagnoses data, we assume that the proportions
of undiagnosed malaria in the < 6 month (2A) and > 6 month
(2B/3) donor groups are, respectively, 0·995 and 0·005 for
P. falciparum
and 0·978 and 0·022 for
proportions and the known incidence in Group 2A, the estimated
incidence can be calculated by using the following formula:
P. vivax
, it was not appropriate to pool the data
P. vivax
P. vivax
. Using these
(I1
÷
I2) = (P1
÷
P2).(2)
Solving for I1:
I1 = [I2
×
(P1
÷
P2)], (3)
where I1 is the incidence of asymptomatic parasitaemia in the
donors who returned from malaria-endemic areas more than
6 months previously (Group 2B/3), and I2 is the incidence in
donors who returned from malaria-endemic areas less than
6 months previously (Group 2A). P1 is the predicted propor-
tion of undiagnosed malaria in donors who returned more
than 6 months previously, and P2 the proportion of un-
diagnosed malaria occurring within 6 months of return from
an endemic area.
Test failure (S)
In the absence of the ability to directly evaluate sensitivity
(i.e. parasitaemia) in the target population (Group 2B/3)
because of the lack of a parasitaemic donor, it was decided
that the most appropriate sensitivity estimate would be that
from acute clinical samples:
Test failure (S) = (1
−
Sensitivity) (4)
Exposure risk (E)
The probability that any undetected parasitaemic donor
would be infected with a particular species of malaria was
derived by analysing the countries visited by the malarial-
risk donors recruited to the study. Each individual visit was
assessed on the basis of
Plasmodium
consulting the World Health Organization (WHO) ‘Inter-
spp. risk assigned by
national Travel and Health’ website (www.who.int./ith/). Visits
to both low and high malarial-risk countries were included.
In the event of visits to countries where more than one species
of Plasmodium was identified as a risk, the visit was counted
as an independent event for each species.
Statistical methods
The confidence intervals for assay specificity, sensitivity and
risk point estimates were derived by using the Wilson score
method (Method 3, in Newcombe [15]).
Results
Assay specificity: non-exposed blood donors
Of the 503 presumably non-exposed blood donors in Group
1, two (0·40%) were RR, but neither had detectable parasitae-
mia. Both RR samples had reactivity on IFAT (one positive,
one equivocal) indicating immunity and were therefore
excluded from the presumed antibody-negative population
for the specificity calculation. After excluding samples with
IFAT reactivity, the assay specificity is 100% (99·2–100·0%).
Malarial-risk donors (Groups 2A, 2B, 3 and 4)
The screening and IFAT results for all four malarial-risk
groups are summarized in Table 1.
The overall incidence of antibody-positive malarial donors,
as defined by their IFAT status, was 10/751 (1·33%). When
compared by exposure group, the incidence within the ‘visi-
tors’ Groups 2A and 2B was not markedly different (0·3% vs.
0·8%). Predictably, however, the combined incidence of all
visitors (Groups 2A and 2B) was lower than that of the ‘resi-
dent’ Group 3 (0·6% vs. 7·9%) and of the ‘previously infected’
Group 4 (0·6% vs. 25%), indicative of a lower exposure risk.
Supplementary testing
No donors in any group had detectable parasitaemia, defined
by concordantly positive results on at least two of the three
supplementary assays.
Clinical sensitivity in acute malaria
The Newmarket EIA detected 106/108 (98·1; 93·5–99·5%)
P. falciparum and 12/12 (100, 75·7–100·0%) P. vivax infections.
Risk analysis
Applying the study data to the model (the full calculation
is shown in Appendix 1), the estimated additional risk of
releasing an infected unit of blood is 1 in 6 109 055

Page 6

© 2005 Blackwell Publishing Ltd. Vox Sanguinis (2005) 88, 98–106
Malarial antibody screening of blood donors
103
(1 078 371–34 608 286) and 1 in 147 043 (25 957–833 011)
for P. falciparum and P. vivax, respectively.
Discussion
As symptomatic donors should be excluded at interview, the
greatest risk to the blood supply from malarial transfusion
transmission is posed by an asymptomatic parasitaemic
donor. Retrospective analysis of donors implicated in trans-
missions has identified that such donors are predominantly
‘semi-immune’, having been raised in malaria-endemic
countries and having experienced repeated malaria infection
in childhood [2,16,17]. They are typically asymptomatic, but
can harbour low-level parasitaemia, often undetectable by
even the most sensitive diagnostic assays [18]. Previous
studies indicate that, because semi-immune individuals retain
high antibody titres [19–21], a sensitive antibody assay
would be expected to identify and exclude such donors. This
was the rationale for selecting a, malarial antibody EIA for
evaluation in the current study.
The lack of a detectable parasitaemic infection in an indi-
vidual of groups 2A or 3, whose blood products could poten-
tially be released under the proposed policy rather than being
routinely discarded, made estimating the sensitivity (for
malaria infection) of the Newmarket EIA in asymptomatic
parasitaemic donors impossible. In response to this limitation,
and consistent with other studies [1,5], our approach was to
assume that the dynamics of the malarial antibody response
in acute infection adequately mirrored that in an asymptomatic
parasitaemic blood donor with delayed or absent symptom
onset. Given this assumption, the assay’s performance in
acute infection (parasitaemic samples) provides a measure
of its ability to detect asymptomatic parasitaemic donors
who present more than 6 months after their last exposure.
Assessed by this standard, the sensitivity of the Newmarket
EIA for P. falciparum and P. vivax infection in samples
confirmed by blood film was an impressive 98·1% and 100%,
respectively. The P. vivax sensitivity, in particular, is
markedly improved over the 50–65% reported in previous
studies [1,5]. This sensitivity estimate calculated in acute
samples is conservative because our proposed policy mandates
a 6-month delay before testing, negating the impact of the
estimated 7–14 day serological window period observed in
early infection [11–13].
In order to optimize our proposed strategy, candidate
assays must have high specificity to ensure minimal exclu-
sion of donors as a result of ‘false-positive’ test results. The
Newmarket assay demonstrated excellent specificity when
evaluated against a presumed antibody-negative donor pop-
ulation. The efficiency of our proposed screening strategy is
also dependent on the prevalence of antibody-positive donors
in the malarial-risk donor population because a donor’s
reinstatement is conditional on a non-reactive result. The
observed incidence of malarial antibodies in all malarial-risk
groups sampled was 1·3%, based on RR samples confirmed
by IFAT. This figure is comparable to the 1·5–1·7% range for
published studies performed in blood donors from the UK [1]
and New Zealand [5]. Based on the observed RR rate of
2·3% (17/751), implementing the revised protocol would be
expected to result in 97·7% of donors being reinstated at least
6 months earlier than the current policy. Assuming that each
donor donates at the ARCBS average rate of twice in a 12-
month period, this represents a potential saving of 17 098
(0·977 × 35 000 × 6/12) additional units of blood (red blood
cells and platelets) available for transfusion annually in
Australia. Such a saving would be of considerable value in
countering the progressive loss of donors/donations associ-
ated with the implementation, in Australia, of safety-related
restrictions related to variant Creutzfeldt–Jacob diseasev (CJD),
West Nile Virus (WNV), Dengue, Severe Acute Respiratory
Syndrome (SARS) and, most recently, increased haemoglobin
thresholds for donor acceptance.
Table 1 Screening enzyme immunoassay (EIA) and indirect fluorescence antibody test (IFAT) results for malarial risk donors


VisitorsResidentsPrevious history of malaria
Group 2A (n = 337)Group 2B (n = 364)Group 3 (n = 38)Group 4 (n = 12)
IR RR
IFAT+a
No.b
par. IRRRIFAT+
No.
par.IRRR IFAT+
No.
par. IRRRIFAT+
No.
par.
Newmarket EIA
%
6
1·8
4
1·2
1c
0·3
07
1·9
6
1·6
3d
0·8
03
7·9
3
7·9
3e
7·9
04430
33·333·325·0
aNumber of repeatedly reactive (RR) donors confirmed by IFAT at the English National Transfusion Microbiology Reference Laboratory (NTMRL).
bNumber confirmed as parasitaemic and defined as positive in two or three confirmatory assays.
cIFAT was not performed on one sample.
dIFAT was equivocal on three samples.
eIFAT was equivocal on one sample.

Page 7

104
C. R. Seed et al.
© 2005 Blackwell Publishing Ltd. Vox Sanguinis (2005) 88, 98–106
One important goal of this study was to estimate the addi-
tional risk exposure for transfusion-transmitted malaria
assuming implementation of our proposed protocol. This is a
critical piece of information as it provides for an evidence-
based risk/benefit analysis. In assessing this risk exposure,
it is important to consider the following. First, the primary
transfusion risk relates to an undetected infection with
P. falciparum, as transmissions involving this species can be
fatal for the recipient in ≈ 10% of cases, whereas transmissions
with other species rarely, if ever, result in death [2]. Second,
transfusion-transmitted malaria cannot be eliminated com-
pletely by current procedures because asymptomatic infections
can remain undiagnosed for longer than the maximum 3-year
restriction period. As a result of the asymptomatic persistence
of parasites, transmission of P. malariae has been documented
as long as 53 years [22], P. vivax 27 years [23] and P. falciparum
13 years [23] after the last exposure. Finally, a review of the
US transmission data by Mungai et al. established that 62%
of donors implicated in a subsequent transmission should
have been deferred from donation if exclusion criteria apply-
ing at the time had been correctly applied [2]. Even when a
malaria-risk donor is identified correctly at interview,
omissions in the application of this information have led to
transmissions. Indeed, this scenario led to the last recorded
transfusion transmission in Australia in 1991 [3].
Assuming 35 000 ‘malarial’ donations screened annually,
risk modelling predicts that implementing our proposed
strategy with the Newmarket EIA would result in an addi-
tional 0·006 P. falciparum- and 0·24 P. vivax-infected dona-
tions per annum. This equates to the release of one additional
infectious donation per 175 years and 4·2 years for P. falci-
parum and P. vivax, respectively. Although any additional
risk exposure is unwelcome, these risks are exceptionally
remote and less than the estimated risk (1 every 3·5 years) of
the release of a human immunodeficiency virus (HIV) infec-
tious unit in Australia [24].
The modelling process required certain assumptions,
which may impact the accuracy of the risk estimates. How-
ever, as appropriate when considering blood safety, we con-
sistently selected the most conservative options. Our model
assumes that the timing of malaria diagnosis, specifically the
proportions apparent before and after 6 months, does not
vary significantly in populations of donors and patients
from the same geographical region. We found no published
evidence to refute the accuracy of this assumption. A further
potential limitation of the model is that the sensitivity
analysis used to predict test failure is derived from samples
obtained from patients in the acute phase of infection, rather
than from group 2A/3 donors directly. As discussed previ-
ously, it is likely that this will underestimate the sensitivity,
leading to an overestimation of the probability of test failure
and consequently an overestimation of the risk. Finally, the
estimates do not consider the risk of transmission of P. ovale
and P. malariae based on their rarity in Australia (< 4% total
notifications). Our analysis of donor travel data failed to
identify any donor visits to countries with a current risk for
either species, confirming the low risk associated with these
species in our donor population.
It is also important to weigh the negative risk, as calculated
by the model, with the potential benefit of implementing
testing. Although this was not quantified in the study, it is
probable that the net effect would be a risk reduction. In sup-
port of this, Chiodini et al. [1] argued that three incidents of
transfusion transmission in the UK could have been avoided
by a dual strategy of testing and history taking.
The efficacy of antibody-based screening strategies is evi-
dent from programmes already in place in Europe. In France,
for example, travel-based questioning, in combination with
IFAT, has been used as part of a targeted screening strategy
since 1983 and continues currently [7]. Donors returning
from malaria-endemic countries are initially deferred for
4 months and then IFAT tested between 4 months and 3 years
with a negative test qualifying the donor for reinstatement.
There has not been a reported transmission using this
approach since 1994, when mandatory reporting of transfu-
sion complications was instituted [6]. Based on the study of
Chiodini et al. [1], the English National Blood Service imple-
mented a similar system in 1997, although it required a 6-
rather than a 4-month delay before testing. Screening was
suspended in 1998, but recommenced in 2001, initially using
parallel testing by an in-house IFAT in combination with the
Newmarket EIA, but more recently relying solely on the latter
assay [25]. However, because the UK has the highest level of
imported malaria in Europe and significant ethnic diversity
in its donor population, the donor selection criteria were
amended to ensure that the highest-risk donor groups, pre-
vious residents and those previously infected, are always
antibody screened prior to reinstatement and after any
additional trips to any malarious area [26].
In summary, the high sensitivity for both P. falciparum
and P. vivax observed for the Newmarket EIA make it an
excellent candidate assay for inclusion in our proposed
testing strategy. The risk analysis confirms that the additional
risk of releasing an infectious donation associated with
implementing our proposed strategy is negligible, being
lower than the estimated risk of releasing an HIV-infected
donation in Australia. When balanced against the signific-
ant number of cellular blood components potentially made
available, we believe that the risk/benefit ratio of our pro-
posed targeted screening protocol is positive and therefore
supports implementation.
Acknowledgements
The authors give sincere thanks to the ARCBS-Perth nursing
staff for donor recruitment/test collection and Juliana

Page 8

© 2005 Blackwell Publishing Ltd. Vox Sanguinis (2005) 88, 98–106
Malarial antibody screening of blood donors
105
Hamzah for assistance with figures/tables. Wendy Davis and
Alexandra Bremner are thanked for statistical assistance.
In addition, the authors acknowledge Dr June Lee and Dr
Anthony Low (ARCBS-Perth) for clinical advice/donor coun-
selling, and Sonya Sultana (ARCBS-Sydney) and Patricia
Lowe (NTMRL-English National Blood Service) for perform-
ing the PCR and IFAT supplemental testing, respectively.
PANBIO Ltd/Newmarket Laboratories (UK) are thanked for
providing their screening EIA for evaluation. This study was
funded, in part, by the National Blood Authority (Australia)
and the National Australia Bank.
References
1 Chiodini PL, Hartley S, Hewitt PE, Barbara JA, Lalloo K, Bligh J,
Voller A: Evaluation of a malaria antibody ELISA and its value
in reducing potential wastage of red cell donations from blood
donors exposed to malaria, with a note on a case of transfusion-
transmitted malaria. Vox Sang 1997; 73:143–148
2 Mungai M, Tegtmeier G, Chamberland M, Parise M: Transfusion-
transmitted malaria in the United States from 1963 through
1999. N Engl J Med 2001; 344:1973–1978
3 Whyte GS: Therapeutic goods and malaria. Med J Aust 1992;
157:439–440
4 Brasseur P, Bonneau JC: [Transfusion-induced malaria: risks,
prevention and cost (1 year’s experience).] Rev Fr Transfus
Immunohematol 1981; 24:597–608
5 Davidson N, Woodfield G, Henry S: Malarial antibodies in
Auckland blood donors. N Z Med J 1999; 112:181–183
6 Silvie O, Thellier M, Rosenheim M, Datry A, Lavigne P, Danis M,
Mazier D: Potential value of Plasmodium falciparum-associated
antigen and antibody detection for screening of blood donors to
prevent transfusion-transmitted malaria. Transfusion 2002;
42:357–362
7 Hassig A, Ambroise-Thomas P, Bruce-Chwatt L, Emmanuel J,
Masvre R, Perrin L, Soulier JP: Which are the appropriate modi-
fications of existing regulations designed to prevent the transmis-
sion of malaria by blood transfusion, in view of the increasing
frequency of travel to endemic areas? Vox Sang 1987; 52:138–148
8 Contreras CE, Pance A, Marcano N, Gonzalez N, Bianco N:
Detection of specific antibodies to Plasmodium falciparum in
blood bank donors from malaria-endemic and non-endemic
areas of Venezuela. Am J Trop Med Hyg 1999; 60:948–953
9 Park CG, Chwae YJ, Kim JI, Lee JH, Hur GM, Jeon BH, Koh JS,
Han JH, Lee SJ, Park JW, Kaslow DC, Strickman D, Roh CS: Sero-
logic responses of Korean soldiers serving in malaria-endemic
areas during a recent outbreak of Plasmodium vivax. Am J Trop
Med Hyg 2000; 62:720–725
10 Mertens G, Vervoort T, Heylen S, Muylle L: Malaria antibody
ELISA insufficiently sensitive for blood donor screening. Vox
Sang 1999; 77:237–238
11 Gillon J: Introduction of malaria antibody ELISA. Vox Sang
1998; 75:80–81
12 Williams HA, Roberts J, Kachur SP, Barber AM, Barat LM,
Bloland PB, Ruebush TK, 2nd, Wolfe EB: Malaria surveillance –
United States, 1995. MMWR CDC Surveill Summ 1999; 48:1–
23
13 Bradley D, Warhurst D, Blaze M, Smith V: Malaria imported into
the United Kingdom in 1992 and 1993. Commun Dis Rep CDR
Rev 1994; 4:R169–R172
14 Charles D, Hart J, Davis WA, Sullivan E, Dowse GK, Davis TME:
Notifications of imported malaria in Western Australia 1990–
2001. Incidence, associated factors and chemoprophylaxis. Med
J Aust 2005; in press
15 Newcombe RG: Two-sided confidence intervals for the single
proportion: comparison of seven methods. Stat Med 1998;
17:857–872
16 MMWR: Probable transfusion-transmitted malaria – Houston
Texas 2003. MMWR 2004; 52:1075–1076
17 de Silva M, Contreras M, Barbara J: Two cases of transfusion-
transmitted malaria (TTM) in the UK. Transfusion 1988; 28:86
18 Ramasamy R: Molecular basis for evasion of host immunity and
pathogenesis in malaria. Biochim Biophys Acta 1998; 1406:10–27
19 Draper CC, Sirr SS: Serological investigations in retrospective
diagnosis of malaria. Br Med J 1980; 280:1575–1576
20 Bruce-Chwatt LJ, Dodge JS, Draper CC, Topley E, Voller A:
Sero-epidemiological studies on population groups previously
exposed to malaria. Lancet 1972; 1:512–515
21 Collins WE, Skinner JC, Jeffery GM: Studies on the persistence of
malarial antibody response. Am J Epidemiol 1968; 87:592–598
22 Gauzzi M, Grazan S: A relapse of quartan malaria after 53 years
latency. Trop Dis Bull 1964; 61:11–12
23 Besson P, Robert JF, Reviron J, Richard-Lenoble D, Gentilini M:
[Two cases of transfusional malaria. Attempted prevention com-
bining an indirect immunofluorescence test with clinical selection
critera.] Rev Fr Transfus Immunohematol 1976; 19:369–373
24 Seed CR, Cheng A, Ismay SL, Bolton WV, Kiely P, Cobain TJ,
Keller AJ: Assessing the accuracy of three viral risk models in
predicting the outcome of implementing HIV and HCV NAT
donor screening in Australia and the implications for future
HBV NAT. Transfusion 2002; 42:1365–1372
25 Kitchen AD, Lowe PHJ, Lalloo K and Chrodimi PL: Evaluation of
a malarial antibody assay for use in the screening of blood and
tissue products for clinical use. Vox Sang 2004; 87:150–155
26 Kitchen A, Mijovic. A, Hewitt P: Transfusion transmitted malaria
– current donor selection guidelines are not sufficient. Vox Sang
2005; in press

Page 9

106
C. R. Seed et al.
© 2005 Blackwell Publishing Ltd. Vox Sanguinis (2005) 88, 98–106
Appendix 1
Risk analysis


Incidence (I)
Group 2A
None of the 337 donors recruited showed evidence of parasitaemia, therefore the estimated incidence of asymptomatic parasitaemia is < 1/337 (or < 0·00296).
Group 2B/3
Assuming the worst case, i.e. the incidence is 1/337 and using this incidence as I1 in eqn 3 {I1 = [I2 × (P1 ÷ P2)]}:
P. falciparum
I1 = [0·00296 × (0·005 ÷ 0·995)] = 0·0000149 (1 in 67 262)
P. vivax
I1 = [0·00296 × (0·022 ÷ 0·978)] = 0·0000666 (1 in 15 026)
Test failure (S)
Test failure (S) = 1 – Sensitivity
P. falciparum= 1 – 0·981 = 0·019
P. vivax= 1 – 0·757 = 0·243 (as the number of samples
assessed was small, n = 12, the lower 95%
sensitivity confidence interval was applied)
Exposure risk (E)
Total visits to endemic countries = 2252
P. falciparum risk visits: n = 1305
P. vivax risk visits: n = 947
Probability of exposure to:
P. falciparum:
= (1305 ÷ 2252) = 0·58
P. vivax:
= (947 ÷ 2252) = 0·42
Additional risk exposure
Risk = I × S × E
P. falciparum= (0·0000149 × 0·019 × 0·58) = 1·64 × 10−7 or 1 in 6 109 055
P. vivax
= (0·0000666 × 0·243 × 0·42) = 6·769 × 10−6 or 1 in 147 043
Predicted additional infected donations per annum
Assume 35 000 ‘malarial risk’ donations per annum
P. falciparum= (35 000 ÷ 6 109 055) = 0·006 infected donations per annum
P. vivax
= (35 000 ÷ 147 043) = 0·238 infected donations per annum