Mollison’s Blood Transfusion in Clinical Medicine - part 9 pdf

The first reported case of transfusion-associated
AIDS (TAA) turned out to be an 18-month-old infant
with severe combined immunodeficiency who had
been transfused repeatedly at birth and had received a
unit of platelets from a donor who subsequently devel-
oped AIDS (Amman et al. 1983). By 1984, 38 cases
of AIDS had been reported in patients with no risk
factors other than a history of transfusion. Nineteen
of the patients were adults who, during the previous
5 years, had received blood components derived from
unpooled donations. In those cases in which all the
donors could be followed up, an individual in a ‘high-
risk’ group could always be identified (Curran et al.
1984). An expanded study of 194 patients showed that
in most cases the high-risk donor was anti-HIV posit-
ive, and in those few cases in which the high-risk
donor lacked anti-HIV, another donor tested positive
(Peterman et al. 1985). A total of 157 525 cases of
AIDS was reported in the USA by the end of 1990, and
5371 (3.4%) of these were attributed to the trans-
fusion of blood or blood components. The median
incubation period, estimated as the time from expos-
ure by transfusion to diagnosis of AIDS, has been
estimated to be 34 months in adults and 22 months in
children (Peterman 1987), although later estimates
indicated a longer period with a median of 7–8 years
for adults and 3–5 years for children (Rogers 1992).
In total, 50% of untreated subjects infected with HIV
by transfusion will develop AIDS within 7 years com-
pared with 33% of subjects infected by other routes
(Ward et al. 1989).

Following the first reports, the number of cases of
TAA in the USA increased rapidly. By the end of 1991,
out of a total of 206 392 cases of AIDS, there were
6060 adults and 472 infants or children who had
acquired the disease by transfusion of blood or blood
products, representing 3% and 13.6% of the total
cases in adults and children respectively (CDC 1992).
The risk of transmission of HIV by blood transfusion
is now trivial in countries without major heterosexual
spread and in which donor education, encouragement
of self-exclusion and screening for HIV antibodies
have been established since 1985. In these countries,
the risk of transmission of HIV is almost solely
attributable to donations given during the window
period (see below).
In western Europe, HIV prevalence among blood
donors has declined progressively since the onset of
systematic testing and, according to the European
Centre for Epidemiologic Monitoring of AIDS,
approximated 1.3 positives per 100 000 donations
in 2002 (www.eurohiv.org). Since 1995, the preval-
ence has remained relatively constant in Belgium,
Scandinavia, Ireland, the Netherlands and the UK, has
decreased regularly in France and Spain, but has
remained above 2 per 100 000 in Italy, Greece and
Portugal (Hamers and Downs 2004). In Eastern
Europe, prevalence has increased markedly since
1995, now exceeding 30 per 100 000 in 2002. The
highest levels are reported in the Ukraine (93), Estonia
(54) followed by Azerbaijan, Georgia and the Russian

Federation.
Because HIV is both cell associated and present in
plasma, all blood components are potentially infec-
tious (Curran 1985). The viral load has been estimated
at between 1.5 × 10
4
tissue culture infective doses in a
250-ml unit of blood from an asymptomatic donor,
to 1.75 × 10
6
from a unit drawn from a symptomatic
person (Ho et al. 1989). The relative importance of the
viral strain, concurrent infection by other blood-borne
agents, cellular receptors for HIV and other genetic
and acquired host factors, both for infection and for
clinical course of HIV in transfusion recipients, has
received a great deal of attention, but is still an area
of intense research (Vicenzi et al. 1997; Keoshkerian
et al. 2003; Zaunders et al. 2004).
Follow-up studies have shown that 90–95% of
patients receiving blood or blood components from
anti-HIV-positive donors have become infected (Ward
et al. 1987; Donegan et al. 1990b). The virus is well
preserved in refrigerated and frozen blood; however,
components that are washed, leucoreduced or cold
stored for several weeks, procedures that diminish
the number of viable leucocytes or the amount of
virus, reduce the likelihood of transfusion transmis-
sion (Donegan et al. 1990a). Neither donor status nor
recipient characteristics affect the likelihood of HIV

transmission (Busch et al. 1990a). However, when
AIDS developed in the donor shortly after donation,
the period of asymptomatic infection in the recipient
was also shortened (Ward et al. 1989). Albumin pre-
parations, immunoglobulins, antithrombin III and
hepatitis B vaccine have not been associated with HIV
infection (Desmyter 1986; Melbye 1986; Morgenthaler
1989; Cuthbertson 1991). Furthermore, when HIV is
added to plasma and the plasma is then fractionated by
the cold-ethanol process, HIV does not appear in the Ig
fraction (Morgenthaler 1989).
CHAPTER 16
720
TAA in infants, and AIDS in infants and children in
general, has a shorter incubation period than in adults
and many of the clinical manifestations of the disease
are different (Rogers 1992). Infants born to anti-HIV-
positive mothers who become infected perinatally and
infants transfused during the first years of life have the
shortest median incubation period (less than 5 years),
and usually develop AIDS in the first year of life. The
increased susceptibility of infants to AIDS may be
related to their immature immune system and to the
larger viral load relative to body size. Children likely to
develop TAA are those who are likely to be transfused:
premature infants and children with haemophilia,
thalassaemia and sickle cell anaemia. In the USA, of
2734 children with AIDS, 250 (9.1%) were transfu-
sion associated and 138 (5.0% of the total) occurred in
children with coagulation factor deficiencies (CDC

1991).
In much of Asia and Africa, the transmission of HIV
by blood transfusion is still an important source of
infection. Reasons for an alarmingly high rate of trans-
mission, reported to be up to 10% of all cases, include:
(1) the high demands for blood for inpatients with
severe anaemia and haemorrhage, mainly in obstetrics,
gynaecology and paediatrics; (2) the prevalence of
HIV infection amongst the donor population, which
can be as high as 20%; (3) the fact that HIV infection is
not confined to a minority of the population who can
be requested to refrain from blood donation; and
(4) the inability of many laboratories to test for HIV or
to perform and control the tests properly. The patient
groups at greatest risk of acquiring HIV-1 or -2 by
blood transfusion in tropical Africa are children with
malaria and anaemia, patients with sickle cell anaemia
(120 000 infants with sickle cell disease are born each
year in Africa), anaemic antenatal patients, women
with severe obstetric haemorrhage and trauma victims
(Fleming 1990).
Transfusion-associated AIDS and haemophilia
To December 1996, 4674 cases of AIDS were reported
in patients with haemophilia, accounting for less
than 1% of the 581 429 AIDS cases reported in
adults and children. Of the 7629 cases of AIDS
reported in children under the age of 13 years, 373
(5%) were recipients of blood or tissue and 231 (3%)
were haemophiliacs (Centers for Disease Control and
Prevention 1997). All but 39 of these infections

occurred prior to testing blood and plasma donors
for HIV.
Studies of patient cohorts and specimens in serum
repositories revealed that more than 90% of severe
haemophiliacs (subjects with less than 1% factor VIII
activity) treated with factor VIII concentrate had been
infected prior to 1984 (Evatt et al. 1985). In a study of
a 16-centre cohort of haemophiliacs in the USA and
Europe, infections first appeared in 1978, peaked
in October 1982 and declined to an estimated four
infections per 100 person-years by July 1984 (Kroner
et al. 1994). For patients who were high-dose recipi-
ents, peak risk appeared even earlier, indicating that
the majority of patients with haemophilia were
infected before the disease was widely recognized and
long before it was attributed to transfusion. The risk
was related to each patient’s mean annual dose of clot-
ting factor concentrate. As clotting factor concentrates
are prescribed on the basis of patient weight or plasma
volume, older patients with severe disease generally
received more concentrate, had more ‘donor expos-
ures’ and seroconverted sooner than did children and
patients with milder disease. When corrected for dose
and severity of disease, the association between age
and early seroconversion disappears. The cumulative
incidence of infection was 96% for high-dose recipi-
ents, 92% for moderate-dose recipients and 56% for
low-dose recipients. Subjects who received only single-
donor products (plasma and/or cryoprecipitate) had
the lowest cumulative incidence of infection, 16%

(Kroner et al. 1994). This experience is consistent with
other reports (Andes et al. 1988; Gjerset et al. 1991).
These startling numbers underscore the potential
public health risks of using transfusion products
manufactured from pools of plasma drawn from
20 000 donors or more.
Although AIDS was first reported in three
haemophiliacs in 1982, studies of serum samples
stored from as far back as 1968 have shown that
the first cases of the development of anti-HIV in
haemophiliacs occurred in 1978 in the USA and in
1981 in the UK (Evatt et al. 1983; Machin et al. 1985;
Ragni et al. 1986). Plasma for preparing the implicated
clotting factor concentrates may have been collected a
year prior to that. The prevalence of HIV seropositiv-
ity and of AIDS varies from one haemophilia centre to
another depending on the source, volume and type of
concentrate used. In a study of 13 haemophilia centres
in western Europe, Canada and Australia involving
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
721
2370 patients with haemophilia A and 434 patients
with haemophilia B, the overall incidence of anti-
HIV was 53.6% (CDC 1987a). The percentage of
haemophiliacs infected with HIV in different countries
has varied from 4% to more than 60%, with the higher
rates in those countries using mainly concentrates
from plasma imported from the USA. Some batches of
factor VIII concentrate from European plasma have
also transmitted HIV (Madhok and Forbes 1990). A

clear correlation was also found between the severity
of haemophilia and HIV seropositivity (Melbye 1986;
UK Haemophilia Centre 1986). Haemophiliacs treated
before 1985 with cryoprecipitate alone have shown
a low risk of HIV infection (Ragni et al. 1986; UK
Haemophilia Centre 1986).
Patients with haemophilia B fared somewhat better
(Evatt et al. 1984; Mannucci et al. 1985; Ragni et al.
1986; UK Haemophilia Centre 1986). The difference
may be due partly to the uneven partitioning of HIV in
infected plasma during fractionation, with HIV sepa-
rating preferentially in the cryoprecipitate fraction
(Aronson 1979; Morgenthaler 1989; Madhok and
Forbes 1990). The source of plasma used for fractiona-
tion is also partly responsible for this difference. In
many countries all the factor IX is prepared locally,
whereas at least part of the factor VIII is imported
from the USA. Approximately 70% of patients in the
USA with haemophilia A and 35% with haemophilia
B developed HIV antibodies before the introduction
of methods for viral inactivation in blood products
(CDC 1987b). In a multicentre study of haemophiliacs
treated in the UK between 1980 and 1984, 896 (44%)
of 2025 patients with haemophilia A were positive for
anti-HIV; 20 (6%) of 324 patients with haemophilia B
and 11 (5%) of 215 patients with von Willebrand’s
disease were seropositive.
Although a large number of severe haemophiliacs in
the UK were treated up to 1983 with unheated factor
VIII concentrate imported from the USA, the incidence

of anti-HIV in haemophiliacs in the UK is much lower
than in countries such as Germany, Spain and the USA,
where the frequency of anti-HIV ranges between 68%
and 94% (Kitchen et al. 1984; 1985). On the other
hand, in Groningen, in the Netherlands, only 1 out
of 18 severe haemophiliacs treated with commercial
factor VIII concentrates between 1978 and 1983
developed HIV antibodies (van der Meer et al. 1986).
There has been no transmission of HBV, HCV
or HIV by US-licensed plasma derivatives since the
introduction of effective virus inactivation procedures
(Tabor 1999).
Prevention of transfusion-transmitted HIV
infection
Reducing the residual risk
The reduction in risk of transfusion-transmitted
HIV over the past 15 years has been dramatic and
reassuring. Nevertheless, enormous public concern
persists. As in almost no other area, blood safety,
and specifically the possibility of HIV transmission,
provokes emotions and measures to further reduce
risks that defy the conventional cost–benefit calculus.
Actions to reduced residual risk fall into three general
categories: (1) measures that can be introduced by
blood collection facilities, such as improved donor
screening, testing, education and exclusion techniques;
(2) enlightened transfusion practice, such as judicious
use of allogeneic blood components, appropriate use
of autologous blood, and alternatives to transfusion
(see Chapter 17); and (3) measures that depend on the

development of new technologies, such as viral inac-
tivation of cellular components and safe substitutes
for blood.
Donor demographics have proved effective at identi-
fying and excluding donors at high risk for infection
and transmission of HIV (Busch et al. 1991a). Donors
with high-risk profiles include men who have had
sexual contact with other men since 1977, intraven-
ous drug users, residents of high prevalence regions,
prisoners, prostitutes, haemophiliacs who have received
‘unsafe’ clotting factor concentrates and the sexual
partners of people in all of these groups. The year 1977
was chosen as the point of reference, because the first
clinical cases of AIDS in the USA were diagnosed retro-
spectively as far back as 1978 (Jaffe et al. 1985). The
rate of seropositivity has been higher in paid plasma
donors than in volunteers. Seven HIV positives were
found out of 35 000 plasma donors attending centres
located outside high prevalence areas (Stramer et al.
1989).
Measures to exclude such subjects from blood
donor rolls are among the most important means
of preventing the spread of AIDS by transfusion.
Nevertheless, predonation medical screening and
donor education are not infallible (Leitman et al.
1989). Between 1% and 2% of donors do not report
CHAPTER 16
722
risks that would disqualify them from blood donation,
and donation incentives such as complimentary labor-

atory testing increases this rate (Glynn et al. 2001).
In a study of blood donors found positive for hepatitis
C antibody, 42% admitted to intravenous drug use on
subsequent questioning, despite denying such use on
predonation screening (Conry-Cantilena et al. 1996).
In an anonymous mail survey of 50 162 volunteer allo-
geneic blood donors, 1.9% of the 34 726 respondents
reported one or more risk factors that should have led
to their deferral at the time of their last donation
(Williams et al. 1997). Refinements in blood donor
screening techniques, such as the use of illustrated risk-
related activities, expansion of the screening question-
naire and interactive computer-based screening, have
been proposed, but data supporting the effectiveness
of such measures are lacking (Mayo et al. 1991). How-
ever, the reasons why some volunteer blood donors
appear to disregard certain screening criteria are un-
known and may have more to do with donor psycho-
logy than with inadequate screening and education.
Donors who are confirmed positive for HIV should
be counselled and referred to specialized centres for
follow-up. Counselling should be performed by spe-
cially trained staff. Appropriate interventions will not
only help the donor to obtain long-term supportive care
and to prevent further spread, but will also aid transfu-
sion services to understand which groups of the popu-
lation are HIV seropositive and still come forward
to donate (Lefrere et al. 1992). Donor education and
selection methods can then be modified accordingly.
Self-exclusion (‘self-deferral’) of donors

Retrospective analysis indicates that the risk of con-
tracting AIDS from the transfusion of blood and blood
components (other than products of plasma fractiona-
tion) prior to focused screening and testing was far
higher than the one infection per million units trans-
fused that had been estimated during the 1983–1985
interval. The risk of HIV transmission from trans-
fusion in San Francisco has been calculated from its
first appearance in 1978 to rise exponentially to a peak
risk of approximately 1.1% per unit transfused in
1983 (Busch et al. 1991a). A retrospective study of
heavily transfused patients with leukaemia in New
York City revealed an overall risk between 0.02% and
0.11% per component transfused (Minamoto et al.
1988). The major blood collectors published a Joint
Statement of Recommendations in January 1983
(American Association of Blood Banks et al. 1983),
and the Public Health Service published recommenda-
tions in March 1983 (CDC 1983) that proposed such
measures as public education, self-deferral for donors
engaging in high-risk activity and confidential unit
exclusion procedures (Pindyck et al. 1984, 1985). These
measures proved unusually effective. An estimated
90% of men in high-risk cat-egories self deferred. By
1984, the risk in San Francisco had dropped to less
than 0.2% per unit. Screening with anti-HIV-1 in 1985
reduced the risk to about 1 in 40 000 units (Busch et al.
1991a).
Routine screening tests for HIV in blood donors
In most countries, screening tests for anti-HIV by

the enzyme immunoassay (EIA) format are now com-
pulsory for all blood donations. If reactive (‘repeat
reactive’), additional tests for HIV using independent
methods are used to confirm the diagnosis of HIV
infection (Stramer et al. 2004). The current algorithm
in the USA requires that anti-HIV-1/2 repeat reactive
specimens be further tested with the HIV-1 Western
blot (see below) and a specific anti-HIV-2 EIA. Early
ELISA assays using disrupted purified virus were
plagued by false-positive reactions. Current genera-
tion screening assays, using recombinant and synthetic
antigens, have reduced false reactives dramatically while
increasing test sensitivity for both the predominant
and the variant viral strains (Busch and Alter 1995).
Less than 10% of reactive assays confirm positive
using this strategy. Western blot is notoriously subjec-
tive and complicated by non-viral bands (Kleinman
et al. 1998). Alternative strategies use a second EIA or
a NAT assay of HIV RNA.
Approximately 25.6 million donations were screened
by the American National Red Cross from September
1999 to 30 June 2003, resulting in 17 090 HIV repeat
reactive blood donations (Stramer et al. 2004). Only
4.8% of these donors (818) were Western blot positive
and approximately 90% (759) of those also tested
positive by NAT. Follow-up testing of the remaining
10% demonstrated that almost all of these donors
represent Western blot false-positive results.
Although some antibodies to core and other anti-
gens of HIV-1 and HIV-2 crossreact (Sazama et al.

1992), currently available tests are designed to detect
both anti-HIV-1 and anti-HIV-2. Such combined
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
723
assays are available as antiglobulin ELISAs and as
sandwich ELISAs and are routinely used in several
countries. Seroconversion in infection with HIV-1 is
detected earlier in these combined assays than by
anti-HIV-1 assays (Gallarda et al. 1994; Fiebig et al.
2003). Most of the combined assays have been found
to be less sensitive for the detection of anti-HIV-2
than most anti-HIV-2 specific tests (Christiansen et al.
1996).
Some samples with anti-HIV are repeatedly positive
in some, but negative in other screening assays. If only
one screening test is used, such samples may give false-
negative results (Hancock et al. 1993). One cause of
such false negatives is that antibodies against subtype
O, a variant found predominantly in West Africa, are
not recognized in all screening assays (Loussert-Ajaka
et al. 1994; Schable et al. 1994). At present group O
prevalence is low in the USA and in western Europe
(Sullivan et al. 2000). False-negative reactions have
also been found to be due to contamination with glove
powder, inhibition by serum proteins, haemoglobin
and certain anticoagulants (Sazama 1995).
Currently available ELISAs for HIV-1 antibodies
detect HIV-1 subtype group 0. Soon after the dis-
covery of an assay for HIV-Ab, transmission of HIV by
blood from seronegative donors had been recognized

(Esteban et al. 1985; Raevsky et al. 1986; Ward et al.
1988; Cohen and Munoz 1989). Studies reported
detection of HIV-1 p24 antigen in analyses of stored
blood specimens from plasma donors as early as 1986,
and confirmed cases of HIV-Ag-positive, Ab-negative
blood in primary HIV infection were reported by 1988
(Allain et al. 1986; Clark et al. 1991; Irani et al. 1991).
The utility of this test as a screening assay was not
so obvious. Antigen tests are positive for only part of
the initial antibody-negative viraemic phase. In some
subjects antigen can be detected as early as 2 weeks
after infection, persisting for between 3 weeks and
3 months, and is no longer detectable when anti-p24
appears in the serum, although it may reappear inter-
mittently during the asymptomatic phase (Allain et al.
1986; Fiebig et al. 2003). Later, antigen may reappear
with a loss of anti-p24. A prospective study of 515 494
units donated at 13 blood centres in the USA failed
to detect a single instance of Ag-positive/Ab-negative
donated blood. A retrospective analysis of 200 000
repository specimens and prospective studies of blood
donors in Europe confirmed these findings (Backer
et al. 1987; Busch et al. 1990a). However, after three
anti-HIV seroconversions followed transfusion of p24
antigen-positive units, testing of donated blood for p24
antigen was mandated in the USA in 1996 (Busch and
Alter 1995). Mathematical models predicted that uni-
versal antigen screening would detect eight additional
potentially infectious units per year. In fact it took
5 years before eight antigen-positive/antibody-negative

units were interdicted (Kleinman et al. 1997; Kleinman
and Busch 2000). Because of this limited usefulness
and troubling frequency of false positives, HIV p24
antigen screening was not adopted widely outside of
the USA. With the adoption of universal NAT for HIV
in the USA in 1999 and its licensure in 2002, HIV-Ag
screening was rendered unnecessary (Busch et al. 2000).
Confirmatory tests: they do not always confirm
The Western blot is the most widely used additional
or ‘confirmatory’ test for HIV. The criteria for the
interpretation of Western blot results have been re-
evaluated several times because of greater sensitivity
and specificity of screening assays and Western blot
reagents, better insight into the serological patterns of
HIV infection, experience with patterns of non-specific
reactivity in low- and high-risk populations and
knowledge of the serological basis of non-specificity
(Sayre et al. 1996). Samples are now considered to
be WB positive demonstrate reaction with the gp41
and gp120/160 env bands or with either of these bands
and the p24 gag band. The earlier requirement for a
reaction with a third gene product (e.g. p31 or p66 pol
bands) has been abandoned and reactivity with more
than one env antigen alone is enough for confirmation
(O’Gorman et al. 1991). If there is reactivity with only
one band, the result of the Western blot is considered
to be indeterminate. The absence of reactivity in the
Western blot indicates that the donor has not devel-
oped anti-HIV (Dodd 1991).
In the original Western blot assay, viral lysate was

used as a source of antigen. In the newer assays, recom-
binant or synthetic viral antigens are applied. These
assays have been found to be both more sensitive and
specific than the original WB (Soriano et al. 1994).
Nevertheless, the assay is still relatively subjective
and beset by indeterminate and false-positive results
when compared with NAT as the ‘gold standard’ or
when investigated with sequential sampling follow-up
(Kleinman et al. 1998; Mahe et al. 2002; Stramer
2004).
CHAPTER 16
724
Detection of HIV DNA and HIV RNA
Donations during the window period constitute the
predominant risk for HIV transmission through
transfusion (Busch et al. 2000). A more sensitive and
specific alternative to testing for p24 antigen is NAT
for HIV DNA in PBMC (Ou et al. 1988) or HIV-1
RNA in plasma by reverse transcription PCR or a
similar amplification assay (Henrard et al. 1992). All
blood in the USA, Japan and most European coun-
tries is tested by NAT in pools of 16–90 specimens.
Ultrasensitive assays can detect fewer than 10 genomic
copies/ml (Busch et al. 2000). However, tests have
been optimized for HIV subtype B and may lack sensi-
tivity when applied to non-B subtypes (Triques et al.
1999). Single-donor assays are inevitable, but await the
development of fully automated combination (HIV/
HCV/HBV) assays.
Current risk of transmission of HIV by blood

transfusion
Although the various measures outlined above have
dramatically reduced the risk of TAA-AIDS, a small
residual risk remains (Delwart et al. 2004; Phelps et al.
2004). Most of this risk results from ‘window period’
donations. Before the introduction of HIV-Ag and
NAT, prospective studies estimated this risk at about
one infection in 60 000 units (Busch et al. 1991b;
Nelson et al. 1992). Subsequent estimates rely on
models based on calculations of HIV incidence and
window period. In the USA, residual risk has been
calculated as 1 per 2 135 000 repeat donors (Dodd
et al. 2002). The incidence rate is approximately
two times greater among first-time donors. In most
European countries where the prevalence of HIV in
blood donors is lower than in the USA, the residual
risk is probably lower still. However, in countries with
a high percentage of infected subjects and where HIV
is spread mainly by heterosexual intercourse, the risk
of transmission of HIV by blood transfusion is still
considerable.
Human T-cell leukaemia viruses types I
and II
The human T-cell leukaemia virus type I (HTLV-I), the
first human retrovirus to be described, was isolated
from cultured cells from a patient with an aggressive
variant of Mycosis fungoides and from a patient with
Sézary syndrome (Poiesz et al. 1980; Gallo et al. 1981).
The virus has subsequently been shown to be identical
to the adult T-cell leukaemia virus (Yoshida et al. 1982;

Watanabe et al. 1984). HTLV-I is the causative
agent of adult T-cell leukaemia (ATL) and is asso-
ciated with a chronic demyelinating neurological
disease called tropical spastic paraperesis (TSP), known
in Japan as HTLV-associated myelopathy (HAM)
(Vernant et al. 1987; Roman and Osame 1988). HTLV-
seropositive individuals appear to have a 0.25% life-
time risk of developing TSP, compared with a 2–5%
risk of developing ATL (Kaplan et al. 1990). The
virus is also associated with lung infections, cancer of
other organs, monoclonal gammopathy, renal failure,
infection with Strongyloides stercoralis, intractable
non-specific dermatomycosis, lymphadenitis and
uveitis. These effects may be due to the immunodefici-
ency induced by HTLV-I infection (Takatsuki 1996).
The association of HTLV-I with Mycosis fungoides is
controversial, as no HTLV-related DNA sequences
could be detected in patients with this disease
(Bazarbachi et al. 1993; Vallejo et al. 1995). In Japan,
only 2.5% of HTLV-I carriers develop ATL (Takatsuki
1996).
HTLV-I and -II belong to the oncovirus subtype of
the retrovirus family and are able to induce polyclonal
proliferation of T lymphocytes in vitro and in vivo.
Like the lentiviruses HIV-1 and -2, these viruses are
lymphotropic and neurotropic, and have the essential
structural genes gag (group antigen), pol (reverse tran-
scriptase) and env (envelope) in addition to regulatory
genes. In HTLV-I, the gag gene codes for the structural
proteins p55/24/19; pol codes for a protein of approx-

imately 100 kDa, and env codes for glycoproteins
gp61/46/21. In HTLV-II the structural proteins are
similar to those in HTLV-I with a high degree of cross-
reactivity; gag encodes the polypeptides p53/24/19;
pol a protein of approximately 100 kDa; and env
codes for the glycoproteins gp61/46/21.
Areas endemic for HTLV have been found, particu-
larly in south-west Japan, with prevalence as high as
15% (Maeda et al. 1984), in the Caribbean with a
1–8% prevalence (Clark et al. 1985), in regions of
Central and South America and in parts of sub-
Saharan Africa (Gessain et al. 1986; Vrielink and
Reesink 2004). Populations in these areas show differ-
ent prevalence rates for anti-HTLV-I, as do emigrants
from these regions (Sandler et al. 1991; Vrielink and
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
725
Reesink 2004). It has been estimated that more than
1 million Japanese people are healthy carriers of
HTLV-I. Carriers of HTLV have also been found
in the USA, especially in Florida and states on the
Pacific, and in France, UK, the Netherlands and many
other countries. The prevalence in blood donors has
been reported to be one in 6250 in the USA (CDC
1990), about one in 30 000 in France (Pillonel et al.
1996), one in 45 000 in the Netherlands (Zaaijer
et al. 1994) and one in 20 000 in London (Brennan
1992).
HTLV-II, the second human retrovirus to be dis-
covered, has a 65% nucleotide sequence identity with

HTLV-I and a significant serological crossreactivity
(Hjelle 1991). However, HTLV-II antibodies are not
detected by all HTLV-I assays. The distinction
between HTLV-I and -II can be made by DNA PCR
(Reesink et al. 1994), and in the recently developed
WB assays in which specific HTLV-I and -II recombin-
ant antigens are used. The relative prevalence of the
two viruses in blood donors in the USA was found to
be about equal (Glynn et al. 2000), but in many other
countries HTLV-I predominates in blood donors.
There is a high prevalence of HTLV-II among i.v. drug
users and their sexual contacts in the USA and other
countries (Vrielink and Reesink 2004). A large pro-
portion of HTLV-II-positive subjects in the USA were
Hispanics and American Indians (Hjelle et al. 1990a;
Sandler et al. 1991). HTLV-II has been found to be
associated with a HAM-like neurological disease
(Hjelle et al. 1992; Murphy et al. 1997). Although
the virus was first found in a patient with hairy
cell leukaemia (Kalyanaraman et al. 1982) and sub-
sequently in other T-cell malignancies, no viral RNA
could be detected in the malignant cells (Manns and
Blattner 1991).
Transmission of human T-cell leukaemia virus
HTLV is mainly transmitted by sexual contact, by the
sharing of infected needles and from mother to child,
particularly by breast-feeding (Kajiyama et al. 1986).
If infected mothers refrain from breast-feeding, trans-
mission of HTLV to their infants is prevented in 80%
of cases (Hino et al. 1996). Infection of infants is also

prevented when the milk is freeze-thawed or heated at
56°C for 30 min (Ando et al. 1986; Hino et al. 1987).
Transmission from mother to fetus has been demon-
strated by culture studies of cord blood lymphocytes in
2 out of 40 cord blood samples from HTLV-I-positive
mothers (Satow et al. 1991).
Human T-cell leukaemia virus and blood
transfusion
HTLV-I has been transmitted by cellular components,
but not by cell-free plasma or plasma derivatives
(Okochi 1985). However, HTLV-RNA is detectable in
plasma from infected subjects. The lack of infectivity
of HTLV-positive plasma may be explained by the
presence of neutralizing antibodies, the fact that it is
integrated in viral DNA and the requirement of cell–
cell interactions for infectivity (Rios et al. 1996).
Antibodies are usually first detectable 14–30 days
after transfusion, although the interval may be as long
as 98 days (Inaba et al. 1989; Gout et al. 1990). Of 85
recipients of anti-HTLV-I-positive cell concentrates in
Japan, 53 (62%) developed antibodies 3–6 weeks after
transfusion: IgM antibodies were present only in the
early stages, whereas IgG antibodies persisted at high
titre throughout the period of follow-up (Sato and
Okochi 1986). Storage of blood appears to decrease
the risk of transmission of HTLV. This may explain
the lower rate of transmission reported in USA trans-
fusion recipients, whose blood may have been stored
for a longer period than the units in Japan and the
Caribbean (Donegan et al. 1994). Antibodies became

detectable in 79.2% of recipients of blood stored for
1–5 days, but in only 55% of recipients of blood stored
11–16 days (Okochi 1989).
Recipients of HTLV-I-infected concentrates may
develop HAM (Gout et al. 1990; Araujo and Hall
2004). It has been estimated that 2–8% of subjects
infected with HTLV-I by blood transfusion will even-
tually develop HAM (Murphy et al. 1997; Araujo and
Hall 2004). ATL developed in two immunosuppressed
patients who had received multiple transfusions 6 and
11 years earlier (Chen et al. 1989). HTLV-II has also
been transmitted by blood transfusion (Hjelle et al.
1990b).
Screening tests for human T-cell leukaemia virus
antibodies
For the detection of HTLV antibodies, ELISAs are
used as well as gelatin particle assays. In the ELISAs for
the detection of anti-HTLV-I, viral lysate has been used.
Although there is considerable (65%) crossreactivity
CHAPTER 16
726
between HTLV-I and -II, the sensitivity of anti-HTLV-
I assays for detecting anti-HTLV-II was found to be
only 55–91% (Wiktor et al. 1991; Cossen et al. 1992).
As infection with HTLV-II is probably associated with
HAM and as many HTLV-positive donors (more than
50% in the USA) are HTLV-II positive, tests designed
to detect both antibodies are now used (US Food and
Drug Administration 1998). In these ELISAs, recom-
binant proteins, including HTLV-I and -II-specific

ones, have been added to lysate or are used exclusively
(Hartley et al. 1991). The sensitivity of such assays has
been claimed to be 100% (Vrielink et al. 1996a).
In Japan, an agglutination test in which gelatin
particles are coated with HTLV-I-antigens has been
developed and used extensively. The sensitivity of
the commercially available agglutination test Serodia
HTLV-I has been claimed to be 100% for detecting
anti-HTLV-I; all of 12 anti-HTLV-II-containing sera
gave positive reactions (Vrielink et al. 1996a). An
ELISA for the combined detection of anti-HIV-1 and
-2 and anti-HTLV-I and -II in which synthetic peptides
of all four viruses are used, has been developed
(McAlpine et al. 1992). In a study 242 samples
from anti-HIV-1/-2 or HTLV-I/-II panels, two HTLV-
II-positive samples and two very weak anti-HIV-1-
positive samples were negative. The specificity of the
test was slightly less than that of specific assays
(Flanagan et al. 1995).
Despite the improved specificity of anti-HTLV-I/-II
screening tests, many repeatedly positive reactions that
cannot be confirmed are still found and all repeatedly
reactive samples must therefore be tested in confirm-
atory assays.
Human T-cell leukaemia virus confirmatory
assays
Confirmatory testing for HTLV-I/-II continues to be
challenging, primarily because of a dearth of licensed
reagents. Western blot and radioimmunoprecipitation
assays (RIPAs) are used (Anderson et al. 1989; Hartley

et al. 1990). For a positive reaction, antibody reactiv-
ity with both a gag (p19 and/or p24) and an env pro-
tein (gp46 and/or gp68) are required (WHO 1990).
All other reaction patterns were considered to be
indeterminate. As the sensitivity of the Western blot
in detecting antibodies against env proteins was low,
many indeterminate results had to be checked in RIPA,
a much more elaborate assay (Lillehoj et al. 1990; Lal
et al. 1992). A report on the transmission of HTLV-I
by blood from a donor with an indeterminate pattern
in the WB and RIPA (p19 and gp68 reactivity) to four
out of six recipients, confirmed by PCR, demonstrated
the insufficient sensitivity of these original confirma-
tion assays (Donegan et al. 1992). A modified WB has
been developed in which both shared (r21e) and speci-
fic (rgp46
I
and rgp46
II
) HTLV-I and -II recombinant
env proteins are used (Lillehoj et al. 1990; Lal et al.
1992). For a positive reaction in this modified Western
blot, a reaction with at least one gag protein (p19
and/or p24) and with env r21e as well as rgp46
I
or
rgp46
II
is required. All other patterns are indeter-
minate. This Western blot, in which a reaction with

HTLV-I or -II can be distinguished, has been found
to be more sensitive and specific than the original
Western blot. Instead of 66, only two positive samples
required confirmation by RIPA and none of 158 inde-
terminate samples (original Western blot) reacted
(Brodine et al. 1993). A recombinant immunoblot
assay (RIBA) in which the same antigens are applied
gave similar results (Vrielink et al. 1996b). NAT has
not been licensed for confirmatory testing.
Cytomegalovirus
Characteristics of the virus and of
cytomegalovirus infection
Cytomegalovirus (CMV) is a large, enveloped, double-
stranded DNA, beta herpes virus that is cell associated,
but may also be found free in plasma and other body
fluids (Drew et al. 2003). CMV has a direct cytopathic
effect on infected cells. The result may lead to neu-
tropenia, some depression of cellular immunity and
inversion of T-cell subset ratios, with a consequent
increase in susceptibility to bacterial, fungal and
protozoa infections in immunosuppressed patients
(Grumet 1984; Landolfo et al. 2003). CMV infection
causes parenchymal damage, such as retinitis, pneu-
monitis, gastroenteritis and encephalitis, and can
result in substantial morbidity and mortality.
CMV can cause primary acute clinical and subclin-
ical infections. Chronic subclinical infections may occur
in which the virus is shed in saliva and urine. CMV
remains latent in a large proportion of infected sub-
jects. The presence of anti-CMV does not guarantee

immunity. As with HIV and HCV infection, specific
antibody is a marker of potential infectivity although
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
727
only in the case of CMV, a relatively small proportion
of seropositive subjects seem to be infectious (Drew
et al. 2003). CMV antibody-positive subjects may
infect others through sexual contact, breast-feeding,
transplacental transmission or transfusion (Tegtmeier
1986). In subjects with antibody, CMV infection may
be reactivated, or the subject may become re-infected
with exogenous strains of CMV.
Primary CMV infection is generally more severe
than is re-infection (co-infection with a different strain)
or reactivation. In view of the difficulty of distinguish-
ing between reactivation and re-infection, the term
recurrent infection has been coined to embrace both.
However, when necessary to distinguish between the
two, donor viral DNA can be distinguished from recipi-
ent viral DNA by restriction enzyme analysis (Glazer
et al. 1979; Chou 1990). In practice, the diagnosis of
recurrent infection is limited to demonstrating a four-
fold increase in antibody titre or the presence of IgM
anti-CMV. In immunosuppressed patients, serological
tests cannot be relied upon for a diagnosis of CMV
partly because the patients may not make antibody
and partly because any anti-CMV detected may have
been derived from transfused blood. Viral culture is
impractical because the virus grows slowly in vitro.
On the other hand, immunofluorescence techniques

to detect viral antigen using monoclonal antibodies on
biopsies or bronchial washings provide results within
hours (Griffiths 1984).
Prevalence of anti-CMV
The frequency of subjects with anti-CMV varies
widely in different populations. Seroprevalence is lower
(30–80%) in developed than in developing countries,
where the figure may reach 100% (Krech 1973;
Preiksaitis 1991). The prevalence of anti-CMV corre-
lates with age and socioeconomic status (Lamberson
1985; Tegtmeier 1986). The frequency of anti-CMV-
positive donors may vary widely within a given coun-
try, for example 25% in southern California and
70% in Nashville, Tennessee (Grumet 1984; Tegtmeier
1986).
Transmission of cytomegalovirus by transfusion
The transmission of CMV by blood transfusion was
first reported in the 1960s (Kaariainen et al. 1966;
Paloheimo et al. 1968; Klemola et al. 1969). CMV is
now known as one of the infectious agents most fre-
quently transmitted by transfusion. The pathogenesis
of transfusion-transmitted CMV infection is not
clearly understood. In most cases CMV appears to
be transmitted in a latent, particulate state only by
cellular blood components, and the virus reactivates
from donor leucocytes after transfusion. Host as well
as donor factors are involved in CMV infection
(Tegtmeier 1989; Preiksaitis 1991). CMV has been
isolated from the mononuclear and polymorpho-
nuclear cells of patients with acute infections. The

specific cell type responsible for carrying the virus
has not been identified, although mononuclear cells
are the favourite candidates as hosts of CMV in latent
infection. Fresh blood appears more likely than stored
blood to transmit CMV infection, although no con-
trolled studies document this impression (Tegtmeier
1986).
Some 3–12% of units have the potential to trans-
mit CMV (Adler 1984), although most authors have
reported a carrier rate of 1% or less (Tegtmeier 1986;
Drew et al. 2003). The discrepancy may be related to
the frequency of donor testing and the sensitivity and
specificity of the assay used. Primary infection rates
depend on the number of transfusions, age of blood,
time of year and immunocompetence of the recipient
(Tegtmeier 1989; Preiksaitis et al. 1988; Preiksaitis
2000). Donors with IgM anti-CMV appear to be
more likely than others to transmit CMV (Lamberson
et al. 1984). One large study found that only 0.5% of
antibody-positive donors have detectable CMV DNA
in their leucocytes (Roback et al. 2003).
At present, no rapid, easy way to identify infectious
subjects exists. Viral excretion in urine is a good index
of infectivity, but blood donors would probably rebel
at this screening strategy. Virus can also be cultured
from saliva, and PCR-based assays are available
for the detection of viral genome in peripheral blood.
Detection of pp65 antigen in leucocytes (pp65 anti-
genaemia) is considered the ‘gold standard’ among
diagnostic tools for diagnosing CMV infection and

initiating antiviral therapy. Both CMV DNA and
immediate early-messenger RNA detection have been
compared with pp65 antigenaemia, but none of them
showed advantages in terms of earlier diagnosis and
better prognosis. PCR’s major advantage is its semi-
automation compared with the immunofluorescence
employed for pp65 antigenaemia. Isolation of CMV
by culture is reliable for the diagnosis of active
CHAPTER 16
728
infection, but is less sensitive and requires more time
for viral detection.
Transfusion-transmitted cytomegalovirus
infection in immunocompetent subjects
Before the era of universal (or near universal) leuco-
reduction, some 30% of anti-CMV-negative recipients
undergoing cardiac surgery involving transfusion
developed infection, as confirmed by virus isolation or
the development of anti-CMV. In addition, some anti-
CMV-positive patients developed recurrent infection.
In almost all cases, the infection is asymptomatic.
Of patients who develop a primary or recurrent
CMV infection following transfusion, fewer than 10%
develop a mononucleosis-like syndrome. This syn-
drome, originally termed the post-perfusion syndrome,
but now referred to as the post-transfusion syndrome,
appears 3–6 weeks after transfusion. Common fea-
tures include fever, exanthema, hepatosplenomegaly,
enlargement of lymph nodes and the presence in
peripheral blood of atypical lymphocytes resembling

those found in infectious mononucleosis (Foster 1966).
Recovery is usually complete. The development of
atypical lymphocytes due to post-transfusion CMV
infection should be distinguished from the develop-
ment of atypical lymphocytes 1 week after transfusion
as a response to allogeneic lymphocytes.
Consequences of transfusion in patients with
impaired immunity
During the past decade, major advances have been
achieved regarding the management of CMV infection
through thedevelopment of new diagnostictechniques
for the detection of the virus and through the perform-
ance of prospective clinical trials of antiviral agents
(Meijer et al. 2003). Nevertheless, in immunosup-
pressed patients, or in fetuses and premature infants
with an immature immune system, CMV infections,
particularly primary infections, still cause severe dis-
ease that can be fatal.
1 The fetus in utero. Following maternal primary
CMV infection, the fetus becomes infected in 30–40%
of cases. Approximately 5–10% of infected infants
develop sequelae such as mental retardation, hearing
loss or chorioretinitis (Stagno et al. 1986).
2 Premature infants. The risk of serious CMV
infection is high when the infant’s birthweight is less
than about 1300 g and when the mother is anti-CMV
negative. In two large prospective studies, 25–30% of
infants with these risk factors, transfused with a total
of 50 ml or more of blood, some of which was anti-
CMV positive, acquired CMV infection, and 25% of

these infants died. Infants transfused with anti-CMV-
negative blood did not develop CMV infection (0 out
of 90) (Yeager et al. 1981; Adler et al. 1983). A lower
incidence of CMV infection (7–9%) has been reported
from two other centres: in infants weighing less than
1500 g born to anti-CMV-negative mothers and trans-
fused with blood, some of which was anti-CMV posit-
ive (Smith et al. 1984; Tegtmeier 1984). All reports
agree that clinically significant CMV infection in
newborn infants develops only when the infant is
premature and of low birthweight, when the mother
lacks anti-CMV and when anti-CMV-positive blood
is transfused (Tegtmeier 1986).
3 Bone marrow transplant recipients frequently
develop primary or recurrent CMV infections that
may prove fatal (Tegtmeier 1986). Blood transfusion
represents the main risk factor for CMV acquisition in
CMV-negative patients receiving bone marrow from
a CMV-negative donor. In a prospective randomized
trial of 97 anti-CMV-negative patients, 57 received
anti-CMV-negative marrow: 32 out of the 57 received
anti-CMV-negative blood components and only one
of these developed a CMV infection; of the 25 who
received blood components unscreened for anti-CMV,
eight developed CMV infection. Among the 40 recipi-
ents of anti-CMV-positive bone marrow, the rate of
CMV infection was no lower in those who received
only anti-CMV-negative blood components (Bowden
et al. 1986). Granulocyte concentrates, which contain
large numbers of leucocytes that can harbour CMV,

reportedly carry the greatest risk of transmitting CMV
infection (Winston et al. 1980; Hersman et al. 1982).
However, when only two prophylactic transfusions
were given, the risk appeared to be no higher in those
who received granulocytes than in those who did not
(Vij et al. 2003).
4 Renal transplant recipients are at high risk of prim-
ary or recurrent CMV infection; the main source of
infection lies in the transplanted kidney (Tegtmeier
1986). In anti-CMV-negative recipients of a kidney
from an anti-CMV-negative donor, blood transfusion
plays a significant role in CMV transmission (Rubin
et al. 1985).
5 Heart and heart–lung transplant recipients may
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
729
develop severe primary CMV infection, which can
lead to opportunistic infections with fungi or bacteria.
In anti-CMV-negative recipients the main sources of
infection are anti-CMV-positive transplanted organs
or organ blood donor units (Preiksaitis et al. 1983).
If both donor and recipient are anti-CMV negative,
blood transfusion is a major source of CMV disease;
infection can be minimized by the use of anti-CMV-
negative red cells and platelets (Freeman et al. 1990).
If a heart or heart–lung from a donor with CMV anti-
bodies is transplanted to an anti-CMV-negative recipi-
ent, prophylactic administration of specific CMV
hyperimmune IVIG seems to lessen the severity of
CMV disease (Freeman et al. 1990), although data in

support of this practice are unimpressive.
6 Following splenectomy due to trauma, patients
receiving massive transfusion may develop serious
CMV infections (Baumgartner et al. 1982; Drew and
Miner 1982).
7 Subjects infected with HIV and especially those
with AIDS, if anti-CMV negative, may acquire prim-
ary CMV infection by transfusion (Jackson et al. 1988;
Sayers et al. 1992). In these patients CMV may cause
sight-threatening infection that may lead to blindness
in up to 25% of patients not receiving antiviral
therapy. Because the rate of reactivation of CMV in
already infected patients is high, it is difficult to deter-
mine the contribution of transfusion-transmitted virus
(Bowden 1995).
8 Liver transplant recipients should also be con-
sidered at risk, especially children or pregnant women.
Now that smaller amounts of blood and blood com-
ponents are needed in liver transplantation, it has been
possible to give only anti-CMV-negative blood and
platelets to anti-CMV-negative recipients of anti-CMV-
negative grafts.
Prevention of transfusion-transmitted
cytomegalovirus infection
Subjects who are at highest risk of severe primary
CMV infection are anti-CMV-negative patients with
impaired immunity. For such patients, preventive
measures are available to reduce transmission of CMV
by transfusion. The selection of CMV-seronegative
donors has proven to be effective, but not infallible

(Tegtmeier 1989; Miller et al. 1991). Seropositivity,
especially the presence of IgM, is a marker of previous
infection and latent, but potentially infectious virus.
However, antibody assays vary in sensitivity and a
small risk of transmission even from seronegative units
remains (Kraat et al. 1993; Bowden et al. 1995).
Window period infections are the most likely source of
antibody screening failures. Although the window
periods for HIV-1, HCV and HBV have been reason-
ably well defined, the length of the CMV-seronegative
window, estimated at 6–8 weeks, is less well character-
ized. Several recent studies using PCR technology
documented CMV DNA in both plasma and cellular
blood components from several weeks before sero-
conversion to several months after seroconversion,
although culture positivity was found for a much
shorter period in white blood cells (WBCs) and not
observed in plasma (Zanghellini et al. 1999).
When only red cells are required, frozen deglycero-
lysed red cells can be used and have not been shown
not to transmit CMV (Brady et al. 1984; Taylor et al.
1986; Sayers et al. 1992).
As CMV is a white cell-associated virus, an altern-
ative approach to preventing CMV transmission
involves filtration removal of leucocytes from red cell
and platelet concentrates (Graan-Hentzen et al. 1989).
Leucoreduction by filtration may fail to prevent CMV
transmission, as 10
5
to 10

6
WBCs may still be trans-
fused and an estimated 1 in 1000 to 1 in 10 000 WBCs
are infected by CMV during latency (Drew et al. 2003).
In 10 patients whose WBCs were CMV antigen and
culture positive before filtration and culture negative
afterwards, 2 out of the 10 did, however, have CMV
DNA detected in leucocytes after filtration (Lipson
et al. 2001). Plasma viraemia, if present, would not be
diminished by leucoreduction and might also account
for CMV transmission following leucoreduced com-
ponents. Although the exact number of residual leuco-
cytes that is sufficiently small to pose no risk of CMV
transmission is unknown and may not exist, a large
prospective study has shown that leucocyte removal
is as safe as selection of anti-CMV-negative donors.
In this study, 502 patients were randomized prior
to bone marrow transplantation to receive filtered,
three-log leucocyte-depleted cellular components or
components from anti-CMV-negative donors. There
was no significant difference in the probability of
CMV infection between the recipients of anti-CMV-
negative or filtered concentrates, although more CMV-
associated disease was observed in the filtration
group (Bowden 1995). The same investigators have
subsequently questioned their finding that filtered red
CHAPTER 16
730
cells are equivalent to CMV-seronegative cells (Nichols
et al. 2003). When granulocyte transfusions are needed,

selection of anti-CMV-negative donors is the only
solution.
The success of leucocyte removal in the prevention
of CMV transmission has raised the question of the
importance of infection with a second strain of CMV
in anti-CMV-positive recipients and the value of trans-
fusing such recipients with leucocyte-depleted concen-
trates. Although co-infection with a second strain of
CMV does occur (Boppana et al. 2001), the clinical
consequences of such infection resulting from trans-
fusion are less clear.
Tests for anti-CMV
Several tests for anti-CMV are available. Complement
fixation used to be the diagnostic reference test, but it
proved too complicated for routine donor screening.
Indirect immunofluorescence, solid-phase fluorescence,
ELISA and particle agglutination assays are also avail-
able. Competitive ELISAs seem to be the most reliable
of the currently available screening tests (Bowden et al.
1987). CMV PCR technology has been used in clinical
diagnostics since the 1980s (Bowen et al. 1997; Lipson
et al. 1998). Conventional PCR requires detection and
confirmatory testing of the amplicon by gel elec-
trophoresis, incorporating an isotopically labelled or
non-radioactive hybridization assay or a nested PCR
to detect targets present in very low copy numbers
(Lipson et al. 1995). A simplified, more rapid PCR-
solid-phase enzyme immunoassay (EIA) plate tech-
nology system has been developed (Davoli et al. 1999).
The quantitative real-time CMV PCR assay using

TaqMan chemistry and an automated sample prepara-
tion system has also been applied to CMV detection
(Piiparinen et al. 2004).
An inexpensive, uncomplicated CMV-Ag (pp65
antigen) assay is available and well suited for most
diagnostic microbiology laboratories (Lipson et al.
1998). Assay for the early antigen pp65 is considered
the ‘gold standard’ for the initiation of antiviral therapy.
Other viruses
Epstein–Barr virus
Infection with Epstein–Barr virus (EBV), a herpes
virus-like CMV, is endemic throughout the world.
EBV can cause primary symptomatic infection (infec-
tious mononucleosis), but most commonly causes
asymptomatic infection followed by latent infection
(Henle 1985). In most countries more than 90% of
blood donors have neutralizing anti-EBV, which coex-
ists with latent virus in B lymphocytes of peripheral
blood and lymph nodes. At least one in 10
7
circulating
lymphocytes of carriers harbour EBV genomes, but
post-transfusion EBV infection is a rare occurrence,
and symptomatic infection is even rarer (Rocchi et al.
1977). The virus is found in three of every 10
4
periph-
eral lymphocytes during acute infection. The majority
of susceptible recipients are young children. Even in
children, the chance of acquiring EBV infection fol-

lowing the transfusion of anti-EBV-positive blood is
minimal, because the donor’s neutralizing antibodies
persist in the recipient’s circulation long after the EBV-
infected lymphocytes disappear. In a study of 25 EBV
antibody-negative patients aged 3 months to 15 years
and transfused with 1–11 units of blood stored for not
more than 4 days, only one developed EBV antibodies
and this patient had no symptoms (Henle 1985).
Of five patients transfused during cardiac bypass,
four who were initially anti-EBV negative produced
antibody that persisted at high titre for many months
postoperatively. Two out of the four patients had
concomitant CMV infection and no heterophile anti-
bodies; one developed transient fever, and the other
hepatitis. Only one of the four patients developed an
infectious mononucleosis-like syndrome with hetero-
phile antibodies (Gerber et al. 1969).
Although most cases of ‘post-transfusion syndrome’
are caused by CMV, two adult patients who were
anti-EBV negative before transfusion developed post-
transfusion syndrome due to EBV (McMonigal et al.
1983). In a survey of some 800 patients, fewer than
8% lacked EBV antibody before transfusion, and only
5% of these developed antibodies following trans-
fusion. These patients suffered no clinical illness or
disturbance of liver function (MRC 1974). Possibly,
the discrepancy between these observations and those
of others may have been related to the use of fresh
blood in the last-named series (Gerber et al. 1969).
Post-transfusion infectious mononucleosis is seen

only rarely in anti-EBV-negative immunocompetent
patients and usually occurs when only a single unit of
blood or blood component, obtained from the donor
during the incubation phase is given within 4 days of
collection. When more than 1 unit is transfused, one
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
731
of the units is almost certain to contain anti-EBV.
In the reported cases, the donors developed symptoms
of mononucleosis 2–17 days after blood donation.
The incubation period in recipients has been 21–
30 days (Solem and Jorgensen 1969; Turner et al.
1972). Transfusion-transmitted EBV infection with
symptomatic mononucleosis has occasionally been
reported in patients transfused for splenectomy (Henle
1985).
Post-transfusion EBV infections may contribute to
the development of lymphomas in severely immuno-
suppressed patients such as haematopoietic graft
recipients. T-lymphocyte suppression allows EBV-
infected B lymphocytes to outlive the passively trans-
fused antibodies and to proliferate (Marker et al.
1979; Hanto et al. 1983).
Other herpes viruses
Herpes simplex and Herpes varicella zoster have never
been shown to be transmitted by blood transfusion;
viraemia occurs only during primary infections, which
usually occur in childhood (Henle 1985).
HHV-6 is a recently characterized human herpes
virus, originally named HBLV (human B lymphotropic

virus) for its ability to infect freshly isolated B cells.
The virus was found in patients with various lympho-
proliferative disorders. HHV-6 can infect monocytes,
macrophages, T cells and megakaryocytes (Ablashi
1987). The virus is cytopathic for selected T-cell lines.
Infection is acquired usually within the first year of life
and the virus was found to be ubiquitous in blood
donors when tested in London and the USA (Briggs
et al. 1988; Saxinger et al. 1988). No transfusion-
associated disease has been reported.
HHV-8, Kaposi’s sarcoma-herpes virus is white cell
associated and may be present in up to 30% of normal
donors. Despite a relatively high HHV-8 seropre-
valence in a Texas blood donor cohort (23%), HHV-8
DNA was not detected in any sample of donor whole
blood using a highly sensitive PCR assay (Hudnall
et al. 2003). Blood components from HHV 8-infected
donors apparently carry little transfusion risk (Engels
et al. 1999).
Human parvovirus B19
HPV B19 infection has long been known to cause
erythema infectiosum (Fifth disease), a common febrile
exanthem of childhood (Anderson et al. 1983). HPV is
also associated with polyarthritis and rash in adults as
a result of antigen–antibody immune complex deposi-
tion in skin and joints (Reid et al. 1985; White et al.
1985). Because of its specific cytotoxic effect on
erythroblasts, HPV B19 can precipitate aplastic crises
in children who have haematological disorders with
shortened red cell survival, such as those with sickle

cell anaemia and other chronic haemolytic anaemias,
particularly hereditary spherocytosis (Pattison et al.
1981; Young and Mortimer 1984). The virus may also
cause thrombocytopenic purpura (Pattison 1987).
Intrauterine infection may cause hydrops fetalis and
spontaneous abortion in early pregnancy (Brown et al.
1984; Anand et al. 1987).
HPV B19 was discovered by an Australian virologist
who noted viral particles in an antigen–antibody line
of detection (plate B, well 19) in an assay for hepatitis
B (Cossart et al. 1975). HPV B19 is a small single-
stranded, non-enveloped, thermostable DNA member
of the Parvoviridae family (Shade et al. 1986; Young
and Brown 2004). The parvoviruses are dependent
on help from host cells or other viruses to replicate.
Parvovirus B19 is the type member of the erythrovirus
genus, which propagates best in erythroid progenitor
cells. The red cell P antigen, a globoside present on a
variety of cells in addition to erythrocytes, has been
documented as the specific receptor for HPV (Brown
et al. 1994). This may account for reports of poly-
arthritis nephropathy, myocarditis and cardiomyo-
pathy (Young and Brown 2004). Persistent infection
with anaemia occurs in immunosuppressed subjects
(Kurtzman et al. 1987). PCR assays have revealed the
presence of viral DNA along with the simultaneous
presence of specific IgG in 0.55–1.3% of normal blood
donors (Candotti et al. 2004), but the long-term
persistence, infectivity and clinical importance of the
virus in these subjects has not been well studied.

Current methods of viral inactivation may not be able
to eliminate HPV B19 completely (Williams et al.
1990a).
Infection with the virus normally occurs via respir-
atory droplets. HPV B19 has been transmitted by
plasma fractionation products derived from large
pools of plasma, particularly by factor VIII and factor
IX concentrates (Santagostino et al. 1997; Blumel et al.
2002). One episode of fulminant hepatitis has been
attributed to the intravenous route of infection
(Hayakawa et al. 2002). Use of recombinant products
CHAPTER 16
732
should eliminate this risk (Soucie et al. 2004).
Transmission of the virus by single-donor components
is unusual, but a red cell unit has been associated with
HPV B19 transmission and possible cardiac involve-
ment in a 22-year-old woman with thalassaemia major
(Zanella et al. 1995).
The classic diagnosis of infection with HPV B19
is based on the detection of IgM or IgG antibodies
with an EIA (Cohen et al. 1983). Alternatively viral
DNA can be detected by PCR (Salimans et al. 1989;
McOmish et al. 1993). HPV B19 has been detected by
PCR in solvent–detergent-treated clotting factor con-
centrates (Lefrere et al. 1994), in heat-treated factor
VIII and IX concentrates and in IVIG (McOmish et al.
1993; Santagostino et al. 1994) and in clotting factor
concentrates prepared by different purification and
inactivation procedures (Zakrzewska et al. 1992;

Blumel et al. 2002). The virus has also been detected in
plasma pools designed for fractionation (in 64 out of
75 pools), in 3 out of 12 albumin preparations, in 3 out
of 15 IVIG and in three out of four IMIG preparations,
as well as in seven out of seven factor VIII prepara-
tions. There is some indication that treatment of
IVIG preparations at low pH may result in removal
of detectable HPV B19 DNA (Saldanha and Minor
1996). Since 2002, major plasma fractionators have
screened plasma units withquantitative measurements
of B19 DNA to reduce the risk of transmission. The
prevalence of HPV B19 in blood donors has been
estimated at 0.03% in the UK (McOmish et al. 1993).
A similar prevalence was found in Japan (1 in 35 000)
but, during an epidemic of erythema infectiosum, the
prevalence was much higher (1 in 4000) (Tsujimura
et al. 1995). B19 prevalence varies according to season
and from year to year (Young and Brown 2004).
A rapid test for HPV B19, suitable for large-scale
screening of donors has been developed (Sato et al.
1996). The test is based on agglutination of blood
group P-positive gluteraldehyde-treated red cells by
the virus for which the P antigen is the receptor (see
Chapter 4). Although only intact viruses can bind to P,
the test has been found to be sensitive and it could be
used to select donors for patients at risk. Whether rou-
tine screening of donors is necessary is an unanswered
question. Although 81.6% of 136 haemophiliacs
studied had anti-HPV B19, B19 DNA was detectable
in none and there were no signs of lasting clinical or

haematological sequelae (Ragni et al. 1996).
Commercial immune globulins are a good source of
antibodies against parvovirus; a persistent B19 infec-
tion responds to a 5- or 10-day course of immuno-
globulin at a dose of 0.4 g per kilogram of body
weight, with a prompt decline in serum viral DNA, as
measured by hybridization methods, accompanied
by reticulocytosis and increased haemoglobin levels
(Kurtzman et al. 1987; Frickhofen et al. 1990).
West Nile Virus
Since its importation into the USA in 1999, West Nile
virus (WNV) has become a significant transfusion-
transmitted infection with a calculated mean risk of
transfusion transmission of 3.02 per 10 000 donations
in high-risk metropolitan areas during epidemic condi-
tions (Biggerstaff and Petersen 2003). Transfusion
transmission was first documented when four organs
harvested from a common multi-transfused cadaver
donor transmitted virus to all four recipients; the
donor’s pretransfusion sample tested negative for
WNV, while one of 63 blood donors tested positive
with a nucleic acid assay and developed IgM antibody
to WNV over the subsequent 2 months (Iwamoto et al.
2003). Using strict case definition criteria, epidemio-
logists document 23 transmissions and another 19
inconclusive investigations of 61 case investigations
during a 12-month period (Pealer et al. 2003). The true
number of transmissions was probably higher by at
least an order of magnitude.
As was the case with HIV and the hepatitis viruses,

blood transfusion of WNV represents a small, but
highly visible portion of a large epidemic. WNV is a
mosquito-borne flavivirus transmitted primarily to
birds and some small mammals. Humans serve as an
incidental host. Approximately 80% of human infec-
tions are asymptomatic; 20% result in a febrile illness
known as West Nile fever. About 1 in 150 patients
develop meningoencephalitis and residual neurolo-
gical deficits have been reported. Although there is no
particular susceptibility to mosquito-borne infection,
elderly and immunosuppressed subjects appear par-
ticularly vulnerable to severe, progressive disease.
The virus is present for a week or more during initial
infection. Most symptomatic subjects describe fever,
headache and malaise, although these are not suffi-
ciently specific for effective donor screening. Viral
shedding may persist for 7–8 weeks after infection,
usually in the presence of specific antibody. The period
of infectivity has not been well defined.
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
733
Of the 23 well-studied patients with transfusion-
transmitted infection, 14 were identified because West
Nile virus-associated illness came to the clinician’s
attention following transfusion. Overall, 15 patients
had recognized illness (13 meningoencephalitis, two
fever). The illness began between 2 and 21 days after
the implicated transfusion. The highly immunosup-
pressed transplant recipients appeared to have the
longest incubation periods (median 13.5 days). Red

cells, platelets and fresh-frozen plasma (FFP) have
all been implicated in transmissions. Sixteen blood
donors were implicated in transmission, with no pre-
dilection for age or gender. Nine out of the sixteen
recalled symptoms compatible with a viral illness
around the time of donation, although all donors
passed the donor screening procedures. Three donors
developed symptoms prior to donation, one on the day
of donation and five post donation. Fever, headache
and weakness were the most common symptoms.
Samples obtained at the time of donation had virus lev-
els less than 80 pfu per millilitre and all were negative
for West Nile virus IgM antibody.
Since screening of blood was undertaken in the USA,
approximately 6 million units were tested during the
June–December 2003 interval, resulting in the removal
of at least 818 viraemic blood donations from the
blood supply. Nevertheless, six cases of WNV trans-
mitted by transfusion occurred because of transfusion
of components containing low levels of virus not
detected by the testing of pooled specimens (Macedo
et al. 2004). All blood donations in the USA are cur-
rently screened by rtPCR, and it is likely that testing of
individual donations will begin in areas of epidemic
transmission and eventually be practised universally
(Custer et al. 2004).
Other flaviviruses such as St Louis encephalitis
virus, Japanese encephalitis virus and dengue virus
are likely to be blood transmissible as well, although
documentation is lacking.

Simian foamy virus
Simian foamy virus (SFV) (spumaretrovirus) is a highly
prevalent primate retrovirus that has been shown to
infect human cells, replicate and produce cell-free
infectious virus. Tropism is broad and includes B and
T lymphocytes, macrophages, fibroblasts, endothelial
cells and kidney cells. Simian retroviruses are spread
in areas of the world such as Central Africa, where
non-human primates are hunted for food and among
people who handle non-human primates as pets or
laboratory animals. In total, 1% of people living in
southern Cameroon were found to harbour antibodies
to SFV; all had reported exposure to fresh primate
blood or body fluids (Wolfe et al. 2004).
Studies from the USA CDC report a prevalence of
infection of 2–5% with simian foamy viruses among
laboratory and zoo workers occupationally exposed
to non-human primates (Heneine et al. 1998; Switzer
et al. 2004). Persistent viraemia has been detected
in peripheral blood lymphocytes (PBLs) and 11 SFV-
infected workers are known to have donated blood.
Six donors were confirmed positive at the time of
donation. SFV transmission through transfusion was
not identified among four recipients of cellular blood
components (two RBC, one filtered RBC, one platelet
concentrate) from one SFV-infected donor (Boneva
et al. 2002). Derivatives containing plasma from that
donor tested negative for SFV.
Evidence of SFV infection included seropositivity,
proviral DNA detection and isolation of foamy virus.

There is no evidence as yet that SFV causes human dis-
ease but, recombination within the host, especially in
immunocompromised hosts that may allow persistent
infection, remains a concern.
Prion proteins
A fundamental feature of prion diseases involves a
normal protein constituent of human tissue, prion
protein (PrP) (Bendheim et al. 1992). PrP can exist
in either a natural cellular form (PrP
C
) or as variant
pathological forms known collectively as ‘scrapie
isoforms’ or PrP
Sc
. Different prion strains result in
distinct clinical and pathological diseases (Prusiner
1998). Both forms of prion protein have an identical
amino acid sequence, but differ in the secondary struc-
ture, beta-pleated structures in the variant forms
instead of the alpha-helical structure of the native pro-
tein (Pan et al. 1993). The structural difference has two
major consequences. First, the beta-pleated conforma-
tion allows PrP
Sc
molecules to form protease-resistant
aggregates. Second, the abnormal protein seems to be
capable of suborning normal protein to the abnormal
form. The molecules with beta-pleated regions cause
key alpha-helical regions of the native protein to
assume the pleated conformation, thus converting

normal protein to abnormal aggregates that cause the
CHAPTER 16
734
spongiform degeneration of the brain that is character-
istic of these diseases.
Prion proteins are linked to the cell plasma mem-
brane by the lipid glycophosphatidyl inositol (GPI).
GPI-linked proteins are released from cells by several
different mechanisms and may be taken up by other
cells, thus facilitating dissemination (Devetten et al.
1997). Prion proteins are taken up selectively by
motile follicular dendritic cells, suggesting that infec-
tive proteins may be present in circulating blood (Klein
et al. 2001).
Creutzfeldt–Jakob disease
Creutzfeldt–Jakob disease (CJD) is a rare, but invari-
ably fatal neurodegenerative disease, with an incid-
ence of about one per million of the UK and USA
population (Collinge and Rossor 1996). CJD may be
‘sporadic,’ arising from a random change in a normal
individual, ‘familial,’ arising from a point mutation
in the DNA coding for the protein, which leads to an
increased susceptibility to production of the abnormal
form, or ‘acquired,’ by the transfer of infectious mater-
ial, as in individuals injected with pituitary-derived
hormones or treated with dura mater transplants
(Brown et al. 1992). A single case of transmission
from a corneal transplant has also been reported. The
proportion of recipients acquiring CJD from growth
hormone varies from 0.3% to 4.4% in different coun-

tries, and acquisition from dura mater varies between
0.02% and 0.05% in Japan, where most cases occurred
(Brown et al. 2000). Patient follow-up after point
exposures to contaminated materials indicate that
clinical latency for iatrogenic CJD may exceed 30 years
(Fradkin et al. 1991; Brown et al. 2000).
Classic CJD typically affects older subjects with
a median age of 60 years. Memory loss is an early
manifestation, but the disease progresses rapidly after
the first symptoms through confusion, motor and cere-
bellar symptoms and dementia. Death often occurs
within a year of the first symptom.
Because of the possibility of transfer of infectious
prions by transfusion, donors have been questioned
to discover (1) whether there is any family history of
CJD; (2) whether they have received pituitary-derived
hormones; or (3) in some countries, whether they have
received corneal grafts or grafts of dura mater. How-
ever, a systematic review of five case–control studies
from the UK, Europe, Japan and Australia, involving
2479 patients, failed to find an association between
CJD and blood recipients. No evidence of transmis-
sion of CJD by blood transfusion exists, despite the
identification of individuals who were exposed to
blood donated by people who later developed this dis-
ease (Wilson et al. 2000). A study of preserved brain
samples of 25 haemophilic patients who have high
exposure to blood transfusions and potentially higher
exposure to blood infected with the agent responsible
for Creutzfeldt–Jakob disease found no evidence of

the disease (Evatt et al. 1998). ‘Look-back’ studies
have not identified any cases of Creutzfeldt–Jakob
disease developing in recipients who received blood
from a donorin whom the disease was later diagnosed.
The risk of CJD transmission through transfusion is
now considered negligible, and plasma pools are no
longer discarded if a contributing donor develops
CJD. Deferrals designed to eliminate ‘high-risk’ donor
categories, such as growth hormone and dura mater
recipients, remain in force.
Variant Creutzfeldt–Jakob disease
A neurodegenerative disease affecting cattle was first
identified in England in 1984 and subsequently branded
‘mad cow disease’. Bovine spongiform encephalopathy
(BSE) is now recognized as a prion-related disorder
with more than 179 000 confirmed clinical cases of
BSE, driven by the recycling of infection through the
inclusion of bovine protein in cattle feed. A ban on
feeding tissues from one ruminant animal to other
ruminant animals, introduced in 1988, brought the
British BSE epidemic under control. However, because
of the long delay between infection and the onset
of clinical signs of disease (5 years on average), the
annual incidence of clinical cases did not peak until
1992. BSE has been found in several other European
countries and in Canada. More rapid and accurate
description of the epidemiology of BSE has been
hampered by the absence of a diagnostic test that
can be applied to live animals to detect those that are
incubating the infection.

The human disease equivalent, new variant
Creutzfeldt–Jakob disease (vCJD), was not identified
until 1996, 7 years after the ban was introduced. A
disease-related isoform of prion resembling that of
BSE is consistent with the conclusion that BSE has
crossed the species barrier, probably from ingestion of
infected beef (Collinge and Rossor 1996; Bruce et al.
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
735
1997). Unlike classic CJD, vCJD usually affects
younger subjects (age 18–53 years) and is charac-
terized by early psychiatric and sensory symptoms.
The disease progresses over 8 months to 3 years, with
progressive dementia, ataxia and myoclonus. Brain
biopsy has a characteristic pattern and Western blot
analysis shows a diagnostic migration pattern of pro-
teinase-treated PrP
Sc
. Pathological PrP
Sc
can be found
in tonsils, spleen lymph nodes and appendix (Collinge
et al. 1996). Prion protein gene analysis has shown
that all cases were homozygous for methionine at
codon 129 (Ironside and Head 2004). More than 150
cases have been found in the UK and cases have been
described in France and Italy. Although the number
of vCJD ‘carriers’ remains unknown, early estimates
of as many as 100 000 cases have in recent years
dwindled to a maximum of just a few hundred cases,

assuming an estimate of 15 to 20 years as the average
incubation period (Valleron et al. 2001). One model
indicates that the primary epidemic in the susceptible
genotype (methionine-methionine (MM)-homozygous
at codon 129 of the prion protein gene) has now
peaked, with an estimate of 40 future deaths (Ghani
et al. 2003). However, the prevalence of vCJD infec-
tion in the UK population, estimated to be as high as
1 per 4000 based on a study of routinely acquired
tonsils and appendices, raises the possibility of a
second and third ‘wave’ of clinical cases in subjects
heterozygous (methionine-valine) and homozygous
for valine at codon 129.
Prion protein has been identified on a wide range
of circulating cells including platelets, myeloid cells,
lymphocytes and red cells (Dodelet and Cashman
1998; Holada and Vostal 2000; Bessos et al. 2001;
Li et al. 2001). Findings in experimental models show
that blood not only contains infective agents of prion
diseases, but that no barrier to transmission exists with
intraspecies transmission, and that the intravenous
route of exposure to prions is fairly efficient (Casaccia
et al. 1989; Cervenakova et al. 2003). The seminal
experiments in sheep have shown transmission of BSE
and scrapie by blood transfusion, and blood for trans-
fusion in these experiments was obtained from sheep
midway through the incubation period (Houston et al.
2000; Hunter et al. 2002). Infectivity has also been
noted in the incubation period and symptomatic phase
in a rodent model of vCJD (Cervenakova et al. 2003).

In 1997, the UK set up a surveillance system
between the national CJD surveillance unit and the UK
national blood services. In 2004, the first probable
transfusion-associated case of vCJD was described
(Llewelyn et al. 2004):
A 62-year-old man received 5 units of red cells for a surgical
procedure. Six and a half years later, the patient developed
irritability and depression, followed by a shuffling gait,
blurred vision, motor dysfunction and cognitive impairment.
Thirteen months after the onset of symptoms, he died of
autopsy-confirmed vCJD. One of the red cell units had been
donated by a 24-year-old subject who developed symptoms
of vCJD 3 years and 4 months later, and subsequently died of
vCJD confirmed at autopsy. Red cells from a second donation
were traced to a patient who died of cancer 5 months later.
Platelets from the donation could not be traced to a recipient.
The clinical presentation of the transfusion recipient was
typical of vCJD, and diagnosis was confirmed by neuropatho-
logical examination. Transfusion took place before universal
leucoreduction. Of particular note in this case is the age of the
recipient, the second oldest patient with vCJD, making it even
less likely that this case was caused by ingestion of tainted
beef.
Statistical analysis, taking account reported vCJD
mortality to date and details of the recipients of vCJD
donations indicate that the probability of recording a
case of vCJD in this population in the absence of trans-
fusion-transmitted infection ranges between about 1 in
15 000 and 1 in 30 000 (Llewelyn et al. 2004).
A case of preclinical vCJD has been reported in a

patient who died from a non-neurological disorder
5 years after receiving a blood transfusion from a
donor who subsequently developed vCJD:
In 1999, an elderly patient received a unit of non-leucodepleted
red blood cells from a donor who developed symptoms of
vCJD 18 months after donation. The donor died in 2001 and
vCJD was confirmed at autopsy. The recipient, who had no
evidence of a neurological disorder, died 5 years after receiv-
ing the transfusion. Western blot analysis of splenic tissue
showed the presence of PrP
res
with the mobility and glyco-
form ratio of the signals similar to those seen in spleen
from patients with clinical vCJD and distinct from those of
sporadic CJD cases.
Protease-resistant prion protein (PrP
res
) was detected
by Western blot, paraffin-embedded tissue blot and
immunohistochemistry in splenic tissue, but not in
the brain. However, animal studies suggest that migra-
tion is inevitable given sufficient time. Prion protein
was also present in a cervical lymph node. Perhaps
the most disturbing observation is that this patient was
CHAPTER 16
736
a heterozygote at codon 129 of prion protein gene
(PRNP), suggesting that susceptibility to vCJD infec-
tion is not confined to the methionine homozygous
genotype. This finding suggests that a second, later

wave of vCJD cases in heterozygotes and even a third
may appear long after the initial peak of the epidemic
(Peden et al. 2004). The chance of observing vCJD
transmission in the absence of a transfusion infection
in this second recipient of blood from a donor with
vCJD is far less than the 1 in 15 000 to 1 in 30 000
chance for the first reported case.
Transmissable spongiform encephalopathy (TSE)
agents are resistant to the range of physical and chem-
ical means that have been used to inactivate viruses in
plasma products. However, a range of studies using
animal TSE models have demonstrated that the pro-
cesses used to purify proteins, including factor concen-
trates, over the course of plasma fractionation can also
contribute significantly to removing both abnormal
prions and infectivity (Lee et al. 2001). Extensive
washing of cellular components may also deplete the
prion content; however, the risk of infection has not
been determined. A prototype prion removal filter has
been designed to remove about four logs of PrP, below
the level of detection in the Western blot assay, and
has prevented infection in a hamster scrapie model
(SO Sowemino-Coker, personal communication).
Bacteria
Transmission of bacteria represents the most frequent
infectious complication of blood transfusion in the
developed world and a major cause of transfusion-
associated mortality (Andreu et al. 2002). The reported
frequency of contamination varies depending upon
the nature of the blood component and the method

of study technique.
Treponema pallidum
Treponema pallidum is a motile spirochaete that
spreads by sexual contact, transfusion, percutaneous
exposure and transmission from mother to infant. The
incubation period from transfusion to clinical pres-
entation varies from 4 weeks to 4.5 months, averaging
9–10 weeks, and the infected recipient usually exhibits
a typical secondary eruption. Donors at any stage of
disease, including late, latent syphilis, can transmit the
infection (Hartmann and Schone 1942).
Transfusion-transmitted syphilis was once consid-
ered a serious problem. Kilduffe and DeBakey (1942)
identified more than 100 cases that had been pub-
lished after 1915, all from direct transfusion. Some
138 cases had been reported by 1941 (De Schryver
and Meheus 1990). Since then, few cases have been
reported in the developed world. The last case pub-
lished in the USA was reported from the Clinical
Center at the National Institutes of Health (Chambers
et al. 1969).
In 1966, a patient was admitted to the Clinical Center with a
diagnosis of lymphoma. His serological test for syphilis (STS)
was negative on admission. He received 5 units of RBCs and
25 units of fresh platelets, none of which reacted in the VDRL
test. Two months later, he developed a maculopapular rash
consistent with secondary syphilis. Serological tests and con-
firmatory assays for syphilis turned positive. The patient was
treated with penicillin and the rash cleared.
Of the 30 donors, 27 tested negative repeatedly, but two

could not be traced and one refused to be retested. The
investigators presumed that one out of those three donors, a
platelet donor, was infectious for syphilis.
The chief reasons for the decline of transfusion-
transmitted syphilis seem to be the almost universal
practice of storing blood at 4°C before transfusion,
universal donor testing and the decline in the preval-
ence of syphilis in many countries since the advent
of penicillin. Other factors that have probably con-
tributed to the decline include the administration of
antibiotics to a large proportion of patients requiring
transfusion and the exclusion of donors with high-risk
sexual practices.
Spirochaetes are unlikely to survive in citrated blood
stored for more than 72 h at 4–6°C (Bloch 1941;
Turner and Diseker 1941). However, organisms can
be detected for as long as 6 days (Selbie 1943). There
is a close relationship between the number of tre-
ponemes added to donor blood and the survival times
of T. pallidum determined by a sensitive assay in rab-
bits; in blood heavily contaminated with spirochaetes
(1.3 × 10
6
and 2.5 × 10
7
/ml blood), surviving tre-
ponema were found at 120 h (Garretta et al. 1977; van
der Sluis et al. 1985).
Several tests are available for screening blood dona-
tions for syphilis, including the automated reagin test

(ART), the rapid plasma reagin test (PRP) and the
Venereal Disease Research Laboratory (VDRL) slide
technique in which non-specific reagins (antibodies)
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
737
are detected and the T. pallidum haemagglutination
assay (TPHA), the fluorescent treponemal antibodies
absorbed with Reiter’s treponeme assay (FTA-Abs)
and ELISAs for the detection of specific antibodies.
Non-specific reagin tests detect antibody to lipoidal
antigen and are often considered to be insufficiently
sensitive as screening tests (Barbara et al. 1982; Young
et al. 1992). Positive screening tests should be tested
for FTA-Abs or some other specific assay in a reference
laboratory. A lack of demonstrable T. pallidum DNA
or RNA suggests that blood donors with confirmed
positive results in STS are unlikely to have circulating
T. pallidum in their blood, and that that their blood
is unlikely to be infectious for syphilis (Orton et al.
2002).
Serological tests cannot prevent all cases of trans-
fusion syphilis because most remain negative in early
primary syphilis, when spirochaetaemia is most prom-
inent (Spangler et al. 1964).
The rationale for continued syphilis testing relies
upon the increasing demands for fresh blood compon-
ents, especially platelets and fresh blood for exchange
transfusion in newborn infants, the very occasional
reported case (Chambers et al. 1969; Soendjojo et al.
1982; Risseeuw-Appel and Kothe 1983) and its ques-

tionable value as a surrogate test to exclude donors
who are in high-risk groups for HIV and HBV infec-
tion. A positive test in a transfusion recipient may
result from antibody acquired passively from the
donor (Rossi et al. 2002). However, passively acquired
antibody rarely remains detectable for more than a few
months after transfusion (Ravitch et al. 1949; Rossi
et al. 2002).
In countries with a high incidence of syphilis, some
consultants recommend that recipients of fresh blood
receive 2 megaunits of penicillin G or its equivalent
(Bruce-Chwatt 1985). In endemic areas, subjects who
have had non-venereal trepanomatosis, such as yaws
or pima caused by T. pallidum pertenue or T. pallidum
carateum, may also have a positive screening test for
syphilis.
Brucella abortus
This organism can survive for months in stored blood
and there are several reports of blood transfusion-
transmitted symptomatic infection, mainly in children
and splenectomized patients (Wood 1955; Tabor 1982).
Antibody-positive donors are common in Mexico,
Greece, Spain and in some rural areas of the USA,
although transfusion-transmitted brucellosis has not
been reported in the USA. Infected donor blood has
very low concentrations of brucella and poses little
risk except for immunosuppressed patients (Fernandez
et al. 1981).
After an incubation period ranging from 6 days to
4 months, recipients of infected blood may develop

undulant fever, headache, chills, excessive sweating,
muscle pains and fatigue. Hepatosplenomegaly, lym-
phadenopathy, leucopenia and arthritis occur and, very
rarely, complications such as purpura, encephalitis or
endocarditis develop (Tabor 1982).
In view of the chronic nature of the disease, persons
with a history of brucellosis should not be used as
donors. However, 80% of infections are asymp-
tomatic. Even in endemic areas screening tests are
not practical; most subjects with high-titre brucella
antibodies do not transmit infection by transfusion
(Fernandez et al. 1981).
Miscellaneous
Lyme disease
The organism responsible for this disease is Borrelia
burgdorferi, a tick-borne spirochaete whose primary
reservoir is the white-footed mouse. Humans are
infected in the nymphal stage of the cycle and white-
tailed deer are infected during the second year of life
of the tick. Most cases of Lyme disease have been
reported in the USA, but thousands of cases have also
been reported in Europe.
The disease has three stages: in the acute stage of
erythema migrans, a skin rash starting at the site of
the tick bite spreads locally and is often accompanied
by ‘flu-like symptoms’. If this stage is not treated
promptly with antibiotics, the second stage progresses
to disseminated infection with cardiovascular and
neurological manifestations. In the third stage,
arthritis develops.

Although no cases of transmission of B. burgdorferi
by blood transfusion have been reported so far, trans-
mission is theoretically possible. Subjects in the primary
stage of infection pose the most risk. However, as these
patients are generally ill and as spirochaetaemia is of
low intensity and of short duration, infectious donors
probably do not present often for blood donation
(Popovsky 1990; Westphal 1991a). The spirochaete
CHAPTER 16
738
has been isolated and cultured from blood as old as
14 days.
Four transfusion recipients of blood components
from a donor who gave blood between the disappear-
ance of erythema migrans and the second stage of
the disease have been studied; none of the recipients
showed signs of disease or developed antibodies against
B. burgdorferi (Halkier et al. 1990).
Mycobacterium leprae
Mycobacterium leprae is known to have been trans-
fused inadvertently from a donor incubating leprosy to
the two recipients without apparent harm. M. leprae
has also been injected intravenously into human
volunteers without the recipients becoming infected,
although the period of follow-up is not known (Tabor
1982).
Tick-borne rickettsial disease
Tick-borne rickettsial diseases are caused by two
groups of intracellular bacteria belonging to the order
Rickettsiales: (1) bacteria belonging to the spotted

fever group of the genus Rickettsia within the family
Rickettsiaceae; and (2) bacteria within the family
Anaplasmataceae, including several genera, such as
Anaplasma and Ehrlichia. Rickettsiosis (Rocky
Mountain spotted fever) has been reported to have
been transmitted by blood from a donor incubating
the disease who subsequently died (Wells et al. 1978).
The rarity of the transmission of rickettsiosis by blood
transfusion probably reflects the clinical illness suf-
fered during most of the period of rickettsial infection
that makes blood donation unlikely (Tabor 1982).
In addition to the spotted fever group rickettsioses,
human anaplasmosis (formerly human granulocytic
ehrlichiosis) has also emerged in Europe (Parola
2004). This disease was first described in the USA
in 1994 and presents most commonly as a febrile
illness occurring in summer or spring. The causative
organism is Anaplasma phagocytophilum, formerly
known as Ehrlichia equi, Ehrlichia phagocytophila
and human granulocytic ehrlichia. The vector in
Europe is I. ricinus, which is also the vector of
Lyme borreliosis. Anaplasma are well preserved in
refrigerated blood components (McKechnie et al.
2000). Transfusion-transmitted ehrlichiosis has been
reported (Eastlund et al. 1999).
Exogenous and various endogenous bacteria and
bacterial products contaminating stored blood or
blood components
Bacteria may affect blood or blood components in one
of the following ways:

1 Bacteria may contaminate solutions or equipment
that are to be used for transfusions but which have not
yet been sterilized. After sterilization the solutions or
equipment may remain contaminated with heat-stable
bacterial products (pyrogens) capable of producing
febrile reactions when they are introduced into the cir-
culation (see Chapter 15). In contaminated equipment
or solutions such as hydroxyethyl starch used in leuca-
pheresis, bacteria may survive ‘sterilization’ or may
contaminate solutions that have previously been steril-
ized, for example when a glass container is cracked
during shipment (Wang et al. 2000). This kind of con-
tamination has become rare.
2 Bacteria originating from skin flora, such as
Staphylococcus epidermidis, Micrococcus species,
Sarcina species and diphtheroids, may gain entrance to
the blood pack during venesection, especially if the site
of venepuncture is scarred. It is virtually impossible to
disinfect the deeper layers of the skin and a skin plug
often enters the blood pack upon collection (Anderson
et al. 1986; Puckett 1986a).
3 Bacteria in the environment (Pseudomonas species,
Flavobacterium species, Bacillus species) may gain
entrance to blood components through minute lesions
in the packs, during collection or processing in open
systems (Szewzyk et al. 1993).
4 The ports of packs of cryoprecipitate or FFP may
become contaminated, if not protected by a secondary
plastic bag, during thawing in waterbaths contami-
nated with pseudomonas (e.g. Pseudomonas cepacia,

P. aeruginosa) (Casewell et al. 1981).
5 Bacteria circulating in the blood of an apparently
healthy donor suffering from asymptomatic bacteraemia
may proliferate in red cell components stored at 4°C or
in platelet concentrates stored at room temperature.
Bacteraemia in donors may be chronic and low grade,
as in the case of the incubation or convalescence periods
of salmonella, Yersinia enterocolitica or Campylobacter
jejuni infections, or acute and transient as occurs
within the first few hours after dental extractions when
the organisms involved are usually Streptococcus viri-
dans, Bacteroides species and, less often, Staphylococcus
aureus. A particularly notorious case reported from the
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
739
Clinical Center at NIH involved a repeat platelet donor
with asymptomatic Salmonella cholerasuis osteomyeli-
tis, whose donations resulted in sepsis in seven patients
and death in one of these before platelets were identi-
fied as the source of infection (Rhame et al. 1973).
Frequency of bacterial contamination of blood
components
Bacteria are usually prevented from growing by the
antibacterial properties of blood. Complement, even
in the absence of specific antibodies may kill bacteria,
which may then be phagocytosed by leucocytes
(Högman et al. 1991; Gong et al. 1994). Nevertheless
positive cultures in samples from blood or blood com-
ponents are found fairly frequently. In an extensive
survey in which approximately 2500 units of red cells

and an equal number of platelet concentrates were
sampled annually, over a period of 10 years the mean
rates of positive cultures were 0.3% and 0.4% respect-
ively (Goldman and Blajchman 1991). Bacterial con-
tamination of whole blood cultured after 2–20 h of
storage at room temperature ranges from 0.34% to
2.2% (Bruneau et al. 2001; de Korte et al. 2001).
Survival of most organisms is reduced by storage at
refrigerated temperatures. However, platelet con-
centrates are stored at room temperature so that
bacterial contamination and continued growth present
an urgent clinical problem. In one early study, 6 out
of 3141 (0.19%) pooled concentrates from random
donors were found to contain bacteria just prior
to transfusion (Yomtovian et al. 1993). In another
prospective study, 7 out of 15 838 (0.04%) concen-
trates were confirmed positive (Blajchman et al. 1994).
A higher percentage (0.79%) was found in samples
taken from freshly prepared concentrates: of 11 posit-
ive cultures only one was confirmed on repeated culture
from the same concentrate (Högman and Engstrand
1996). Only 5 out of 17 928 (0.03%) single-donor
concentrates obtained by apheresis contained bacteria
(Barrett et al. 1993). A 6-year experience using a semi-
automated system for routine platelet cultures in
Denmark reported an initial positive reaction in 84
samples (0.38%) from 22 057 platelet units. Growth
was confirmed in 70 of these (Munksgaard et al.
2004). The risk of collecting a contaminated unit of
platelets is estimated at 1 in 2000 (Chiu et al. 1994).

A prospective study using a conservative case
definition found the rate of transfusion-transmitted
bacteraemia (in events/million units) to be 9.98 for
single-donor platelets, 10.64 for pooled platelets and
0.21 for RBC units (1/500 000); for fatal reactions, the
rates were 1.94, 2.22 and 0.13 respectively (Kuehnert
et al. 2001). From a 12-year retrospective analysis
of transfusion data from a single institution, symp-
tomatic sepsis from transfused platelets was estimated
to occur at between 1 in 5000 (pooled whole-blood-
derived platelets) and one in 15 000 single-donor
apheresis platelet transfusions (Ness et al. 2001).
The incidence of clinical reactions due to contamin-
ated blood is much lower than that expected from the
reported rates of positive bacterial cultures. Although
some contaminated components, especially those
containing large numbers of organisms or endotoxin,
result in immediate, catastrophic reactions including
shock and death, others may be quite mild with little
more than low-grade fever or chills (Yomtovian et al.
1993). As many patients who receive platelet trans-
fusions already have an infection, febrile reactions or
symptoms of septicaemia may not be ascribed to the
transfusion (Chiu et al. 1994). Furthermore many of
these patients are being treated with antibiotics that
blunt detection and diagnosis of septic reactions.
Autologous blood
As might be anticipated, autologous donations are
also susceptible to bacterial contamination. At least
five such cases have been reported. The majority of

the infected units contained Y. enterocolitica, and the
donors recalled in retrospect gastrointestinal episodes
in the weeks prior to donation (Haditsch et al. 1994).
Some properties of organisms that grow in
stored blood
Organisms isolated from refrigerated blood are usu-
ally Gram-negative rods capable of metabolizing
citrate (Pittman 1953). Many strains of organisms
isolated from contaminated stored blood cause clot
formation by consuming citrate, yet another reason to
examine all units carefully at the time of issue (Braude
et al. 1952; Pittman 1953). Blood that is heavily con-
taminated with organisms is not necessarily haemo-
lysed. Among bacteria isolated from stored blood, 25%
produce no haemolysis (Braude et al. 1952). Many
organisms that grow in stored blood are psychrophilic
– organisms that grow preferentially in the cold.
CHAPTER 16
740
Unless blood is heavily contaminated with organ-
isms, microscopic examination is an unreliable method
of detecting infection. After serial dilutions of an
inoculum of a coliform organism to blood, 24 × 10
5
organisms/ml could be detected readily but 24 × 10
4
/ml
(one organism in every 100 fields) could be detected
only with difficulty (Walter et al. 1957). On the other
hand, when an aliquot of 0.3 ml was cultured for 24 h,

contamination could be detected when the number of
organisms added was as low as 24/ml.
The effect of antibiotics on bacteria in stored blood
has been investigated. Chlortetracycline, oxytetracy-
cline and polymyxin B in concentrations as low as
10 mg in 500 ml of blood (i.e. 20 µg/ml) would all pre-
vent the inocula of either cold- or warm-growing bac-
teria from multiplying in human blood either in the
cold or at room temperature. Nevertheless, routine
addition of antibiotics to stored blood cannot be
recommended. First, antibiotics cannot be autoclaved
and would therefore have to be added later, and this
addition might itself contaminate the collection.
Second, no antibiotics are effective against all organ-
isms. Third, there is the risk of immunizing patients to
particular antibiotics or of inducing a hypersensitivity
reaction in a patient already immunized.
Importance of maintaining refrigeration
After the first 24 h or so of storage, strict maintenance
of refrigeration at a temperature of 4°C becomes
essential. When blood has to be transported for any
considerable distance the use of refrigerated containers
can maintain blood below 10°C for as long as 48 h.
Some transfusion centres routinely keep blood at
approximately 20°C for up to 24 h after collection
before it is processed. The storage at room temper-
ature does not seem to affect the yields of factor VIII in
plasma subsequently frozen and fractionated or the
shelf-life of platelets or red cells (Pietersz et al. 1989).
Leucocytes in fresh blood have a bactericidal effect for

a few hours after collection and before processing into
blood components with removal of the buffy coats.
From experiments in which defined numbers of colony-
forming units of different bacteria were added to fresh
blood, leucocytes were shown to effect clearing of bac-
teria, the efficiency varying according to the bacterial
species: Staphylococcus epidermidis and Escherichia
coli were cleared more readily than S. aureus, which
was engulfed by the leucocytes and then released after
causing cell death (Högman et al. 1991). The clearing
effect of leucocytes sometimes requires as long as 24 h.
Some Gram-negative bacteria are killed by plasma fac-
tors, presumably antibodies and complement. Certain
anaerobic bacteria such as Propionibacterium species
do not grow in stored blood, regardless of the presence
or absence of white cells. Yersinia enterocolitica was
cleared temporarily, but reappeared within a few hours;
if leucocytes were removed from the blood after 5 h,
the units remained sterile. Although some bacteria are
killed intracellularly, others kill the cells, which then
disintegrate releasing the bacteria (Högman et al. 1991).
Bacteria contaminating different blood
components
The type of bacterial contamination largely depends
on the type of blood component:
1 Red cells involved in cases of bacterial infection
have been contaminated mainly with Pseudomonas
fluorescens, P. putida and Y. enterocolitica. Yersinia is
normally sensitive to complement and is phagocytosed
by leucocytes in fresh blood. However, it is capable of

producing a virulence plasmid that expresses a surface
protein rendering the organism resistant to comple-
ment and intracellular killing after phagocytosis (Lian
et al. 1987). Yersinia grows well at 4°C, uses citrate
as a source of energy and, owing to its lack of
siderophores, requires iron for optimum growth. Red
cell concentrates provide an ideal culture medium. The
organism produces a potent endotoxin during storage.
Transfusion of contaminated red cells can cause fatal
septicaemia. In almost all serious reactions due to
Y. eneterocolitica, the red cell concentrates had been
stored for more than 3 weeks. Phagocytes containing
live bacteria may persist after a donor has otherwise
contained a yersinia infection. During storage of red
cell concentrates the leucocytes may disintegrate,
releasing live bacteria, which can then multiply
(Högman et al. 1992). Leucocyte-depleted red cell
concentrates, previously inoculated with Y. eneteroco-
litica remain sterile, whereas growth develops after
2–3 weeks in leucocyte-containing concentrates (Gong
et al. 1994). Between April 1987 and the end of
February 1991, eight fatalities associated with the
transfusion of red cells contaminated with bacteria
were reported to the FDA; seven of these were due to
Y. enterocolitica. During the same period, 10 cases
of Y. enterocolitica caused by contaminated red cell
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
741
transfusions were reported in the USA, three cases
in France, one in Belgium and one in Australia. The

patients developed fever and hypotension within 50 min
of the start of transfusion. In the 10 cases that occurred
in the USA, the blood had been stored in the cold for a
mean of 33 days (range 25–42 days). Although reports
of septicaemia due to yersinia are rare, the numbers
have increased in the past 10 years. If patients are tak-
ing antibiotics, infection by blood contaminated with
yersinia can be asymptomatic (Jacobs et al. 1989).
2 Platelet concentrates stored at 20–24°C have caused
fatal bacterial sepsis when contaminated with any of
several Gram-negative or -positive organisms such as
staphylococci, streptococci, Serratia species, flavo-
bacteria and salmonellae (Goldman and Blajchman
1991). One plateletpheresis donor with subclinical
Salmonella cholerae suis osteomyelitis caused sepsis
in seven platelet recipients (Rhame et al. 1973).
3 Cryoprecipitates and FFP can become contaminated
with P. cepacia and P. aeruginosaduring thawing in con-
taminated waterbaths (Goldman and Blajchman 1991).
4 Any of the above (1–3) can become infected if
the exterior of blood packs is massively contaminated
during manufacture by bacteria such as Serratia
marcescens. Septicaemia has been reported in recipi-
ents of such contaminated units (Heltberg et al. 1992).
Effect of transfusing bacterially
contaminated blood
Shock following the injection of bacteria is pre-
sumably due to bacterial toxins, although an immune
reaction between naturally occurring antibodies and
bacteria may also produce ill effects. The transfusion

of contaminated blood may produce immediate col-
lapse, followed by profound shock and hyperpyrexia;
haemorrhagic phenomena due to disseminated intra-
vascular coagulation are common; in a series of
25 cases, mortality was 58% (Habibi et al. 1973).
A review of more than 30 reports of post-transfusion
sepsis due to contaminated red cells, platelets or a
few other blood components showed that septic shock
developed in many cases and that the overall mortality
rate was 26%. Common signs and symptoms reported
were fever, chills, vomiting, tachycardia and hypo-
tension, often during the transfusion, but sometimes
developing a few hours later. In the case of
Pseudomonas cepacia, septicaemia or wound infection
developed several days or weeks after transfusion
(Goldman and Blajchman 1991). However, atypical
presentations have been well documented in prospect-
ive studies and suggest that clinical syndromes are
under-reported (Yomtovian et al. 1993).
Investigations following the suspected
transfusion of injected blood
Any blood remaining in the container should be
cultured at various temperatures, including 4°C and
20°C, in appropriate culture media. A negative culture
excludes the possibility that the blood was heavily
infected at the time of transfusion. A positive culture
does not enable one to say with certainty whether the
blood was contaminated before the transfusion or
became contaminated during or after transfusion, or
upon sampling. Techniques to exclude contamination

upon sampling have been described (Puckett 1986b).
It is still useful to culture implicated containers that
have been left at room temperature for 24 h. Most
sterile units remain sterile or at worst lightly con-
taminated. In fatal cases, blood for culture should be
collected from the body (e.g. by cardiac puncture) as
soon as possible after death. Blood contaminated with
a coliaerogenes bacterium cultured positive, although
when pseudomonads were involved, blood collected at
autopsy was sterile (Pittman 1953).
Specific identification of the contaminating organ-
ism can be important for determining the extent of
donor follow-up and the donor counselling message
(Haimowitz 2005).
Inspection of red cells: the dark side of
transfusion
Colour change in stored red cell concentrates may
indicate that the unit has been contaminated with bac-
teria. In 13 units of red cells that supported the growth
of Y. enterocolitica, a darkening related to haemolysis
and a decrease in PO
2
was observed in the primary
container (Kim et al. 1992). The attached sample seg-
ments, which were sealed from the main unit, remained
sterile and did not darken. The colour change was
apparent in all the contaminated units by day 35, or
1.5 to 2 weeks after the bacteria were first detected in
cultures of the blood. A keen observer can identify
units grossly contaminated with Y. enterocolitica by

comparison of the colour of the segment tubing
with that of the unit (Kim et al. 1992). Review of
CHAPTER 16
742
photographs on file at the Centers for Disease Control
revealed this colour change in 2 units of blood that
caused transfusion-transmitted sepsis (Enterobacter
agglomerans and an unidentified Gram-negative bacil-
lus), which demonstrates that the colour change is not
limited to contamination by Y. enterocolitica.
Screening for bacterial contamination
Avoidance of bacterial contamination is preferable
to detecting organisms downstream. Examination
of the donor’s arms and rigorous skin preparation
reduce quantitative bacterial growth by as much as
50% (Goldman et al. 1997) (see Chapter 1). Diversion
of an aliquot of the initial blood collection removes
some of the organisms that populate the phlebotomy
site (de Korte et al. 2002). Screening collected blood for
bacterial contamination is complicated by the small
initial inoculum (sampling error), differential growth
rates of various species, and the potential for introduc-
tion of organisms at the time of the screening procedure.
Despite the absence of convenient and inexpensive
detection methods, routine surveillance of platelet
preparations for bacterial contamination has been
instituted in several countries. The two most com-
monly used methods are culture based. A sensitive and
specific automated, continuous monitoring culture
system measures carbon dioxide production as a

marker of bacterial growth (Brecher et al. 2001,
2004). Both aerobic and anaerobic cultures are pos-
sible. The procedure increases the storage interval
before platelets can be released and cultures may turn
positive after platelets have been issued or transfused.
A second system monitors consumption of oxygen
after 24 h at room temperature in aliquots of platelet
concentrates as a marker of microbial growth (Rock
et al. 2004). The disadvantages include inability to
detect anaerobic organisms, reliance on analysis at a
single time point and limited ability to detect slow-
growing bacteria. However only one example of a
fatality associated with an anaerobic organism has
been reported (McDonald et al. 1998). Both methods
have interdicted contaminated transfusions.
Malaria
Transmission of malaria by blood transfusion
Malaria can be transmitted by the transfusion of any
blood component likely to contain even small numbers
of red cells; platelet and granulocyte concentrates,
fresh plasma and cryoprecipitate have all been incrim-
inated. An inoculum containing as few as 10 parasites
can cause Plasmodium vivax malaria (Bruce-Chwatt
1972).
Despite intensive eradication programmes, the
prevalence of malaria in tropical and subtropical
countries is extremely high, with about 300 million
cases reported each year, and more than 1 million
annual fatalities attributed to malaria in tropical
Africa (Wyler 1983). The increasing resistance of the

mosquitoes to insecticides and the increasing appear-
ance of drug-resistant strains, particularly of P. falci-
parum and the recurrence of malaria in areas where
the disease had previously been eradicated, have greatly
increased the incidence of malaria in recent years.
Malaria parasites of all species can remain viable in
stored blood for at least 1 week (Hutton and Shute
1939). Cases of P. falciparum malaria have been trans-
mitted by blood stored for 14 days (Grant et al. 1960)
and for 19 days (De Silva et al. 1988). Malaria para-
sites may survive longer in blood stored in adenine-
containing solutions (WHO 1986). In a review of
nearly 2000 cases of transfusion malaria due to
P. malariae, in the great majority of cases the blood
had been stored for less than 5 days and cases follow-
ing the transfusion of blood stored for 2 weeks were
rare (Bruce-Chwatt 1974). The parasites survive well
in frozen blood (Kark 1982). Plasma that has been
frozen or fractionated has never been known to trans-
mit malaria.
In non-malarious countries, mainly because of delay
in diagnosis, post-transfusion malaria has a relatively
high fatality rate. Malaria is particularly serious in
pregnant women and in splenectomized or immuno-
suppressed patients (Kark 1982; Bruce-Chwatt 1985).
Transfusion-transmitted malaria usually responds
to conventional drug therapy. In severe cases in which
the diagnosis has been delayed and in fulminant cases,
the patient may benefit from exchange transfusion
in the period before antimalarial drugs have had time

to become effective (Yarrish et al. 1982; Kramer et al.
1983; Files et al. 1984). However, no controlled trials
have been performed and antimalarial medical therapy
is effective in most cases (Mordmuller and Kremsner
1998).
The red cells in a renal allograft and in a bone mar-
row graft have been known to transmit P. falciparum
INFECTIOUS AGENTS TRANSMITTED BY TRANSFUSION
743
malaria (Kark 1982; Dharmasena and Gordon-Smith
1986). The plasmodia can also be transmitted by
transplacental passage.
Frequency of post-transfusion malaria
Estimates of the frequency of post-transfusion malaria
vary from less than 0.2 cases per million units of blood
transfused in non-endemic countries to 50 or more
cases per million in some endemic countries (Bruce-
Chwatt 1985). In countries where malaria is non-
endemic the frequency of reported transfusion malaria
is relatively low. The UK has reported eight cases in
50 years, the USA 26 cases in 10 years from 1972 to
1981, and France 110 cases in 20 years (Bruce-Chwatt
1985), although between 1980 and 1986 only 14 cases
were observed. On the other hand, nine cases of
transfusion-transmitted malaria were reported to the
Centers for Disease Control (CDC) in the USA in
1982, although only 11 cases in the following 4 years
(Westphal 1991b). In Canada, only three cases of
transfusion-transmitted malaria were reported between
1994 and1999, a rate of 0.67 cases per million dona-

tions (Slinger et al. 2001). The rate of transfusion-
transmitted malaria in the USA since 1992 has
fluctuated from 0.18 to 0.6 per million units trans-
fused (Mungai et al. 2001). In many malarious areas
where reporting is inefficient, the frequency of post-
transfusion malaria is unknown.
Frequency with which the different parasites
are involved
In the period 1950–72, P. malariae was responsible
for about 50% of cases of transfusion malaria and
P. vivax for 20% (Bruce-Chwatt 1974). In the period
1973–80 a global survey showed a large relative
increase of cases due to P. vivax, which, with a few
cases of P. ovale, accounted for 42% of all cases;
cases due to P. malariae fell to 38% whereas those due
to P. falciparum increased four-fold (Bruce-Chwatt
1982). In more than one-half of the cases reported,
the species involved was not identified. The relative
increase in cases due to two of the parasites has been
attributed to the failure of donors from highly endemic
areas to comply with rules laid down by transfusion
services (Guerrero et al. 1983; Shulman et al. 1984).
The simultaneous transmission of two different strains
of malaria (P. falciparum and P. malariae) by trans-
fusion has been described (Aymard et al. 1980). Of
93 cases of transfusion-transmitted malaria reported
to the CDC from 1963 to 1999, 33 (35%) were due to
P. falciparum, 25 (27%) were due to P. vivax, 25
(27%) were due to P. malariae, 5 (5%) were due to
P. ovale, 3 (3%) were mixed infections and 2 (%) were

due to unidentified species; 10 out of the 93 patients
(11%) died.
Incubation period
The incubation period of transfusion malaria depends
on the numbers and strain of plasmodia transfused, on
the host and on the use of antimalarial prophylaxis;
with P. falciparum and P. vivax it is between 1 week
and 1 month, but with P. malariae it may be many
months (Bruce-Chwatt 1974).
Period for which donors remain infectious
P. falciparum infections, prevalent in West Africa, are
usually eliminated within 1 year but have been known
to persist for as long as 5 years after the last exposure
to infected mosquitoes (Guerrero et al. 1983). P. falci-
parum transmitted from asymptomatic blood donors
can cause fulminant infection and death in an already
ill, non-immune recipient in whom the diagnosis,
being unexpected, is usually delayed. P. vivax infec-
tions, prevalent on the Indian subcontinent, may
relapse for up to 2.5 years after infection but only
rarely after more than 3 years. With P. ovale a clinical
attack may occur up to 4 years after returning from an
endemic area. In immune carriers who have lived in
endemic areas for most of their lives, P. falciparum or
P. vivax may reappear long after the limits given above.
P. malariae infections persist for longer than those of
any of the other malaria parasites. Transmission of
malaria has been known to occur up to 46 years after
the last exposure to infection (Miller 1975).
Rules to prevent transmission of malaria in

non-endemic areas
In non-endemic areas, eligibility of potential blood
donors who have visited endemic areas is restricted.
Regulations and standards are not harmonized and
not always consistent. Two different approaches have
been adopted: the USA relies solely on the deferral of
subjects who have been in endemic areas, whereas in
CHAPTER 16
744