0095-1137/10/$12.00
doi:10.1128/JCM.02264-09
Copyright © 2010, American Society for Microbiology. All Rights Reserved.
Association of Schistosomiasis with False-Positive HIV Test Results
in an African Adolescent Population
䌤
Dean B. Everett,
1,2,3* Kathy J. Baisely,
1,2,3Ruth McNerney,
1Ian Hambleton,
1,2,3Tobias Chirwa,
1,2,3David A. Ross,
1John Changalucha,
2Deborah Watson-Jones,
1,2,3Helena Helmby,
1David W. Dunne,
4David Mabey,
1and Richard J. Hayes
1London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, United Kingdom
1; National Institute for
Medical Research, Mwanza Centre, P.O. Box 1462, Mwanza, Tanzania
2; African Medical and Research Foundation,
Lake Zone Programme, P.O. Box 1482, Mwanza, Tanzania
3; and University of Cambridge, Department of
Pathology, Tennis Court Road, Cambridge CB2 1QP, United Kingdom
4Received 19 November 2009/Returned for modification 6 January 2010/Accepted 18 February 2010
This study was designed to investigate the factors associated with the high rate of false-positive test results
observed with the 4th-generation Murex HIV Ag/Ab Combination EIA (enzyme immunoassay) within an
adolescent and young-adult cohort in northwest Tanzania. (4th-generation assays by definition detect both HIV
antigen and antibody.) The clinical and sociodemographic factors associated with false-positive HIV results
were analyzed for 6,940 Tanzanian adolescents and young adults. A subsample of 284 Murex assay-negative
and 240 false-positive serum samples were analyzed for immunological factors, including IgG antibodies to
malaria and schistosoma parasites, heterophile antibodies, and rheumatoid factor (RF) titers. Conditional
logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs). False-positive
HIV test results were associated with evidence of other infections. False positivity was strongly associated with
increasing levels of
Schistosoma haematobium
worm IgG1, with adolescents with optical densities in the top
quartile being at the highest risk (adjusted OR
ⴝ
40.7, 95% CI
ⴝ
8.5 to 194.2 compared with the risk for those
in the bottom quartile). False positivity was also significantly associated with increasing
S. mansoni
egg IgG1
titers and RF titers of
>
80 (adjusted OR
ⴝ
8.2, 95% CI
ⴝ
2.8 to 24.3). There was a significant negative
association between Murex assay false positivity and the levels of
S. mansoni
worm IgG1 and IgG2 and
Plasmodium falciparum
IgG1 and IgG4. In Africa, endemic infections may affect the specificities of
immuno-assays for HIV infection. Caution should be used when the results of 4th-generation HIV test results are
interpreted for African adolescent populations.
HIV remains one of the most serious challenges to world
health, particularly in sub-Saharan Africa, where the national
adult prevalence in eight countries currently exceeds 15%.
During 2007, 1.9 million people were newly infected with the
virus, and approximately half of those were young people
un-der the age of 25 years (35, 36).
Since their introduction in the 1980s, immunoassay-based
se-rological tests for HIV have enabled the simple rapid detection of
infection and have been widely implemented as screening tools
for both diagnosis and epidemiological monitoring. Concerns
about the specificity of serological tests for HIV when sera from
sub-Saharan Africa were tested were first raised in the 1980s.
False-positive reactions in enzyme immunoassays (EIAs) that
used viral lysate antigens were attributed to polyclonal B-cell
activation as a result of malaria and other endemic infections
(3–5, 7, 16, 19, 26, 33). The reduced specificity was addressed in
part by the use of recombinant proteins and synthetic peptides
instead of human viral lysates in later-generation EIAs.
Pressure to detect early HIV infections has resulted in the
development of EIAs of enhanced sensitivity. Several
pub-lished studies have suggested that these 4th-generation assays,
which by definition detect both HIV antigen and HIV
anti-body, are both highly sensitive and specific (1, 6, 25). However,
we have previously reported a lack of specificity in one of those
assays, the Murex HIV Ag/Ab Combination EIA (13). In a
study of over 7,000 adolescents and young adults in
northwest-ern Tanzania, the assay had a specificity of 91.5% and a
pos-itive predictive value of only 7.9%. We describe here the
clin-ical and biologclin-ical factors associated with the false positivity of
that assay. We believe that this is the first report to identify the
immunological factors associated with false-positive HIV test
results in an adolescent and young adult population in
sub-Saharan Africa.
MATERIALS AND METHODS
Study population and study design.A community-randomized trial of a mul-ticomponent sexual health intervention (MEMA kwa Vijana [MkV]) was con-ducted among young people in 20 distinct communities within the Lake Victoria region of Tanzania between 1999 and 2002. The design and the results of the trial have been reported previously (18, 31). At the follow-up survey in 2002, serum samples from 7,333 individuals aged 16 to 27 years were tested for HIV infection by using the Murex HIV Ag/Ab Combination EIA (Murex Biotech, Dartford, Kent, United Kingdom). The participants were interviewed about their sexual behavior and underwent clinical assessment by use of a syndromic approach for signs and symptoms of sexually transmitted infections (STIs). Serum samples were tested for lifetime exposure toTreponema pallidum(syphilis) and herpes simplex virus type 2 (HSV-2) antibodies. In addition, the participants were tested for possible Schisto-soma haematobiumexposure by using a red blood cell (RBC) urine dipstick test (Becton Dickinson, Oxford, United Kingdom) and treated according to Tanzanian guidelines. An RBC dipstick result of 3⫹or 4⫹was presumed to be a positive result
* Corresponding author. Present address:
Malawi-Liverpool-Well-come Trust Clinical Research Programme, P.O. Box 30096, Chichiri,
Blantyre 3, Malawi. Phone: 265 1875918. Fax: 265 1875774. E-mail:
Dean.Everett@liverpool.ac.uk.
䌤
Published ahead of print on 24 February 2010.
1570
on May 16, 2020 by guest
http://jcm.asm.org/
forSchistosoma haematobiumin all participants except female participants who were menstruating. In those cases, the participant was treated for schistosomiasis if she reported noticing blood in her urine when she was not menstruating. Pregnant females who were diagnosed with possible schistosomiasis were referred to the nearest health center for treatment after delivery. The participants were treated syndromically for malaria if they presented with fever.
In the cohort of 7,333 individuals, 674 (9.2%) serum samples were initially reactive by the Murex EIA, according to the manufacturer’s criteria; however, only 53 were found to be HIV positive on confirmatory testing (13), giving an overall HIV prevalence of 0.7%. Although females were more likely than males to be confirmed to be HIV positive (1.5% of females versus 0.2% of males), similar proportions of males (8.1%) and females (8.4%) had false-positive results. To investigate the fac-tors associated with the high proportion of false-positive results, a random collection of 240 and 284 Murex EIA-negative sera, sampled to provide an equal number of males and females, was taken for immunological analysis. Negative sera were se-lected from a subgroup that had been tested by an additional HIV assay for a separate validation study, which included all samples that were positive for syphilis. Sera, which had been frozen at⫺20°C, were transported to the Department of Pathology, University of Cambridge, Cambridge, United Kingdom, and tested for IgG antibodies to malaria and schistosoma parasites, heterophile antibody, and rheumatoid factor (RF) titer.
Laboratory methods.Sera were tested for HIV-1 and HIV-2 by the Murex HIV Ag/Ab Combination EIA; reactive specimens were sent to the United Kingdom Central Public Health Laboratory for confirmation of the results. Full details of the HIV testing algorithm and confirmatory testing have been de-scribed previously (13); but in brief, three further EIAs, a p24 antigen test, an in-house PCR withpol-specific primers, and a Western blot assay, were used as part of the Central Public Health Laboratory confirmatory algorithm. A speci-men was defined as false positive if it did not contain HIV-1 RNA.
IgG isotype analysis was performed by in-house enzyme-linked immunosor-bent assays to determine the levels of IgG isotypes 1 to 4 toS. mansonisoluble worm and egg antigens,S. haematobiumsoluble worm antigen, andPlasmodium falciparumantigen.S. haematobiumsoluble egg antigen was unavailable at the time of testing. An in-house test was also performed to determine the presence of heterophile antibody. The details of the IgG assays have been described previously (14, 28).S. mansoniantigens were obtained from a Puerto Rican isolate ofS. mansonimaintained in the outbred Tucks original strain of mice (Harlan United Kingdom Ltd., Bicester, United Kingdom) and albino freshwater snails (Biomphalaria glabrata) in Cambridge (11).S. haematobiumsoluble worm antigen andP. falciparumcrude schizont extract were supplied by collaborating laboratories of the Department of Pathology, University of Cambridge.
The RF titer was measured by the agglutination method with a Waaler Rose rheumatoid factor kit (bioMe´rieux, Basingstoke, United Kingdom). Lifetime exposure to syphilis was examined by using aTreponema pallidumparticle ag-glutination (TPPA) test (Serodia TPPA test; Fujirebio Inc., Tokyo, Japan). Sera were tested for antibodies to HSV-2 by using a monoclonal enzyme immunoassay (HSV-2 IgG; Kalon Biological Ltd., Ash Vale, United Kingdom).
Statistical analyses.Data were double-entered into and verified by use of the Dbase-IV program and analyzed by using Stata (version 10.0) software (Stata Corporation, College Station, TX). First, logistic regression was used to examine univariate associations of sociodemographic and clinical factors with false-posi-tive Murex EIA results for HIV-negafalse-posi-tive members of the cohort. The analysis was restricted to individuals for whom complete sociodemographic and clinical data were available. Second, since the prevalence of schistosomiasis and other infections was likely to differ among the 20 communities involved in the trial, conditional logistic regression was used to examine the association with each factor, controlling for community. All factors whose adjusted univariate associ-ation with false positivity reached significance at aPvalue of⬍0.10 were in-cluded in a multivariate model; factors that remained independently associated with the outcome at aPvalue of⬍0.05 were retained.
Next, immunological factors (IgG antibodies to malaria and schistosomiasis, heterophile antibody, and rheumatoid factor titer) associated with false-positive Murex EIA results were analyzed for a subsample of 524 HIV-negative serum samples by using conditional logistic regression. The analysis was conditioned on community. Since females and those with TPPA test-positive results were over-sampled among the Murex EIA-negative controls, the analysis was adjusted for sex, age, and TPPA test results. For each antibody test, the results were con-verted from continuous optical density (OD) values into an ordered, four-cate-gory variable for analysis, based on the quartiles of the OD distribution for the samples Murex EIA-negative results.
Immunological variables that had a significant linear association with false posi-tivity (P-value trend⬍0.05), adjusted for age, sex, TPPA test result, and community, were considered for inclusion in a multivariate model. Many of the antibody tests
were colinear, so the multivariate model was built in two stages. First, immunological variables that had a positive linear association with false positivity by the Murex EIA were added to the model, one at a time, in order of decreasing strength of association in the univariate analysis. Variables that improved the model fit by aPvalue of
⬍0.10 by use of the likelihood ratio test were retained. Second, immunological variables that were inversely associated with false positivity were added to the multivariate model, one at a time, by using the same approach. Lastly, the final model was reached by excluding variables one at a time until all remaining immu-nological variables were significant at the level of aPvalue of⬍0.05.
Ethical approval.Ethical clearance for the main study was obtained from the Tanzanian Medical Research Coordinating Committee and the Ethics Commit-tee of the London School of Hygiene and Tropical Medicine.
RESULTS
Of 7,280 HIV-negative participants, complete questionnaire
and clinical data were available for 6,940 (95.3%), and those
data were included in the analysis of sociodemographic and
clinical risk factors (6,366 Murex EIA-negative participants
and 574 participants with false-positive Murex EIA results;
Table 1). The mean (standard deviation [SD]) age was 18.1
(1.3) years; 58.5% of the participants were males.
In the univariate analysis, the sociodemographic factors
as-sociated with false-positive results included ethnic group, being
unemployed, and marital status (although that factor had less
of an association among those who were previously married).
However, after controlling for community, none of the
socio-demographic factors was significantly associated with
false-positive results. In the univariate analysis of clinical symptoms,
a diagnosis of possible schistosomiasis and higher scores by the
urine RBC dipstick test were significantly associated with false
positivity. False-positive results were less prevalent in
partici-pants with respiratory tract infections. Those associations
re-mained after community was controlled for. In the multivariate
model, a diagnosis of possible schistosomiasis was an
indepen-dent risk factor for false-positive results (adjusted odds ratio
[OR]
⫽
1.37, 95% confidence interval [CI]
⫽
1.11 to 1.70). In
addition, false-positive results were inversely associated with
respiratory tract infection (adjusted OR
⫽
0.56, 95% CI
⫽
0.31
to 1.00). There was no significant association between a
diag-nosis of malaria and a false-positive result or between HSV-2
or TPPA serology and false-positive results in either the
uni-variate or the adjusted analysis.
A subsample of 524 HIV-negative participants (284 Murex
EIA negative and 240 Murex EIA false positive; Table 2) were
included in the analysis of the immunological factors
associ-ated with false-positive results; their mean (SD) age was 18.0
(1.3) years, and 50.0% were males. In the univariate analysis,
false-positive results were significantly associated with
increas-ing levels of
S. mansoni
egg IgG1 and IgG3, increasing
S.
haematobium
worm IgG1, and an RF titer of
ⱖ
80 (Table 2). In
addition, false-positive results were significantly associated
with decreasing levels of
S. mansoni
worm IgG1 to IgG3,
S.
haematobium
worm IgG3, and
P. falciparum
IgG1 to IgG4.
In the final multivariate model, after age, sex, the TPPA test
result, and community were controlled for, independent
im-munological risk factors for false-positive Murex EIA results
were increasing levels of
S. mansoni
egg IgG1 and
S.
haema-tobium
worm IgG1 and an RF titer of
ⱖ
80 (Table 3). In
addition, false-positive results were significantly associated
with decreasing levels of
S. mansoni
worm IgG1 and IgG2 and
P. falciparum
IgG1 and IgG4.
on May 16, 2020 by guest
http://jcm.asm.org/
DISCUSSION
We found that a clinical diagnosis of urinary schistosomiasis
was strongly associated with false-positive results by the Murex
EIA in a cohort of young people in northwestern Tanzania.
Furthermore, among a sample of participants, we observed an
extremely strong association between false-positive results and
specific immune responses to schistosomiasis, especially
in-creased titers of IgG1 antibody to the
S. haematobium
worm
antigen. We also found a significant independent association
between an increased
S. mansoni
egg IgG1 titer and
false-TABLE 1. Association of sociodemographic and clinical factors with false-positive Murex EIA results for HIV in adolescents in Mwanza
region, Tanzania
aVariable
No. (%b
) of participants with the
following Murex EIA results: Univariate OR
(95% CI) P
c Adjusted ORd
(95% CI) P
e Negative
(n⫽6,366)
False positive (n⫽574)
Sociodemographic
Age (yr)
⬍
18
2,483 (39.0)
207 (36.1)
1
0.57
1
0.43
18
1,794 (28.2)
167 (29.1)
1.12 (0.90, 1.38)
1.10 (0.89, 1.37)
19
1,188 (18.7)
113 (19.7)
1.14 (0.90, 1.45)
1.15 (0.90, 1.46)
ⱖ
20
901 (14.1)
87 (15.2)
1.16 (0.89, 1.50)
1.22 (0.94, 1.60)
Sex
Female
2,632 (41.3)
247 (43.0)
1
0.43
1
0.47
Male
3,734 (58.7)
327 (57.0)
0.93 (0.79, 1.11)
0.94 (0.79, 1.12)
Ethnic group
Sukuma
5,063 (79.5)
484 (84.3)
1
0.04
1
0.33
Chaga
938 (14.7)
66 (11.5)
0.74 (0.56, 0.96)
0.82 (0.61, 1.11)
Zinza
269 (4.2)
16 (2.8)
0.62 (0.37, 1.04)
0.68 (0.39, 1.16)
Jita
96 (1.5)
8 (1.4)
0.87 (0.42, 1.80)
0.98 (0.46, 2.06)
Education
Less than grade 7
616 (9.7)
57 (9.9)
1
0.84
1
0.76
Grade 7
5,750 (90.3)
517 (90.1)
0.97 (0.73, 1.29)
0.96 (0.71, 1.28)
Occupation
Agriculture/fishing
4,774 (75.0)
409 (71.3)
1
0.04
1
0.14
Professional/business
510 (8.0)
52 (9.1)
1.19 (0.88, 1.61)
1.17 (0.86, 1.60)
Unemployed
707 (11.1)
85 (14.8)
1.40 (1.10, 1.80)
1.32 (1.02, 1.70)
Other
375 (5.9)
28 (4.9)
0.87 (0.59, 1.27)
0.90 (0.60, 1.35)
Marital status
Never married
5,471 (85.9)
496 (86.4)
1
0.08
1
0.11
Married now
836 (13.1)
77 (13.4)
1.02 (0.79, 1.31)
1.04 (0.80, 1.33)
Previously married
59 (0.9)
1 (0.2)
0.19 (0.03, 1.35)
0.20 (0.03, 1.48)
Clinical factors
Malaria
f600 (9.4)
58 (10.1)
1.08 (0.81, 1.44)
0.60
1.13 (0.84, 1.51)
0.42
Schistosomiasis
f1,223 (19.2)
145 (25.3)
1.42 (1.17, 1.73)
⬍
0.001
1.37 (1.11, 1.70)
0.003
Respiratory tract infections
244 (3.8)
12 (2.1)
0.54 (0.30, 0.96)
0.02
0.56 (0.31, 1.00)
0.03
Fungal infections
754 (11.8)
60 (10.5)
0.87 (0.66, 1.15)
0.31
0.87 (0.65, 1.15)
0.31
Pelvic inflammatory disease
102 (1.6)
11 (1.9)
1.20 (0.64, 2.25)
0.58
1.25 (0.66, 2.38)
0.50
Intestinal worms
580 (8.8)
43 (7.5)
0.84 (0.61, 1.16)
0.29
0.96 (0.69, 1.33)
0.79
TPPA test seropositive
155 (2.4)
12 (2.1)
0.86 (0.47, 1.55)
0.60
0.86 (0.47, 1.56)
0.60
HSV seropositive
980 (15.4)
83 (14.5)
0.93 (0.73, 1.18)
0.55
0.97 (0.76, 1.23)
0.78
Urine dipstick test result
Negative
4,647 (73.0)
384 (66.9)
1
0.01
1
0.04
⫹
174 (2.7)
12 (2.1)
0.83 (0.46, 1.51)
0.77 (0.42, 1.40)
⫹⫹
139 (2.2)
16 (2.8)
1.39 (0.82, 2.36)
1.29 (0.75, 2.19)
⫹⫹⫹
565 (8.9)
70 (12.2)
1.50 (1.15, 1.96)
1.44 (1.09, 1.91)
⫹⫹⫹⫹
841 (13.2)
92 (16.0)
1.32 (1.04, 1.68)
1.27 (0.99, 1.62)
a
Data are for 6,940 HIV-negative cohort members for whom complete questionnaire and clinical data were available. b
Column percentages. c
Pvalue for univariate ORs. d
Sociodemographic variables were conditioned on community. Except for urine dipstick test results, the clinical variables were adjusted for schistosomiasis and respiratory tract infections, conditioned on community. Urine dipstick test results were adjusted for respiratory tract infections, conditioned on community.
e
Pvalue for adjusted ORs. f
See text for diagnostic criteria.
on May 16, 2020 by guest
http://jcm.asm.org/
[image:3.585.48.544.87.583.2]TABLE 2. Association of immunological characteristics with false-positive Murex EIA results for HIV
in adolescents in Mwanza region, Tanzania
aFactorb
No. (%c
) of participants with the
following Murex EIA results: OR (95% CI) OD range for
uninfected controlse Negative
(n⫽284)
False positive
(n⫽240 Univariate Adjusted
d
S. mansoni
worm IgG1
Pt
f⫽
0.007
Pt
⫽
0.016
⫺
0.006–0.022
72 (25.4)
74 (30.8)
1
1
⫺
0.001–0.019
0.022–0.059
70 (24.6)
59 (24.6)
0.82 (0.51, 1.32)
0.65 (0.38, 1.11)
0.059–0.167
71 (25.0)
82 (34.2)
1.12 (0.71, 1.77)
0.90 (0.54, 1.51)
0.167–0.704
71 (25.0)
25 (10.4)
0.34 (0.20, 0.60)
0.32 (0.16, 0.62)
S. mansoni
worm IgG2
Pt
⬍
0.001
Pt
⬍
0.001
⫺
0.03–0.06
73 (25.7)
113 (47.1)
1
1
⫺
0.016–0.024
0.06–0.106
73 (25.7)
37 (15.4)
0.33 (0.20, 0.54)
0.28 (0.16, 0.49)
0.106–0.196
68 (23.9)
51 (21.3)
0.48 (0.30, 0.77)
0.49 (0.29, 0.83)
0.196–0.697
70 (24.6)
39 (16.3)
0.36 (0.22, 0.59)
0.31 (0.18, 0.55)
S. mansoni
worm IgG3
Pt
⫽
0.049
Pt
⫽
0.034
⫺
0.002–0.1285
71 (25.0)
92 (38.3)
1
1
⫺
0.003–0.053
0.1285–0.224
71 (25.0)
42 (17.5)
0.46 (0.28, 0.75)
0.44 (0.25, 0.76)
0.224–0.362
72 (25.4)
49 (20.4)
0.53 (0.33, 0.85)
0.48 (0.28, 0.82)
0.362–1.092
70 (24.6)
57 (23.8)
0.63 (0.39, 1.00)
0.60 (0.34, 0.96)
S. mansoni
worm IgG4
Pt
⫽
0.17
Pt
⫽
0.17
⫺
0.014–0.0055
71 (25.0)
45 (18.8)
1
1
⫺
0.007–0.013
0.0055–0.053
72 (25.4)
60 (25.0)
1.31 (0.79, 2.18)
1.24 (0.71, 2.17)
0.053–0.197
70 (24.6)
74 (30.8)
1.67 (1.02, 2.74)
1.62 (0.93, 2.82)
0.197–1.141
71 (25.0)
61 (25.4)
1.36 (0.82, 2.25)
1.42 (0.78, 2.60)
S. mansoni
egg IgG1
Pt
⬍
0.001
Pt
⬍
0.001
⫺
0.011–0.04
71 (25.0)
37 (15.4)
1
1
0.01–0.034
0.04–0.1205
71 (25.0)
40 (16.7)
1.08 (0.62, 1.88)
0.89 (0.48, 1.65)
0.1205–0.296
72 (25.4)
58 (24.2)
1.55 (0.91, 2.62)
1.47 (0.82, 2.61)
0.296–1.096
70 (24.6)
105 (43.8)
2.88 (1.75, 4.74)
2.86 (1.64, 4.99)
S. mansoni
egg IgG2
Pt
⫽
0.28
Pt
⫽
0.16
0.005–0.169
71 (25.0)
61 (25.4)
1
1
0.025–0.081
0.169–0.292
72 (25.4)
50 (20.8)
0.81 (0.49, 1.33)
0.84 (0.49, 1.44)
0.292–0.4485
70 (24.6)
52 (21.7)
0.86 (0.53, 1.42)
0.96 (0.55, 1.65)
0.4485–1.244
71 (25.0)
77 (32.1)
1.26 (0.79, 2.02)
1.41 (0.84, 2.36)
S. mansoni
egg IgG3
Pt
⬍
0.001
Pt
⬍
0.001
0.001–0.067
71 (25.0)
18 (7.5)
1
1
0.008–0.072
0.067–0.15
72 (25.4)
26 (10.8)
1.42 (0.72, 2.82)
1.56 (0.76, 3.21)
0.15–0.401
71 (25.0)
74 (30.8)
4.11 (2.23, 7.58)
4.12 (2.14, 7.95)
0.401–1.07
70 (24.6)
122 (50.8)
6.87 (3.79, 12.46)
6.86 (3.58, 13.16)
S. mansoni
egg IgG4
Pt
⫽
0.81
Pt
⫽
0.017
⫺
0.008–0.018
72 (25.4)
33 (13.8)
1
1
⫺
0.005–0.047
0.018–0.1325
70 (24.6)
97 (40.4)
3.02 (1.81, 5.06)
2.39 (1.37, 4.17)
0.1325–0.486
71 (25.0)
73 (30.4)
2.24 (1.33, 3.80)
2.54 (1.42, 4.52)
0.486–0.998
71 (25.0)
37 (15.4)
1.14 (0.64, 2.02)
2.04 (0.99, 4.19)
S. hematobium
worm IgG1
Pt
⬍
0.001
Pt
⬍
0.001
⫺
0.005–0.0235
71 (25.0)
13 (5.4)
1
1
0.002–0.042
0.0235–0.0615
71 (25.0)
37 (15.4)
2.85 (1.40, 5.80)
3.24 (1.51, 6.93)
0.0615–0.166
72 (25.4)
57 (23.8)
4.32 (2.18, 8.58)
5.29 (2.47, 11.31)
0.166–0.805
70 (24.6)
133 (55.4)
10.38 (5.37, 20.04)
12.01 (5.60, 25.77)
S. hematobium
worm IgG2
Pt
⫽
0.65
Pt
⫽
0.20
⫺
0.001–0.0545
71 (25.0)
74 (30.8)
1
1
⫺
0.003–0.025
0.0545–0.1055
71 (25.0)
50 (20.8)
0.68 (0.42, 1.10)
0.52 (0.30, 0.88)
0.1055–0.196
72 (25.4)
50 (20.8)
0.67 (0.41, 1.08)
0.48 (0.27, 0.82)
0.196–1.079
70 (24.6)
66 (27.5)
0.90 (0.57, 1.44)
0.72 (0.42, 1.22)
S. hematobium
worm IgG3
Pt
⫽
0.002
Pt
⬍
0.001
⫺
0.031–0.088
72 (25.4)
93 (38.8)
1
1
⫺
0.006–0.034
0.088–0.1695
70 (24.6)
54 (22.5)
0.60 (0.37, 0.96)
0.50 (0.30, 0.84)
Continued on following page
on May 16, 2020 by guest
http://jcm.asm.org/
positive results, although that effect was not as strong.
Inter-estingly, we observed a significant inverse association between
false-positive Murex EIA results and the levels of some IgG
isotypes, including those of
S. mansoni
worm IgG1 and IgG2
and
P. falciparum
IgG1 and IgG4. The reasons for these
in-verse associations are unclear, and any attempt to suggest
possible explanations would be purely speculative.
Schistosomiasis is endemic in many parts of sub-Saharan
Africa and disproportionately affects adolescents. In many
ar-eas, most children have been infected by the age of 14 years
(27). In our study population, 73% of the samples had
S.
haematobium
IgG1 OD levels more than 2 standard deviations
above the mean for uninfected European controls. A survey of
6,897 schoolchildren aged 7 to 20 years in the Mwanza region
of Tanzania found the prevalence of
S. haematobium
infection
to be 56.5% and that of
S. mansoni
infection to be 10.9% (24).
The schools with the highest
S. mansoni
prevalence were
lo-cated less than 5 km from the shore of Lake Victoria, whereas
S. haematobium
was found in all localities. The wider
geo-graphical distribution of
S. haematobium
is mainly due to the
distribution of its intermediate hosts.
[image:5.585.46.539.79.512.2]We also found a strong association between an RF titer of
ⱖ
80 and false-positive Murex EIA results. RF is an
autoanti-body against the Fc component on IgG and is associated with
TABLE 2—
Continued
Factorb
No. (%c) of participants with the
following Murex EIA results: OR (95% CI) OD range for
uninfected controlse Negative
(n⫽284)
False positive
(n⫽240 Univariate Adjusted
d
0.1695–0.290
71 (25.0)
47 (19.6)
0.51 (0.32, 0.83)
0.48 (0.28, 0.83)
0.290–0.806
71 (25.0)
46 (19.2)
0.50 (0.31, 0.81)
0.41 (0.24, 0.70)
S. hematobium
worm IgG4
Pt
⫽
0.14
Pt
⫽
0.048
⫺
0.018–0.014
73 (25.7)
55 (22.9)
1
1
⫺
0.001–0.019
0.014–0.054
70 (24.6)
73 (30.4)
1.38 (0.86, 2.24)
1.33 (0.77, 2.29)
0.054–0.219
70 (24.6)
84 (35.0)
1.59 (0.99, 2.55)
1.28 (0.74, 2.21)
0.219–0.730
71 (25.0)
28 (11.7)
0.52 (0.30, 0.92)
0.45 (0.23, 0.86)
P. falciparum
IgG1
Pt
⫽
0.002
Pt
⬍
0.001
⫺
0.005–
⬍
0.079
72 (25.4)
98 (40.8)
1
1
⫺
0.007–0.017
0.079–0.150
70 (24.6)
55 (22.9)
0.58 (0.36, 0.92)
0.45 (0.26, 0.76)
0.150–0.231
71 (25.0)
32 (13.3)
0.33 (0.20, 0.55)
0.29 (0.16, 0.51)
⬎
0.231–0.731
71 (25.0)
55 (22.9)
0.57 (0.36, 0.91)
0.47 (0.27, 0.80)
P. falciparum
IgG2
Pt
⫽
0.015
Pt
⫽
0.023
⫺
0.028–0.089
72 (25.4)
92 (38.3)
1
1
⫺
0.004–0.048
0.089–0.1505
70 (24.6)
50 (20.8)
0.56 (0.35, 0.90)
0.62 (0.37, 1.04)
0.1505–0.261
71 (25.0)
44 (18.3)
0.48 (0.30, 0.79)
0.50 (0.29, 0.86)
0.261–0.979
71 (25.0)
54 (22.5)
0.60 (0.37, 0.95)
0.59 (0.35, 1.00)
P. falciparum
IgG3
Pt
⫽
0.048
Pt
⫽
0.050
0.008–0.38
72 (25.4)
112 (46.7)
1
1
0.005–0.045
0.38–0.458
71 (25.0)
26 (10.8)
0.24 (0.14, 0.40)
0.19 (0.10, 0.34)
0.458–0.52
73 (25.7)
27 (11.3)
0.24 (0.14, 0.40)
0.16 (0.09, 0.29)
0.52–1.06
68 (23.9)
75 (31.3)
0.71 (0.46, 1.10)
0.71 (0.42, 1.21)
P. falciparum
IgG4
Pt
⬍
0.001
Pt
⬍
0.001
⫺
0.006–0.012
77 (27.1)
119 (49.6)
1
1
0.0–0.004
0.012–0.024
68 (23.9)
38 (15.8)
0.36 (0.22, 0.59)
0.29 (0.16, 0.48)
0.024–0.056
68 (23.9)
45 (18.8)
0.43 (0.27, 0.69)
0.33 (0.19, 0.57)
0.056–0.551
71 (25.0)
38 (15.8)
0.35 (0.21, 0.56)
0.29 (0.17, 0.51)
Heterophile antibody
Pt
⫽
0.42
Pt
⫽
0.66
⫺
0.012–0.006
73 (25.7)
51 (21.3)
1
1
0.003–0.039
0.006–0.025
69 (24.3)
63 (26.3)
1.31 (0.80, 2.14)
0.90 (0.52, 1.55)
0.025–0.095
71 (25.0)
64 (26.7)
1.29 (0.79, 2.11)
0.87 (0.51, 1.50)
0.095–0.980
71 (25.0)
62 (25.8)
1.25 (0.76, 2.05)
0.88 (0.51, 1.53)
Rheumatoid factor titer
fP
⬍
0.001
P
⬍
0.001
⬍
80
161 (94.7)
161 (79.7)
1
1
ⱖ
80
9 (5.3)
41 (20.3)
4.56 (2.14, 9.68)
4.69 (2.03, 10.81)
a
Data are for a subsample of 524 HIV-negative participants (284 Murex EIA negative and 240 Murex EIA false positive). b
Categories were defined on the basis of the quartiles of the OD results for the participants with true-negative results. c
Column percentages. d
Adjusted for age, sex, and TPPA test result, conditioned on community. e
Estimated normal range of ODs for uninfected European controls, calculated as the mean⫾2⫻SD on the basis of data from the Department of Pathology, University of Cambridge.
f
Pt,Pvalue for linear trend.
on May 16, 2020 by guest
http://jcm.asm.org/
autoimmune diseases, such as rheumatoid arthritis. However,
it can also be produced in individuals with viral, bacterial, and
other parasitic infections (30, 34). It occurs naturally in up to
4% of healthy European populations (20) and in up to 30% of
individuals in some ethnic groups (2, 20). Cross-reactivity
be-tween RF and the results of diagnostic tests for malaria has
been reported (17, 21).
Cross-reactivity between HIV-1 peptides and antibody to
S.
mansoni
has been demonstrated in sera from children from
Kenya infected with
S. mansoni
and confirmed to be HIV
negative (23) Given the high prevalence of schistosomiasis in
many African populations and the strong association between
S. haematobium
IgG1 antibody and false-positive Murex EIA
results that we found in our study, a high proportion of
false-positive Murex EIA results and, perhaps, other HIV assays in
similar settings in sub-Saharan Africa may be attributable to
cross-reactivity with schistosoma antibody. This warrants
fur-ther investigation.
Our study was a detailed investigation of the possible
mech-anisms for false positivity following the low specificity that we
observed with the Murex EIA in this population. We had
chosen this EIA for use as the confirmatory test in a serial
testing strategy (13). While the rate of false-positive results
gave considerable cause for concern, we do not believe that
this problem is restricted to the Murex EIA.
Lower-than-ex-pected specificities were also observed in this study population
with other 4th-generation EIAs (unpublished data), including
the screening test used in our algorithm, the Uni-Form II
Ag/Ab EIA (bioMe
´rieux, Marcy l’Etoile, France), although not
to the same degree. Problems of low specificity with HIV EIAs
have also been reported elsewhere (8, 22, 32), and more
re-cently, problems with false positive results by simple rapid
point-of-care tests for HIV have been found (15).
[image:6.585.44.538.88.479.2]While specificity should be improved when tests are used in
combination, the high rates of false-positive results associated
with the 4th-generation EIAs may lead to a considerable rise in
TABLE 3. Final multivariate model for the immunological characteristics associated with false-positive Murex EIA results for HIV in
adolescents in Mwanza region, Tanzania
Factor No. (%
a
) of participants with the following Murex EIA results:
Adjusted ORb(95% CI) Negative (n⫽170) False positive (n⫽202)
S. mansoni
worm IgG1
Pt
c⫽
0.002
⫺
0.006–0.022
34 (20.0)
61 (30.2)
1
0.022–0.059
45 (26.5)
49 (24.3)
0.31 (0.12, 0.83)
0.059–0.167
47 (27.7)
75 (37.1)
0.29 (0.10, 0.87)
0.167–0.704
44 (25.9)
17 (8.4)
0.09 (0.02, 0.37)
S. mansoni
worm IgG2
Pt
⬍
0.001
⫺
0.03–0.06
42 (24.7)
93 (46.0)
1
0.06–0.106
42 (24.7)
34 (16.8)
0.28 (0.11, 0.71)
0.106–0.196
42 (24.7)
42 (20.8)
0.30 (0.12, 0.77)
0.196–0.697
44 (25.9)
33 (16.3)
0.11 (0.03, 0.34)
S. mansoni
egg IgG1
Pt
⫽
0.033
⫺
0.011–0.04
35 (20.6)
30 (14.9)
1
0.04–0.1205
40 (23.5)
34 (16.8)
0.57 (0.18, 1.80)
0.1205–0.296
52 (30.6)
46 (22.8)
0.62 (0.21, 1.87)
0.296–1.096
43 (25.3)
92 (45.5)
2.72 (0.80, 9.16)
S. hematobium
worm IgG1
Pt
⬍
0.001
⫺
0.005–0.0235
36 (21.2)
7 (3.5)
1
0.0235–0.0615
39 (22.9)
33 (16.3)
3.79 (1.04, 13.82)
0.0615–0.166
46 (27.1)
46 (22.8)
6.58 (1.62, 26.67)
0.166–0.805
49 (28.8)
116 (57.4)
40.67 (8.51, 194.22)
P. falciparum
IgG1
Pt
⬍
0.001
⫺
0.005–
⬍
0.079
37 (21.8)
82 (40.6)
1
0.079–0.150
42 (24.7)
42 (20.8)
0.32 (0.13, 0.80)
0.150–0.231
45 (26.5)
28 (13.9)
0.17 (0.06, 0.47)
⬎
0.231–0.731
46 (27.1)
50 (24.8)
0.24 (0.09, 0.62)
P. falciparum
IgG4
Pt
⫽
0.030
⫺
0.006–0.012
39 (22.9)
100 (49.5)
1
0.012–0.024
38 (22.4)
29 (14.4)
0.25 (0.10, 0.65)
0.024–0.056
42 (24.7)
40 (19.8)
0.41 (0.16, 1.03)
0.056–0.551
51 (30.0)
33 (16.3)
0.33 (0.13, 0.88)
Rheumatoid factor titer
P
⬍
0.001
⬍
80
161 (94.7)
161 (79.7)
1
ⱖ
80
9 (5.3)
41 (20.3)
8.24 (2.79, 24.31)
a
Column percentages. b
Adjusted for age, sex, TPPA test result, and all variables in the table, conditioned on community. c
Pt,Pvalue for linear trend.
on May 16, 2020 by guest
http://jcm.asm.org/
the cost of testing, since positive test results by a particular
assay require confirmation when that assay is used as part of a
testing algorithm. This would be of special concern in settings
in which financial resources are limited or in which the capacity
to employ technically difficult confirmatory tests such as the
Western blot assay are limited. In our study, further
confirma-tory testing at a United Kingdom reference center by a number
of different assays, including a PCR directed toward the
pol
gene, was needed to correct the false-positive results obtained
by the EIAs.
Although the specificity of the 4th-generation Murex EIA
was very low for our study population of adolescents and young
adults, we have used the same assay to test older adults in the
Mwanza region and found its specificity to be within the
man-ufacturer’s reported range (12). The prevalence of HIV
infec-tion in these older groups is much higher, generally ranging
from 15 to 30%, whereas it was
⬍
1% in the younger cohort
evaluated in the present study. However, although the
preva-lence of HIV infection will affect the assay’s positive predictive
value, it should not affect its specificity. The prevalence of
schistosomiasis may be lower in older adults, or there may be
age-related changes in antibody responses that make
cross-reactivity less likely (29). Schistosoma-specific antibody
re-sponses in areas where schistosomiasis is endemic have been
shown to be dependent on the age of the individual and the
intensity of infection (28). In studies designed to investigate
resistance to reinfection with
S. mansoni
in areas of endemicity
after treatment, the levels of IgG2 against parasite egg
poly-saccharide antigens were found to be the highest in younger
children and declined with age, whereas other responses, such
as the IgE response against adult worm antigens, tended to rise
with age (9, 10).
There are several limitations to our analysis. Due to the
colinearity between the immunological tests, it was difficult to
disentangle the independent effects of each immunological
factor. This was a cross-sectional analysis, and so we cannot
determine the causality of the associations observed. Although
we did a comprehensive range of IgG isotype assays, we did not
have data for
S. haematobium
egg IgG and did not look at
other antibody responses, such as IgA and IgE responses.
Lastly, we examined the association of 18 immunological
vari-ables with false-positive Murex EIA results, and such multiple
statistical testing can produce spurious results. However, a
large proportion of the immunological factors were strongly
associated with false-positive results, suggesting that this was
not a chance finding. If we apply a Bonferroni correction for
multiple comparisons, eight of the univariate associations are
still statistically significant (P
⬍
0.0028).
Although the association with schistosomiasis IgG antibody
and the RF titer is unlikely to provide the complete picture, it
is clear that in our study population of Tanzanian adolescents
with a low prevalence of HIV infection, the 4th-generation
Murex EIA performed poorly (13). The p24 antigen
compo-nent of the 4th-generation tests may partially explain this (22).
The differences in serological response between African
pop-ulations and those in other regions and the prevalence of
parasitological coinfection and the subsequent immune
re-sponse will affect the performance of HIV assays. Clearly, the
best method for resolution of problems of low specificity would
be for manufacturers to design and evaluate tests specific for
the African environment.
In summary, among adolescents and young adults in
north-western Tanzania, we found a high prevalence of false-positive
results by the 4th-generation Murex EIA. Individuals with high
levels of IgG1 antibody to
S. haematobium
soluble worm and
S.
mansoni
soluble egg antigens and an RF titer
ⱖ
80 were more
likely to have false-positive Murex EIA results. Further
re-search on the cross-reactivity of HIV diagnostic tests with
endemic infections in sub-Saharan Africa is warranted.
ACKNOWLEDGMENTS
We thank the Ministry of Health, Tanzania; the National AIDS
Control Programme, Tanzania; and the director general of the
Na-tional Institute for Medical Research, Tanzania, for permission to
conduct and publish the findings of the study. We also thank Frances
Jones of the Cambridge University Department of Pathology, the staff
of the National Institute for Medical Research, Tanzania, and the
African Medical and Research Foundation, Tanzania, for their
sup-port and assistance with carrying out this study.
The study was supported by the European Commission, Irish Aid,
and the United Kingdom Department for International Development.
We have no conflicts of interest to declare.
D.B.E. designed the study reported here, supervised the original
laboratory tests, performed many of the specific testing detailed in the
study, and prepared the first draft of the manuscript. He is the
guar-antor. K.J.B. carried out the analyses and contributed to the writing of
the draft manuscript. R.J.H. provided guidance on the study design,
statistical methods, and interpretation of the assay results. R.M.
pro-vided scientific input, guidance on the study design, and interpretation
of the assay results. I.H. and T.C. prepared the initial data sets.
D.A.R., J.C., and R.J.H. were coprincipal investigators of the study
from which the samples were derived; and they, together with D.W.-J.
and D.M., provided scientific input into the design and interpretation
of the results of the specific study reported here. H.H. provided
im-munological guidance for the study. D.W.D. provided scientific input
and supervised the IgG work in Cambridge. All coauthors read and
commented on drafts of the paper.
REFERENCES
1.Abbott Laboratories.2002. Murex HIV Ag/Ab Combination: improved de-tection of seroconversion to HIV. Report A 92308. Abbott Laboratories, North Chicago, IL.
2.Adebajo, A. O., T. E. Cawston, and B. L. Hazleman.1995. Specificity of IgM antiglobulins produced in rheumatoid arthritis and tropical infections. Clin. Rheumatol.14:37–40.
3.Baba, T. W., V. Liska, R. Hofmann-Lehmann, J. Vlasak, W. Xu, S. Ayehunie, L. A. Cavacini, M. R. Posner, H. Katinger, G. Stiegler, B. J. Bernacky, T. A. Rizvi, R. Schmidt, L. R. Hill, M. E. Keeling, Y. Lu, J. E. Wright, T. C. Chou, and R. M. Ruprecht.2000. Human neutralizing monoclonal antibodies of the IgG1 subtype protect against mucosal simian-human immunodeficiency virus infection. Nat. Med.6:200–206.
4.Behets, F., A. Disasi, R. W. Ryder, K. Bishagara, P. Piot, M. Kashamuka, M. Kamenga, N. Nzila, M. Laga, and G. Vercauteren.1991. Comparison of five commercial enzyme-linked immunosorbent assays and Western immuno-blotting for human immunodeficiency virus antibody detection in serum samples from Central Africa. J. Clin. Microbiol.29:2280–2284.
5.Biggar, R. J., P. L. Gigase, M. Melbye, L. Kestens, P. S. Sarin, A. Bodner, P. Demedts, W. J. Stevens, L. Puluku, C. Delacollette, and W. A. Blattner.1985. ELISA HTLV retrovirus antibody reactivity associated with malaria and immune complexes in healthy Africans. Lancetii:520–523.
6.Binsbergen, J. V., A. Siebelink, A. Jacobs, W. Keur, F. Bruynis, M. van der Graaf, J. van der Heijden, D. Kambel, and J. Toonen.1999. Improved performance of seroconversion with a 4th generation HIV antigen/antibody assay. J. Virol. Methods82:77–84.
7.Bredberg-Råde´n, U., J. Kiango, F. Mhalu, and G. Biberfeld.1988. Evalua-tion of commercial enzyme immunoassays for anti-HIV-1 using East African sera. AIDS2:281–285.
8.Bruzzone, B., F. Bisio, A. Ventura, N. Nigro, L. M. Miguel, F. A. Mayinda Mboungou, A. C. Nzagou, P. Mayembo, F. Uberti, A. De Maria, G. Icardi, and C. Viscoli.2008. HIV serological screening in a population of pregnant women in the Republic of Congo: suitability of different assays. Trop. Med. Int. Health13:900–903.
on May 16, 2020 by guest
http://jcm.asm.org/
9.Butterworth, A. E., R. Bensted-Smith, A. Capron, M. Capron, P. R. Dalton, D. W. Dunne, J. M. Grzych, H. C. Kariuki, J. Khalife, D. Koech, et al.1987. Immunity in human schistosomiasis mansoni: prevention by blocking anti-bodies of the expression of immunity in young children. Parasitology94(Pt 2):281–300.
10.Dunne, D. W., A. E. Butterworth, A. J. Fulford, H. C. Kariuki, J. G. Langley, J. H. Ouma, A. Capron, R. J. Pierce, and R. F. Sturrock.1992. Immunity after treatment of human schistosomiasis: association between IgE antibod-ies to adult worm antigens and resistance to reinfection. Eur. J. Immunol.
22:1483–1494.
11.Dunne, D. W., S. Lucas, Q. Bickle, S. Pearson, L. Madgwick, J. Bain, and M. J. Doenhoff.1981. Identification and partial purification of an antigen (omega 1) from Schistosoma mansoni eggs which is putatively hepatotoxic in T-cell deprived mice. Trans. R. Soc. Trop. Med. Hyg.75:54–71.
12.Everett, D. B., K. Baisley, J. Changalucha, A. Vallely, D. Watson-Jones, C. Cook, L. Knight, D. A. Ross, K. Mugeye, S. McCormack, C. J. Lacey, U. Jentsch, and R. J. Hayes.2009. Suitability of simple human immunodefi-ciency virus rapid tests in clinical trials in community-based clinic settings. J. Clin. Microbiol.47:1058–1062.
13.Everett, D. B., H. A. Weiss, J. Changalucha, A. Anemona, T. Chirwa, D. A. Ross, D. Watson-Jones, J. V. Parry, R. Hayes, and D. C. Mabey.2007. Low specificity of the Murex fourth-generation HIV enzyme immunoassay in Tanzanian adolescents. Trop. Med. Int. Health12:1323–1326.
14.Fitzsimmons, C. M., T. J. Stewart, K. F. Hoffmann, J. L. Grogan, M. Yazdan-bakhsh, and D. W. Dunne.2004. Human IgE response to the Schistosoma haematobium 22.6 kDa antigen. Parasite Immunol.26:371–376.
15.Gray, R. H., F. Makumbi, D. Serwadda, T. Lutalo, F. Nalugoda, P. Opendi, G. Kigozi, S. J. Reynolds, N. K. Sewankambo, and M. J. Wawer.2007. Limitations of rapid HIV-1 tests during screening for trials in Uganda: diagnostic test accuracy study. BMJ335:188.
16.Greenwood, B. M., and H. C. Whittle.1981. Immunology of medicine in the tropics. Edward Arnold Publishers, London, United Kingdom.
17.Grobusch, M. P., U. Alpermann, S. Schwenke, T. Jelinek, and D. C. War-hurst.1999. False positive rapid tests for malaria in patients with rheumatoid factor. Lancet353:297.
18.Hayes, R. J., J. Changalucha, D. A. Ross, A. Gavyole, J. Todd, A. I. Obasi, M. L. Plummer, D. Wight, D. C. Mabey, and H. Grosskurth.2005. The MEMA kwa Vijana project: design of a community randomised trial of an innovative adolescent sexual health intervention in rural Tanzania. Con-temp. Clin. Trials26:430–442.
19.Hecht, F. M., M. Busch, B. Rawal, M. Webb, E. Rosenberg, M. Swanson, M. Chesney, J. Anderson, J. Levy, and J. O. Kahn.2002. Use of laboratory tests and clinical symptoms for identification of primary HIV infection. AIDS
16:1119–1129.
20.Husby, G., J. T. Gran, and A. Johannessen. 1988. Epidemiological and genetic aspects of IgM rheumatoid factors. Scand. J. Rheumatol. Suppl.
75:213–218.
21.Iqbal, J., A. Sher, and A. Rab.2000. Plasmodium falciparum histidine-rich protein 2-based immunocapture diagnostic assay for malaria: cross-reactivity with rheumatoid factors. J. Clin. Microbiol.38:1184–1186.
22.Jnaoui, K., N. Lamarti, A.-T. Paˆques, and P. Goubau.2007. Prozone effect with the VIDAS HIV DUO Ultra assay. J. Clin. Virol.38:273–274. 23.Khalife, J., J. M. Grzych, R. Pierce, J. C. Ameisen, A. M. Schacht, H.
Gras-Masse, A. Tartar, J. P. Lecocq, and A. Capron.1990. Immunological crossreactivity between the human immunodeficiency virus type 1 virion infectivity factor and a 170-kD surface antigen of Schistosoma mansoni. J. Exp. Med.172:1001–1004.
24.Lwambo, N. J.2005. Editorial: schistosomiasis and soil transmitted helmin-thiasis in Tanzania. Tanzania Health Res. Bull.7:iii.
25.Ly, T. D., S. Laperche, C. A. Brennan, A. S. Vallari, A. Ebel, J. C. Hunt, D. Martin, D. Daghfal, G. Schochetman, and S. Devare.2004. Evaluation of the sensitivity and specificity of six HIV combined p24 antigen and antibody assays. J. Virol. Methods122:185–194.
26.M’Boup, S., P. J. Kanki, and F. S. Mhalu.1991. Laboratory diagnosis of HIV-1 and HIV-2 in Africa. AIDS5:93–101.
27.Mott, K. E.2004. Schistosomiasis, p. 349–391.InC. J. L. Murray, A. D. Lopez, and C. D. Mathers (ed.), The global epidemiology of infectious diseases, vol. 4. WHO, Geneva, Switzerland.
28.Naus, C. W. A., F. M. Jones, M. Z. Satti, S. Joseph, E. M. Riley, G. Kimani, J. K. Mwatha, C. H. Kariuki, J. H. Ouma, N. B. Kabatereine, B. J. Venner-vald, and D. W. Dunne.2003. Serological responses among individuals in areas where both schistosomiasis and malaria are endemic: cross-reactivity between Schistosoma mansoni and Plasmodium falciparum. J. Infect. Dis.
187:1272–1282.
29.Naus, C. W. A., G. Kimani, J. H. Ouma, A. J. C. Fulford, M. Webster, G. J. van Dam, A. M. Deelder, A. E. Butterworth, and D. W. Dunne.1999. De-velopment of antibody isotype responses to Schistosoma mansoni in an immunologically naive immigrant population: influence of infection dura-tion, infection intensity; and host age. Infect. Immun.67:3444–3451. 30.Newkirk, M. M.2002. Rheumatoid factors: what do they tell us? J.
Rheu-matol.29:2034–2040.
31.Ross, D. A., J. Changalucha, A. I. Obasi, J. Todd, M. L. Plummer, B. Cleophas-Mazige, A. Anemona, D. Everett, H. A. Weiss, D. C. Mabey, H. Grosskurth, and R. J. Hayes.2007. Biological and behavioural impact of an adolescent sexual health intervention in Tanzania: a community-randomized trial. AIDS21:1943–1955.
32.Salinas, A., M. Gorgolas, and M. Fernandez-Guerrero.2007. Refrain from telling bad news: patients with leishmaniasis can have false-positive HIV test results. Clin. Infect. Dis.45:139–140.
33.Schupbach, J., and M. Tanner.1986. Specificity of human immunodeficiency virus (LAV/HTLV-III)-reactive antibodies in African sera from southeast-ern Tanzania. Acta Trop.43:195–206.
34.Tighe, H., and D. A. Carson.1997. Rheumatoid factors, p. 241–249.InW. N. Lilley, E. D. Harris, S. Ruddy, and C. B. Sledge (ed.), Textbook of rheuma-tology. The W. B. Saunders Company, Philadelphia, PA.
35.UNAIDS.2007. AIDS epidemic update. Report UNAIDS/07.27E / JC1322E. UNAIDS, Geneva, Switzerland.
36.UNAIDS.2008. Report on the global HIV/AIDS epidemic. Report UN-AIDS/08.25E / JC1510E. UNAIDS, Geneva, Switzerland.
