Abstract Book Vol. 2 [International Conference on AIDS (14th: 2002: Barcelona, Spain)]

108 Abstracts WePeC6100-WePeC6103

XIV International AIDS Conference

-03010, Alicante, Spain, Tel.: [phone], Fax: [phone], E-mail: [email protected]

Presenting author: Rebecca Bunnell, CDC-Uganda, PO Box 49, Entebbe,
Uganda, Tel.: [phone], Fax: [phone], E-mail: [email protected]

WePeC6100 The impact of making HIV infection a notifiable I WePeC6102 Describing heterogeneity in HIV prevalence in
disease on requests for HIV antibody testing in Malawi using sentinel surveillance data
Alb rtn C-, dE aI a e EEt

Mlii,/ U 2.9 %,au uauW.. A ob u noI~l I Fl&
G.C. Jayaraman1, J. Preiksaitis2, B. Larke3. 1Health Canada, Provincial
Laboratory for Public Health, 3030 Hospital Drive, NW Calgary, Alberta,
Canada; 2Provincial Laboratory for Public Health, Edmonton, Canada;
3Canadian Blood Services, Whitehorse, Canada
Background: HIV became reportable in the province of Alberta, Canada on May
1, 1998. Routine prenatal screening for HIV began on September 1, 1998. We
assessed whether fears that HIV reporting would deter people from testing were
reflected in the actual number of HIV screenings performed and to evaluate the
impact of the routine prenatal screening program on HIV testing.
Methods: The Provincial Laboratory for Public Health (PLPH) database provided
the number of HIV tests conducted prior to and since HIV became reportable.
The Canadian Blood Services (CBS) database provided the number of HIV tests
for routine prenatal HIV screening program. Further analyses using gender and
age at testing were also conducted.
Results: Among males, the number of HIV tests increased from 2,230 ~ 167
tests per month in the 16 months prior to HIV notifiability, to 2,440 ~ 185 tests
since this time (p=0.001). Among females,
3,366 ~ 341 tests per month were performed in the 16 months prior to HIV reporting. Between May 1, to August 31, 1998, the numbers increased to 4,208 -
104 tests per month (p>0.0001). Since
September 1, 1998, the numbers of tests have declined to 3,075 ~ 244 tests per
roonth (p>0.0001). Inclusion of prenatal HIV screenings almost double the figures
from 3,534 ~ 461 tests per month to
6,960 ~ 490 tests per month Septemer 1, 1998 (p>0.0001). There has been a
significant decline over time in the proportion of women refusing prenatal screening for HIV.
However there has been an increase in the rates of confirmed HIV positive cases
among the women tested under this program.
Conclusions: Despite HIV reporting, the total number of HIV screening tests has
increased significantly over time. Routine prenatal screening program has almost
certainly increased coverage for women of childbearing age. Of concern are the
increased rates of HIV infection during pregnancy and further studies are required
to understand this trend, if it persists.
Presenting author: Gayatri Jayaraman, Provincial Laboratory for Public Health,
3030 Hospital Drive, NW, Calgary, Alberta, Canada, Tel.: [phone], Fax:
[phone], E-mail: [email protected]
WePeC6101 1Using voluntary counselling and testing data to
monitor prevalence and incidence in the Uganda
AIDS epidemic
R. Bunnell1, F. Baryarama1, R. Ransom1, J. Kalule2, E. Tumuhairwe2,
C. Hitimana-Lukanika2, J. Mermin1. 'CDC-Uganda, CDC-Uganda, PO Box 49,
Entebbe, Uganda; 2AIDS Information Centre, Kampala, Uganda
Background: Demographic and risk information is often routinely collected as
part of voluntary counseling and testing (VCT) and could provide important
surveillance data at little to no cost. We analyzed trends in HIV prevalence and
estimated incidence among VCT clients in Uganda.
Methods: Data from clients tested between 1992 and 2000 at four AIDS Information Centre VCT sites were analyzed. Repeat testers and clients reporting
possible HIV-related illnesses were excluded. Age and gender-specific incidence
was estimated using a birth cohort analysis with adjustment for differences in
expected mortality among HIV-infected and uninfected clients.
Results: Data from 203,299 clients were analyzed. Overall, HIV prevalence declined from 23% to 13% (men: 17%-9%; women: 30%-17%). Among men, declines were observed within all age groups except those above 40 years. Among
women, large declines were observed in the younger age groups, but not in
women over 30 years old. For both male and female cohorts, the highest incidence was among those born before 1967.
Age in years Observed Observed Expected Average annual HIV
in 1992 HIV prevalence HIV prevalence HIV prevalence incidence per 100
in 1992 in 2000 in 2000* uninfected persons,
1992-2000
Women 15-49 yrs 30.8 (9228) 25.5 (5179) 15.2 1.9
Men 15-49 yrs 17.6 (11249) 11 (7095) 7.9 0.5
Total 23.6 (20477) 17.1 (12274) 11.1 1.0
*if no incidence occurred between 1992 and 2000.
Conclusion: In Uganda's mature HIV epidemic, peak prevalence and incidence
are in adults aged 35 and above, reinforcing the need for care and prevention
interventions targeting this group. High prevalence and incidence within groups
who are not captured in antenatal clinic data demonstrates the utility of analyzing
routinely collected VCT data for surveillance and for monitoring shifts in high risk
groups.

S.O.M. Manda. University of Malawi, Department of Statistics, University of
Auckland, Private Bag 92019, Auckland, Malawi
Background In Malawi, the National AIDS Control Programme (NACP) annually
undertakes a HIV sero-surveillance of women attending antenatal clinics (ANC)
in selected sites across the country. Combining such data to provide accurate
estimates of the HIV prevalence rates in the country is a challenging problem. In
this presentation, a hierarchical model structure for the site prevalence rates is
proposed.
Methods A Bayesian random effects logistic regression model is used to model
the HIV prevalence rates in the sites. It is applied to the 7,361 pregnant women
sampled from 19 sentinel sites across the country in 2001. Blood samples from
these women were tested for HIV as well as syphillis. Risk factors available for
consideration were the tree regions (North, Centre and South), locality (urban,
semi-urban and rural) and syphillis prevalence rate.
ResultsThe overall HIV prevalence in the adult population was estimated at median value 15.8% with 95% credible interval (C.1) [12.1, 20.0]. The degree of heterogeneity in the site-specific HIV prevalence rate was 0.60 [0.41, 0.91] which
is quite large. Central and southern regions had estimated univariate odds ratios
(OR) 0.80 [0.39, 1.67] and 1.52 [0.78, 2.91]; urban and rural localities 2.54 [1.53,
3.87] and 2.35 [1.34, 3.54]; and syphillis 1.10 [1.10, 1.16]. On multivariate analysis the corresponding ORs estimates were 0.95 [0.57, 1.46] and 1.51 [0.93, 2.42];
2.29 [1.35, 4.05] and 2.06 [1.34, 3.36]; and 1.04 [0.98, 1.10] with the degree of
heterogeneity of site-specific effect estimated at median value 0.32.
Conclusions The results show that there are substantial variation in the HIV
prevalence rates among sites. However, consideration of risk factors particulary
locality partially diminishes the degree of heterogeniety in the site-specific risks.
The results also confirm that urbanization is associated with high HIV rates in
many developing countries.
Presenting author: Samuel Manda, Department of Statistics, University of Auckland, Private Bag 92019, Auckland, New Zealand, Tel.: [phone] x8811,
Fax: [phone], E-mail: [email protected]
WePeC6103 The absence of HIV infection contrasts with high
prevalences of sexual transmitted infections (STI)
in Toliary, Madagascar
S. Xueref1, J. Holianjavony2, R. Daniel3, H. Capart4, M.A. Trabaud5,
P Vanhems1. 1 Laboratoire dEpidemiologie et de Sante Publique, Domaine
Rockefeller, Lyon, France; 2Centre Hospitalier Regional, Toliary, Madagascar;
3Direction Inter Regionale de Developpement Sanitaire, Toliary Madagascar;
4Medecins du Monde, Toliary Madagascar; 5Laboratoire d'Hygiene et de
Virologie, Lyon, France
Background: Madagascar has a low prevalence of HIV infection in the general population (0.15%). Regional data concerning STI are lacking. Objectives
were to assess in the Southwest capital Toliary: seroprevalences of HIV, syphilis,
Chlamydia trachomatis antibodies and hepatitis B virus (HBV) surface antigenemia (AgHBs) and viremia, and their associate risk factors.
Methods: A cross sectional study enrolled all patients consulting 6 dispensaries
during summer 2000 (396 pregnant women [PW] in prenatal care and 121 patients in a STI dispensary [C-STI]). Two different EIA tests were used for HIV1&2
antibodies, another for AgP24 detection. Syphilis screening was based on RPR,
TPHA tests and specific IgM determination and C.trachomatis on specific IgG
quantification. AgHBs and bDNA were quantified. Patients completed a confidential questionnaire. Multiple logistic regression was done to identify independent
factors associated with a positive result of these tests.
Results: No case of HIV infection was detected. Syphilis was screened among
8.2% of PW and 14.0% of C-STI; 4.4% of PW and 4.3% of C-STI had syphilitic
IgM. C.trachomatis IgG were detected among 25.8% of PW and 30.8% of C-STI.
AgHBs was observed among 11.5% of PW and 12.9% of C-STI, HBV viremia
among 2.3% of PW and 5.7% of C-STI.
For PW, low educational status, urban living, being jobless were risk factors
whereas the declaration of absence of condom using was protective for syphilis
and C.trachomatis; main risk factors for HBV infection were high educational status, rural living and age over 16.
Conclusions: This study confirmed the low HIV prevalence in this region in
2000, despite many risk factors. The distribution of HBV differed from syphilis and
C.trachomatis: HBV might be more often contracted during close contact in childhood than later from sexual route. Preventive and curative actions are urgently
needed for syphilitic and HBV infections. Efforts on HIV surveillance should be
strengthened.
Presenting author: serge xueref, paho-who, 525 23rd street nw, 20037 washington, washington dc, United States, Tel.: [phone], Fax: [phone]
10, E-mail: [email protected]