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

XIV International AIDS Conference

Abstracts ThPeC7559-ThPeC7563 475

ThPeC7559 Mycobaterial infection in PLWA of an urban
hospital clinic of Guatemala city
B. Samayoa Herrera1, T. Velasquez-Porta1, D. Lau', S. Lickez', E. Arathoon2
1DIGI-Universidad de San Carlos, la avenida 11-19 zona 1, Guatemala, Ciudad,
Guatemala; 2Asociacion de Salud Integral, Guatemala, Guatemala
Background: To determine the prevalence and associated factors of mycobacterial infections in people living with HIV/AIDS (PLWA).
Methods: A consecutive sample of 237 PLWA that attended to Clinica Familiar
"Luis Angel Garcia", from February to November 2000. The laboratory diagnosis was performed by direct observation of smears stained with Ziehl Neelsen
method, cultivation in Lowenstein Jensen, and presumptive identification by niacin
test. The analysis of data were effected by means of frequencies and a bivariate
analysis using a Chi-squared test, with the statistic program EPI Info 6.0, at a
significance level of 0.05.
Results: The general prevalence of mycobacterial infection was 11.6% with direct observation of the smears, 18.2% by cultivation of the samples, and 21.6%
by the combination of both techniques; 41 strains (84%) were identified as Mycobacterium tuberculosis and 4 strains (8%) as Mycobacterium sp.; analphabet or uncompleted elementary school time spend in prison, (OR=2.64; IC 95%=
0.96-5.72; OR=2.43; IC95%= 0.94-6.21; p=0.03 respectively) and masculine gender (OR=2.17; IC 95%= 0.92-5.72; p=0.06) were not significant factors, but we
considered important to point out. Current empiric antimycobacterial treatment
(OR=2.17; IC95%= 0.93-6.13; p=0.04) and normal X-ray results (OR=0.49; 95%=
0.21-1.12; p=0.06) were also found as factors associated to mycobacterial infection in these patients. Other demographic or clinical factors were not statistically
significant inthis sample (p > 0.05).
Conclusions: The mycobacterial diagnosis in PLWA should be performed with
direct observation of the smears combine with cultivation of the samples. The
prevalence of tuberculosis obtained in this study reflects a problem that demands
immediate actions and improvements in the mycobacterial diagnostic services for
PLWA, in hospitalary centers in developing countries
Presenting author: Blanca Samayoa, la avenida 11-19 zona 1, Guatemala, Ciudad, Guatemala, Tel.: [phone], Fax: [phone], E-mail: cflag@amigo.
net.gt
ThPeC7560 Low rates of tuberculin skin testing among persons
with newly diagnosed HIV infection, US, 1995-1997
L.M. Lee1, S.E. Buskin2, A. Morse3, S. Costa1, J. Beil4, M. Lobato'. 'Centers
for Disease Control and Prevention, CDC, 1600 Clifton Rd., NE, MS E-47, Atlanta
GA 30333, United States; 2Public Health- Seattle and King County Seattle,
United States; 3Louisiana Department of Health, New Orleans, United States;
4New Jersey Department of Health and Senior Services, Trenton, United States
Background: HIV-infected persons are 100 times more likely to develop active tuberculosis (TB) compared with uninfected persons if latent TB infection is present
and prophylaxis is absent. In 1995, recommendations stated that persons with
newly diagnosed HIV should undergo a tuberculin skin test (TST) following HIV
diagnosis and annually thereafter. Our objective was to assess missed opportunities for prevention of TB, including a missed TST.
Methods: We selected a population-based random sample of 1093 persons diagnosed and reported with HIV from June 1995 through June 1997 in Seattle/King
County WA, New Orleans LA, and Jersey City NJ. We excluded persons who died
within 6 months of diagnosis because of insufficient follow-up time. To determine
the proportion of persons receiving a TST following HIV diagnosis, we conducted
medical record reviews at the reporting provider, all other known providers within
each jurisdiction, and the local TB control office.
Results: Of completed reviews on 869 (80%) persons, 73% were male; 50%
were black, 28% white, 17% Hispanic, 5% other/unknown; 76% were US born,
10% foreign born, 14% unknown. Excluding 21 active TB cases and 7 previous
TST positives, 52% had a TST following HIV diagnosis; 83.3% of those within 1
year. Median time between HIV diagnosis and TST was 1 month, mean (standard deviation) was 5.7 (8.9) months. There was no difference in the proportion
with TST by sex (50.9% males, 49.4% females, p=0.69) or race (56.4% white,
48.9% black, 44.4% Hispanic, 51.3% other/unknown, p=0.11). Among persons
with known birth country (n=758), there was no difference in proportion with TST
between US and non-US born (53.0% and 53.4%, respectively). Of 437 persons
with TST, 33 (7.5%) were positive.
Conclusions: Half of people newly diagnosed with HIV had a TST following HIV
diagnosis, with little variation by demographic group. This signals a need to improve TB prevention beginning with better coverage of TST screening.
Presenting author: Lisa Lee, CDC, 1600 Clifton Rd., NE, MS E-47, Atlanta GA
30333, United States, Tel.: [phone], Fax: [phone], E-mail: Imlee
@cdc.gov

ThPeC7561 The relationship between HIV in TB patients and
HIV in the general population - a tool for estimating
HIV prevalence
J.L. Wennberg. Harlem Hospital Center-Columbia University 501 W 123rd St.
#19H, New York, NY 10027, United States
Background: Many papers present data on TB as the presenting illness for HIV,
but no systematic examination of this data has been done to estimate HIV preva
lence. This paper proposes to estimate HIV prevalence in the general population
from the HIV prevalence in TB patients and the incidence of TB in that population.
Methods: A mathematical model was developed. It assumed that the prevalence
rate of Latent Tuberculosis Infection (LTBI) was the same for both HIV infected
and uninfected persons. A literature search yielded data that described the prevalence of HIV in TB patients (HIVTB). This was combined with data on HIV in the
general population (HIVGP) from UNAIDS and data on TB incidence (TBI) from
WHO. The ratio of HIVTB/HIVGP was calculated and plotted vs. TBI. A least
squares trend-line was determined.
Results: The model gives HIVTB/HIVGP = TBI-1 x LTBI x Constant. The least
square trend-line for the graph is HIVTB/HIVGP = 264.89 TB-0.8864. In a practical
example, the graph was used to estimate the HIV prevalence in Kosovo from data
of HIV in TB patients measured in an inpatient sample.

incidence of TB per 100,000
TB incidence vs ratio of HIV prevalence in TB patients to area of HIV prevalence.
Conclusions: There is good agreement between the mathematical model and
historical experimental data for the relationship between HIV prevalence in TB
patients and the general population. While further confirmation is required, this
can be an effective tool for estimating HIV prevalence for healthcare resource
allocation particularly in resource poor areas.
Presenting author: Jens L. Wennberg, 501 W. 123rd St. #19H, New York, NY
10027, United States, Tel.: [phone], E-mail: [email protected]
ThPeC7562I Characteristics, needs, and HIV and TB knowledge
among HIV-infected and HIV uninfected patients
undergoing treatment for tuberculosis (TB) in
Harlem, NY
Y. Hirsh-Moverman, W.M. EI-Sadr, P. Colson. Harlem Hospital/Columbia
University Infectious Disease Division, Room 3101A, Harlem Hospital, 506
Lenox Avenue, New York, N.y 10037, United States
Background: Co-infection with HIV is a major issue in the management of TB disease. However, the characteristics, needs, and TB/HIV knowledge of HIV-infected
patients (pts) have not been well defined in contrast to those among pts without
HIV infection who are undergoing TB treatment.
Methods: We assessed pts characteristics, and TB and HIV knowledge in pts
started on TB treatment at Harlem Hospital and evaluated effect of these characteristics on completion of TB treatment.
Results: Of 112 patients with TB, 38% were HIV-infected, 66% men, 93% AfricanAmerican or Latino, with mean age 41 years, and 14% were married. Pts with
HIV infection were less likely to be employed (2.4% vs.15.7%; p=0.029) and to
be foreign-born (19.0% vs. 35.7%; p=0.097) and more likely to have a history of
homelessness (57.1% vs. 35.7%; p=0.044), to report current drug use (47.6%
vs. 25.7%; p=0.030) and to be participants in drug treatment programs(38.1% vs.
5.7%; p<0.001). Of 24 TB knowledge questions, 28.6% of HIV infected and 4.4%
of HIV uninfected pts answered correctly >75% of the questions (p=0.042). Of
12 questions regarding HIV knowledge, 69.0% of HIV-infected and 44.3% of HIVuninfected pts answered correctly >75% of the questions (p=0.019). Of confirmed
TB cases, 93.1% of HIV-infected pts and 96.8% of pts without HIV completed TB
treatment (p=1.000).
Conclusions: HIV-infected pts with TB face significant barriers to completion of
TB treatment including unemployment, homelessness, and drug use. However,
they have higher knowledge level of TB and HIV than pts without HIV infection.
Inspite of the barriers identified, TB treatment completion was similar among HIVinfected and uninfected pts.
Presenting author: Wafaa EI-Sadr, Infectious Disease Division, Room 3101A,
Harlem Hospital, 506 Lenox Avenue, New York, N.y. 10037, United States, Tel.:
[phone], Fax: [phone], E-mail: wmel @columbia.edu
ThPeC7563 Recurrent tuberculosis in gold miners in South
Africa: relapse or re-infection?
S. Charalambous', A.D. Grant2, R. Warren3, V.M. Moloil, D. Sehloho', P. van
Helden4, R.E. Chaisson5, K.M. De Cock2, G.J. Churchyard1. 'Aurum Health
Research, Aurum Health Research, PO. Box 87, Welkom, 9469, South Africa;
2London School of Hygiene and Tropical Medicine, London, United Kingdom;
3University of Stellenbosch Medical School, Tygerberg, South Africa; 4 University
of Stellenbosch Medical School, Welkom, South Africa; 5Johns Hopkins
University Baltimore, United States
Background: Tuberculosis (TB) is the most important cause of morbidity in HIVinfected persons worldwide. The incidence of TB, and recurrence of TB, among