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

476 Abstracts ThPeC7564-ThPeC7567

XIV International AIDS Conference

South African gold miners remain high despite well-implemented control programmes. Determination of the relative contributions of relapse and reinfection
to recurrent TB is important to improving TB control.
Methods: Mineworkers cured of a first episode of TB (negative sputum microscopy at end of treatment) were recruited to a prospective cohort study and followed to detect recurrent TB. All sputum specimens were cultured on LowensteinJensen medium and frozen at -70~C. DNA from all MTB isolates was genotypically classified using probe IS6110 according to the international standardized
protocol. Fingerprints from initial and subsequent episodes of TB were compared
to determine whether recurrence was due to reinfection (isolates genotypically
unrelated) or relapse (isolates identical). HIV status and CD4 count were determined at the time of the initial episode.
Results: The overall recurrence rate among all cases of TB was 8.1/100py. Subsequent analysis is restricted to patients who were culture-positive at both first
and second episode; these patients had a recurrence rate of 4.9/100py (median
follow up 0.88y). The recurrence rate was higher in HIV-infected vs. -uninfected
patients (7.5/100py vs. 3.2/100py, P=0.06) and among the HIV-infected, higher
with decreasing CD4 count (>500/ll:0/100py, 200-500/1: 7/100py, <200/i1:
16/100py; x2 6.33, P=0.042). In a preliminary analysis, paired fingerprints have
been analysed in 17 of 46 recurrences; 14 were HIV-infected (7 relapses, 7 reinfections); 3 HIV-uninfected (all relapses).
Conclusion: Recurrence rates of TB are increased in HIV-infected individuals
and are higher in those who are more immunosuppressed. Both reinfection and
relapse are important mechanisms of recurrent TB in miners in South Africa.
Presenting author: Salome Charalambous, Aurum Health Research, P.O. Box
87, Welkom, 9469, South Africa, Tel.: +27 - 57 - 900-4570, Fax: [phone],
E-mail: salomec @aghs.co.za
ThPeC7564 Does HIV viral load change at the time of an
episode of tuberculosis?
J.H. Day1, A.D. Grant2, K.L. Fielding2, C. Sefuthi1, K.M. De Cock2,
R.E. Chaisson3, R.J. Hayes2, G.J. Churchyard1. 'Aurum Health Research Unit,
Welkom, South Africa; 2LSHTM, London, United Kingdom; 3Johns Hopkins
University, Baltimore, United States
Background: Tuberculosis (TB) is the most important disease among HIVinfected persons world-wide. TB-associated immune activation may enhance HIV
replication and it is proposed that this may accelerate the progression of HIV
disease. HIV viral load (VL) is reported to rise at the time of episodes of acute
infectious disease. Our aim was to determine the effect of a TB episode on VL.
Methods: 1400 consenting HIV-infected mineworkers in South Africa had plasma
samples stored 6-monthly, and at the time of any episode of TB. VL was estimated
on baseline (>3 m pre-TB), episode (within 21d of starting treatment) and final
(>5 m post-TB) samples from individuals who developed TB. Linear regression
was used to model final VL as a function of baseline VL and follow-up time. VL
measured at the time of the TB episode was compared to predicted VL estimated
from the model.
Results: Baseline and final VL values from 17 patients who had TB were used to
develop the model. 11 patients had a VL sample at the time of the TB episode;
their mean age was 42 yrs, median CD4 count 292/[1, mean baseline log (VL)
4.72 (95%CI 4.27, 5.17) copies/ml and median time from baseline to TB episode
32 weeks. Observed mean log (VL) at the time of TB was 4.97 (95%CI 4.52, 5.41)
copies/ml, compared with mean log (VL) predicted by the model of 4.73 (95%CI
4.43, 5.02) copies/ml, difference in log VL of 0.24 (95%Cl: -0.21, 0.69) between
observed and predicted.
Conclusions: The observed VL was higher than VL predicted by the model at
the time of the TB episode, but the difference was not clinically or statistically significant, although confidence intervals were wide. High VL observed at the time of
an episode of TB may be a reflection of high VL prior to the episode, rather than
the effect of the acute TB episode.
Presenting author: Katherine Fielding, Department of Infectious and Tropical
Diseases, London School of Hygiene and Tropical Medicine, Keppel Street, London, United Kingdom, Tel.: [phone], Fax: [phone], E-mail:
katherine.fielding @lshtm.ac.uk
ThPeC7565 HIV/AIDS and Tuberculosis: Study of Trends and
Relative Risks(A Case of Mumbai City in India)
N. Saggurti, G. Rama Rao. India
Background: The present paper aims to study (i) the trends in prevalence of
HIV, HIV associated tuberculosis (TB) morbidity, and (ii) to examine the current
impact of HIV on TB according to different socio-demographic, and behavioural
characteristics.
Methods: To assess the trends of HIV prevalence, the incidence rates among
different populations are modeled together with proportion of populations in each
risk group. The results are extrapolated with age-gender adjustments. Data from

HIV surveillance center, and Tuberculosis Control Programme are used for this
purpose. Further, to obtain the risk of TB among HIV populations and to estimate
the proportion of TB cases that could be attributable to HIV, among both HIV infected and the overall populations, the data from a case-control study was used.
Odds ratio and Attributable risk are the technical tools used to analyse the data.
Results: There is a significant increase in the estimated cases of HIV in the
city over the years. The HIV seroprevalence among tuberculosis patients have in
creased considerably from 2.56 in 1988 to 22.7 by the year 1996. Further, the risk
of developing tuberculosis among HIV positive subjects of age 25-34 years is 4.2
times, while the risk is low for 45+ years with 1.38 times. It is quite evident from
the present study that the proportion of tuberculosis cases that are attributable to
HIV among HIV infected is 68 per cent, while among overall population, it is 17 per
cent. Further a significant difference noticed on the impact of HIV on tuberculosis
with occupation, gender, and behavioural characteristics.
Conclusion: Increase in HIV has shown a profound effect on the increase in tuberculosis morbidity over the period. The serious concerns were felt from the year
1991 in the city, the time exactly the increase in the number of HIV cases are noticed. These results provide statistical evidence to the situation of HIV+TB in the
AIDS capital city of India.
Presenting author: Niranjan Saggurti, Department of Population and Development, International Institute for Population Sciences, Govandi Station Road, Deonar, Mumbai - 400 088, India, Tel.: [phone], Fax: [phone], Email: [email protected]
ThPeC7566 Prevalence of Latent Tuberculosis Infection in
HIV-Infected and Uninfected Sowetan Adults, South
Africa
J. Nachega1, T. Adendorff2, R. Msandiwa2, J. Coetzee2, G. Gray 2, J. Mc Intyre
2, R.E. Chaisson3. 1Johns Hopkins University, Johns Hopkins University, 615N
Wolfe Street, Suite 5515, Baltimore, MD, United States; 2Perinatal HIV Research
Unit, Chris Hani Baragwanath Hopsital, Witwatersrand University Soweto,
Johannesburg, South Africa; 3Johns Hopkins Univesrity, Baltimore, United States
Background: An estimated 1 billion people in the world have been infected with
Mycobacterium tuberculosis. Most infected individuals were exposed as children
and have a reactive tuberculin test (TST) and 5% lifetime risk of tuberculosis
(TB) reactivation. Dual HIV and TB eidemic as in South Africa has dramatically
increase that risk to up 10%/year.
Methods: 305 HIV-infected and 53 HIV-uninfected individuals randomly sampled were enrolled in this cross-sectional study. TST was done by injecting 2
TU (RT23) in the middle third of the left forearm, and the result was read after
48-72 hours. We used the US Centers for Disease Control (CDC) cutoffs recommendation for positive TST (5mm and 10 mm of induration in HIV-infected and
uninfected individuals respectively). Statistical analysis was performed using Epi
Info 2000 software from CDC.
Results: HIV-1 infected individuals were similar to HIV-1 uninfected individuals
in mean age (29~1 vs. 30.1~0.9; P = 0.9) and sex (99% and 98% were female
respectively; P = 0.8). Among HIV-1-seropositive individuals, 150/305 (49%) had
a TST reactivity> 5mm compared with 19/33 (58%) of the seronegative subjects
adults (P< 0.01). The mean diameter of induration among all HIV-1 seropositive
subjects was 8.6 ~ 0.6 mm compared with 10.0 ~ 0.2 mm among all seronegative subjects (P < 0.01). Among HIV-1 infected individuals with TST positive and
negative, the mean CD4+ lymphocytes counts were 534/mm3 [95%CI 451-620]
as compared to 251/mm3 [95%CI 200-330] respectively.
Conclusions:TST remains a useful method for identifying a significant proportion
of HIV-infected individuals with latent TB infection at high risk of TB reactivation.
These individuals should be considered for TB preventive therapy especially in
a setting with limited access to highly active antiretroviral therapy such as South
Africa. In addition, HIV-infected individuals are likely to have false negative TST
with lower CD4+ lymphocytes counts.
Presenting author: Jean Nachega, Johns Hopkins University, 615N Wolfe
Street, Suite 5515, Baltimore, MD, United States, Tel.: [phone], Fax:
[phone], E-mail: [email protected]
ThPeC7567 Treatment of latent Tuberculosis infection in
intravenous drug users co-infected by HIV and M.
Tuberculosis
M. Balague, P. Hernandez2, F S.nchez3, J.L. Ldpez-Colomos3, P. Garca de
Olalla1, J.A. Cayl&., J. Garcia-Vidal4, A. Marco4, R. Guerrero4. 1Public Health
Institute, Barcelona, Spain; 2Toxicology Unit. Hospital del Mar, Barcelona, Spain;
3Infectious Diseases Unit. Hospital del Mar, Barcelona, Spain; 4County Sheriffs
Department, Barcelona, Spain
Background: In Spain, where more than 50.000 persons co-infected by HIV plus
M. tuberculosis are calculated, tuberculosis (TB) still remains the main AIDSrelated opportunistic infection. Up to now there were not studies validating shorttherapies to treat Latent TB Infection (LTBI) in intravenous drug users (IVDU). The
aim of this work was to compare safety and effectiveness of two therapy strategies
for LTBI to avoid active TB among HIV-infected and IVDU people.
Methods: Community assays in HIV-infected and IVDU people if they had a current or previously documented positive tuberculin skin test (> 5 mm), a chest
radiograph without evidence of active TB and no history of previous adequate
treatment for LTBI or hepatic disorder. Treatment arms and strategies are: Di
rectly observed therapy with a) Isoniazide, 5 mg/Kg per day for 9 months (9H); or
b) Rifampin, 10 mg/Kg/day + Pyrazinamide, 25 mg/Kg/day for 2 months (2RZ).
Results: Along the first year of the study, 120 co-infected individuals were included (37% of those evaluated), 59 in 9H and 61 in 2RZ. The male/female ratio
was 4/1. The distribution by age (37 ~ 6 year old), the median of CD4+ (687
cell/mL) and the viral load of the HIV by PCR (1.9 log) were similar in both treatment arms. Three patients from 2RZ and 14 from 9H gave up (p=0.03). Four