Bridging the Gap: Conference Record [Abstract book, International Conference on AIDS (12th: 1998: Geneva, Switzerland)]

620 Abstracts 33142-33146

12th World AIDS Conference

be developed to communicate that this product would not be a substitute for
condom use.
Lessons learned: Any microbicide will need to be available in a variety of
formulations so that it would be acceptable where perceptions of pleasurable sex
differ. The men in this study were supportive of the concept of a microbicide, in
general, however the farmers seemed to be more guarded in their enthusiasm.
Men's and women's views should be included in product development efforts in
order to ensure that potential products will be acceptable.
33142 Nonoxynol-9 in preventing heterosexual transmission of
HIV - Using multivariate methods in reanalyzing previous
and planning future studies
Knut M. Wittkowski1, E. Susser2, K. Dietz3. 1NY ACAD Med-Ctr Urban EPID
Studies 1216 5th Ave, New York, NY 10029; 2HIV-Center for Clinical & Beh.
Studies New York NY, USA; 3Eberhard-Karls-Univ Dept. Med. Biometry
Tuebingen, Germany
Objectives: Accounting simultaneously for a variety of barrier methods and risk
factors is paramount in addressing the efficacy of spermicides for the control of
heterosexual HIV transmission. Here we demonstrate how the choice of statistical
methods for multivariate data may explain seemingly contradictory results and
lack of power in previous analyses of studies on the efficacy of nonoxynol-9 in
HIV prevention.
Methods: To assess the efficacy of spermicides and condoms, used separately
and together, we applied three different models: linear (Pearl-index), exponential (maximum likelihood, ML), and monotonic (a recently developed marginal
likelihood for multivariate ordinal data, MrgL). To assess power, required sample
size for efficacy studies is computed accounting for (1) over-reporting of barrier
method use, (2) dependent method failure rates, (3) partially effective "placebos",
and (4) non-linear increase of cumulative incidence.
Results: In our reanalysis of coital log data from a longitudinal study of 224
Cameroonian prostitutes, using both barrier methods together, is seen less efficacious than expected under a multiplicative failure model. N9 and condoms (with
or without N9) are 100% (C/95 = [43%; 100%]) and 92% (C/95 = [79%; 100%])
effective, respectively. Confounding or choice of specific model assumptions were
unlikely to account for these results.
Discussion: Our results are compatible with previous estimates based on
contraception studies and generally compatible with previous reports on analyses
of these same data. We were able to explain peculiarities in these reports by
virtue of the statistical methods employed. Lack of power in several studies may
be attributed to overly optimistic assumptions made in sample size computations.
Conclusions: This reanalysis highlights the importance of choosing appropriate statistical methods both in sample size computations and analysis of both
observational studies and clinical trials. The results explain lack of power in some
of these studies and support the view that spermicides are efficacious in reducing
the risk of heterosexual transmission of HIV.
[1] (1989) AIFO 4: 3; (1988/92) JASA 83: 1163/87: 258; (1995) AIDS 9: 310; (1998) AJPH 88:
in press
133143 Nonoxynol-9 enhances rectal infection by herpes simplex
virus in mice
David Phillips, V.R. Zacharopoulos. Population Council, 1230 York Ave, New
York, NY USA
Background: Although N9-containing products are commonly used during rectal
intercourse, the effect of N9 on rectal infection has been given little attention. We
used mice to study the effect of N9 and OTC N9 products on the likelihood that
mice would become infected by herpes simplex following rectal inoculation.
Methods: Mice were starved for 24 h, and the rectum was rinsed with 50 tL
PBS. Twenty /IL of the test formulation was introduced into the rectum using
a P20 Pipetman. Five min later, mice were inoculated intrarectally with 10/1L
of herpes simplex virus 2 (HSV-2) in MEM cell culture medium. Infection was
determined by inflammation and lesions in the rectal area.
Results: Rectal inoculation of mice resulted in dose dependent infection. At
a dose of virus that resulted in only 5% infection of control (PBS treated) mice,
90 to 100% of the animals which had been pretreated rectally with 2% N9, or
the OTC vaginal spermicides Gynol II, or K-Y PLUS became infected. The N9
containing lubricant Forplay also significantly enhanced infection. On the other
hand a 2% formulation of the sulfated polysaccharide carrageenan protected all
of the animals from HSV-2 infection. Examination of rectal tissue from animals
10 min following rectal treatment with 2% N9 or Gynol II with light and electron
microscopy revealed that N9 had resulted in the loss of epithelial cells from the
basement membrane. Epithelial cells were found scattered over the surface of
the connective tissue or free in the rectal lumen.
Conclusion: N9-containing products placed in the rectum of mice prior to inoculation with HSV-2 increase the probability of infection by destroying the rectal
epithelium. This suggests that the use of N9-containing products rectally may
increase the risk of infection by HSV-2 or other sexually transmitted pathogens.

33144 Screening and selection of active anti-HIV agents for the
development of vaginal microbicides
Lee E. Calypool1, G.F. Doncel2, L.B. Allen3, K.A. Keith3, A.D. Brazier3,
R.W. Buckheit, Jr.3. 11611 N. Kent Street, Suite 806 Arlington, VA; 2CONDRAD
Program, Norfolk, VA; 3Southern Research Institute Birmingham AL, USA
Background: The development of vaginal products which protect against the
sexual transmission of HIV/AIDS and other STDs, and can be used by women at
their own discretion, is urgently needed.
Problem: Potentially active agents to be used in microbicidal products need to
be screened for anti-HIV activity and prioritized for further testing and development.
Methods: A comprehensive testing algorithm was developed to evaluate the
activity of potentially active agents against 1) cell-free lymphocytotropic HIV strain
RF, 2) cell (H-9) -associated lymphocytotropic HIV strain RF, 3) binding of virus to
target cells using lymphocytotropic HIV strain IIIB binding to MT-2 lymphoid cells,
4) binding of virus to target cells using macrophage-tropic HIV strain BaL or ADA
binding to MAGI-CCR-5 cells, and 5) cell-to-cell transmission of virus using chronically infected H9 cells (HIV strain SK1) with ME180 cervical epithelioid target
cells. All agents are first tested for virucidal activity against cell-free HIV. Agents
which are positive are then tested for virucidal activity against cell-associated
HIV. Agents which are negative for activity against cell-free virus are then tested
for antiviral activity in the virus binding and cell-to-cell transmission assays.
Results: A broad-spectrum but well-defined testing algorithm has been developed and adopted for the testing of potentially active microbicidal agents. Several
agents have been assessed. Nonoxynol-9, the positive control for virucidal activity, reduced the titer of the viral stock by a mean of more than 4 log units at 0.032%
(n = 6), and dextran sulfate (MW 40 Kd), the positive control for inhibition of viral
binding to lymphoid cells, reduced binding by 50% at 0.000022% (n = 7). Other
agents which are strongly positive for virucidal or antiviral activity are prioritized
for further development depending on their cytotoxicity, potential for inflamation
and irritation, and activity against other STD pathogens. Agents which are negative for virucidal and antiviral activity are eliminated from further consideration.
(Supported by NICHD and USAID through the CONRAD Program.)
Conclusions: A rigorous screening algorithm to identify and prioritize potentially active agents with virucidal or antiviral activity against both lymphocytotropic
and macrophage-tropic strains of HIV is an essential and useful component in
the development of vaginal microbicides.
33145 1 Safety of multiple daily applications of COL-1492, a
nonoxynol-9 vaginal gel, among female sex workers
Lut Van Damme1, V. Chandeying2, G. Ramjee3, H. Rees4, P. Sirivongrangson5,
M. Laga1, J.H. Perriens6. 'Institute of Tropical Medicine Nationalestraat 155
2000, Antwerp, Belgium; 2Prince of Songka University Hat Yai Songkla5 Ministry
of Public Health VD Division, Bangkok, Thailand; 3Centre for Epidemiological
Research (CERSA), Durban;4 Reproductive Health Research Unit,
Johannesburg, Rep. of South Africa; 6UNAIDS, Geneva, Switzerland
Objectives: To assess the safety of multiple daily applications of COL-1492, a
bio-adhesive vaginal gel containing 52.5 mg nonoxynol-9 per application.
Methods: The reported study is part of a phase II/11I trial assessing the effectiveness of COL-1492 in preventing HIV infection in women. HIV seronegative
female sex workers were invited to participate in this randomised, double-blind,,
placebo-controlled trial in Bangkok, Hat Yai, Durban and Johannesburg. Participants were asked to insert the product every morning and before each act of
sexual intercourse if they had cleaned their vagina. They were instructed to complete a coital log chart (data on sexual acts, condom and product use). Visits were
scheduled on days 0, 14, 28 and thereafter every month. Pelvic examinations,
including colposcopy following the WHO manual, were performed. Blood was
taken for HIV and syphilis serology and samples for the detection of C. albicans,
T Vaginalis, N. gonorrhoeae and C. trachomatis were obtained.
Results: Only blinded results on colposcopy will be presented since the other
outcomes (visual pelvic examination, STD assessment and acceptability) are part
of the phase III study endpoints. One hundred and fifty women were randomised
(Bangkok 47, Hat Yai 53, Durban 44, Johannesburg 6) to receive either COL-1492
or its matching placebo. Preliminary analysis showed that the frequency of lesions with an epithelial disruption, i.e. ulcer and abrasion, was low: 2 ulcers and
5 abrasions in group A, 1 and 6 respectively in group B. In survival analysis there
was no difference in the probability of any colposcopy lesion (with or without
epithelial disruption) between the two groups (p = 0.99).
Conclusions: The multiple daily use of COL-1492 in a population of female
sex workers was found to be safe. A phase III trial, without colposcopy, assessing
the effect on HIV and STD transmission, is now ongoing in four centres (Hat Yai,
Durban, Cotonou and Abidjan).
|33146 PRO 2000 gel, a potential topical microbicide for HIV
prevention, can block infection by other sexually
transmitted disease pathogens

Albert T. Profy1, A.J. Sonderfan1, D.I. Bernstein2, T. Chancellor', E.R. Kern3
L.R. Stanberry1. 1 Proncept Inc 840 Memorial Drive Cambridge MA 02139;
2Children's Hospital Medical Center, Cincinnati, OH; 3University of Alabama at
Birmingham, Birmingham, AL, USA
Background: PRO 2000 is a 5-kD naphthalene sulfonate polymer that shows
broad activity against HIV-1 in vitro. The compound is being developed as a top