See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/5757017
SAVVY (C31G) gel for prevention of HIV infection in women: A Phase III,
double-blind, randomized, placebo-controlled trial in Ghana
Article  in  PLoS ONE · February 2007
DOI: 10.1371/journal.pone.0001312 · Source: PubMed
CITATIONS READS
144 35
11 authors, including:
Kavita Nanda Baafuor Opoku
FHI 360 Kwame Nkrumah University Of Science and Technology
108 PUBLICATIONS   4,740 CITATIONS    41 PUBLICATIONS   315 CITATIONS   
SEE PROFILE SEE PROFILE
William K Ampofo Wes Rountree
Noguchi Memorial Institute for Medical Research Duke University Medical Center
134 PUBLICATIONS   1,356 CITATIONS    46 PUBLICATIONS   860 CITATIONS   
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Most at Risk Population Study View project
Longitudinal Analysis of EQA/PT for Flow Cytometry Lymphocyte Subset Analysis View project
All content following this page was uploaded by Baafuor Opoku on 21 May 2014.
The user has requested enhancement of the downloaded file.
SAVVYH (C31G) Gel for Prevention of HIV infection in
Women: A Phase 3, Double-Blind, Randomized, Placebo-
Controlled Trial in Ghana
Leigh Peterson1, Kavita Nanda1*, Baafuor Kofi Opoku2, William Kwabena Ampofo3, Margaret Owusu-Amoako3, Andrew Yiadom Boakye2, Wes
Rountree1, Amanda Troxler1, Rosalie Dominik1, Ronald Roddy4, Laneta Dorflinger1
1 Family Health International, Durham, North Carolina, United States, 2 Komfo Anokye Teaching Hospital, Kwame Nkrumah University of Science &
Technology, Kumasi, Ghana, 3 Noguchi Memorial Institute for Medical Research, University of Ghana, Accra, Ghana, 4 Duke Clinical Research Institute,
Durham, North Carolina, United States
Objective. The objective of this trial was to determine the effectiveness of 1.0% C31G (SAVVY) in preventing male-to-female
vaginal transmission of HIV infection among women at high risk. Methodology/Principal Findings. This was a Phase 3,
double-blind, randomized, placebo-controlled trial. Participants made up to 12 monthly visits for HIV testing, adverse event
reporting, and study product supply. The study was conducted between March 2004 and February 2006 in Accra and Kumasi,
Ghana. We enrolled 2142 HIV-negative women at high risk of HIV infection, and randomized them to SAVVY or placebo gel.
Main outcome measures were the incidence of HIV-1 and HIV-2 infection as determined by detection of HIV antibodies from
oral mucosal transudate specimens and adverse events. We accrued 790 person-years of follow-up in the SAVVY group and 772
person-years in the placebo group. No clinically significant differences in the overall frequency of adverse events, abnormal
pelvic examination findings, or abnormal laboratory results were seen between treatment groups. However, more participants
in the SAVVY group reported reproductive tract adverse events than in the placebo group (13.0% versus 9.4%). Seventeen HIV
seroconversions occurred; eight in participants randomized to SAVVY and nine in participants receiving placebo. The Kaplan-
Meier estimates of the cumulative probability of HIV infection through 12 months were 0.010 in the SAVVY group and 0.011 in
the placebo group (p = 0.731), with a hazard ratio (SAVVY versus placebo) of 0.88 (95% confidence interval 0.33, 2.27). Because
of a lower-than-expected HIV incidence, we were unable to achieve the required number of HIV infections (66) to obtain the
desired study power. Conclusions/Significance. SAVVY was not associated with increased adverse events overall, but was
associated with higher reporting of reproductive adverse events. Our data are insufficient to conclude whether SAVVY is
effective at preventing HIV infection relative to placebo. Trial Registration. ClinicalTrials.gov NCT00129532
Citation: Peterson L, Nanda K, Opoku BK, Ampofo WK, Owusu-Amoako M, et al (2007) SAVVYH (C31G) Gel for Prevention of HIV infection in Women: A
Phase 3, Double-Blind, Randomized, Placebo-Controlled Trial in Ghana. PLoS ONE 2(12): e1312. doi:10.1371/journal.pone.0001312
INTRODUCTION METHODS
Topical microbicides are designed to inhibit the sexual transmis- The protocol for this trial and supporting CONSORT checklist
sion of Human Immunodeficiency Virus (HIV) and other disease are available as supporting information; see Checklist S1 and
pathogens. They offer a female-controlled prophylactic option in Protocol S1.
situations where male condom use cannot be negotiated. Marketed
chemical spermicides, which show some activity against sexually Participants
transmitted pathogens in vitro, have been evaluated as topical
We conducted this randomized, double-blind, placebo-controlled
microbicides. However, no clinical studies have yet demonstrated
trial between January 2004 and March 2006 in Accra and
that these products can prevent HIV infection, and spermicides
Kumasi, Ghana. We enrolled HIV-antibody negative, non-
with the surfactant nonoxynol-9 (N-9) have caused mucosal
pregnant women 18 to 35 years old who were at risk of HIV
erosion and ulceration, which may increase the risk of HIV
transmission [2–4].
Several in vitro studies [5] have suggested that C31G (SAVVYH,
Cellegy Pharmaceuticals [formerly BIOSYN, Inc.], Huntington Academic Editor: Keymanthri Moodley, University of Stellenbosch, South Africa
Valley, Pennsylvania) has a potent effect on enveloped HIV by Received September 5, 2007; Accepted September 6, 2007; Published December
disrupting the outer membrane. However, because its mechanism of 19, 2007
action is similar to N-9 [6;7], some have raised concerns about the
Copyright: 
 2007 Peterson et al. This is an open-access article distributed
safety of SAVVY. In pre-clinical studies, SAVVY demonstrated under the terms of the Creative Commons Attribution License, which permits
minimal toxicity at 0.001% concentration, as measured by the cell unrestricted use, distribution, and reproduction in any medium, provided the
proliferation assay, and no toxicity to mammalian cells at 0.0003%, original author and source are credited.
as measured by the MTT assay [5]. In addition, 5 Phase 1 clinical Funding: This study was supported by funds from the United States Agency for
studies, including a total of 121 women and 24 men exposed to International Development (USAID) under Cooperative Agreement CCP-A-00-95-
SAVVY, have been conducted using three concentrations of 00022-02 and GPO-A-00-05-00022-00. Cellegy Pharmaceuticals (formerly BIOSYN,
Inc.) (Huntington Valley, Pennsylvania) donated the investigational product
SAVVY (0.5%, 1.0%, 1.7%). No serious or severe adverse events (SAVVY and placebo-gel applicators) evaluated in this study.
related to use of 1.0% SAVVY gel were reported during these studies
[8–10]. We investigated the safety and effectiveness of 1.0% SAVVY Competing Interests: The authors have declared that no competing interests
exist.
in preventing HIV infection in a population of young, sexually active
women at high risk for acquiring HIV infection. * To whom correspondence should be addressed. E-mail: KNanda@fhi.org
PLoS ONE | www.plosone.org 1 December 2007 | Issue 12 | e1312
SAVVY Gel for HIV Prevention
infection because of having a mean of three or more coital acts per informed consent form stated that: ‘‘We do not know the effects
week and two or more sexual partners in the 3 months before and safety of the gel during pregnancy. Pregnant women may not
screening. Study participants were recruited from areas within join this study. If you become pregnant during the study you
each city that were considered high HIV transmission areas, should tell your study doctor or nurse right away. Your study gel
including markets, bars, and hotels. Although we did not will be stopped and the study doctor will discuss your choices with
specifically ask as part of the clinical trial procedures if the you.’’
participants were sex workers, most exchanged sex for money. At each monthly follow-up visit, participants underwent OMT
Special ethical considerations were taken into account because of HIV and pregnancy testing, AE assessment, STI risk reduction
the potential vulnerability of this population. We included women counseling, and study product and condom re-supply. Participants
who agreed to: use study gel as directed and follow study responded to structured questionnaires on their interval sexual
procedures; report self-medication with antibiotics during study behavior (including coital activity, and gel and condom use),
participation; and avoid use spermicides or other vaginal experience using the gel, and were reminded of study concepts
contraceptives or lubricants during the study. We excluded discussed during the informed consent process. If clinically
women who: were intending pregnancy; had a history of latex indicated, participants underwent physical examination and STI
allergy; were injection drug users; or had gynecological conditions testing. Study staff documented whether product use was
that could affect the safety or effectiveness of the study gel. interrupted temporarily or permanently for any of the following
reasons: participant ran out of supplies, investigator withdrew
Ethics study product in the interest of the safety and well being of the
participant, positive pregnancy test result, or confirmed HIV
We developed strategies to protect the confidentiality and
infection. Pregnant women were allowed to resume study product
autonomy of the participants, increase/ensure comprehension of
use after pregnancies had ended. Study staff referred participants
the informed consent and research methods, and promote access
infected with HIV during the study to local HIV-related
to resources and services during and after the trial. The study
psychological, social, and medical services (such as viral load,
protocol and informed consent forms were approved by 1) the
CD4 level, and HIV resistance testing) as well as antiretroviral
Committee on Human Research, Publications and Ethics, School
drug therapy when needed. If a participant missed a scheduled
of Medical Sciences, University of Science & Technology, Kumasi,
follow-up appointment, study staff made up to 3 attempts to
Ghana, 2) Noguchi Memorial Institute for Medical Research IRB,
contact that participant and reschedule the appointment (ideally to
University of Ghana, Legon, Ghana, and 3) the Protection of
occur within 2 weeks of the original appointment). If the
Human Subjects Committee, Family Health International, USA.
participant could not be located after 3 attempts, no further
All participants provided written informed consent in their
efforts were made to find her, but her file remained open until
preferred language. Illiterate participants were read the informed
study closeout. If the participant did not return to the study before
consent forms verbatim in the presence of a witness, and provided
the study was closed, she was considered lost to follow-up. The
a mark or thumbprint in lieu of signature.
‘‘lost to follow-up’’ designation was not made for any participant
until the closing date of the study.
Interventions
During recruitment, study staff explained the general purpose of Objective
the study and the eligibility requirements. If eligible, women were
The objective of this trial was to determine the effectiveness of
referred to one of two study clinics in Kumasi or Accra. At the
1.0% C31G in preventing male-to-female vaginal transmission of
screening visit, women were interviewed to confirm understanding
HIV infection among women at high risk.
and willingness to comply with study requirements, completed
written informed consent, received HIV pretest and condom
counseling, and underwent oral mucosal transudate (OMT) rapid Outcomes
HIV testing. All participants received HIV post-test counseling, The primary measure of effectiveness was infection with HIV-1 or
physical and pelvic examinations (including vaginal wet mount to HIV-2, measured by detecting antibodies in oral mucosal
support the diagnosis for bacterial vaginosis, trichomoniasis, or transudate (OMT) (OraQuickH ADVANCE Rapid HIV-1/2
vaginal candidiasis), urine pregnancy tests, and STI (syphilis, Antibody Test, Orasure Technologies) and confirmed by an
TM
gonorrhea, and chlamydia) tests. Women with reactive OMT enzyme-linked immunosorbent assay (ELISA) (Genetic Systems
rapid HIV tests received ELISA to confirm HIV status. We asked HIV-1/HIV-2 Plus O ELISA from BioRad) and/or Western Blot
TM
potential participants to return 4 weeks after their screening visit (Genetic Systems HIV-1 Western Blot, BioRad) from a finger
to receive the results of their STI and confirmatory HIV tests, if prick or serum specimen. We evaluated safety by comparing the
applicable. incidence of adverse events (AEs) including pelvic exam findings
At this second visit, participants received a detailed explanation and sexually transmitted infections (STIs).
of study procedures, signed or marked a consent form for
enrollment, received HIV counseling, provided urine for preg- Sample Size
nancy testing, provided another OMT sample for HIV testing, We estimated that a sample size of 2142 participants (1,071 in
and if eligible were randomized to receive either SAVVY or each treatment group) would give us 80% power to detect a 50%
placebo. Study staff gave each eligible participant her first month’s difference in the HIV infection rate (two-sided log-rank test,
supply of study product after she had been randomized. Clinic a = 0.05 significance level) between the two groups. We assumed
staff counseled participants to use the gel vaginally before each act the rate of HIV infection in the control group to be 5/100 person-
of sexual intercourse (and to insert a second dose if more than one years and loss to follow-up to be 20%; approximately 66 total HIV
hour had elapsed between the first application and sexual infections were needed to achieve the desired power. Our protocol
intercourse), emphasized that the gel had unknown effectiveness included plans for assessing (in a blinded manner) whether
for preventing HIV, distributed condoms, and counseled partic- additional participants would be needed to observe the required
ipants to use condoms for all sexual contacts with all partners. The 66 events.
PLoS ONE | www.plosone.org 2 December 2007 | Issue 12 | e1312
SAVVY Gel for HIV Prevention
Randomization and Blinding ments, would review AEs and primary safety and HIV
We randomized enrolled participants into either the SAVVY or seroconversion data twice, after approximately 16 and 33 events,
respectively. However, testing for early evidence of effectiveness
placebo arm using a 1:1 allocation ratio. A statistician not
was scheduled only to occur at the second of these two planned
otherwise involved with the study developed the allocation
looks and later when the target total number of events (66) was
sequence using a computer random number generator and
obtained (i.e., the first look at the HIV seroconversion data was to
randomly varied permuted-blocks of 12, 18, and 24. Six label
review the data for signs of harm) .
colors (three SAVVY and three placebo) were used to differentiate
Monitors (trained in Good Clinical Practice) visited sites regularly
the otherwise identically packaged gels. Randomization was
to review study eligibility, informed consent, protocol compliance,
stratified by study site. We used sequentially numbered, sealed
laboratory procedures, source documents, product accountability,
opaque envelopes to assign participants to one of six color groups
and AEs. We attempted to get original hospital records, when
after they had qualified for the study and signed the consent form.
available, for serious adverse events (SAEs) and deaths.
The randomization envelopes were maintained in a secure office
and were not available to study staff until the moment of
randomization. Each randomization envelope was used only once. RESULTS
Participants, field study staff, monitors, statisticians, and other FHI Recruitment, Participant Flow, Numbers Analyzed
staff involved in the trial were not aware of which gel colors were
and Baseline Characteristics
associated with SAVVY or placebo. Both SAVVY and placebo
Participants were recruited and enrolled into this study for
gels were clear, with similar viscosity and pH, dispensed in 3.5 mL
15 months, starting February 2004. A total of 3,490 women were
doses with identical applicators.
screened, from which we enrolled 2,142 participants into the study
The placebo gel was formulated to minimize any possible
(Figure 1). The primary reasons potential participants were not
effects—negative or positive—on study endpoints. It was isotonic
enrolled were because they failed to return for enrollment, were
to avoid epithelial cell swelling or dehydration, and formulated at a
HIV-infected at baseline, or were pregnant. Overall loss to follow-
pH of 4.4 but with minimal buffering capacity. When mixed with
up after enrollment was approximately 15% (n = 310). The loss-to-
an equal volume of semen, the placebo gel induced only a trivial
follow-up rate was highest during the first month of study
decrease in semen pH (from 7.8 to 7.7). The placebo gel contained
participation, during which 40% of the overall loss to follow-up
hydroxyethylcellulose as a gelling agent, and its viscosity was
(123 of 310 participants) occurred. Among these were 103
comparable to that of SAVVY. Hydroxyethylcellulose does not
participants who never returned after their enrollment visit and
have anti-HIV properties. The gel also contained sorbic acid as a
were therefore excluded from the primary safety analyses.
preservative; sorbic acid has no anti-HIV activity and is readily
The first scheduled DMC safety review was repeated because
metabolized by human cells.
errors were found in the initial HIV seroconversion data due to
false positive ELISA testing. At the third DMC safety review
Statistical Methods meeting, when there were 16 seroconversions in the database, we
For the primary effectiveness analysis we compared the probability informed the DMC that the study statistician estimated that
of HIV infection for the SAVVY and placebo gel groups using a two- approximately 3,500 additional participants (beyond the 2,142
sided site-stratified, exact log-rank test (StatXact). We calculated planned sample size) would be needed to achieve the required
Kaplan-Meier estimates of HIV infection probabilities by treatment number of HIV infections (66) to obtain the desired study power.
group, pooled across sites. We used a proportional hazards In light of this formidable challenge, we requested that the DMC
regression model to estimate the hazard ratio, along with a 95% review the available outcome data by group (along with results of
confidence interval, comparing the SAVVY group to the placebo conditional power calculations under a range of assumptions about
group for HIV incidence, controlling for site. Each HIV infection the true effect size) and provide a recommendation as to whether
onset date was estimated as the midpoint between the date of the first the trial should continue or be terminated early. The DMC
positive test result and the previous, negative test date. A right recommended that the trial be stopped; we thus decided to
censoring time of 380 days was applied. Because the trial was prematurely terminate the study in November 2005.
terminated well before reaching the number of HIV infections Due to premature termination of the study, 660 participants
targeted for pre-planned tests of effectiveness (i.e., before any of the exited the study before completing their Month 12 visit (Figure 1).
type I error was to be spent), p-values for analyses of effectiveness Twelve additional participants (0.6%) discontinued early (6 in each
should be interpreted as descriptive statistics. We calculated exact treatment group), including one participant in each of the
95% confidence intervals for the relative risk of pre-specified priority treatment groups who died during follow-up. Other than the
AEs within system organ classes under a Poisson assumption for the deaths, which were unrelated to product use, no participants
event rates in each treatment group (StatXact). We compared withdrew early due to a medical reason. Most enrolled participants
proportions of women with any pelvic exam findings or STDs in both groups were young (mean age 22.7), unmarried, and had
between treatment groups with a two-sided Mantel-Haenszel Chi- fewer than 9 years of education. Other demographic character-
Square Test stratified by site at the 0.05 significance level. istics and medical history were also similar (Table 1).
All primary and most secondary analyses were either conducted
on the Effectiveness Population which is the subset of the Intent- Sexual Behavior
to-Treat (ITT) Population for whom at least one post-enrollment At enrollment participants reported a mean coital frequency of 9–
HIV evaluation is available, or the Safety Population which is the 10 acts per week (SAVVY 9.5; placebo 8.9) and a mean of 5.8
subset of the ITT Population who ever returned after enrollment. different partners in the last 30 days (in both groups). Approxi-
The Evaluable Population includes the same participants as the mately 2% of the participants (SAVVY 1.8%; placebo 2.5%)
Effectiveness Population but excludes all data collected from a reported having anal sex in the previous 30 days. During follow-
participant after her first documented interruption of product use. up, participants consistently reported a mean of 7 coital acts per
Our data monitoring plan specified that the independent Data week, with a mean of 5–6 sexual partners in the previous 30 days.
Monitoring Committee (DMC), with access to treatment assign- Self-reported condom use increased from 40% during the last
PLoS ONE | www.plosone.org 3 December 2007 | Issue 12 | e1312
SAVVY Gel for HIV Prevention
Figure 1. Participant Trial Flow Diagram (P-Y, person years).
doi:10.1371/journal.pone.0001312.g001
coital act before screening to over 89% at the first follow-up visit, receiving placebo), no abnormal outcomes were reported. A total
remaining relatively constant throughout follow-up. of 40 live births (all normal) were reported to have occurred
Although pregnant participants are included in the intent-to- among study participants.
treat analysis, we asked these participants to stop gel use upon a
positive pregnancy test (they could return to product use after a Product Use
negative pregnancy test). A total of 383 participants in the SAVVY Participants reported that they used the gel for an average of 75% and
group and 386 participants in the placebo group became pregnant 77% of coital acts in the SAVVY and placebo groups, respectively,
at least once during the study. The Kaplan-Meier pregnancy and reported condoms use for 89% and 90% of coital acts in the
probability at 12 months was 42.5 per 100 person-years in the SAVVY and placebo groups, respectively. They reported using both
SAVVY group and 43.7 per 100 person-years in the placebo gel and condoms for 70.1% and 71.7% of acts in the SAVVY and
group. Of the women who became pregnant during the study, placebo groups, respectively. Participants reported using neither gel
most (79%) were pregnant only once. However, approximately nor condoms for 6.2% and 5.3% of acts in the SAVVY and placebo
19% were pregnant twice, and almost 2% were pregnant three groups, respectively. Gel use decreased with time in both groups
times during the course of the study. The median amount of time (from 86–87% during Month 1 to 67–71% during Month 12),
without gel use due to pregnancy (among the participants who whereas condom use was more consistent. From these data, we
became pregnant) was approximately 2 months, resulting in 151 calculated that SAVVY or placebo gel was used alone (i.e., without a
person-years (or 10% of the total person-time) off product due to condom) for approximately 5% of all vaginal acts and that condoms
pregnancy. Of the 942 total pregnancies detected during the study were used alone (i.e., without gel) for approximately 16% of all vaginal
(among 769 participants), pregnancy outcome information was acts. We also calculated that SAVVY or placebo gel was used for
available for 606 pregnancies. Other than 13 spontaneous 81% of vaginal acts during which a condom was also used, but for
abortions (4 in participants receiving SAVVY and 9 in participants only 43–47% of acts where no condom was used (Table 2).
PLoS ONE | www.plosone.org 4 December 2007 | Issue 12 | e1312
SAVVY Gel for HIV Prevention
Table 1. Baseline Characteristics (ITT Population) Table 2. Overall Estimates of Gel and Condom Use at Follow-
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Up by Treatment Groups*
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SAVVY (N = 1073) Placebo (N = 1069)
Overall
Characteristic
Age 22.7 (3.6) 22.7 (3.6) SAVVY N = 1021 Placebo N = 1008
Marital status Mean percentage of vaginal sex acts in the last 7 days with study gel
Unmarried, not cohabiting 946 (88.2) 945 (88.4) Mean (SD) 75.0 (24.0) 77.0 (23.1)
Education Range (median) 0–100 (81.6) 0–100 (83.0)
# 9 years 831 (77.4) 812 (76.0) Mean percentage of vaginal sex acts in the last 7 days with a condom
Pregnancy history Mean (SD) 88.6 (16.5) 89.5 (15.9)
Ever pregnant 782 (72.9) 782 (73.2) Range (Median) 0–100 (94.9) 0–100 (96.0)
Number of pregnancies 2.1 (1.4) 2.1 (1.4) Mean percentage of vaginal sex acts in the last 7 days with both the study
gel and a condom
Number of vaginal deliveries 0.9 (1.0) 0.9 (0.9)
Mean (SD) 70.1 (26.5) 71.7 (25.8)
Contraceptive use
Range (Median) 0–100 (76.5) 0–100 (77.8)
Condom 496(46.2) 516 (48.3)
Mean percentage of vaginal sex acts in the last 7 days with neither a
None 423 (39.4) 397 (37.1)
condom nor study gel
Oral 122 (11.4) 125 (11.7)
Mean (SD) 6.2 (11.8) 5.3 (10.7)
Injectable 23 (2.1) 22 (2.1)
Range (Median) 0–100 ( 0.0) 0–100 ( 0.0)
IUD 5 (0.5) 5 (0.5)
Mean percentage of vaginal sex acts in the last 7 days with study gel only
Other 4 (0.4) 4 (0.4) (without a condom)
History of STI 182 (17.0) 199 (18.6) Mean (SD) 5.1 (17.6) 5.2 (17.9)
Previous spermicide use 91 (8.5) 77 (7.2) Range (median) 0–100 ( 0.0) 0–100 ( 0.0)
Douching 578 (53.9) 555 (51.9) Mean percentage of vaginal sex acts in the last 7 days with a condom only
(without study gel)
Data reported as N(%) or mean (SD); SD = standard deviation, STI = sexually
Mean (SD) 16.4 (32.4) 16.5 (32.7)
transmitted infection
doi:10.1371/journal.pone.0001312.t001 Range (Median) 0–100 ( 0.0) 0–100 ( 0.0)
Mean percentage of vaginal sex acts in the last 7 days with study gel and a
condom
Safety: Adverse Events Mean (SD) 80.8 (35.8) 80.8 (36.2)
Approximately one-third of participants in both groups had adverse Range (Median) 0–100 ( 100) 0–100 ( 100)
events (33% in the SAVVY group and 29% in the placebo group), Mean percentage of vaginal sex acts in the last 7 days with study gel when a
with no significant differences in the overall frequency of AEs condom is not used
between treatment groups. The most frequently reported AEs Mean (SD) 43.1 (47.3) 46.9 (47.9)
included: malaria, abdominal pain, headache, genital pruritus, Range (Median) 0–100 ( 0.0) 0–100 (29.3)
vaginal candidiasis, general pain, bacterial vaginitis, and vaginal
discharge. No significant differences in frequency of AEs occurred *For each participant and variable of interest (e.g., percentage of vaginal acts
between treatment groups. However, more participants in the where study gel was used in the last 7 days prior to the follow-up visit), we first
calculated the participant’s mean value of the variable of interest across all of
SAVVY group had reproductive tract AEs, with an incidence of
their follow-up visits. (Follow-up visits where women reported 0 or a missing
14.5 per 100 person-years in the SAVVY group versus 10.9 per 100 number of vaginal sex acts in the last 7 days are excluded from the calculation
person-years in the in the placebo group (rate ratio = 1.33, 95% of a participant’s mean value.) The mean, SD, range, and median values of the
confidence interval 0.99, 1.80, [Table 3]). The most frequently occur- distributions of these mean values were then obtained for each treatment
group.
ring AEs with the reproductive tract were genital pruritus, vaginal
doi:10.1371/journal.pone.0001312.t002
candidiasis, vaginal discharge, bacterial vaginitis, and vulvovaginitis.
In post hoc analyses that examined subgroups defined by self-
reported frequency of gel use, we found a greater treatment group
difference in the incidence of reported reproductive system and participants had pelvic examinations (63 in the SAVVY group and
breast disorders (17.3 per 100 person-years in the SAVVY group and 58 in the placebo group). No significant differences were seen
10.1 per 100 person-years in the placebo group) among the subgroup between treatment groups for bacterial vaginosis, trichomoniasis,
of participants whose self-reported gel use was at or below the overall or vaginal candidiasis on saline wet mounts.
median than among the subgroup of participants whose self-reported Twenty-two SAEs were reported; 15 by participants in the
gel use was above the overall medium. Among participants whose SAVVY group and 7 by participants in the placebo group. Only
self-reported gel use was at or below the overall median the rate ratio one, gonorrhea, was classified by the investigator as possibly
was 1.71 (95% confidence interval = 1.13–2.61). Similarly, the rate related to the study product (the participant had received placebo).
ratio was 1.64 (95% confidence interval = 1.09–2.50) for the Two deaths occurred during the course of the study. One was
subgroup of participants with self-reported sexual frequency at or suspected to be due to possible sickle cell crisis (the participant had
below the overall median (17.4 per 100 person-years in the SAVVY been randomized to SAVVY), and one was due to viral hepatitis
group and 10.6 per 100 person-years in the placebo group). complicated by hepatic encephalopathy (the participant had been
Participants underwent pelvic examination and STI testing randomized to placebo). No participant was asked to discontinue
during follow-up only if clinically indicated. During follow-up, 121 study drug due to an AE.
PLoS ONE | www.plosone.org 5 December 2007 | Issue 12 | e1312
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SAVVY Gel for HIV Prevention
Table 3. Selected Priority Adverse Events
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Rate Ratio (95% CI),
System Organ Class/ Preferred Term SAVVY (N = 1027 ) Placebo (N = 1012 ) SAVVY vs. Placebo
No. of No. of % of IR* (per 100 No. of No. of % of IR* (per 100
Events Women Women person years) Events Women Women person years)
Reproductive system and breast disorders 140 107 10.4 14.5 120 80 7.9 10.9 1.33 (0.99, 1.80)
Amenorrhoea 3 3 0.3 0.4 0 0 0.0 0.0 INF
Cervicitis 0 0 0.0 0.0 1 1 0.1 0.1 0.00
Dysmenorrhoea 6 6 0.6 0.8 4 4 0.4 0.5 1.47
Dyspareunia 1 1 0.1 0.1 1 1 0.1 0.1 0.98
Genital abscess 1 1 0.1 0.1 2 2 0.2 0.3 0.49
Genital pruritus female 28 27 2.6 3.5 23 20 2.0 2.6 1.32
Genitourinary tract gonococcal infection 0 0 0.0 0.0 1 1 0.1 0.1 0.00
Menorrhagia 1 1 0.1 0.1 2 2 0.2 0.3 0.49
Menstruation irregular 7 7 0.7 0.9 7 6 0.6 0.8 1.14
Ovulation pain 1 1 0.1 0.1 1 1 0.1 0.1 0.98
Pelvic inflammatory disease 5 5 0.5 0.6 4 4 0.4 0.5 1.23
Post coital bleeding 1 1 0.1 0.1 1 1 0.1 0.1 0.98
Vaginal abscess 3 3 0.3 0.4 1 1 0.1 0.1 2.94
Vaginal burning sensation 0 0 0.0 0.0 3 2 0.2 0.3 0.00
Vaginal candidiasis 29 28 2.7 3.6 30 28 2.8 3.7 0.98
Vaginal discharge 16 16 1.6 2.0 13 11 1.1 1.4 1.43
Vaginal erythema 4 3 0.3 0.4 0 0 0.0 0.0 INF
Vaginal haemorrhage 0 0 0.0 0.0 1 1 0.1 0.1 0.00
Vaginal laceration 0 0 0.0 0.0 2 2 0.2 0.3 0.00
Vaginal pain 2 2 0.2 0.3 1 1 0.1 0.1 1.96
Vaginitis bacterial 19 19 1.9 2.4 16 14 1.4 1.8 1.33
Vulvovaginal ulceration 1 1 0.1 0.1 1 1 0.1 0.1 0.98
Vulvovaginitis 11 11 1.1 1.4 5 5 0.5 0.6 2.16
Vulvovaginitis trichomonal 1 1 0.1 0.1 0 0 0.0 0.0 INF
Gastrointestinal disorders 85 73 7.1 9.7 94 76 7.5 10.4 0.93 (0.67, 1.30)
Pregnancy, puerperium and perinatal 7 7 0.7 0.9 13 12 1.2 1.6 0.57 (0.19, 1.57)
conditions
Renal and urinary disorders 15 12 1.2 1.5 18 17 1.7 2.2 0.69 (0.30, 1.53)
*Incidence Rate
doi:10.1371/journal.pone.0001312.t003
Effectiveness: HIV Incidence DISCUSSION
The overall HIV incidence rate was 1.09 per 100 person years (95%
confidence interval 0.63, 1.74). In the Effectiveness Population, we Interpretation and Overall Evidence
observed 17 HIV seroconversions, 8 in the SAVVY group and 9 in We stopped this study prematurely following recommendations of
the placebo group. The Kaplan-Meier estimates of the cumulative an independent DMC because the HIV incidence among enrolled
probability of HIV infection at 12 months were 0.010 in the SAVVY participants was substantially lower than expected; we therefore
group and 0.011 in the placebo group (log-rank p = 0.731), with a could not evaluate the effectiveness of SAVVY in preventing HIV
hazard ratio (SAVVY versus placebo) of 0.88 (95% confidence as intended. Due to the small number of events available for
interval 0.33, 2.27). Of these 17 seroconversions, 9 were in Accra and analysis, this trial was unable to meet the objective of accessing the
8 occurred in Kumasi. Fifteen seroconversions (7 SAVVY; 8 effectiveness of SAVVY in preventing male-to-female transmission
placebo) occurred in participants younger than 25 years. Ten (5 of HIV. To date, no randomized clinical studies have identified a
SAVVY; 5 placebo) of the seroconverters had baseline coital microbicide that is effective at preventing HIV infection [11].
frequency at or below the median and 7 (3 SAVVY; 4 placebo) had Lower-than-expected HIV incidence has been an increasing
baseline coital frequency above the median. problem in HIV prevention trials and the lower-than-expected
When we repeated the analysis in the Evaluable Population, we HIV incidence seen in this trial may have been due in part at least
observed 11 HIV infections (4 in the SAVVY group; 7 in the 3 factors: 1) a greater-than-expected effect of trial risk reduction
placebo group). The Kaplan-Meier estimates of the cumulative measures such as condom use, 2) participants who join clinical
probability of HIV infection at 12 months were 0.007 in the trials may be more inclined to safer behavior than those in their
SAVVY group and 0.012 in the placebo group, with a hazard community who do not participate, and 3) an inaccurate estimate
ratio of 0.57 (95% confidence interval of 0.17 to 1.94). of trial incidence due to changes in the local epidemic. The
PLoS ONE | www.plosone.org 6 December 2007 | Issue 12 | e1312
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SAVVY Gel for HIV Prevention
incidence rate for this trial was estimated from our experience in actual negative association between treatment effect and frequency
earlier trials in a similar population and was not specifically of intercourse, is unclear.
measured in each population before starting the study. One limit of our study is generalizability. Although we did not
Ensuring high product adherence is another important specifically recruit sex workers, women in our study averaged a
challenge in the successful conduct of HIV prevention studies. coital frequency of 7 acts per week and 5–6 different partners in
Low product use due to such factors as pregnancy, missed visits, or the last month. Thus conclusions should be interpreted cautiously
participant’s choice compromises study power. Although partic- with regard to other populations.
ipants reported using the gel during approximately 76% of coital The availability of study product did not appear to result in
acts, our calculations show that participants were more likely to increased risk taking behaviors among participants. During
use gel if a condom was also used than when a condom was not enrollment, participants reported a mean of 9 coital acts per week
used (.80% gel use with condoms versus ,50% gel use without with a mean of 6 sexual partners in the previous 30 days. During
condoms). Since the percent of gel use among risky acts (i.e., with follow-up, participants reported a mean of 7 coital acts per week,
an HIV-infected partner) without condom use is a major with a mean of 5–6 sexual partners in the previous 30 days.
determinant of the overall level of effectiveness that can be Reported condom use during follow-up increased from 40%
observed for a truly efficacious product, unless gel use can be during the last coital act prior to screening to over 89% at the first
increased, studies designed to detect a 50% effect may not be able follow-up visit and remained relatively constant throughout the
to identify truly efficacious products[12]. We observed a entire follow period.
pregnancy probability at 12 months of 42.5 per 100 person-years
in the SAVVY group and 43.7 per 100 person-years in the placebo Conclusion
group. If further studies demonstrate that SAVVY is an effective As a new HIV prevention approach, microbicides could be used
contraceptive method in other populations, the lack of a difference with other prevention strategies such as condoms to reduce the
between the observed pregnancy rates for the SAVVY and number of people who become infected with HIV. In our study,
placebo groups in our study may indicate that the level of gel SAVVY was not associated with increased adverse events overall
adherence was even lower than the participants reported. relative to placebo, but was associated with higher reporting of
The likelihood of pregnancy in reproductive-age women having reproductive adverse events. Our data are insufficient to conclude
multiple sexual encounters is quite high even with relatively high whether SAVVY is effective at preventing HIV infection relative
condom use. Using Wilcox’s estimates of the probability of to placebo. Phase III trials with larger numbers of seroconver-
pregnancy resulting after one or more unprotected act for each sions are needed to determine the effectiveness, safety, and
day of the menstrual cycle (assuming 30 days per cycle and 12 cycles feasibility of using microbicides for prevention of HIV infection in
per year), we expect that if women have two days with unprotected women.
acts on average during each cycle (e.g., 90% condom use, 20 acts per
cycle, and no other contraceptive method use), the 12-month
cumulative pregnancy probability would be 51%. Therefore, the SUPPORTING INFORMATION
pregnancy probability of 43% at 12 months observed during this Checklist S1 CONSORT Checklist
study is not inconsistent with what the participant’s reported–89% Found at: doi:10.1371/journal.pone.0001312.s001 (0.05 MB
condom use and a coital frequency of 30 acts per month. Unless DOC)
adequate and well-controlled safety studies of microbicides in
Protocol S1 Trial Protocol
pregnant women are conducted prior to the start of future HIV
Found at: doi:10.1371/journal.pone.0001312.s002 (0.74 MB
prevention studies thereby allowing women who become pregnant to
DOC)
remain on study product, investigators should consider mandating
that all female participants who have reproductive potential must be
using an effective non-barrier contraceptive such as depot ACKNOWLEDGMENTS
medroxyprogesterone acetate (DMPA), combined oral contraceptive We wish to acknowledge the study participants in Accra and Kumasi,
pills (COCs), an intrauterine device (IUD), or a contraceptive Ghana. We’d also like to acknowledge the study staff at Komfo Anokye
Teaching Hospital, Kwame Nkrumah University of Science & Technol-
implant at least at the time of enrollment.
ogy, Kumasi, Ghana and Noguchi Memorial Institute for Medical
We did not observe a significant difference in the overall rate of Research, University of Ghana, Accra, Ghana.
adverse events between participants receiving SAVVY and those The trial and its reporting complied with the CONSORT Guidelines
receiving placebo. The frequency of adverse events in the [1]. This study was conducted under an IND to the US Food and Drug
reproductive tract/breast disorders (specifically symptoms of Administration. We conducted this study in accordance with Good Clinical
vaginal or vulvar irritation) appeared higher in the SAVVY Practice (GCP) as established by the International Conference on
group, but further subgroup analyses revealed no evidence that the Harmonisation.
effect of SAVVY use was more pronounced among women with
more frequent exposure to the product. Whether the observed Author Contributions
increased risk for these events among the SAVVY users in the
Conceived and designed the experiments: RR WR RD LD LP. Analyzed
subset of women who had fewer than average sexual acts and/or the data: WR. Wrote the paper: KN LP. Other: Study medical monitor:
fewer gel uses is a chance finding, a result of confounding due to KN Served as project manager for the study: AT Recruited and followed
subgrouping by a post-randomization factor, or evidence of an study participants: WA MO AB BO.
REFERENCES
1. Moher D, Schulz KF, Altman DG (2001) The CONSORT statement: revised 3. Niruthisard S, Roddy RE, Chutivongse S (1991) The effects of frequent
recommendations for improving the quality of reports of parallel-group nonoxynol-9 use on the vaginal and cervical mucosa. Sex Transm Dis 18: 176–9.
randomised trials. Lancet 357: 1191–4. 4. Van Damme L, Ramjee G, Alary M, Vuylsteke B, Chandeying V, et al. (2002)
2. Roddy RE, Cordero M, Cordero C, Fortney JA (1993) A dosing study of Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in
nonoxynol-9 and genital irritation. Int J STD AIDS 4: 165–70. female sex workers: a randomised controlled trial. Lancet 360: 971–7.
PLoS ONE | www.plosone.org 7 December 2007 | Issue 12 | e1312
SAVVY Gel for HIV Prevention
5. Biosyn, Inc. Investigator’s Brochure: 1% C31G. 2005. 9. Mauck CK, Creinin MD, Barnhart KT, Ballagh SA, Archer DF, et al. (2004) A
6. Bax R, Douville K, McCormick D, Rosenberg M, Higgins J, et al. (2002) phase I comparative postcoital testing study of three concentrations of C31G.
Microbicides–evaluating multiple formulations of C31G. Contraception 66: Contraception 70: 227–31.
365–8. 10. Mauck CK, Frezieres RG, Walsh TL, Schmitz SW, Callahan MM, et al. (2004)
7. Krebs FC, Miller SR, Catalone BJ, Welsh PA, Malamud D, et al. (2000) Sodium Male tolerance study of 1% C31G. Contraception 70: 221–5.
dodecyl sulfate and C31G as microbicidal alternatives to nonoxynol 9: 11. Balzarini J, Van Damme L (2007) Microbicide drug candidates to prevent HIV
comparative sensitivity of primary human vaginal keratinocytes. Antimicrob infection. Lancet 369: 787–97.
Agents Chemother 44: 1954–60. 12. Trussell J, Dominik R (2005) Will microbicide trials yield unbiased estimates of
8. Mauck CK, Weiner DH, Creinin MD, Barnhart KT, Callahan MM, et al. microbicide efficacy? Contraception 72: 408–13.
(2004) A randomized Phase I vaginal safety study of three concentrations of
C31G vs. Extra Strength Gynol II. Contraception 70: 233–40.
PLoS ONE | www.plosone.org 8 December 2007 | Issue 12 | e1312
View publication stats