This study assessed the safety and tolerability of VivaGel® versus placebo gel when used vaginally twice a day for 14 days in 18 to 24 year old women in San Francisco and Kisumu. GU AEs were common and occurred in both arms; 83% of participants in the VivaGel® arm and 58% of participants in the placebo arm experienced at least one grade 1 or 2 GU AE. Overall, GU AEs judged to be associated with product use were more common among participants in the VivaGel® arm than the placebo arm. These AEs were predominantly mild in nature (grade 1) and were self-limiting. All grade 2 GU AEs that were associated with product use occurred in four women in the VivaGel® arm at the San Francisco site. Superficial colposcopic abnormalities were also common in both arms and these findings were detected more commonly among participants in the VivaGel® arm. Vaginal flora was not adversely affected by study product and there were no significant changes to colonization with lactobacillus spp. Changes in serum chemistry and renal and liver function tests occurred infrequently during follow-up across both treatment arms and are unlikely to be related to study treatment since it was also found that there was no evidence of systemic absorption of SPL7013. There were no grade 3 or 4 AEs. Two participants in the VivaGel® arm discontinued themselves from study product during follow-up due to a self-reported GU AE. No participant was discontinued from study product by the study staff due to an AE.
The proportion of women with a GU AE associated with product use was greater among those randomized to the VivaGel® arm in comparison to the placebo arm among participants enrolled at the Kisumu site but not at the San Francisco site. The Kisumu site enrolled twice as many participants as the San Francisco site, and with a 2
1 randomization scheme, the number of women assigned to the placebo arm was relatively small at each site. Adherence in the VivaGel® group was higher than in the placebo group at the Kisumu site, contributed in part by the early closure of the study and the discontinuation of study product in a greater proportion of those randomized to the placebo arm. Thus it is possible that exposure to fewer doses of placebo in comparison to VivaGel® among participants at the Kisumu site may have contributed to the difference in the incidence of GU AEs. Both VivaGel® and the placebo are carbopol-based aqueous gels. Thus, if carbopol or an excipient in the gel led to GU AEs, it is possible that the differences in GU AEs observed between the VivaGel® and placebo arms may in part be due to differential exposure to the investigational product between the study arms in Kisumu. Alternatively, young women in Kisumu might be more susceptible to epithelial irritation in the genital tract perhaps in relation to differences in vaginal flora or increased immune activation in their genital mucosa 
. Furthermore, imperfect adherence to product per protocol in the VivaGel® (74%) and placebo (53%) arms may have resulted in an underestimation of the true risk of AEs associated with twice daily product use over 14 days.
Results of this trial are consistent with previously published phase 1 vaginal microbicide studies among sexually abstinent populations in that GU AEs were common and generally mild. In the phase 1 trial of once daily dosing for seven days of SPL7013, GU AEs considered potentially product-related were all mild and reported by 5 out of 25 (20%) women receiving VivaGel® (0.5%, 1.0% or 3.0%) and 2 (17%) of 12 women receiving placebo gel 
. In published reports of other candidate vaginal microbicides the proportion of participants who experienced a GU AE has ranged from 14% to 87% 
. To our knowledge, this is one of the first published studies to use the newly developed toxicity table to grade GU AEs in microbicide clinical trials which could theoretically have contributed to the high proportion of women found to experience GU AEs in this study. It is unlikely that the number of GU AEs reported was due to the young age of our participants and this is supported by a recently published study that found that older age was positively correlated with the proportion of participants with a genital finding in another microbicide trial 
. The high rate of reported AEs and colposcopic findings in the placebo arm may suggest that the vehicle itself may have caused mild irritation. In addition, participants in San Francisco reported a greater incidence of AEs in comparison to participants in Kisumu. This finding could have arisen from a difference in perception of AEs, for example women in San Francisco reported more abnormal vaginal discharge than women in Kisumu, or may represent biological differences between these distinct populations. Regardless, this finding suggests the importance of conducting multisite trials in different populations to help evaluate product safety.
The strengths of this study include the double-blind placebo controlled design, and the inclusion of biomarkers to corroborate self-reported use of study product and sexual abstinence. The study population was limited to women aged 18–24 years as they are in greatest need of a microbicide to prevent HIV and STIs; however, these findings may not be generalizable to older age groups. By simultaneously conducting this study in the U.S. and Kenya we have collected safety data in populations of varying characteristics, an important step in the clinical development of any vaginal microbicide.
GU AEs and colposcopic findings consistent with mild genital epithelial irritation and inflammation occurred more commonly in the VivaGel® arm compared with placebo gel when the products were used twice daily for 14 days. Whether the findings from this study translate into safety concerns for this product cannot be ascertained entirely from this study, but should be assessed in conjunction with all the available preclinical and clinical data.