Abstract
OBJECTIVE
Products containing nonoxynol-9 have been used as spermicidal contraceptives for many years, but limited data have been published describing the long-term effects of nonoxynol-9 use on the vaginal microbial ecosystem. This longitudinal study was conducted to examine the effects of nonoxynol-9 on the vaginal ecology.
METHODS
Vaginal swabs were obtained from 235 women enrolled in a randomized clinical trial before initiation of use of 1 of 5 different formulations of nonoxynol-9 for contraception, and up to 3 more samples were gathered over 7 months of use. The swab samples were evaluated in a single laboratory. The prevalence of several constituents of the normal vaginal flora was evaluated. The associations between nonoxynol-9 dosage, formulation, average product use per week, and number of sex acts per week were calculated.
RESULTS
The changes in prevalence of vaginal microbes after nonoxynol-9 use were minimal for each of the different nonoxynol-9 formulations. However, when both nonoxynol-9 concentration and number of product uses are taken into account, nonoxynol-9 did have dose-dependant effects on the increased prevalence of anaerobic gram-negative rods (odds ratio [OR] 2.4, 95% confidence interval [CI] 1.1–5.3), H2O2-negative lactobacilli (OR 2.0, 95% CI 1.0–4.1), and bacterial vaginosis (OR 2.3, 95% CI 1.1–4.7).
CONCLUSION
This study demonstrated that most nonoxynol-9 users experienced minimal disruptions in their vaginal ecology. There were no differences between the different formulations evaluated with respect to changes in vaginal microflora. However, independent of the nonoxynol-9 formulation, there was a dose-dependent effect with increased exposure to nonoxynol-9 on the risk of bacterial vaginosis and its associated flora.
LEVEL OF EVIDENCE
II-2
Products containing nonoxynol-9 have been used as vaginal spermicides for decades and are a widely used method of family planning. In the past several years, these products have also been evaluated as topical microbicides for prevention of sexually transmitted infections including human immunodeficiency virus (HIV).1 Most of these studies were done in sex workers and did not demonstrate that women who used nonoxynol-9 had a decreased risk of sexually transmitted infection. Over the past decade, the importance of an intact vaginal ecosystem in resistance to urogenital infection has been highlighted. Most of the published data describing the effects of nonoxynol-9 use on the vaginal microbial ecosystem have assessed short-term nonoxynol-9 exposures, and their effects on the vaginal ecosystem have raised some concern. Women with vaginal colonization by lactobacilli are less likely to have2 and acquire3 sexually transmitted infections such as herpes simplex virus 24 and human papilloma virus5 and are also less likely to acquire bacterial vaginosis.3 Two cross-sectional studies and 1 longitudinal study have shown an association between the presence of vaginal lactobacilli and decreased prevalence or incidence of HIV.6–8 Whether nonoxynol-9 affects lactobacilli colonization and the integrity of the vaginal ecosystem remains unclear.
Some in vitro studies have suggested that use of spermicides containing nonoxynol-9 would decrease colonization by lactobacilli,9 whereas others have reported that most peroxide-producing strains of lactobacilli are resistant to nonoxynol-9.10 The mechanism, however, by which nonoxynol-9 might alter the vaginal ecology in vivo is not clear, and whether these vaginal flora alterations are clinically relevant has not been elucidated. There are studies that have shown an adverse association between nonoxynol-9 use and the maintenance of vaginal flora,9 although other clinical studies have not confirmed this association.11 Some evidence suggests that nonoxynol-9 does not increase the risk of either vulvovaginal candidiasis11,12 or bacterial vaginosis,13 and that vaginal epithelia remain intact at lower doses of nonoxynol-9.12 However, because certain doses of nonoxynol-9 seem to be at least transiently disruptive to vaginal ecology,14,15 this study was designed to investigate the effect of long-term use and average weekly dose of 5 different formulations of nonoxynol-9 on vaginal flora in sexually active women in mutually monogamous relationships.
MATERIALS AND METHODS
This study was conducted at 3 sites between June 1998 and August 2002 as a planned substudy of a multicenter, randomized trial of the contraceptive effectiveness and safety of nonoxynol-9. The institutional review boards at each site approved the trial before commencement. All participants underwent an informed consent process. The details of the multicenter trial design and primary results have been published.16 Briefly, sexually active women from 18–40 years of age with no history of subfertility and at low risk for sexually transmitted infections were eligible. At screening, all women underwent an interview, pelvic examination including Pap test, wet preparation, and urine pregnancy test. As a part of the larger multicenter trial, each participant was randomly assigned to use 1 of the following 5 nonoxynol-9 formulations as her only method of contraception for 7 months: a 52.5-mg gel, a 100-mg gel, a 150-mg gel, a 100-mg film, and a 100-mg suppository. Follow-up visits 2 through 4 were scheduled at weeks 4, 17, and 30. At each scheduled visit, the participant obtained 2 vaginal specimens by twice inserting the tip of a swab about one-half to 1 inch into her vagina and rotating it for 10 seconds. The swabs were placed into a specimen transport tube and sent by research personnel to a central laboratory for analysis.
Upon enrollment, the site study personnel telephoned a central randomization center to obtain an assignment to 1 of the 5 nonoxynol-9 formulations. The randomization scheme was computer-generated at Family Health International before the start of enrollment and was stratified by site with randomly ordered block sizes. Treatment assignment was not concealed from the participants or the clinical examiners. However, laboratory personnel were masked to the treatment assignment.
One swab was rolled across a glass slide and air dried for preparation of a Gram stain for assessment of vaginal flora. The slides were interpreted by a standard method for the diagnosis of bacterial vaginosis described by Nugent et al.17 A score of 0 to 3 was interpreted as consistent with normal, Lactobacillus-predominant flora, a score of 4 to 6 corresponded to intermediate flora, and a score of 7–10 indicated bacterial vaginosis. The other swab was placed in Amies transport media and transported to the laboratory for culture of Escherichia coli, Lactobacillus, Candida, Enterococcus, and Staphylococcus. In addition to a 5% sheep blood agar plate, a Rogosa agar plate was inoculated for recovery of lactobacilli and incubated in an anaerobic chamber. Lactobacilli were identified to the genus level by Gram stain morphology and production of lactic acid as assessed by gas chromatography. All lactobacilli were tested for the production of hydrogen peroxide in a qualitative assay using tetramethylbenzidine agar. Anaerobic gram-negative rods were detected on Brucella agar after 4–5 days of incubation under anaerobic conditions.
The sample size was estimated based on previous studies evaluating the effect of the diaphragm plus Conceptrol gel (Ortho-McNeil Pharmaceutical, Inc., Raritan, NJ) on vaginal microflora, which demonstrated an increase in the prevalence of E. coli from 15% before spermicide use to 48% after spermicide use.18 Based on a 2-tailed McNemar’s χ2 test at the 0.05 significance level, a sample size of 40 women in each group, or a total of 200 women, would have 80% power to demonstrate similar increases in the prevalence of E. coli and other organisms between baseline and follow-up. The sample size was increased to a convenient number of 255 women, which allowed for a 22% discontinuation rate. A sample size of 47 in each group would have 80% power to detect a minimal between-group difference of 22% in the prevalence of E. coli after spermicide use.
All statistical analyses were performed using Stata statistical software, release 8.0 (StataCorp, College Station, TX), and all statistical tests were evaluated at the .05 significance level. Differences in the baseline characteristics of the women randomly assigned to the 5 nonoxynol-9 formulation groups were evaluated using analysis of variance or Pearson’s χ2, as appropriate. Vaginal colonization by hydrogen peroxide-producing and hydrogen peroxide-negative Lactobacillus, Gardnerella vaginalis, E. coli, Enterococcus, anaerobic gram-negative rods, and yeast was defined as detection of growth of the organism from agar culture. Differences in the prevalence of vaginal microbes and a Gram stain diagnosis of bacterial vaginosis at each visit among the nonoxynol-9 formulation groups were evaluated using Pearson’s χ2. Paired differences in prevalence between visits within an nonoxynol-9 formulation group were evaluated using McNemar’s χ2 test.
Average weekly nonoxynol-9 dose was calculated by multiplying the number of times the participant used the product, as obtained from the subjects’ coital diaries, by the dose of nonoxynol-9 in that product and then dividing by the number of weeks since the prior visit. Generalized estimating equations were used to identify factors that were associated with the prevalence of vaginal microbial colonization and Gram stain diagnosis of bacterial vaginosis after spermicide use. Generalized estimating equations were also used to identify organisms associated with the prevalence of self-reported symptoms after spermicide use. Generalized estimating equations are a regression method that is able to model the marginal expectation of repeated, correlated outcomes. An exchangeable working correlation matrix was specified that assumes a common within-group correlation, and modified sandwich estimates of the variance were calculated, which produces valid standard errors even if the within-group correlation structure is not correctly specified. Sandwich estimates of the variance combine the variance estimate for the specified model with the variance matrix constructed from the data. The sandwich estimate is then modified to consider the sums of the observations for each individual. Statistical inference was based on the generalized Wald test statistic.19 Multivariate generalized estimating equations models were developed using forward stepwise regression. Variables were retained in the model if the Wald test statistic had a P value of 0.05 or less. Models for the prevalence of each organism after spermicide use were adjusted for baseline colonization status. The population averaged odds ratios derived from the generalized estimating equations models are presented along with the 95% confidence intervals and P values.
RESULTS
A total of 269 women were enrolled and 235 (87%) completed between 2 and 4 visits. The median time between the enrollment visit and visits 2, 3, and 4 were 4, 17, and 30 weeks, respectively. Of the 235 women with follow-up, 46 (20%) were assigned to use the 52.5-mg gel, 43 (18%) the 100-mg gel, 46 (20%) the 150-mg gel, 51 (22%) the 100-mg film, and 49 (21%) the 100-mg suppository. Sixty-four percent of the participants used their assigned spermicide with every coital act. The demographic characteristics of the 235 participants with follow-up are displayed in Table 1. The most common contraceptive method used by the women in all groups was the oral contraceptive pill (used by 75% of study participants at some point in the past). However, a range of contraceptive methods were used by these women, including implants, depomedroxyprogesterone acetate, condoms (male and female), intrauterine devices, spermicides, diaphragm, periodic abstinence, and withdrawal. There was no statistically significant difference in the types of contraceptives used across the 5 groups. Reports of urogenital symptoms and infections were not statistically significantly different among the groups, and few (1–3%) had symptoms within the week before enrollment.
Table 1.
Participant Characteristics
| Characteristic | Gel A (n = 46) | Gel B (n = 43) | Gel C (n = 46) | Film (n = 51) | Suppository (n = 49) | P |
|---|---|---|---|---|---|---|
| Mean age ± standard deviation (y) | 28.3 ± 5.7 | 27.7 ± 5.4 | 28.0 ± 5.1 | 27.1 ± 5.0 | 27.2 ± 5.1 | .8 |
| Race/ethnicity | .7 | |||||
| White, non-Hispanic | 28 (61) | 21 (49) | 26 (57) | 30 (59) | 26 (53) | |
| African-American, non-Hispanic | 13 (28) | 20 (47) | 16 (35) | 20 (39) | 18 (37) | |
| Hispanic | 3 (7) | 1 (2) | 1 (2) | 1 (2) | 2 (4) | |
| Other | 2 (4) | 1 (2) | 3 (7) | 0 | 3 (6) | |
| Education | .96 | |||||
| No degree/high school/GED | 13 (28) | 14 (32) | 11 (24) | 17 (33) | 12 (24) | |
| Associate degree (< 4 y college) | 14 (30) | 16 (37) | 17 (37) | 17 (33) | 16 (33) | |
| College (4 y) | 13 (28) | 11 (26) | 11 (24) | 11 (22) | 14 (29) | |
| Postgraduate | 6 (13) | 2 (5) | 7 (15) | 6 (12) | 7 (14) | |
| Marital status | .5 | |||||
| Married, living with partner | 15 (33) | 13 (30) | 23 (50) | 16 (31) | 18 (37) | |
| Single, living with partner | 11 (24) | 15 (35) | 10 (22) | 12 (24) | 12 (24) | |
| Single, not living with partner | 20 (43) | 15 (35) | 13 (28) | 23 (45) | 19 (39) | |
| Current smoking | .7 | |||||
| No | 32 (70) | 32 (74) | 36 (78) | 40 (78) | 40 (82) | |
| Yes | 14 (30) | 11 (26) | 10 (22) | 11 (22) | 9 (18) | |
| Gravidity | .3 | |||||
| 0 | 15 (33) | 15 (35) | 16 (35) | 18 (35) | 22 (45) | |
| 1 | 15 (33) | 5 (12) | 7 (15) | 14 (27) | 9 (18) | |
| 2 | 7 (15) | 8 (19) | 9 (20) | 6 (12) | 10 (20) | |
| ≥3 | 9 (20) | 15 (35) | 14 (30) | 13 (25) | 8 (16) | |
| Parity | .08 | |||||
| 0 | 23 (50) | 19 (44) | 19 (41) | 25 (49) | 34 (69) | |
| 1 | 13 (28) | 7 (16) | 10 (22) | 12 (24) | 8 (16) | |
| ≥2 | 10 (22) | 17 (40) | 17 (37) | 14 (27) | 7 (14) |
GED, general equivalency diploma.
Values are n (%) unless otherwise specified.
The proportion of women with colonization by the individual microorganisms at baseline and subsequent visits are listed in Table 2. There was an overall difference at visit 2 in the percentage of women with an increase the prevalence of yeast, and there was a difference in the presence of H2O2-negative lactobacilli at visit 3 as well with Enterococcus at visit 4. In the 52.5-mg gel group, there was a statistically significant increase in Enterococcus from the baseline visit to visit 2, but this resolved by visit 3. In the 100-mg film group, the number of H2O2-negative lactobacilli decreased significantly between visits 2 and 3, and this effect was persistent.
Table 2.
Microflora Prevalence by Visit Number and Product Formulation
| Microflora | n | 52.5-mg Gel | n | 100-mg Gel | n | 150-mg Gel | n | 100-mg Film | n | 100-mg Suppository | P |
|---|---|---|---|---|---|---|---|---|---|---|---|
| H2O2 + Lactobacillus | |||||||||||
| Baseline | 46 | 29 (63) | 43 | 24 (56) | 46 | 27 (59) | 51 | 29 (57) | 49 | 31 (63) | .9 |
| Visit 2 | 46 | 35 (76) | 43 | 24 (56) | 46 | 29 (63) | 51 | 27 (53) | 49 | 30 (61) | .2 |
| Visit 3 | 35 | 24 (69) | 32 | 22 (69) | 34 | 22 (65) | 38 | 24 (63) | 37 | 27 (73) | .9 |
| Visit 4 | 21 | 15 (71) | 21 | 17 (81) | 26 | 17 (65) | 22 | 15 (68) | 22 | 15 (68) | .8 |
| H2O2– Lactobacillus | |||||||||||
| Baseline | 46 | 8 (17) | 43 | 7 (16) | 46 | 6 (13) | 51 | 9 (18) | 49 | 6 (12) | .9 |
| Visit 2 | 46 | 8 (17) | 43 | 5 (12) | 46 | 5 (11) | 51 | 12 (24)* | 49 | 8 (16) | .4 |
| Visit 3 | 35 | 6 (17) | 32 | 6 (19) | 34 | 11 (32) | 38 | 1 (3)* | 37 | 8 (22) | .03 |
| Visit 4 | 21 | 3 (14) | 21 | 3 (14) | 26 | 4 (15) | 22 | 1 (5) | 22 | 4 (18) | .7 |
| Gardnerella vaginalis | |||||||||||
| Baseline | 46 | 14 (30) | 43 | 22 (51) | 46 | 15 (33) | 51 | 23 (45) | 49 | 24 (49) | .1 |
| Visit 2 | 46 | 13 (28) | 43 | 18 (42) | 46 | 13 (28) | 51 | 24 (47) | 49 | 20 (41) | .2 |
| Visit 3 | 35 | 11 (31) | 32 | 14 (44) | 34 | 10 (29) | 38 | 17 (45) | 37 | 14 (38) | .6 |
| Visit 4 | 21 | 7 (33) | 21 | 8 (38) | 26 | 7 (27) | 22 | 9 (39) | 22 | 8 (36) | .9 |
| Escherichia coli | |||||||||||
| Baseline | 46 | 17 (37) | 43 | 21 (49) | 46 | 13 (28) | 51 | 18 (35) | 49 | 16 (33) | .3 |
| Visit 2 | 46 | 22 (48) | 43 | 18 (42) | 46 | 21 (46) | 51 | 21 (41) | 49 | 24 (49) | .9 |
| Visit 3 | 35 | 13 (37) | 32 | 10 (31) | 34 | 10 (29) | 38 | 9 (24) | 37 | 18 (49) | .2 |
| Visit 4 | 21 | 4 (19) | 21 | 6 (29) | 26 | 8 (31) | 22 | 8 (35) | 22 | 8 (36) | .8 |
| Enterococcus | |||||||||||
| Baseline | 46 | 24 (52)* | 43 | 25 (58) | 46 | 23 (50) | 51 | 28 (55) | 49 | 26 (53) | .95 |
| Visit 2 | 46 | 33 (72)* | 43 | 24 (56) | 46 | 26 (57) | 51 | 26 (51) | 49 | 31 (63) | .3 |
| Visit 3 | 35 | 18 (51) | 32 | 16 (50) | 34 | 17 (50) | 38 | 22 (58) | 37 | 22 (59) | .9 |
| Visit 4 | 21 | 15 (71) | 21 | 8 (38) | 26 | 12 (46) | 22 | 12 (52) | 22 | 17 (77) | .04 |
| Anaerobic gram-negative rods | |||||||||||
| Baseline | 46 | 37 (80) | 43 | 37 (86) | 46 | 32 (70) | 51 | 39 (76) | 49 | 38 (78) | .4 |
| Visit 2 | 46 | 35 (76) | 43 | 41 (95) | 46 | 36 (78) | 51 | 40 (78) | 49 | 43 (88) | .08 |
| Visit 3 | 35 | 28 (80) | 32 | 29 (91) | 34 | 28 (82) | 38 | 27 (71) | 37 | 30 (81) | .4 |
| Visit 4 | 21 | 13 (62) | 21 | 18 (86) | 26 | 21 (81) | 22 | 16 (70) | 22 | 20 (91) | .1 |
| Yeast | |||||||||||
| Baseline | 46 | 6 (13) | 43 | 6 (14) | 46 | 6 (13) | 51 | 5 (10) | 49 | 7 (14) | .97 |
| Visit 2 | 46 | 10 (22) | 43 | 8 (19) | 46 | 3 (7) | 51 | 2 (4) | 49 | 5 (10) | .03 |
| Visit 3 | 35 | 5 (14) | 32 | 6 (19) | 34 | 2 (6) | 38 | 1 (3) | 37 | 4 (11) | .2 |
| Visit 4 | 21 | 4 (19) | 21 | 3 (14) | 26 | 4 (15) | 22 | 0 | 22 | 1 (5) | .2 |
| Staphylococcus aureus | |||||||||||
| Baseline | 46 | 0 | 43 | 1 (2) | 46 | 1 (2) | 51 | 0 | 49 | 3 (6) | .2 |
| Visit 2 | 46 | 1 (2) | 43 | 1 (2) | 46 | 1 (2) | 51 | 1 (2) | 49 | 1 (2) | 1.0 |
| Visit 3 | 35 | 0 | 32 | 1 (3) | 34 | 0 | 38 | 1 (3) | 37 | 0 | .5 |
| Visit 4 | 21 | 0 | 21 | 0 | 26 | 0 | 22 | 0 | 22 | 1 (5) | .4 |
| Nugent score | |||||||||||
| Baseline | 45 | 42 | 45 | 51 | 47 | .6 | |||||
| Normal (0–3) | 24 (53) | 16 (38) | 26 (58) | 24 (47) | 25 (53) | ||||||
| Intermediate (4–6) | 11 (24) | 16 (38) | 10 (22) | 12 (24) | 9 (19) | ||||||
| Bacterial vaginosis (7–10) | 10 (22) | 10 (24) | 9 (20) | 15 (29) | 13 (28) | ||||||
| Visit 2 | 46 | 43 | 45 | 51 | 48 | .4 | |||||
| Normal (0–3) | 27 (59) | 19 (44) | 23 (51) | 19 (37) | 26 (54) | ||||||
| Intermediate (4–6) | 11 (24) | 10 (23) | 14 (31) | 18 (35) | 13 (27) | ||||||
| Bacterial vaginosis (7–10) | 8 (17) | 14 (33) | 8 (18) | 14 (27) | 9 (19) | ||||||
| Nugent score | |||||||||||
| Baseline | 45 | 42 | 45 | 51 | 47 | .6 | |||||
| Normal (0–3) | 24 (53) | 16 (38) | 26 (58) | 24 (47) | 25 (53) | ||||||
| Intermediate (4–6) | 11 (24) | 16 (38) | 10 (22) | 12 (24) | 9 (19) | ||||||
| Bacterial vaginosis (7–10) | 10 (22) | 10 (24) | 9 (20) | 15 (29) | 13 (28) | ||||||
| Visit 2 | 46 | 43 | 45 | 51 | 48 | .4 | |||||
| Normal (0–3) | 27 (59) | 19 (44) | 23 (51) | 19 (37) | 26 (54) | ||||||
| Intermediate (4–6) | 11 (24) | 10 (23) | 14 (31) | 18 (35) | 13 (27) | ||||||
| Bacterial vaginosis (7–10) | 8 (17) | 14 (33) | 8 (18) | 14 (27) | 9 (19) | ||||||
| Visit 3 | 34 | 33 | 32 | 37 | 36 | .9 | |||||
| Normal (0–3) | 17 (50) | 15 (45) | 16 (50) | 16 (43) | 16 (44) | ||||||
| Intermediate (4–6) | 10 (29) | 8 (24) | 9 (28) | 10 (27) | 10 (28) | ||||||
| Bacterial vaginosis (7–10) | 7 (21) | 10 (30) | 7 (22) | 11 (30) | 10 (28) | ||||||
| Visit 4 | 22 | 20 | 25 | 23 | 21 | .7 | |||||
| Normal (0–3) | 14 (64) | 11 (55) | 12 (48) | 11 (48) | 10 (48) | ||||||
| Intermediate (4–6) | 3 (14) | 7 (35) | 8 (32) | 6 (26) | 5 (24) | ||||||
| Bacterial vaginosis (7–10) | 5 (23) | 2 (10) | 5 (20) | 6 (26) | 6 (29) | ||||||
Values are n (%) unless otherwise specified.
Indicates significant paired change between visits within the given group.
Symptomatology was uncommon in this study. Thirty women were colonized with Candida spp after nonoxynol-9 use, and a statistically significant proportion of these women complained of vaginal burning (P = .002), vaginal itching (P = .004), and vulvar burning (P = .04). Pelvic pain and vaginal bleeding were not associated with vaginal flora disturbances, but dyspareunia was associated with the presence of Gardnerella vaginalis (P = .03). There were no significant complaints that were associated with a diagnosis of bacterial vaginosis.
There was a decrease in the prevalence of yeast in those women who used the film and suppository nonoxynol-9 formulation, even after controlling for baseline colonization status. The prevalence of anaerobic gram-negative rods was increased in those who used the 100-mg gel, but the prevalence of Enterococcus was decreased in this same group.
Colonization at baseline was the most significant risk factor for colonization with the same microbe later at follow-up (P < .001 for each). African-American women had a statistically significantly increased risk for Gardnerella vaginalis colonization (odds ratio [OR] 2.7, 95% confidence interval [CI] 1.6–4.5), anaerobic gram-negative colonization (OR 2.5, 95% CI 1.2–4.9), and bacterial vaginosis (OR 3.6, 95% CI 1.9–6.7) when compared with women of other races. African-American race was also associated with a decreased risk of colonization by H2O2-producing lactobacilli (OR 0.5, 95% CI 0.3–0.8). Vaginal micro-flora did differ between sites, but these differences were attributable to differences in ethnic distribution by site. Thus, after statistical adjustment for race, site-specific differences in vaginal microflora were no longer significant. Similarly, the type of nonoxynol-9 delivery system had limited impact on microbe colonization (Table 3) after accounting for baseline microbe colonization.
Table 3.
Odds Ratios of Microbe Colonization by Nonoxynol-9 Delivery System, Adjusted for Baseline Colonization
| Microbe/Nonoxynol-9 Formulation | Odds Ratio (95% Confidence Interval) |
|---|---|
| H2O2+ Lactobacilli | |
| 52.5-mg gel | 1.0 (referent) |
| 100-mg gel | 0.8 (0.3–2.0) |
| 150-mg gel | 0.7 (0.4–1.4) |
| 100-mg film | 0.5 (0.2–1.1) |
| 100-mg suppository | 0.7 (0.3–1.5) |
| H2O2– Lactobacilli | |
| 52.5-mg gel | 1.0 (referent) |
| 100-mg gel | 0.9 (0.4–1.9) |
| 150-mg gel | 1.2 (0.5–2.9) |
| 100-mg film | 0.8 (0.4–1.7) |
| 100-mg suppository | 1.3 (0.6–2.9) |
| Gardnerella vaginalis | |
| 52.5-mg gel | 1.0 (referent) |
| 100-mg gel | 1.1 (0.5–2.6) |
| 150-mg gel | 0.8 (0.4–1.9) |
| 100-mg film | 1.4 (0.7–3.0) |
| 100-mg suppository | 1.0 (0.4–2.2) |
| Escherichia coli | |
| 52.5-mg gel | 1.0 (referent) |
| 100-mg gel | 0.7 (0.4–1.3) |
| 150-mg gel | 1.0 (0.6–1.8) |
| 100-mg film | 0.7 (0.4–1.4) |
| 100-mg suppository | 1.4 (0.8–2.6) |
| Enterococcus | |
| 52.5-mg gel | 1.0 (referent) |
| 100-mg gel | 0.5 (0.2–0.9) |
| 150-mg gel | 0.6 (0.3–1.2) |
| 100-mg film | 0.5 (0.3–1.0) |
| 100-mg suppository | 0.9 (0.4–1.9) |
| Anaerobic gram-negative rods | |
| 52.5-mg gel | 1.0 (referent) |
| 100-mg gel | 3.6 (1.3–10.4) |
| 150-mg gel | 1.7 (0.7–4.0) |
| 100-mg film | 1.0 (0.5–2.3) |
| 100-mg suppository | 2.3(0.9–5.6) |
| Bacterial Vaginosis gram stain | |
| 52.5-mg gel | 1.0 (referent) |
| 100-mg gel | 2.0 (0.8–5.4) |
| 150-mg gel | 1.2 (0.5–2.4) |
| 100-mg film | 1.4 (0.6–3.5) |
| 100-mg suppository | 0.7 (0.4–1.4) |
| Yeast | |
| 52.5-mg gel | 1.0 (referent) |
| 100-mg gel | 0.9 (0.4–2.4) |
| 150-mg gel | 0.4 (0.1–1.1) |
| 100-mg film | 0.1 (0.02–0.5) |
| 100-mg suppository | 0.4 (0.1–0.96) |
To test the hypothesis that increased weekly exposure to nonoxynol-9 accounted for changes in the vagina microflora, average weekly nonoxynol-9 exposures were categorized based on quartiles as< 100 mg, 100–174 mg, 175–281 mg, and > 281 mg. There was a dose-dependant relationship between average weekly nonoxynol-9 exposure and an increase in the prevalence of anaerobic gram-negative rods and bacterial vaginosis, even after adjustment for baseline colonization (Table 4). The increased risk for bacterial vaginosis and anaerobic gram-negative rods associated with nonoxynol-9 exposure was present even after adjustment for race (OR 2.2, 95% CI 1.1–4.4; OR 2.4, 95% CI 1.1–5.3, respectively).
Table 4.
Odds Ratios of Microbe Colonization by Nonoxynol-9 Dose, Adjusted for Baseline Colonization
| Microbe/Average Nonoxynol-9 Dose per Wk | Odds Ratio (95% Confidence Interval) | P |
|---|---|---|
| H2O2+ Lactobacilli | ||
| < 100 mg | 1.0 (referent) | |
| 100–174 mg | 1.2 (0.7–2.1) | .6 |
| 175–281 mg | 0.9 (0.5–1.7) | .8 |
| > 281 mg | 0.7 (0.4–1.4) | .3 |
| H2O2– Lactobacilli | ||
| < 100 mg | 1.0 (referent) | |
| 100–174 mg | 1.2 (0.7–2.2) | .7 |
| 175–281 mg | 0.9 (0.4–1.8) | .7 |
| > 281 mg | 2.0 (1.0–4.1) | .04 |
| Gardnerella vaginalis | ||
| < 100 mg | 1.0 (referent) | |
| 100–174 mg | 1.8 (1.0–3.2) | .06 |
| 175–281 mg | 1.6 (0.9–2.8) | .1 |
| > 281 mg | 1.8 (0.9–3.4) | .08 |
| Escherichia coli | ||
| < 100 mg | 1.0 (referent) | |
| 100–174 mg | 1.0 (0.6–1.7) | .9 |
| 175–281 mg | 0.9 (0.5–1.5) | .7 |
| > 281 mg | 1.3 (0.8–2.1) | .4 |
| Enterococcus | ||
| < 100 mg | 1.0 (referent) | |
| 100–174 mg | 1.3 (0.7–2.2) | .4 |
| 175–281 mg | 0.9 (0.5–1.5) | .6 |
| > 281 mg | 1.1 (0.6–2.1) | .7 |
| Anaerobic gram-negative rods | ||
| < 100 mg | 1.0 (referent) | |
| 100–174 mg | 1.2 (0.6–2.2) | .6 |
| 175–281 mg | 1.4 (0.7–2.6) | .3 |
| > 281 mg | 2.4 (1.1–5.3) | .02 |
| Bacterial vaginosis gram stain | ||
| < 100 mg | 1.0 (referent) | |
| 100–174 mg | 1.7 (0.9–3.3) | .1 |
| 175–281 mg | 1.6 (0.8–3.2) | .2 |
| > 281 mg | 2.3 (1.1–4.7) | .02 |
| Yeast | ||
| < 100 mg | 1.0 (referent) | |
| 100–174 mg | 1.7 (0.9–3.3) | .1 |
| 175–281 mg | 1.6 (0.8–3.2) | .2 |
| > 281 mg | 2.3 (1.1–4.7) | .02 |
The mean number of coital acts, product usages, and average milligram dose of nonoxynol-9 per week are shown in Table 5. With the exception of those participants in the 100-mg gel group, there was no difference in number of coital acts per week among the groups. This group also accounts for the difference found in product use per week. Average milligram dose per week was predictably increased with the higher-dose nonoxynol-9 formulations and also increased in the group who used the 100-mg nonoxynol-9 gel (because that group had more coital acts per week).
Table 5.
Average Coital Acts, Product Use, and Nonoxynol-9 Use Per Week
| Formulation | n | Coital Acts/Wk | P | Product Use/Wk | P | Nonoxynol-9 Dose/Wk (mg) | P |
|---|---|---|---|---|---|---|---|
| 52.5-mg gel | 46 | 2.1 ± 1.3 | 1.9 ± 1.2 | 101.4 ± 63.3 | |||
| 100-mg gel | 43 | 2.9 ± 2.4 | 2.8 ± 2.3 | 280.5 ± 233.6 | |||
| 150-mg gel | 46 | 2.1 ± 1.2 | 2.0 ± 1.2 | 296.2 ± 176.5 | |||
| 100-mg film | 51 | 2.1 ± 1.6 | 1.9 ± 1.5 | 194.6 ± 150.0 | |||
| 100-mg suppository | 49 | 2.1 ± 1.4 | 1.9 ± 1.4 | 190.9 ± 143.8 | |||
| .06 | .03 | < .001 |
Values are mean ± standard deviation unless otherwise specified.
DISCUSSION
This longitudinal investigation did not show dramatic effects of nonoxynol-9 on the vaginal ecologic system. The most significant risk factor for being colonized with an organism after spermicide use was baseline colonization, not nonoxynol-9 exposure. However, even after adjustment for colonization status at baseline and for race, an increase in nonoxynol-9 exposure was associated with a statistically significant increase in colonization by anaerobic gram-negative rods, H2O2-negative lactobacilli, and bacterial vaginosis. Furthermore, the vehicle itself, gel, film, and suppository, was not generally associated with changes in the vaginal ecology independent of nonoxynol-9 dose. Rather, it was the average nonoxynol-9 exposure per week, which reflects the number of sex acts with spermicide, as well as the dose of each application used that increased the risk for vaginal flora disturbances.
The clinical significance of this dose-dependant relationship is less clear than its statistical significance; symptomatology was uncommon in this trial. The laboratory diagnosis of bacterial vaginosis, as opposed to the clinical diagnosis, was used in this study because this testing can be standardized across sites. Not all of the women with a Gram stain consistent with bacterial vaginosis would have fulfilled the clinical criteria for the diagnosis of bacterial vaginosis. Nevertheless, an increase in the dose of nonoxynol-9 was associated with vaginal ecologic disruptions, and whether these changes predispose women to HIV or other virulent sexually transmitted infections remains an important question. Even if the dose-dependant changes in the vaginal flora did not produce symptoms in the study population, it is possible that the changes themselves create an environment that facilitates infection with sexually transmitted infections. Nonoxynol-9 use would clearly be limited if this were the case.
The population studied here is representative of people who are in long-standing, mutually monogamous relationships. In this sample, nonoxynol-9 was not generally disruptive to the vaginal flora. Because of the low incidence of sexually transmitted infections in the study population, associations between nonoxynol-9, vaginal ecologic disruption, and sexually transmitted infection transmissibility could not be delineated. However, studies have demonstrated an association of nonoxynol-9 use with increased HIV acquisition in the sex worker population,20,21 where increased weekly dose may have been a factor. An association with the presence of bacterial vaginosis and increased HIV prevalence has been described.6 It is possible that the increase in bacterial vaginosis that is found with increased nonoxynol-9 dose in this study is an intermediate state for the increased HIV incidence that has been found on other studies. Further research in this area is necessary to test this hypothesis.
Eighty-seven percent of the participants completed 75% of the planned follow-up visits. Weaknesses of the study include that the both product use and frequency of coital acts were dependant upon information extracted from participant diaries. There was no objective way to verify this information. Another weakness of the analysis is the lack of control group for comparison.
Overall, this trial supports the safety of nonoxynol-9 as a contraceptive with respect to its effects on the vaginal ecosystem. Raymond et al16 recently showed that these 5 nonoxynol-9 formulations are safe and effective forms of contraception, with increased dose being associated with increased efficacy. Although we found that there was clearly an effect of increased dose of nonoxynol-9 on the vaginal ecosystem, this effect was present only among the most frequent users.
Footnotes
Funding support: Family Health International With funds From the National Institute of Child Health and Human Development contract number N01-HD-73271.
References
- 1.Wilkinson D, Ramjee G, Tholandi M, Rutherford G. Nonoxynol-9 for preventing vaginal acquisition of sexually transmitted infections by women from men. Cochrane Database Syst Rev. 2002;(4):CD003939. doi: 10.1002/14651858.CD003936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hillier SL, Krohn MA, Nugent RP, Gibbs RS. Characteristics of three vaginal flora patterns assessed by gram stain among pregnant women. Vaginal Infections and Prematurity Study Group. Am J Obstet Gynecol. 1992;166:938–44. doi: 10.1016/0002-9378(92)91368-k. [DOI] [PubMed] [Google Scholar]
- 3.Hawes SE, Hillier SL, Benedetti J, Stevens CE, Koutsky LA, Wolner-Hanssen P, et al. Hydrogen peroxide-producing lacto-bacilli and acquisition of vaginal infections. J Infect Dis. 1996;174:1058–63. doi: 10.1093/infdis/174.5.1058. [DOI] [PubMed] [Google Scholar]
- 4.Cherpes TL, Meyn LA, Krohn MA, Lurie JG, Hillier SL. Association between acquisition of herpes simplex virus type 2 in women and bacterial vaginosis. Clin Infect Dis. 2003;37:319–25. doi: 10.1086/375819. [DOI] [PubMed] [Google Scholar]
- 5.Watts DH, Fazzari M, Minkoff H, Hillier SL, Sha B, Glesby M, et al. Effects of bacterial vaginosis and other genital infections on the natural history of human papillomavirus infection in HIV-1-infected and high-risk HIV-1-uninfected women. J Infect Dis. 2005;191:1129–39. doi: 10.1086/427777. [DOI] [PubMed] [Google Scholar]
- 6.Cohen CR, Duerr A, Pruithithada N, Rugpao S, Hillier S, Garcia P, et al. Bacterial vaginosis and HIV seroprevalence among female commercial sex workers in Chiang Mai, Thailand. AIDS. 1995;9:1093–97. doi: 10.1097/00002030-199509000-00017. [DOI] [PubMed] [Google Scholar]
- 7.van De Wijgert JH, Mason PR, Gwanzura L, Mbizvo MT, Chirenje ZM, Iliff V, et al. Intravaginal practices, vaginal flora disturbances, and acquisition of sexually transmitted diseases in Zimbabwean women. J Infect Dis. 2000;181:587–94. doi: 10.1086/315227. [DOI] [PubMed] [Google Scholar]
- 8.Martin HL, Richardson BA, Nyange PM, Lavreys L, Hillier SL, Chohan B, et al. Vaginal lactobacilli, microbial flora, and risk of human immunodeficiency virus type I and sexually transmitted disease acquisition. J Infect Dis. 1999;180:1863–8. doi: 10.1086/315127. [DOI] [PubMed] [Google Scholar]
- 9.McGroarty JA, Tomeczek L, Pond DG, Reid G, Bruce AW. Hydrogen peroxide production by Lactobacillus species: correlation with susceptibility to the spermicidal compound nonoxynol-9. J Infect Dis. 1992;165:1142–4. doi: 10.1093/infdis/165.6.1142. [DOI] [PubMed] [Google Scholar]
- 10.Moncla BJ, Hillier SL. Why nonoxynol-9 may have failed to prevent acquisition of Neisseria gonorrhoeae in clinical trials. Sex Transm Dis. 2005;32:491–494. doi: 10.1097/01.olq.0000170444.13666.e9. [DOI] [PubMed] [Google Scholar]
- 11.Barbone F, Austin H, Louv WE, Alexander WJ. A follow-up study of methods of contraception, sexual activity, and rates of trichomoniasis, candidiasis, and bacterial vaginosis. Am J Obstet Gynecol. 1990;163:510–4. doi: 10.1016/0002-9378(90)91186-g. [DOI] [PubMed] [Google Scholar]
- 12.Mauck CK, Weiner DH, Ballagh SA, Creinin MD, Archer DF, Schwartz JL, et al. Single and multiple exposure tolerance study of polystyrene sulfonate gel: a phase I safety and colposcopy study. Contraception. 2004;70:77–83. doi: 10.1016/j.contraception.2004.02.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.McGroarty JA, Reid G, Bruce AW. The influence nonoxynol-9-containing spermicides on urogenital infection. J Urol. 1994;152:831–3. doi: 10.1016/s0022-5347(17)32584-3. [DOI] [PubMed] [Google Scholar]
- 14.Stafford MK, Ward H, Flanagan A, Rosenstein IJ, Taylor-Robinson D, Smith JR, et al. Safety study of nonoxynol-9 as a vaginal microbicide: evidence of adverse effects. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;17:327–31. doi: 10.1097/00042560-199804010-00006. [DOI] [PubMed] [Google Scholar]
- 15.Watts DH, Rabe L, Krohn MA, Aura J, Hillier SL. The effects of three nonoxynol-9 preparations on vaginal flora and epithelium. J Infect Dis. 1999;180:426–37. doi: 10.1086/314881. [DOI] [PubMed] [Google Scholar]
- 16.Raymond EG, Chen PL, Luoto J. Spermicide Trial Group Contraceptive effectiveness and safety of five nonoxynol-9 spermicides: a randomized trial. Obstet Gynecol. 2004;103:430–9. doi: 10.1097/01.AOG.0000113620.18395.0b. [DOI] [PubMed] [Google Scholar]
- 17.Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol. 1991;29:297–301. doi: 10.1128/jcm.29.2.297-301.1991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Gupta K, Hillier SL, Hooton TM, Roberts PL, Stamm WE. Effects of contraceptive method on the vaginal microbial flora: a prospective evaluation. J Infect Dis. 2000;181:595–601. doi: 10.1086/315267. [DOI] [PubMed] [Google Scholar]
- 19.Hardin JW, Hilbe JM. Generalized estimating equations. Boca Raton (FL): Chapman & Hall/CRC; 2003.
- 20.Van Damme L, Chandeying V, Ramjee G, Rees H, Sirivongrangson P, Laga M, et al. Safety of multiple daily applications of COL-1492, a nonoxynol-9 vaginal gel, among female sex workers. COL-1492 Phase II Study Group AIDS. 2000;14:85–8. doi: 10.1097/00002030-200001070-00010. [DOI] [PubMed] [Google Scholar]
- 21.Kreiss J, Ngugi E, Holmes K, Ndinya-Achola J, Waiyaki P, Roberts PL, et al. Efficacy of nonoxynol 9 contraceptive sponge use in preventing heterosexual acquisition of HIV in Nairobi prostitutes. JAMA. 1992;268:477–82. [PubMed] [Google Scholar]