Abstract
This report summarises the evidence presented to a technical consultation on the safety of N-9 and its effectiveness for protection against pregnancy, sexually transmitted infections and HIV, organized by the World Health Organization Department of Reproductive Health and Research with CONRAD in October 2001. Key conclusions from the meeting include: 1. Although Nonoxynol-9 has been shown to increase the risk of HIV infection when used frequently by women at high risk of infection, it remains a contraceptive option for women at low risk. 2. Nonoxynol-9 offers no protection against sexually transmitted infections such as gonorrhoea or chlamydia. 3. There is no evidence that condoms lubricated with nonoxynol-9 are any more effective in preventing pregnancy or infection than condoms lubricated with silicone, and such condoms should no longer be promoted. However, it is better to use a nonoxynol-9 lubricated condom than no condom at all. 4. Nonoxynol-9 should not be used rectally.
Résumé
Ce rapport résume les données présentées lors d’une consultation technique sur la sécurité du N-9 et son efficacité pour éviter une grossesse, des infections sexuellement transmissibles (IST) et le VIH, organisée par le Département de santé et recherche génésiques de l’OMS et CONRAD en octobre 2001. Voici les principales conclusions de la consultation: 1. Bien que le Nonoxynol-9 accroisse le risque d’infection à VIH quand il est utilisé fréquemment par des femmes au haut risque, il demeure une option contraceptive pour les femmes à faible risque. 2. Le Nonoxynol-9 ne protège pas des IST comme la gonococcie et la chlamydiose. 3. Rien n’indique que les préservatifs lubrifiés au nonoxynol-9 soient plus efficaces pour prévenir les grossesses ou les infections que les préservatifs lubrifiés avec de la silicone, et ces préservatifs ne devraient plus être préconisés. Néanmoins il vaut mieux utiliser un préservatif lubrifié au nonoxynol-9 que pas de préservatif. 4. Le nonoxynol-9 ne doit pas être utilisé par voie rectale.
Resumen
Este informe resume las pruebas presentadas en una consulta técnica organizada por el Departamento de Salud e Investigación Reproductiva de la OMS con CONRAD, en octubre de 2001, acerca de la seguridad de N-9 y su efectividad para proteger contra el embarazo, las infecciones de transmisión sexual y el VIH. Las principales conclusiones de la reunión son: 1. Aunque nonoxynol-9 aumenta el riesgo de infección de VIH cuando lo usan mujeres que corren un alto riesgo de infección, sigue siendo una opción anticonceptiva para mujeres en bajo riesgo. 2. El nonoxynol-9 no ofrece ninguna protección contra las infecciones de transmisión sexual tales como la gonorrea o la clamidia. 3. No hay evidencia que los condones lubricados con nonoxynol-9 sean más efectivos que los condones lubricados con silicona para prevenir el embarazo o las infecciones, por lo tanto no se debe promocionar mas dichos condones. No obstante, es mejor usar un condón lubricado con nonoxynol-9 que no usar ningún condón. 4. No se debe usar nonoxynol-9 en el recto.
The spermicide nonoxynol-9 (N-9) has been widely available as a contraceptive for many years and has been shown to be effective against HIV in laboratory studies. If it also provided effective protection against HIV in clinical studies, N-9 could be made rapidly available to women who require protection. The World Health Organization Global Programme on AIDS (GPA) and the Joint United Nations Programme on HIV/AIDS (UNAIDS) sponsored a clinical trial of a gel containing N-9 to assess its effectiveness in protecting against HIV. Preliminary results from the study were presented in July 2000 at the 13th International AIDS Conference in Durban, South Africa, and showed, contrary to expectation, that the HIV incidence was higher in women using N-9 than in women using a comparison product. While a disappointment with regard to the rapid deployment of an effective microbicide, these results also raised questions about the safety of N-9 when used for its main indication, protection against unwanted pregnancy.
After presentation of the preliminary results from the study in July 2000, the World Health Organization (WHO) was approached to provide an assessment of the scientific information regarding the safety and effectiveness of N-9 when used for family planning purposes. This summary should permit Member States to assess the risks and benefits of N-9 use among women in their country who may be at risk of HIV infection from inadequately protected sexual activity. Accordingly, the WHO Department of Reproductive Health and Research (RHR) convened a Technical Consultation in October 2001, in partnership with the CONRAD Program, Arlington VA, USA, to review the available evidence and provide advice to member states on the use of N-9. The Consultation included experts from developed and developing countries with experience in product development, safety assessment and public health, and representatives from collaborating agencies. Reviews of key issues were commissioned prior to the meeting and are summarised in this report. The meeting also considered the submitted manuscripts from recently completed studies directly relevant to the safety and effectiveness of N-9.
This report summarises the evidence presented to the meeting on the safety of N-9 and its effectiveness for protection against pregnancy, sexually transmitted infections (STIs) and HIV. The meeting concluded with recommendations on the use of N-9 and identified key areas of uncertainty where more research was urgently required.
Review of evidence
| • | Use of nonoxynol-9 containing products in developed and developing countries | ||||
A background paper was prepared by Kirsten Vogelsong, WHO, and Jeff Spieler, USAID, on the use of N-9 containing products in developed and developing countries. It identified a total of 18 spermicidal products containing N-9 in different formulations (gels, creams, vaginal suppositories and film) each containing between 50 and 150 mg of N-9. In addition, spermicides containing other active ingredients, such as octoxynol-9, benzalkonium chloride and menfegol, are also distributed in developed and developing countries. The private sector sales and market share of different types of spermicide in developing countries were difficult to establish. Public sector sales through donors revealed very uneven distribution of spermicides and large variations year to year in procurement.
According to DHS data, fewer than 1% of women in Asia reported ever using spermicides. These methods are more commonly used in Latin America than in other parts of the world. While between 5% and 10% of women reported ever using spermicides, current use was less than 1%. The highest reported use was Trinidad and Tobago where 16.6% of women reported ever use and 3.5% current use. Reported use by unmarried women and women under age 25 years was higher than among married women.
In the United States, based on data from the National Survey of Family Growth (NSFG) in 1995, 4.6% of women who used any contraceptive method used spermicide (alone or with a barrier method). It is not known what proportion of spermicide use refers to different formulations, but all spermicides in the USA contain N-9. The majority of the reported use is in combination with a physical barrier method, such as the diaphragm or cervical cap. Current use of these products fell by more than 50% since the earlier round of the NSFG conducted in 1988 where 10.6% of respondents reported current use. The trend of decreasing spermicide use in the USA has also been seen in decreasing private sector sales in other parts of the world. Comparable data on trends in the use of N-9-lubricated condoms were not available. NSFG data did not show a decreasing trend in use with increasing age, nor preferential use among unmarried women.
The characteristics of spermicides make them particularly attractive to unmarried women and young people. They are available over-the-counter (OTC), in addition to traditional family planning service delivery points, which distribute provider-dependent methods. Spermicides thus represent one of the few options to reduce the risk of unwanted pregnancy for young unmarried women who have poor access to provider-dependent methods.
Although the analysis of survey data and private and public sector distribution provide some information on the extent of use of spermicides, they are not sufficiently specific to answer all questions. In particular, data are lacking on coital frequency and the use of spermicides, and other characteristics of typical spermicide users.
| • | Toxicology of nonoxynol-9 | ||||
A summary of issues in the toxicological assessment of microbicides, and N-9 in particular, was presented by Ralph Heywood, consultant toxicologist, UK. The main challenge in the toxicological assessment of topical microbicides in general, and N-9 in particular, is to balance any irritation, which may be associated with increased risk of lesions and susceptibility to infection, against efficacy. Toxicological assessments allow evaluation of the former. Specifically, it is necessary to determine a dose sufficient to have the desired efficacy while not increasing risk. The balance is further complicated by the fact that the mechanism of action against each outcome (sperm, bacteria, and viruses) differs, as does the minimally effective dose. All products, even inert devices, cause some degree of epithelial disruption and may alter the vaginal microflora. They thus have the potential to increase penetration by infectious elements. This is particularly an issue for surfactants such as N-9.
A related toxicological concern is clearance: how long the product remains in the vaginal vault after initial application and the effects of retained material. Such issues have not been adequately studied for N-9, which was never tested using current standards of toxicological assessment.
In addition to the specific toxicological questions regarding N-9, there is inadequate information of the safety of the excipients used to formulate the product. This applies to other surfactants as well as to other OTC products, including vaginal and rectal lubricants.
| • | Clinical safety data on nonoxynol-9 | ||||
A summary of clinical safety data on N-9 was prepared by Christine Mauck, CONRAD, USA. Safety data with respect to epithelial disruption were reviewed from a total of 16 studies on four different N-9 formulations: suppository, sponge, gel, and film. The studies differed with respect to their purpose, when they were conducted, the doses and concentrations of N-9 studied, formulation, frequency and duration of use, sample sizes, target populations, comparison products, rules regarding intercourse, and the means of assessing endpoints. Cross-study comparisons were therefore difficult.
A summary of the evidence regarding the clinical safety of N-9 showed a trend towards a greater frequency of disruption with greater frequency of use and higher doses of N-9. This trend was most evident in studies of suppositories and the sponge (although data on the sponge are quite limited), where the incidence of epithelial disruption was greater with more frequent use. In general, the literature supports the conclusion that infrequent use of products containing low doses of N-9 is probably safe.
The summary also revealed major obstacles to interpreting the available data on the clinical safety of N-9. In particular, it is unclear what should be considered normal with respect to signs of epithelial disruption and it is not possible to distinguish between epithelial changes resulting from sexual intercourse and the impact of the study product, despite randomised studies in commercial sex workers masked to the exact product used. There is poor correlation between clinical findings and self-reported symptoms, and clinical findings and laboratory markers of inflammation. In addition, the clinical significance of signs and symptoms of disruption is not known, particularly with regard to any increased risk of acquiring an STI or HIV. Furthermore, it is not known how signs and symptoms of disruption correlate with product acceptability and use.
Studies of other surfactants such as benzalkonium chloride and menfegol have shown that these products are also associated with epithelial disruption.
| • | Effectiveness of nonoxynol-9 for pregnancy prevention | ||||
A summary of the available data on the contraceptive effectiveness of N-9 was presented by James Trussell, Princeton University. Most studies that have evaluated the contraceptive effectiveness of N-9 were poorly conducted and few were randomized controlled trials. In particular, not all studies had comparison groups, loss to follow-up was often high, and efficacy under conditions of perfect use could not be reliably assessed due to lack of data on consistent and correct use over the full duration of the study. More reliable estimates are available for typical use of the products. In addition, the dose of the product under study, the frequency of product use, and the composition of the target populations (age, level of sexual activity and risk) varied across studies, making comparisons almost impossible.
The typical use pregnancy rate data on spermicides used alone come from 31 clinical trials with exposures ranging from 3 to 67 months, which reported Pearl pregnancy rates, and four clinical trials and six surveys, which reported cumulative 12-month life-table pregnancy rates (Table 31-3 in Contraceptive Technology). The typical use pregnancy rates (expressed either as pregnancies per 100 woman-years or as 12-month cumulative life table rates) ranged from less than 2% in nine studies, 2–5% in four studies, 5–10% in eight studies, 10–15% in three studies, 15–30% in eleven studies, and between 37% and 59% in six studies. Not one of these studies met modern standards of study design, execution, analysis or reporting.
The only study using modern standards of design, execution and analysis involved 765 women randomly assigned to Vaginal Contraceptive Film (VCF) or Conceptrol foaming tablets. The study subjects were in the age range 18–35 years (58% less than 25 years), had no history suggestive of subfecundity, and had a high coital frequency (71% reported more than 10 acts of intercourse per month). The percentages of women becoming pregnant within 6 months under conditions of typical use were 24.9% (95% confidence interval 20–30%) for VCF and 28.0% (23–34%) for Conceptrol. The corresponding consistent use pregnancy rates were 24.1% (19–30%) and 27.9% (22–34%), respectively. The contraceptive efficacy for the two products was similar, but was associated with high rates of pregnancy in this young, sexually active population. It is difficult to interpret these results since there was no comparison group using a method with known efficacy.
The only study to evaluate the ability of Advantage-24 contraceptive gel to inhibit sperm transport and survival involved the application of gel 15–30 minutes, 12 hours, and 24 hours before intercourse. The authors concluded that effectiveness may decrease when it is applied more than 15–30 minutes before intercourse. However, postcoital tests were not performed until 8–12 hours after intercourse and the results could have been worse if postcoital tests had been performed earlier. In addition, there was no control time or assessment in the study, other than the use of gel applied 15–30 minutes before intercourse. Moreover, it is not known how well the postcoital test among normal fertile couples can reflect typical or perfect use pregnancy rates.
One study of spermicidally-lubricated condoms reported a 2.1% typical-use 12-month probability of pregnancy, but 26% of the men were in the age range 40–44 years, 27% in the range 45–49 years and 24% were over 50 years. The study subjects had a low coital frequency and their female partners were well beyond the period of peak fecundity. In addition, only pregnancies leading to a live birth were reported.
In summary, the data support the conclusion that use of spermicide alone reduces risk of pregnancy compared with use of no product, despite the limitations of the studies. Although spermicides in the USA are mainly used with barrier methods, the increased benefit of such combinations has not been well established. However, existing data do not support any added benefit of male condoms lubricated with N-9 compared with other lubricants. Data quantifying the contraceptive effects of N-9 in various formulations and doses, used with and without barriers, are clearly needed and studies to address this deficit are currently under way.
| • | Effectiveness of nonoxynol-9 for prevention of sexually transmitted infections | ||||
A review of the effectiveness of N-9 for prevention of sexually transmitted infections was given by Ward Cates (Family Health International). This was supplemented with data from two recently completed but (at the time of the meeting) not yet published trials presented by the principal investigators, Ron Roddy (Family Health International) and Lut Van Damme (Institute of Tropical Medicine, Antwerp, and CONRAD). The Roddy study was an open-label, randomised trial comparing 100 mg N-9 gel plus condoms to condoms alone among 1251 women attending STI clinics in Cameroon. The relative risk in the gel group compared with the condom group was 1.5 (1.0–2.3) for gonococcal infection and 1.0 (0.7–1.4) for chlamydial infection. There were 5 new cases of HIV infection in the gel group and 4 in the condom group. The Van Damme study was a triple-blind randomized controlled trial comparing a 52.5 mg N-9 gel (COL-1492) with Replens®, an OTC vaginal moisturiser. The study population consisted of 892 sex workers from Benin, Côte d’Ivoire, South Africa and Thailand. There was little effect of N-9 on the incidence of gonococcal (RR 1.2 [0.9–1.6]) or chlamydial infection (RR 1.2 [0.8–1.6]). However the incidence of HIV infection among N-9 users was 48% higher than among placebo users (14.7 compared with 10.3 cases per 100 woman-years, adjusted relative risk 1.5 [1.0–2.2]). For women who reported using more than 3.5 gel applications per working day the risk of HIV infection among COL-1492 users was almost twice (RR 1.8 [1.0–3.2]) the risk among placebo users. The risk did not differ between the two treatment arms among women using the gel less frequently. Higher frequency of use of the study products was also associated with a higher incidence of genital lesions in both study groups, with a greater incidence ratio among N-9 users compared with placebo. These results were dominated by one of the centres which contributed 69% of the HIV seroconversions observed during the study. Considerable discussion at the meeting focused on the validity of self-reported condom use and coital diaries on which the dose-response analyses were based. Potential limitations of the study, such as the low follow-up rate and the use of Replens® as the control product, were also discussed.
A systematic review of randomized controlled trials presented by David Wilkinson, University of South Australia, confirmed that there was no evidence of any protective effect of N-9 against STIs. Using Cochrane procedures for meta-analysis of randomized trials, data from nine studies including almost 5000 women were included. Confidence limits for all outcomes (gonorrhoea, trichomonas, bacterial vaginosis, and candidiasis) all included 1.0.
The limitations of currently available data were recognized at the meeting. In particular, most of the data on the effectiveness of N-9 in protecting against STIs and HIV come from studies among sex workers or other high risk groups. Few data are available from women who use the product at low frequencies, who are at low risk of infection, or from women who use N-9 alone. In addition it is not known whether adding N-9 to barrier methods such as the cervical cap or the diaphragm has any protective effect against sexually transmitted infections.
| • | Safety of rectal use of nonoxynol-9 | ||||
David Phillips, Population Council, New York, presented data on the safety of rectal use of N-9. Rectal application of two N-9 containing products (K-Y Plus and ForPlay®) was evaluated in both a mouse model and human subjects (n=4); two formulations not containing N-9 (carrageenan and methyl cellulose) were used as controls. In both studies, rectal lavage (saline) revealed significant sloughing of sheets of epithelium. However, no rectal epithelial cells were found in the baseline lavage specimens, any of the lavages involving non-N-9-containing formulations, or, most importantly, in rectal lavage specimens collected 8 hours post N-9 product use, suggesting that any damage had been repaired. Nonetheless, the possibility of increased risk of infection soon after the application of products containing N-9 seems quite likely, even for lubricants not advertised to be microbicidal. More research is urgently required on the safety of rectal use of N-9 and other lubricants, as well as further understanding of the time course for the rectal damage to repair.
| • | The regulator’s perspective | ||||
Helen Rees, University of the Witwatersrand and former chair, South Africa Medicines Control Council gave an overview of the problems of developing country regulators when faced with the new data on the safety of N-9. She emphasised the complexity of regulating OTC products with respect to labelling and instructions for use. Products available and effective for one indication, for example contraception, can have unintended adverse effects on other outcomes such as STIs and HIV infections. One solution for products such as N-9 which may be associated with risk under particular conditions of use, might be to change the OTC status or to specify instructions according to risk. However, because risk involves both individual behaviour (number of sexual partners, frequency of sexual intercourse, whether or not condoms are used) as well as the local prevalence of infection, risk differs for different settings and for different people. Furthermore, recommendations for product use, even as a contraceptive, should depend on the availability of other products, which also varies. As a result, drug approval based on one or two well-conducted studies may not be sufficient for universal guidelines. In fact, guidelines for product use, in this case N-9 containing products, might need to be country-specific. Lack of post-marketing surveillance of OTC products in general, further exacerbates any evaluation of N-9 containing products.
Conclusions
Effectiveness of nonoxynol-9 for pregnancy prevention
| 1. | Nonoxynol-9 (N-9) when used alone is moderately effective as a contraceptive. | ||||
| 2. | For women who choose to use N-9 alone in preference to other methods, it is better than no contraceptive method at all. | ||||
| 3. | When used with a female mechanical barrier method (for example a cervical cap or diaphragm) N-9 is more effective than when used alone. | ||||
| 4. | N-9 is available in a number of different formulations (film, sponge, gel, suppository, and foam) and doses. It is not known whether contraceptive effectiveness differs according to formulation or dose. | ||||
| 5. | There is no evidence that condoms lubricated with N-9 are more effective in preventing pregnancy than lubricated condoms without N-9. | ||||
| 6. | Limited evidence suggests that the contraceptive effectiveness of the diaphragm and cervical cap may be moderately more effective when used with a spermicide than without. | ||||
Safety of nonoxynol-9
| 1. | N-9 is a surfactant which disrupts the cell membrane. It has been shown to be an irritant in animal as well as human models. | ||||
| 2. | N-9 has been shown to cause epithelial disruption in the vagina and rectum. | ||||
| 3. | Increasing frequency of use of N-9 suppositories increases the risk of vaginal epithelial disruption. No other studies designed to look specifically at the effect of frequency of use with other formulations of N-9 have been published. However observational studies have shown increased rates of epithelial disruption with greater frequency of use of N-9. | ||||
Effectiveness of nonoxynol-9 for prevention of sexually transmitted infections including HIV
| 1. | There is good evidence that N-9 does not reduce the risk of sexually transmitted infections (STIs) or HIV among sex workers nor women attending STI clinics. | ||||
| 2. | Most of the clinical evidence on the risk of HIV infection with use of N-9 comes from studies conducted among women who were sex workers or women attending STI clinics. | ||||
| 3. | No studies on the effectiveness of N-9 to prevent infection with STI or HIV have been conducted in other groups of women. | ||||
Recommendations
These recommendations are based on the above evidence and on consideration of the balance of risks and benefits to individual users. Family planning programmes in different countries and communities must consider these recommendations in the light of their particular circumstances, including the prevalence of HIV and sexually transmitted infections, and the range of available contraceptive options.
Nonoxynol-9 for STI prevention
N-9 should not be used for the purpose of STI or HIV prevention. Condoms should always be used to prevent infection.
Nonoxynol-9 for contraception
Among women at low risk of HIV infection, the use of N-9 remains a contraceptive option. Although its effectiveness is low compared with other contraceptive methods, it is generally easily available, can be obtained OTC without requiring a medical consultation or prescription, and is a method under the control of the woman.
| 1. | Since high frequency use of N-9 products may cause epithelial damage and increase the risk of HIV infection, women who have multiple daily acts of intercourse should be advised to choose another method of contraception. | ||||
| 2. | Women at high risk of HIV infection should not use N-9 for contraception, since N-9 may increase the risk of infection. | ||||
Rectal use of nonoxynol-9
N-9 should not be used rectally.
Use of nonoxynol-9-lubricated condoms
There is no published scientific evidence that N-9-lubricated condoms provide any additional protection against pregnancy or STIs compared with condoms lubricated with other products. Since adverse effects due to the addition of N-9 to condoms cannot be excluded, such condoms should no longer be promoted. However, it is better to use N-9-lubricated condoms than no condoms.
Key areas of uncertainty and the need for further research
| 1. | The safety concern about N-9 may also apply to other spermicide products marketed for contraception. Their safety should be formally assessed as a matter of urgency. This also applies to vaginal or rectal lubricants which contain N-9. | ||||
| 2. | The safety of lubricants containing only ingredients thought to be inactive which are used vaginally and/or rectally needs to be assessed as a matter of urgency. | ||||
| 3. | There is an urgent need to develop safe lubricants for rectal use. | ||||
| 4. | The precise clinical significance of epithelial disruption observed in safety studies is not known. | ||||
| 5. | There are currently no published studies on the effect of N-9 among women with HIV infection. | ||||
| 6. | Continued work is urgently required to develop safe and effective microbicides. | ||||
Acknowledgements
The financial support of the United States Agency for International Development is gratefully acknowledged by WHO in the report.
Notes
☆ Excerpted from the report with kind permission of WHO. Citation[1]
References
- Full report and references available at: www.who.int/reproductive-health/rtis/N-9_meeting_report.doc