Vaginal washing increases HIV infection risk; results for dry sex less clear-cut

This article is more than 15 years old. Click here for more recent articles on this topic

A meta-analysis of data on vaginal practices and HIV infection from ten African cohorts has found that both vaginal washing with soap and wiping the vagina with cloths, tissues or paper were associated with an increased risk of acquiring HIV, researchers reported at the Fifth International AIDS Society conference in Cape Town last week. The use of products to dry or tighten the vagina, often referred to as ‘dry sex’, did not however have a statistically significant association with HIV infection.

Globally, there is a wide variety of vaginal practices and products used by women to clean, tighten, dry or warm the vagina. Women’s efforts to change their genital environment can undermine the body’s innate defences against pathogens. In particular, vaginal practices have been linked to disruption of the vaginal epithelium and loss of lactobacilli (i.e. an increase in bacterial vaginosis, which is also associated with an increased risk of HIV infection).

In resource-limited settings with strong imbalances in power between men and women, women may be motivated to adopt vaginal practices which enhance men’s sexual pleasure in order to maintain economically essential relations with husbands or other sexual partners. Other reasons to engage in these practices include genital hygiene, self-treatment of vaginal discharge, and pregnancy prevention.


multivariate analysis

An extension of multivariable analysis that is used to model two or more outcomes at the same time.

bacterial vaginosis

A condition caused by the overgrowth of certain species of the bacteria that are normally present in the vagina.


Expresses the risk that, during one very short moment in time, a person will experience an event, given that they have not already done so.

hazard ratio

Comparing one group with another, expresses differences in the risk of something happening. A hazard ratio above 1 means the risk is higher in the group of interest; a hazard ratio below 1 means the risk is lower. Similar to ‘relative risk’.

statistical significance

Statistical tests are used to judge whether the results of a study could be due to chance and would not be confirmed if the study was repeated. If result is probably not due to chance, the results are ‘statistically significant’. 

Partly because of the wide variety of vaginal practices in different communities, and because of differences in the way in which practices have been recorded by researchers, epidemiological studies have so far produced conflicting findings on whether vaginal practices are associated with an increased risk of women acquiring HIV.

An international team of researchers therefore took the data from ten previous studies, and conducted an individual patient data meta-analysis. Most meta-analyses involve combining the end results from a number of different studies, whereas the approach taken here involved re-coding and re-analysing the data for each individual who had participated in the ten separate studies. This allowed for vaginal practices to be recorded according to a standardised classification system and for new analyses to be carried out.

The analysis from ten cohorts in six sub-Saharan African countries includes 16,307 women, of whom 865 contracted HIV. Six cohorts were of women in the general population; four were of commercial sex workers or women working in high-risk settings. There was a wide variation in the prevalence of both vaginal practices and bacterial vaginosis in these cohorts.

Compared to women who had no vaginal practices or who cleaned the vagina with water alone, women who cleaned with soap had an increased risk of acquiring HIV. In multivariate analysis (which controlled for age, marital status and number of sexual partners - the only factors found to skew the result), the hazard ratio was 1.22 (95% confidence interval 1.00-1.49).

Women who used cloths, tissues or paper to wipe the vagina and remove moisture also had an increased risk of infection. In multivariate analysis, the hazard ratio was 1.41 (95% confidence interval 1.03-1.77). The researchers suggest that some items used intravaginally may be more harmful than others and could cause abrasions.

As the lower confidence interval for the last two factors was above 1.00, the results are considered statistically significant (i.e. unlikely to be produced by chance alone).

However in the case of the use of products to dry or tighten the vagina, the hazard ratio was 1.23 (95% confidence interval 0.97-1.56). The lower confidence interval of 0.97 means that this result is not statistically significant, but Matthew Chersich, presenting, said that the findings did suggest that these practices are harmful.

The researchers hypothesised that vaginal practices may cause an increase in bacterial vaginosis, which in turn leads to an increased risk of HIV acquisition. However, when bacterial vaginosis was added to the multivariate model, it did not substantially affect the results.

Matthew Chersich said that although the meta-analysis had attempted to re-classify previous studies’ data using a standardised classification, the questions used in the studies were often ambiguous. He called for more research which asked more precise questions about different practices and products. He also recommended evaluation of interventions encouraging women to switch to less harmful practices, such as washing with water alone.

Concentrating on the finding that washing with soap is associated with an increased risk of infection, he noted: “The effect size is modest, but as this is a very common practice, at a population level a small effect would have important implications”.

Further information

A powerpoint presentation by Matthew Chersich and a webcast of this conference session are available on the IAS 2009 website.