Advancing Sexual and Reproductive Health and Rights
International Family Planning Perspectives
Volume 30, Number 3, September 2004

In West Africa, Vaginal Discharge Is Not Usually Caused by Gonorrheal or Chlamydial Infection

In West Africa, women with vaginal discharge who are not sex workers and do not have abdominal pain are unlikely to have a cervical gonorrheal or chlamydial infection.1 According to an analysis of data from more than 700 women in five countries, only 2% were infected with gonorrhea and 3% with chlamydia. The prevalence of these infections among women who still experienced discharge after treatment for vaginal infections (based on the assumption that this was causing the symptom in most of the women) was no higher than the prevalence before this treatment, leading the researchers to conclude that vaginal discharge does not seem to be associated with gonorrheal and chlamydial infections in this population.

Data for analysis came from women visiting 11 health facilities in Benin, Burkina Faso, Ghana, Guinea and Mali whose main symptom was vaginal discharge. Women were excluded if they were sex workers, had abdominal pain or were pregnant. After providing demographic, behavioral and health information, the participants received a pelvic examination, and cervical specimens were collected and tested for gonorrhea, chlamydia and trichomonas. Because vaginal discharge is often caused by bacterial vaginosis, trichomoniasis or candidiasis, all of the women were treated with a single 2 g dose of metronidazole and with clotrimazole vaginal cream for three days; half were asked to return a week later if their discharge had not improved, while the other half were asked to return a week later regardless of their symptoms. At the return visit, all women in the former group and those in the latter group who had experienced no or only a partial response were treated for gonorrhea and chlamydia with ciprofloxacin and doxycyclin.

Half of the 726 women included in analyses were 21-29 years old. The median duration of vaginal discharge was 30 days, and one-third of women had previously been treated. Most women (82%) had had one sexual partner in the three months before the visit, and 17% had had a new sexual partner during that time.

Overall, only 2% of the women had a gonorrheal infection; the proportion who were infected was significantly lower in Ghana (0%), Benin (1%) and Burkina Faso (1%) than in Guinea (4%) or Mali (4%). The prevalence of gonorrhea was significantly higher among women who had pain during urination than among those who did not (4% vs. 1%), and was also higher among women who experienced pelvic pain during examination than among other women (7% vs. 2%).

Just 3% of women were infected with chlamydia, a proportion that did not vary signifcantly across countries. Women who had had no formal education were significantly more likely to have a chlamydial infection than were those who had had at least some education (7% vs. 3%). Prevalence rose from 0% among women who had had no partners in the past three months to 3% among those who had had one and 16% among those who had had two or more. In addition, infection with chlamydia was more common among women who had pain on intercourse than among those who did not (6% vs. 2%), and was more frequent among women for whom pus was found on the cervical swab than among other women (7% vs. 3%).

In all, 4% of women had one or both infections. Overall prevalence was significantly lower in Ghana and Benin (2%)—where large-scale interventions for sex workers have been operating for several years—than in Burkina Faso, Mali and Guinea (6%), where such large-scale programs do not exist. The risk factors were the same as those for chlamydial infections.

The proportion of women returning one week after the initial visit was significantly higher in the group told to come back regardless of symptoms than in the group told to come back only if the discharge had not improved (62% vs. 38%). In both groups, however, women who did not have cervical infections at the initial visit were as likely as those who did to return.

At the return visit, women asked to come back if their symptoms had not improved and women asked to come back regardless were similarly likely both to have had a complete response (39% and 47%, respectively) and to have had a partial response (59% and 53%) on the basis of self-reports. Within each treatment group, rates of responses did not differ significantly between women who did and did not have cervical infections at the initial visit. Likewise, in terms of objective responses, within each treatment group, women who had a cervical infection were as likely as women who did not to have a vaginal discharge at the return visit.

The strategy of asking women to return if treatment for vaginal infections did not improve their symptoms and treating all who returned resulted in treatment of 60% of women with cervical infections (sensitivity) and avoided treatment of 64% of women without such infections (specificity); however, only 5% of women treated actually had cervical infections (positive predictive value). The corresponding values for the strategy of asking women to return regardless of response and treating those who had no or only a partial response were 23%, 66% and 4%. In addition, the set of risk factors commonly used for presumptive treatment of cervical infections in women with vaginal discharge had respective values of 38%, 75% and 6%.

The findings of this and other studies suggest that the symptom of vaginal discharge in West African women is not usually caused by gonorrheal or chlamydial infections, the authors say. Furthermore, continuing to search for predictive factors for these infections among women in the region with vaginal discharge is "futile" in light of the low prevalence observed. They suggest instead that "women who are not [sex workers] who present with vaginal discharge without abdominal pain should be treated only for agents of vaginitis." They also call for intensified efforts to reduce gonorrheal and chlamydial infections among sex workers, noting that data suggest that such efforts eventually lead to lower rates of cervical infection among women in the general population.

—S. London


1. Pépin J et al., Low prevalence of cervical infections in women with vaginal discharge in west Africa: implications for syndromic management, Sexually Transmitted Infections, 2004, 80(3):230-235.