Vaginal infections in diabetic women: is empiric antifungal therapy appropriate?

Diabetes mellitus is reported to predispose women to vaginal candidiasis and hence patients attending busy diabetic clinics are often treated empirically with antifungal agents for genital symptoms.

To investigate the etiology of vaginal infections in diabetic women and to determine appropriateness of empiric antifungal therapy for symptomatic women.

Prospective study performed on consecutive patients attending two diabetic clinics.

Diabetic clinics serving developing communities belonging to two ethnic groups.

Two hundred and one women (101 African and 100 Indian) comprising 90 women with symptoms of pruritus vulvae and/or vaginal discharge and 111 asymptomatic women.

Candidiasis (50% vs. 24%; P = 0.012) and bacterial vaginosis (28% vs. 8%; P = 0.017) occurred significantly more frequently in symptomatic African women compared to asymptomatic African women. Among Indian women bacterial vaginosis (28% vs. 5%; P = 0.026) occurred more frequently in symptomatic women. The prevalence of trichomoniasis was similar in symptomatic and asymptomatic women of both ethnic groups. Mixed vaginal infections were detected significantly more often in African compared to Indian women (24% vs. 5%; P = 0.03).

In diabetic women with genital symptoms, an attempt at diagnosis should be made prior to commencement of therapy. However, in busy clinics with overstressed facilities where investigations cannot be performed, the use of empiric antifungal therapy alone is not appropriate and consideration should be given to the use of an antifungal plus a nitroimidazole agent which would be effective for both trichomoniasis and bacterial vaginosis.