In developing countries without access to laboratory testing, the syndromic management of cervicitis in women with lower genital tract symptoms results in overtreatment. We evaluated the validity of using different age cut-offs to predict the risk of cervical infection in women attending a clinic for sexually transmitted diseases in Cape Town. Specimens from 161 women with lower genital tract symptoms were examined microbiologically for evidence of cervicitis (gonorrhoea and/or chlamydial infection). Women with cervicitis (median 21 years) were significantly younger than those not infected (median 23 years). Retrospective simulation using the age cut-offs (< 45 years (current approach), < 30 years, and < 25 years) to predict cervicitis showed that specificity could be increased (overtreatment reduced) from 17% to 50% if presumptive diagnoses of cervicitis were restricted to women aged under 25 years. Lowering the age cut-off would severely limit sensitivity, however, and as many as 9% and 20%, respectively, of infected women would go untreated if the age cut-offs < 30 and < 25 years were used. The use of a lower age cut-off would therefore serve neither individual nor public health in the presumptive diagnosis and treatment of cervicitis in a high risk population, and would result in a substantially lower detection rate.