Chlamydial cervicitis: testing the practice guidelines for presumptive diagnosis.

OBJECTIVE: To test the recommendation from the Canadian guidelines for sexually transmitted diseases (STDs) that mucopurulent endocervical discharge and 10 or more polymorphonuclear leukocytes (PMNs) per high-power field of a Gram-stained endocervical smear or, when Gram staining is not possible, the presence of endocervical discharge and one of edema, erythema or induced mucosal bleeding of the cervix can be considered diagnostic for chlamydial cervicitis.
METHODS: A total of 596 consecutive women attending 2 family planning clinics for routine care underwent vaginal speculum examination and were tested for Chlamydia trachomatis and Neisseria gonorrhoeae. PMN counts from Gram-stained endocervical smears and the presence or absence of putative indicators of chlamydial infection were recorded.
RESULTS: The prevalence of chlamydial cervicitis was 6.2% (37/596), and no women tested positive for N. gonorrhoeae. Presumptive diagnosis of chlamydial cervicitis based on the guidelines criteria of mucopurulent endocervical discharge and 10 or more PMN per high-power microscopic field had a sensitivity and specificity of 18.9% and 97.0% respectively, a positive predictive value of 29.2% and a positive likelihood ratio (LR) of 6.2 (p = 0.003). Presumptive diagnosis based on endocervical discharge with edema, erythema or induced mucosal bleeding of the cervix had a sensitivity and specificity of 43.2% and 80.0% respectively, a positive predictive value of 12.5% and a positive LR of 2.2 (p = 0.002). In the presence of bacterial vaginosis or vaginitis, the LR for the criteria of mucopurulent endocervical discharge and 10 or more PMN per high-power field was 5.4 (p = 0.04), whereas the LR was 4.3 (p = 0.10) if bacterial vaginosis and vaginitis were absent.
CONCLUSIONS: In this setting, the practice of making a presumptive diagnosis of chlamydial cervicitis on the basis of the criteria given in the Canadian STD guidelines was not supported.