Abstract
Our objective is to compare the efficacy of using Papanicolaou (PAP)-stained cervical cytology smears with a standardized method of interpreting Gram-stained vaginal smears for the diagnosis of bacterial vaginosis (BV) in pregnancy.
High vaginal smears were Gram-stained and examined by a single observer to characterize 3 grades of vaginal flora and diagnose BV. Cervical smears were PAP-stained and examined for characteristic patterns of vaginal flora including evidence of BV by either a number of cytotechnicians or a single cytopathologist. The results of the 2 methods were compared. Seven hundred and forty-seven women attending an antenatal clinic in a district general hospital who consented to have a smear of vaginal secretions and cervical cytology in early pregnancy. The main outcome measure is the diagnosis of BV by different methods in a pregnant population.
Compared with the Gram-stain method for the diagnosis of BV, there was good agreement between PAP-stain interpretation by a single observer but the agreement was not as good with PAP-stain interpretation by multiple cytotechnicians. When the grades were consolidated to normal (grade I) and abnormal flora (grades II and III), compared to Gram-stained smears, PAP cytology undertaken by several cytotechnicians had a sensitivity of 80.7% and a specificity of 90.7%. The sensitivity and specificity increased to 87% and 97%, respectively, when the PAP-stained smears were read by a single cytopathologist. Using kappa scores, only those readings made by a single cytopathologist were reliable.
The setting in a cytopathology laboratory comprises multiple cytotechnicians, so that PAP-stain analysis of vaginal smears for the diagnosis of BV is likely to provide results which are less reliable than those obtained by Gram staining. The latter should be the first choice and every effort should be made to set up this service.
Get full access to this article
View all access options for this article.