A randomized trial of three methods of obtaining Papanicolaou smears.

The relationship between technique of obtaining Papanicolaou smears, presence of endocervical cells, and rate of cervical neoplasia was studied by comparing an endocervical and ectocervical nylon brush (Bayne brush), Ayre spatula plus endocervical brush, and spatula plus cotton-tipped swab in a randomized, prospective trial involving 11,061 patients. Eligible patients had a cervix and were not pregnant. Clinic records of patients with abnormal cytology were reviewed to determine the pathologic diagnosis. Whether pathology was defined as including condyloma, dysplasia, and cancer; dysplasia and cancer; moderate dysplasia, severe dysplasia, and cancer; or just severe dysplasia and cancer, no significant difference was found in the rates of pathology between the three techniques. Endocervical cells were identified in 89.5% of smears obtained with the Bayne brush, in 91.5% with the spatula plus endocervical brush, and in 71.1% with the spatula plus cotton-tipped swab (P less than .001). Among smears obtained with the spatula plus swab, the rate of any pathology was higher in smears that contained endocervical cells than in smears in which endocervical cells were absent (2.0 versus 0.6%; P = .009). After correction for the influence of age, there remained predictive value with the presence of endocervical cells. Once corrected for the influence of age, the rate of pathology and abnormal cytology in smears obtained with the spatula plus brush or the Bayne brush was not dependent upon the presence of endocervical cells.(ABSTRACT TRUNCATED AT 250 WORDS)