BACKGROUND: There is no consensus concerning the frequency of screening Pap tests in postmenopausal women. High false-positive rates for abnormal test results in this group increase cost, risk, and discomfort of invasive procedures, and anxiety. In postmenopausal women, the majority of abnormal Pap test results are composed of atyptical squamous cells of uncertain significance (ASCUS) and data conflict as to whether hormone replacement therapy (HRT) has any effect on its development.
POPULATION STUDIED: From 20 outpatient clinical centers, 2561 postmenopausal women were recruited from screening interviews of 68,561 women. The criteria for entering the study were age younger than 80 years, an intact uterus, and a normal Pap test result at baseline. The mean age was 66.7 years, and all had coronary artery disease.
STUDY DESIGN AND VALIDITY: This is a 2-part study using a prospective cohort design to calculate the incidence of cytologic abnormalities and a randomized double-blinded placebo-controlled design to evaluate the effect of HRT. Cervical tests were performed annually in participating clinics, with all tests evaluated at the same central pathology laboratory. Abnormal Pap test results were managed in the conventional manner. For each cytologic abnormality the final histologic diagnosis was considered normal only if all follow-up tests were normal or proved negative by colposcopy. In a second aspect of the study, the participants were randomly assigned to receive placebo or the combination of oral conjugated equine estrogens 0.625 mg plus medroxy- progesterone acetate 2.5 mg daily.
OUTCOMES MEASURED: The primary outcomes were the incidence over 2 years of new cervical cytologic abnormalities (ASCUS, atypical glandular cells of undetermined significance [AGCUS], low-grade squamous intraepithelial lesion [LGSIL], and high-grade squamous intraepithelial lesion [HGSIL]), and final histologic diagnoses, with the main outcome being the actual diagnosis of cervical cancer.
RESULTS: The incidence of new cytologic abnormalities in the first year was 78 women (3%), and in the second year there were an additional 32 women (1.4%). The total incidence of abnormal Pap results for the 2 years following a normal test was 110, or 2.3%. Most of Pap abnormalities were reported as ASCUS (67.3%) or AGCUS (21%); 10.9% were LGSIL; and 0.9% were HGSIL. More important, no women in the first year were found to have high-grade cervical intraepithelial neoplasia (CIN). In the second year, one patient had a CIN grade 1 or 2 lesion. Therefore, the positive predictive value of any abnormal test identified 1 year after an initial normal test was 0% (95% confidence interval [CI], 0%-5%) and within 2 years was 0.9% (95% CI, 0%-3%). Hormone therapy did not show any significant effect on the incidence of cervical cytologic abnormalities.
Although a small percentage of Pap tests performed 1 and 2 years after a normal Pap test result in postmenopausal women will show minor abnormalities, the positive predictive value for discovering high-grade lesions is only 0.9%. In other words, 99.1% of the positive Pap test results in this population will be falsely positive. Therapy with hormone replacement did not affect these results. Therefore, routine cervical tests should not be performed within 2 years of normal cytology in postmenopausal women, regardless of whether they are taking HRT. Most clinicians already recommend extending Pap tests to 2 or 3 years for this population, and this study lends confidence to this trend in practice. For postmenopausal women with no previous Pap screening or with previous abnormal test results, the frequency of screening should continue to depend on the individual case or risk factors. Women choosing to have more frequent routine screening should be informed that they are much more likely to undergo further unnecessary diagnostic studies for false-positive results.