Objective: To determine if repeating a smear improves detection of high-grade pre-invasive or invasive disease of the cervix compared with women who do not have a repeat smear at the time of the initial colposcopic assessment for referral of an abnormal smear.
Design: Retrospective cohort study.
Setting: All Wales Colposcopy database.
Population: The cohort included women referred for colposcopy with an abnormal smear result, who were seen between 1st January 2002 and 31st December 2002.
Method: The management of the group of women who had a repeat smear was compared with the group who did not have a repeat smear test at their first visit. Confidence interval analysis was used to compare results between the two groups.
Main outcome measure: The positive predictive value and sensitivity of the repeat smear to predict outcome, compared to the referral smear in cases where a smear was not repeated at the colposcopy clinic.
Results: 3505 cases (54.0%) did not have a repeat smear at initial colposcopy, whereas 2990 (46%) did. There was a significantly higher rate of referral for low-grade dyskaryosis in the group who had repeat smear testing. The median time interval from the referral smear to the first colposcopy was 49 days for both groups. The specificity and positive predictive value for high-grade lesions were significantly higher in the repeat smear results. The sensitivity was significantly lower, and a high-grade lesion would have been missed with the repeat smear in 14.0% of all referrals.
Conclusions: Colposcopists appeared more inclined to repeat smears for low-grade referrals. The time interval from the referral smear to colposcopy did not appear to influence this decision. A repeat smear test at the first colposcopy visit will detect less high-grade pre-invasive or invasive disease than the initial referral smear.