Value of duplicate smears in cervical cytology

The cytologic accuracy, sources of error and value of duplicate samples were studied in 87 women simultaneously biopsied for cervical intraepithelial neoplasia. Separately evaluated by two laboratories, correlation with biopsy was 70% for the first cytology and 67% for the second. Discrepant cytologies were reciprocally reviewed in error analysis. The sources of error were cytologic sampling (10.9%), evaluation (5.8%) and both (4.0%). Attention is directed to the high proportion of cases (15%) in which sampling error was a factor; awareness of this observation is necessary in patient management. The duplicate second cytology was of substantial value in 7% of the cases and helpful in an additional 3.5%; both figures are lower than in recent reports. If errors in evaluation or evaluation sampling are held to a 10% level, routine duplicate studies may not be justified.