The definition of minimal relevant Pap smear abnormality is crucial for balancing the beneficial effects of screening (prevented mortality) with negative side-effects (the high positivity rate). After inflammation ceased to be defined as a borderline abnormal smear outcome in The Netherlands in 1996, the proportion of these smears dropped from 10% to less than 2%. Because this may have caused a loss in smear sensitivity, we analysed the changes in the incidence of cervical cancer after a negative Pap smear. All negative smears made at ages 30-64 in 1990-1995 (n = 1,546,252) and 1998-2006 (n = 3,552,716), registered in the national registry of histo- and cytopathology (PALGA), were followed for up to 9 years. During follow-up of the 1990-1995 smears, 377 women developed cervical cancer within 5,232,959 woman-years at risk, while during the follow-up of the 1998-2006 smears, 619 women developed cervical cancer within 11,210,675 woman-years at risk. The cumulative incidence after the definition change was not significantly higher than before: e.g. at 6 years, the cumulative incidence for smears made in 1990-1995 was 46 per 100,000 (95% CI: 41-52), and for smears in 1998-2006 was 48 per 100,000 (95% CI: 43-54), p = 0.59. The hazard ratio for 1998-2006 compared to 1990-1995 adjusted for age, number of previous negative smears and history of abnormalities was 0.90 (95% CI: 0.78-1.03). In The Netherlands, a setting with high-quality cytological screening, treating smears with only signs of inflammation as negative leads to a considerably lower positivity rate without increasing the risk for cervical cancer after a negative smear.
(c) 2008 Wiley-Liss, Inc.