A study of Pap test history and histologically determined cervical cancer in NSW women, 1997–2003

Abstract

To determine the risk of a histological diagnosis of cervical cancer following a given Pap test result and for a given history of Pap test results in a screened population across the full spectrum of possible cytological results. All the Pap screening results held on the New South Wales Pap Test Register for 1997-2003 (five million tests for 1.87 million women) were analysed using Cox proportional hazards regression to estimate the odds of having a histologically determined cervical cancer for a given Pap test result and test result history. The hazard ratios of having cervical cancer in relation to Pap test result histories were estimated: (i) in regard only to the last Pap test result adjusting for age, frequency of Pap testing and proportion of high grade (>or=cervical intraepithelial neoplasia 2 [CIN2]) abnormalities found in a woman's total recorded test result history; and (ii) with regard to the last Pap test result against the highest grade of cytological abnormality found prior to the last Pap test result. The hazard ratios are for a cancer diagnosis occurring before the next Pap test and were adjusted for age, quintile of socioeconomic status of residence, frequency of past Pap testing and proportion of high-grade abnormalities detected in each woman's prior Pap test history. The adjusted hazard ratios were then applied to the tabulated proportions of referent women with negative cytology in each broad age group, and for all women, to estimate the '1 in n' odds of being diagnosed histologically with cervical cancer for a given last Pap test result, and by a given last Pap test result for various prior Pap test result histories. After adjusting for age, socioeconomic status, frequency of previous Pap testing and proportion of past high-grade screen-detected abnormalities, the adjusted hazard ratio of having a subsequent cervical cancer diagnosis for women with a negative Pap test result was 1 in 5,546, compared with 1 in 833 for a low-grade epithelial abnormality (atypia, CIN1), 1 in 56 for a high-grade epithelial abnormality and 1 in seven for a cytological prediction of cervical cancer. These odds estimates were significantly modified by age and by the highest Pap test result prior to the most recent Pap test result: the higher the age and, less consistently, the higher the previous highest Pap test result for a given last Pap test result, the shorter the odds of having a subsequent histological diagnosis of cervical cancer. The results presented here will enable clinicians to inform their patients of their chances of being diagnosed with cervical cancer for a given Pap test result, and for some combinations of the last Pap test result and highest recorded prior Pap test result.