Implementation and Operational Research: Age Distribution and Determinants of Invasive Cervical Cancer in a "Screen-and-Treat" Program Integrated With HIV/AIDS Care in Zambia

Abstract

Background: Cervical cancer screening efforts linked to HIV/AIDS care programs are being expanded across sub-Saharan Africa. Evidence on the age distribution and determinants of invasive cervical cancer (ICC) cases detected in such programs is limited.

Methods: We analyzed program operations data from the Cervical Cancer Prevention Program in Zambia, the largest public sector programs of its kind in sub-Saharan Africa. We examined age distribution patterns by HIV serostatus of histologically confirmed ICC cases and used multivariable logistic regression to evaluate independent risk factors for ICC among younger (≤35 years) and older (>35 years) women.

Results: Between January 2006 and April 2010, of 48,626 women undergoing screening, 571 (1.2%) were diagnosed with ICC, including 262 (46%) HIV seropositive (median age: 35 years), 131 (23%) HIV seronegative (median age: 40 years), and 178 (31%) of unknown HIV serostatus (median age: 38 years). Among younger (≤35 years) women, being HIV seropositive was associated with a 4-fold higher risk of ICC [adjusted odds ratio = 4.1 (95% confidence interval: 2.8, 5.9)] than being HIV seronegative. The risk of ICC increased with increasing age among HIV-seronegative women and women with unknown HIV serostatus, but among HIV-seropositive women, the risk peaked around age 35 and nonsignificantly declined with increasing ages. Other factors related to ICC included being married (vs. being unmarried/widowed) in both younger and older women, and with having 2+ (vs. ≤1) lifetime sexual partners among younger women.

Conclusions: HIV infection seems to have increased the risk of cervical cancer among younger women in Zambia, pointing to the urgent need for expanding targeted screening interventions.

Figure 1

Histograms of the ages at diagnosis of ICC stratified by HIV serostatus among women screened in the CCPPZ.

Figure 2

Probability of ICC by age and HIV serostatus among women screened in the CCPPZ. Predicted probability curves and 95% confidence intervals are shown for women with HIV seropositive, HIV seronegative, and serostatus unknown. Small histograms are overlaid in the top panel to show the density of overall population screened in each subgroup by their age (which forms the denominator of the probability analysis).