Pregnancy and cervical cancer: a retrospective study of the associations of age at first pregnancy and parity with non-invasive and invasive cervical lesions among HIV-negative women in Senegal

Cervical cancer is the leading cause of cancer-related deaths in Senegal. Senegal’s cervical cancer prevention strategy has prioritized vaccinating young girls with human papillomaviruses (HPV) vaccines, but not screening of older women. There is limited knowledge about the roles of age at first pregnancy (AFP) and parity in developing cervical intra-epithelial lesions (CIN) and invasive cervical cancer (ICC) in this population. We investigated the associations of AFP and parity with CIN and ICC using data from four studies on cervical carcinogenesis conducted in Senegal between 1998 and 2011. Eligibility criteria included not being HIV positive, 18 years or older, not pregnant at visit, and having either histology or cytology results. Missing exposure and covariate data were imputed with Multiple Imputation by Chained Equations using R statistical software. We conducted multinomial logistic regression adjusted for age, age at first sex, birth control method, lifetime number of male sexual partners, marital status, smoking history, study, site, and visit year to evaluate the associations of CIN and ICC with AFP; live births for parity; and AFP and parity. We did a sub-analysis among those who were positive for any HPV type. Among the 5,588 women included in this study, the median AFP was 18 years, and the median number of live births was 5. AFP was significantly associated with CIN in the 12-14 (OR=2.01, 95% CI:1.06-3.83) and 15-16 (OR=1.72, 95% CI:1.02-2.88) groups compared to the 21-24 group. Number of live births was associated with CIN and ICC in both the complete and HPV restricted populations, with lower risks in nulliparous group and higher risks in the 3-4, 5-6, and 7+ groups compared to the 1-2 live births group. Parity in the 3-4 live births group was significantly associated with CIN and ICC in the complete population, OR=2.20, 95% CI:1.20-4.02 and OR=1.73, 95% CI:1.10-2.71 respectively; and with ICC in the HPV restricted population, OR=2.26, 95% CI:1.12-4.60. These associations of parity with CIN and with ICC were similar when evaluated along with AFP. Early AFP increases risk of CIN and higher parity increases risk of ICC among HIV-negative women in Senegal. The effect of AFP and parity on CIN and ICC are independent of each other when assessed with confounding. These findings can inform future cancer prevention and screening strategies.