Parity, age at first birth, and risk of invasive cervical cancer: Meta-analyses

PURPOSE: The roles of parity and age at first birth as risk factors for cervical cancer remain unclear, although several studies in different populations have examined this question. We performed meta-analyses to predict relative risk of invasive cervical cancer as functions of parity and of age at first birth by fitting a variety of models. METHODS: We fit models relating the logarithm of relative risk for cervical cancer to parity and to age at first birth. Parity models were based on parity-specific relative risk estimates from 21 studies published between 1985 and 2001. Age at first birth models were based on corresponding data from 11 studies published during the same time period. Some models included covariates indicating whether individual studies adjusted for human papilloma virus (HPV) status, number of sex partners, age at first birth (parity models only), and/or parity (age at first birth models only). RESULTS: Relative risk for cervical cancer was found to increase with parity, at a declining rate (i.e., the greatest increase in risk was between parity 0–1, with consistently smaller increases thereafter). Relative risk for first birth at ages greater-or-equal, slanted17 was significantly lower than relative risk for first birth at age 13, but no significant relationship was found between relative risk and age at first birth for first birth at ages greater-or-equal, slanted17. Overall, indicators of whether studies adjusted for covariates were unable to explain the large variations among study results. CONCLUSION: Risk of invasive cervical cancer was associated with increasing parity, with the greatest increase in risk with the first birth. Risk was decreased for first birth at age 17 compared to 13; for first birth at ages greater-or-equal, slanted17, we were unable to derive a quantitative relationship between age at first birth and relative risk for cervical cancer.