A large part of our portfolio within this Project has focused on the effects on cancer risk of exogenous hormones. Descriptive analyses using data from SEER documented increases in endometrial cancer incidence after 2002 when results from the Womens Health Initiative (WHI) Trial were published. We hypothesized that this reflected widespread decreases in continuous estrogen plus progestin (E+P) MHT use (the therapy linked with increased breast cancer risk within WHI), given that this is an exposure that we as well as others have documented as leading to reductions in endometrial cancer risk in overweight and obese women. In contrast, in another analysis, we found long-term sequential E+P use (which involves substantial exposure to unopposed estrogens) is associated with increased risk;this association was restricted to thin-to-normal weight women, presumably reflecting their lower endogenous estrogen levels. In an additional investigation focused on ovarian cancer, we found that both sequential and continuous E+P usage was associated with ovarian cancer risk increases. In fact, in a descriptive study, we noted an accelerated decline in ovarian cancer incidence among women 50 years and older in age-period-cohort models following the marked reduction in MHT use that occurred after publication of the WHI results. The notable gender discrepancy in rates of liver cancer has suggested that hormones influence risk, leading to an interest in the effects of MHT use. However , in the LCPP, we found no evidence that liver cancer risk was related to MHT use, although we had limited information of the specific types of preparations used. Some studies, including one that we conducted, have suggested that certain hormonal and reproductive factors may affect the subsequent risk of lung cancer. This has prompted some attention regarding the effects of MHT. Although several studies have noted risk alterations associated with usage, in the NIH-AARP Study we found no evidence that risk of any type of lung cancer was affected by use of either estrogens alone or estrogen plus progestin therapies.In addition to MHT, we have also been concerned regarding the long-term use of fertility drugs. In an extended followup of our large U.S. infertility cohort, we saw no relationship of clomiphene use to either ovarian or endometrial cancers. However, women exposed to 12 or more clomiphene cycles were at an increased risk of invasive breast cancers. In an evaluation of the long-term effects of in vitro fertilization (IVF) undertaken in collaboration with investigators at one of Israelis largest HMOs, we saw no significant associations with breast, endometrial or ovarian cancer risk, but a significant reduction in cervical cancer, presumably reflecting increased surveillance and treatment of precursor conditions among women availing themselves of reproductive assistance. We also collaborated on a study in Denmark of children born to women with a history of infertility (83). There was some evidence of increases in risk of leukemia in childhood and endocrine glands tumors in young adulthood, although mechanisms underlying the associations remained unclear.The relationship of oral contraceptives to various cancers remains of interest. In a large study of women with all stages of cervical carcinogenesis, we demonstrated that OC use was associated with an increased risk of cervical precancer compared to HPV infection, but not with an increased risk of cancer compared to precancer, suggesting that hormone exposure may be particularly important at preinvasive stages. In 1999, the International Agency for Research on Cancer (IARC) reviewed the existing literature on OC use and HCC and concluded that there was sufficient evidence of a causal relationship. However, the number of studies included in the review was small and the number of cases per study modest. In the LCPP, which involved large numbers, we found no evidence that OC use was related to an increased risk of HCC. The increasing recognition of the importance of chronic inflammation in the etiology of ovarian cancer has prompted an interest in risk associated with usage of non-steroidal anti-inflammatory drug (NSAID) usage. In the NIH-AARP study, we evaluated aspirin use and ovarian cancer risk, but did not find an association, possibly due to limited information on use patterns. In a large pooled analysis of individual data within OCAC, a significant reduction in ovarian cancer risk was associated with regular aspirin use, with evidence that the reduced risk was strongest for daily low-dose (100 mg) usage. This suggested that the same aspirin regimen proven to protect against cardiovascular events and associated with risk reduction of several cancers might have chemopreventive implications for ovarian cancer.The majority of risk factors for HCC cause chronic inflammation;thus we hypothesized that use of NSAIDs might be related to reduced risk. In analyses within the NIH-AARP study, we found that aspirin, but not non-aspirin NSAID use, was significantly inversely associated with HCC risk. As most aspirin use in the population was on a daily basis, the result suggested that consuming an 80 mg dose for cardiovascular chemoprophylaxis might also lead to a reduction in HCC risk.In an analysis that we conducted in the SEER-Medicare dataset, we demonstrated that metabolic syndrome is a risk factor for HCC. As high cholesterol levels are one of the defining conditions of metabolic syndrome, we sought to determine whether use of cholesterol-lowering drugsstatins--would decrease risk. In an analysis within the Henry Ford HMO in Detroit, Michigan, we did indeed find that persons who took statins were at significantly decreased risk of developing HCC. A subsequent analysis within the U.K.s Clinical Practice Research Datalink (CPRD) confirmed the inverse association between statin use and liver cancer risk. Analyses restricted to higher-risk individuals (i.e., those with pre-existing liver disease and those with diabetes) found similarly strong inverse associations, suggesting that the observed risk reduction associated with statins was unlikely to reflect confounding by contraindication (concerns about hepatotoxicity with the use of statins that may result in biased prescribing patterns), and that statins may be beneficial even among persons at high-risk for liver cancer.A number of prior studies had suggested that use of metformin, an anti-diabetic drug, is inversely associated with development of liver cancer. Most of these studies however, have compared metformin use to that of all other anti-diabetes medications. However, anti-diabetic medications are strongly linked to diabetes duration and severity. To assess whether the apparent protective effect of metformin was due to it being a first line therapy, we conducted an analysis in the CPRD that compared HCC cases with diabetes to controls with diabetes. Our analysis found that metformin was not strongly inversely associated with the development of HCC, and has offered important methodologic insights related to the use of appropriate comparison populations while studying cancer risk associated with pharmacologic exposures.
Showing the most recent 10 out of 23 publications
