PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Int J Cancer. Author manuscript; available in PMC 2010 August 15.
Published in final edited form as:
PMCID: PMC2767328
NIHMSID: NIHMS109108

Family history of hormonal cancers and colorectal cancer risk: A case-control study conducted in Ontario

Abstract

Aggregation of cancers among families with highly penetrant genetic mutations such as hereditary nonpolyposis colorectal cancer is well-described. However, there is a paucity of data regarding familial aggregation of hormonal cancers (cancers of the breast, endometrial, ovarian, and prostate) and colorectal cancer (CRC) in the general population. We investigated the association between having a first-degree family history of breast, endometrial, ovarian, or prostate cancer and CRC risk. Population-based CRC cases and controls were recruited by the Ontario Familial Colorectal Cancer Registry (OFCCR). Logistic regression was conducted to obtain odds ratio (OR) estimates and 95% confidence intervals (95% CIs). First-degree family history of breast cancer was associated with a modest, borderline statistically significant increased CRC risk (age-, sex-adjusted OR = 1.2, 95% CI = 1.0, 1.5). The magnitude of CRC risk was greatest if more than one first-degree kin had breast cancer (age-, sex-adjusted OR = 1.7, 95% CI = 1.0, 2.0), as well as if the kin was diagnosed at >50 years of age (age-, sex-adjusted OR = 1.4, 95% CI = 1.1, 1.8). Family history of ovarian cancer was associated with reduced CRC risk (multivariate-adjusted OR = 0.6, 95% CI = 0.3, 1.0). While statistically significant increases in CRC risk were observed in the age-, sex-adjusted OR estimates for family history of endometrial and prostate cancers, the associations were no longer significant after multivariate-adjustment. In conclusion, individuals with a first-degree kin with breast cancer may have a modest increased risk for CRC compared to individuals without.

Keywords: colorectal neoplasms, cancer, family history, familial aggregation, case-control study

INTRODUCTION

Colorectal cancer (CRC) is the third most common type of cancer for men and women, and the second and third leading cause of cancer death for men and women, respectively, in Canada.1 The modest five-year survival rate of CRC at 62%1 emphasizes the importance of preventive measures such as screening and public education regarding CRC risk factors.

The association between genetic factors and CRC has been well studied. However, while the association between rare, highly penetrant genetic conditions, such as familial adenomatous polyposis and hereditary nonpolyposis colorectal cancer (HNPCC), and CRC are well-established,2 the possible association between other hereditary factors such as common, low penetrant genes and CRC risk remains unclear.3,4 Similarly, while a family history of CRC is a known CRC risk factor,512 the literature regarding CRC risk posed by a family history of other types of cancers remains inconclusive.

Previous epidemiological studies have investigated the association between first-degree family history of various types of cancer and CRC risk.6, 1021 Overall, the results of these studies have varied depending on the cancer type investigated. However, statistically significant associations were reported between increased colon cancer risk and a family history of breast, uterine, ovarian, and prostate cancer.19 In addition, family histories of intestinal15 and stomach18 cancers have been associated with increased CRC risk. Given the growing evidence that sex hormones may play a role in CRC risk,2230 of particular interest in this current study was the possible association between a family history of hormonal cancers, namely breast, prostate, ovarian, and endometrial cancer, and CRC risk.

We conducted a population-based case-control study investigating the association between having a first-degree family history of breast, endometrium, ovarian, or prostate cancer and CRC risk, utilizing the data collected by the OFCCR,31,32 a population-based resource.

MATERIAL & METHODS

The methods of the OFCCR have previously been described in detail.31, 32 A brief summary of the registry’s methods that are most relevant to the current study are as follows.

Case and Control Selection

Incident, pathology confirmed CRC cases (International Classification of Diseases, 9th revision (ICD-9)33 codes 153.0, 153.1, 153.2, 153.3, 153.4, 153.5, 153.6, 153.7, 153.8, 153.9, 154.0, 154.1, 154.2, 154.3, 154.8) diagnosed from July 1, 1997 to June 30, 2000 were obtained from the Ontario Cancer Registry. These individuals were Ontario residents who were alive at diagnosis. Cases were mailed self-administered questionnaires in two stages. First, the Family History Questionnaire (FHQ) was mailed to all cases, then based on responses on the FHQ, cases were categorized into high, intermediate, and low familial CRC risk. Individuals were classified as high risk if they met the Amsterdam criteria for HNPCC,34 and classified as intermediate risk if they met one of several criteria, two of which included: 1) Having two relatives with HNPCC cancers (including endometrial and ovarian cancers) and at least two of these three individuals being first-degree relatives; and 2) Having a family member with an HNPCC cancer diagnosed at ≤35 years of age. In the second stage of questionnaire mailing, all high and intermediate risk cases and 25% of randomly selected low risk cases were re-contacted and mailed the Personal History Questionnaire (PHQ) and the Diet Questionnaire (DQ). Only cases who completed the PHQ were included since the many epidemiological factors (e.g., age, sex, and other confounding variables) that were necessary for analysis were collected using the PHQ. Of the 2124 cases mailed the PHQ, approximately 72% completed it. Among these cases, 25 were subsequently found to be mismatch repair (MMR) gene mutation carriers, and 1 case was a MutY human homologue (MYH) gene mutation carrier. These carriers were excluded from analyses as these mutations are part of known, highly penetrant genetic syndromes. In total, 1512 cases were included in this study, with 59% of them being female. The average age of cases was 61 years.

Population-based controls were recruited to sex- and age-frequency match the cases within 5–10 year age groups using two methods: 1) Random-digit dialing; and 2) the Ministry of Finance Property Assessment Database. Controls were mailed all three questionnaires (FHQ, PHQ, DQ) at once. Questionnaires were mailed to 2745 controls, and of these, approximately 71% completed the PHQ. In total, 1953 controls were in this study, with 47% of them being female. The average age of controls was 59 years.

Family History

Family history for each cancer type was self-reported by the participants who completed the FHQ. However, for a very small number of deceased cases (n = 46), a proxy (e.g., a spouse or relative) completed the questionnaire. The FHQ asked if a first-degree kin ever had cancer or tumors, and if so, what type of cancer or tumor the kin had, the age of diagnosis, and the date of diagnosis. These four questions were asked regarding participants’ biological mother, father, siblings, and children. While information on cancer in other family members was also collected by the FHQ (using more general questions), only the first-degree family history was used in this study as it is the most reliable and consistently reported3539 and it was asked specifically for each first degree kin.

Family history of cancer was defined in three different ways in our analyses, for a total of three exposure variables per cancer type. Note that these three variables are not independent. The three family history variables were: 1) Presence of first-degree family history of the specified cancer (Yes or No); 2) The number of first-degree kin with the specified cancer (None, 1 kin, or >1 kin); and 3) The age at which the first-degree kin was diagnosed with the specified cancer (No kin diagnosed, ≤50 years of age, or >50 years of age). It was hypothesized that individuals with >1 kin with a family history of cancer, and those with kin diagnosed with cancer at ≤50 years of age would have the greatest risk for CRC.

Other Variables

Information on many suspected or established CRC risk factors (potential confounding variables) was collected using the PHQ. The PHQ asked participants about their health (bowel screening; medication; menstruation, pregnancy, and menopause; diet; physical activity; alcohol consumption; smoking; weight and height; demographic information) and the DQ asked more detailed information about dietary intake.

Statistical Analyses

Multivariate unconditional logistic regression analysis was conducted to obtain odds ratio (OR) estimates for the association between CRC risk and each of the four family history variables (breast, endometrial, ovarian, and prostate cancer among first degree kin). Two multivariate logistic models were created for all family history cancer types: age-, sex-adjusted and multivariate-adjusted. In order to identify possible confounders that should be adjusted for in the multivariate models, 15 risk factors for CRC (family history of CRC, CRC screening, body mass index, use of non-steroidal anti-inflammatory drugs, alcohol consumption, cigarette smoking, fiber intake, total fat intake, red meat consumption, fruit consumption, vegetable consumption, physical activity in 20s, and physical activity in 30/40s, number of first-degree relatives, and familial risk of CRC) were individually evaluated as potential confounders using the dichotomized definition (Yes/No) of breast, endometrial, ovarian, and prostate cancer family history and applying the 10% change in OR method.40 The number of first-degree female relatives was tested as a potential confounder for family histories of breast, endometrial, and ovarian cancer, while the number of first-degree male relatives was tested as a potential confounder for family history of prostate cancer. Use of post-menopausal hormone replacement therapy and oral contraceptives were tested as potential confounders in a female only subset of the dataset. All potential confounders excluding number of first-degree relatives (all, females, and males), fiber intake, total fat intake, and familial risk of CRC, were defined and categorized as in our previous paper.41 The three first-degree relative variables, fiber intake, and total fat intake were treated as categorical variables based on tertile distributions among controls, while familial risk of CRC was dichotomized into high/intermediate and low/sporadic risk categories. Potential confounders that yielded greater than 10% change in the age-, sex-adjusted OR were deemed to be confounders,40 and were subsequently adjusted for in multivariate models. Variables that were not identified as confounders were not included in the final multivariate model which was designed to be the most parsimonious model.

The Cochran-Armitage trend test was conducted to evaluate the significance of the trends for the number of first-degree kin diagnosed with cancer and the age of the first-degree kin’s cancer diagnosis variables. The exact test two-sided p-values are reported.

Since oversampling of intermediate familial risk cases may have created a selection bias (the criteria used to select intermediate risk cases include two of our exposure variables, family histories of endometrial cancer and ovarian cancer), we investigated this possibility, and cases and controls were stratified according to familial risk, and the age-, sex-adjusted OR and multivariate-adjusted OR estimates for family history of endometrial cancer and ovarian cancer were compared. No differences were found (p > 0.05). Similarly, results from a likelihood ratio statistic test investigating the possibility of effect modification by familial risk by adding the product term40 in the multivariate models were not statistically significant for both family history of endometrial cancer and ovarian cancer. Hence, no further adjustments were made, and all cases were combined for analyses.

The possibility of effect modification between sex and family history of each cancer type was also investigated using the likelihood ratio statistic. If the results of the likelihood ratio statistic indicated the presence of statistically significant (p < 0.05) effect modification, stratified analyses by sex were conducted.

To investigate the possibility of differences in the association between family history of cancer and CRC risk by CRC sub-site, case-case age-, sex-adjusted OR estimates for family history of breast, endometrial, ovarian, and prostate cancers comparing distal CRC cases (ICD-933 codes 153.2, 153.3, 154.0, 154.1, 154.2, 154.3, and 154.8) to proximal CRC cases (ICD-933 codes 153.0, 153.1, 153.4, 153.5, 153.6, and 153.7) were calculated. If the case-case age-, sex-adjusted OR was statistically significant (p < 0.05), cases were stratified according to cancer sub-site and compared to all controls. Cases with ICD-933 codes 153.8 (other specified sites of large intestine) and 153.9 (colon, unspecified) were excluded for the sub-site analyses as the specific sites of these cancers were unclear.

RESULTS

Sex was not a statistically significant effect modifier in the association between family history of breast cancer and CRC risk (p = 0.47). The case-case age-, sex-adjusted OR estimate comparing distal cases to proximal cases was 0.8 (95% CI = 0.6, 1.0), suggesting a borderline statistically significant difference in the association between CRC and family history of breast cancer by CRC sub-site. Thus, stratified analysis by CRC sub-site was conducted and presented. Since no additional confounders were identified in the association between family history of breast cancer and CRC risk, further multivariate analyses were not conducted. Table 1 shows the frequency distributions of cases and controls, age-, sex-adjusted OR and 95% CI estimates, and p-values for trend tests for family history of breast cancer and CRC risk. The result from the analysis comparing all cases and controls indicates a modest increase in CRC risk among those with a first-degree family history of breast cancer (age-, sex-adjusted OR = 1.2, 95% CI = 1.0, 1.5). Individuals with more than one first-degree kin with breast cancer had a higher risk for CRC (age-, sex-adjusted OR = 1.7, 95% CI = 1.0, 2.9) compared to those with only one kin (age-, sex-adjusted OR = 1.1, 95% CI = 0.9, 1.4). In addition, individuals with a kin diagnosed at >50 years of age had a higher risk for CRC (age-, sex-adjusted OR = 1.4, 95% CI = 1.1, 1.8) compared to those with a kin diagnosed ≤50 years of age (age-, sex-adjusted OR = 1.2, 95% CI = 0.9, 1.8). Significant trends were observed for both the number of kin with breast cancer (p = 0.03) and the age of kin’s breast cancer diagnosis (p < 0.01). CRC risk associated with family history of breast cancer was higher among proximal cases (age-, sex-adjusted OR = 1.4, 95% CI = 1.1, 1.8) than distal cases (age-, sex-adjusted OR = 1.1, 95% CI = 0.8, 1.4). Similarly, significant trends were observed for the number of kin with breast cancer and the age of kin’s breast cancer diagnosis for proximal cases, but not for distal cases.

Table 1
Distribution of colorectal cancer cases, controls, and age- and sex-adjusted odds ratio (ASOR) for family history of breast cancer among first-degree relatives.

Sex was a statistically significant effect modifier in the association between family history of endometrial cancer and CRC risk (p = 0.02), and the case-case age-, sex-adjusted OR estimate comparing distal cases against proximal cases was 0.5 (95% CI = 0.3, 0.8), suggesting a difference in the association between CRC risk and family history of endometrial cancer by CRC sub-site. Table 2 shows the frequency distributions of cases and controls, age-, sex-adjusted OR, multivariate-adjusted OR and 95% CI estimates, and p-values for trend tests for family history of endometrial cancer and CRC risk. Family history of endometrial cancer was associated with an increased CRC risk, but this was not statistically significant (multivariate-adjusted OR = 1.3, 95% CI = 0.8, 2.1). However, note that significant trends were observed for the number of kin with endometrial cancer (p < 0.01) and the age of kin’s endometrial cancer diagnosis (p < 0.01). A family history of endometrial cancer was associated with higher proximal CRC risk (multivariate-adjusted OR = 2.0, 95% CI = 1.1, 3.7) than distal CRC risk with family history of endometrial cancer was also higher among men (multivariate-adjusted OR = 0.7, 95% CI = 0.4, 1.4). CRC risk associated iate-adjusted OR = 2.4, 95% CI = 1.0, 5.7) than women (multivariate-adjusted OR = 1.0, 95% CI = 0.5, 1.8). Consistently, significant trends were observed for both the number of kin with endometrial cancer and the age of kin’s endometrial cancer diagnosis (p < 0.01) among men, but only the latter had a significant trend among women (p = 0.03).

Table 2
Distribution of colorectal cancer cases, controls, age- and sex-adjusted odds ratio (ASOR), and multivariate-adjusted odds ratio (MVOR) for family history of endometrial cancer among first-degree relatives.

Sex was a statistically significant effect modifier in the association between family history of ovarian cancer and CRC risk (p = 0.03). However, the case-case age-, sex-adjusted OR estimate comparing distal cases against proximal cases was not statistically significant (age-, sex-adjusted OR = 0.6, 95% CI = 0.3, 1.2). Table 3 shows the frequency distributions of cases and controls, age-, sex-adjusted OR, multivariate-adjusted OR and 95% CI estimates, and p-values for trend for family history of ovarian cancer and CRC risk. The age-, sex-adjusted OR did not suggest CRC risk having a statistically significant association with family history of ovarian cancer (age-, sex-adjusted OR = 1.4, 95% CI = 0.9, 2.2), however, the multivariate-adjusted OR indicated a marginally significant reduced CRC risk (multivariate-adjusted OR = 0.6, 95% CI = 0.3, 1.0). No significant trends were observed for the number of kin with ovarian cancer and the age of kin’s ovarian cancer diagnosis. Results stratified by sex indicated that the association between family history of ovarian cancer and decreased CRC risk was statistically significant for women (multivariate-adjusted OR = 0.4, 95% CI = 0.2, 0.9), but not men (multivariate-adjusted OR = 0.9, 95% CI = 0.3, 2.2). However, note that significant trends for number of kin with ovarian cancer and the age of kin’s ovarian cancer diagnosis were observed for men, but not women; this is likely due to our small frequency counts for women.

Table 3
Distribution of colorectal cancer cases, controls, age- and sex-adjusted odds ratio (ASOR), and multivariate-adjusted odds ratio (MVOR) for family history of ovarian cancer among first-degree relatives.

Sex was not a statistically significant effect modifier in the association between family history of prostate cancer and CRC risk (p = 0.85). Similarly, the case-case age-, sex-adjusted OR estimate comparing distal cases to proximal cases was 0.9 (95% CI = 0.6, 1.2), suggesting no difference in the association between CRC risk and family history of prostate cancer by CRC sub-site. Table 4 shows the frequency distributions of cases and controls, age-, sex-adjusted OR, multivariate-adjusted OR, and 95% CI estimates, and p-values for trend tests for family history of prostate cancer and CRC risk. Our results suggest that family history of prostate cancer is not significantly associated with CRC risk (multivariate-adjusted OR = 1.0, 95% CI = 0.8, 1.3). However, CRC risk was higher among those with more than one kin diagnosed with prostate cancer (multivariate-adjusted OR = 2.1, 95% CI = 0.9, 4.8) compared to just one kin (multivariate-adjusted OR = 0.9, 95% CI = 0.7, 1.2), and also if the kin was diagnosed at >50 years of age (multivariate-adjusted OR = 1.1, 95% CI = 0.8, 1.5) compared to ≤50 years of age (multivariate-adjusted OR = 0.1, 95% CI = 0.01, 1.0). Only the trend for the age of kin’s prostate cancer diagnosis was statistically significant (p = 0.04).

Table 4
Distribution of colorectal cancer cases, controls, and age- and sex-adjusted odds ratio (ASOR) for family history of prostate cancer among first-degree relatives.

DISCUSSION

Family history of breast cancer among first degree kin was associated with a modest, marginally statistically significant increased CRC risk. This association was also stronger if there was more than one kin diagnosed with breast cancer, and if the kin was diagnosed at >50 years of age. The association between family history of breast cancer and CRC also appeared to be stronger among proximal CRC cases compared to distal cases. Family history of ovarian cancer had a marginally statistically significant inverse association with CRC risk. While the age-, sex-adjusted ORs for family history of endometrial cancer and prostate cancer showed statistically significant associations with CRC risk, these associations were no longer significant after multivariate adjustment.

Comparison of findings to previous reports

Our findings that a family history of breast cancer is associated with a moderate increase in CRC risk is consistent with some previous epidemiological reports,16, 19 but not others.6, 1011, 1314, 1718 Alternatively, a couple of previous studies did not find an association between a family history of CRC and risk for breast cancer among first-degree relatives.8, 42 Our result of a strong association between having more than one kin diagnosed with breast cancer (compared to none) and CRC risk suggests that families with multiple breast cancer diagnoses may be genetically and/or environmentally susceptible to both CRC and breast cancer. Meanwhile, the strong association observed between having a kin diagnosed with breast cancer at >50 years of age (compared to none) and CRC risk is in contrast to our hypothesis that a stronger association may exist if the kin is diagnosed with breast cancer at ≤50 years of age, given that cancers arising from genetic predisposition tend to have earlier onset. Our finding that breast cancer diagnosis occurring relatively later in life is strongly associated with CRC risk – even after adjusting for age – suggests that the familial clustering of breast cancer and CRC observed in our study may be distinct from cancer clustering observed in HNPCC families. In contrast with another study,19 we found a moderate association between family history of breast cancer and proximal CRC risk and no association observed for distal CRC risk.

Given the limited power for our analyses of family history of endometrial cancer and CRC risk, our results must be interpreted with caution. We found a statistically non-significant positive association between CRC risk and family history of endometrial cancer, which is consistent with previous findings that investigated family histories of endometrial11 and uterine13, 18 cancers. However, one previous study did report that family history of endometrial cancer is associated with reduced colon cancer risk.6 Our finding that CRC risk is higher among men reporting a family history of endometrial cancer, compared to women, is in agreement with those previously reported.6, 19 It is difficult to compare our finding that family history of endometrial cancer was associated with higher CRC risk for proximal cases than distal cases as the only other study that investigated risk by sub-site was by Slattery & Kerber,19 and they reported that colon cancer risk was higher for distal than proximal cases reporting a family history of uterine cancer among men, but the risk was higher for proximal than distal cases among women.

Our results regarding family history of ovarian cancer and CRC risk must be interpreted with caution given the limited statistical power. Our multivariate-adjusted OR estimate suggests that a family history of ovarian cancer is associated with reduced CRC risk; however, this finding contradicts most previous studies. It is possible that this was a spurious finding since to our knowledge, only a handful of studies investigated this association with two studies13,18 reporting a non-significant positive association with family history of ovarian cancer and CRC risk and another study19 reporting a statistically significant positive association among women only. Only one study reported a non-significant negative association among women.6 A negative association between a family history of ovarian cancer and CRC suggests the possibility that (a) factor(s) – genetic or environmental – increasing the susceptibility of ovarian cancer in a family can actually be associated with reduced risk for CRC.

We did not find a statistically significant association between family history of prostate cancer and CRC risk. Results from previous case-control studies are mixed with some reporting statistically significant,16, 19 and non-significant18,19 increased risk, as well as no risk6, 11 associated with family history of prostate cancer. In addition, a previous study investigating the alternative association of family history of CRC and risk for prostate cancer did not find a statistically significant association.42 In summary, the association between family history of prostate cancer and CRC risk remains unclear.

We can speculate on possible factors contributing to the discrepancies between our study results and those in the literature, as well as the inconsistency of findings in the literature. First, residual confounding may have resulted in inflated results in studies that did not adjust for appropriate confounding variables. Slattery & Kerber19 reported several significant associations between family histories of hormonal cancers and colon cancer risk. However, while the authors controlled for a limited number of variables (e.g., age, sex), they did not explore the possibility of confounding by lifestyle and dietary variables and thus residual confounding may have biased their results away from the null 19 In our study, we tested for 15 potential confounders and then adjusted for these if they were deemed to be confounders in our datasets; in addition, age and sex were always adjusted for in all our models. Second, some studies10,17 that investigated family history of cancers and CRC risk had relatively small sample sizes for their calculations and thus may have had inadequate statistical power to detect a significant association, or their relative risk estimates may not have been stable. Third, some studies in the literature only evaluated colon cancer risk rather than CRC risk, as we did in the present study. It is possible that rectal cancer is more strongly associated with family history of hormonal cancers and that results from studies only investigating colon cancer risk attenuates the CRC risk estimated by the overall literature. Previous results suggest that rectal cancer is more strongly associated with family history of hormonal cancers, with one study reporting a stronger (although non-significant) association between family histories of breast, uterine, ovarian, and prostate cancers with rectal cancer risk compared to colon cancer risk.18

Hormones and breast, ovarian, endometrial, prostate, and colorectal cancers

Studies suggest that hormonal cancers and CRC may share common biological etiologies such as exogenous sources of hormone. Several reviews concluded that hormone replacement therapy is associated with modest increased risk for breast and ovarian cancers, and that specifically unopposed estrogen therapy is associated with increased risk for endometrial cancer, but that hormone replacement therapy significantly reduces the risk for colon cancer.4345 Meanwhile, the use of oral contraceptives is associated with increased risk for breast cancer in some studies, but decreased risk for endometrial, ovarian, and colorectal cancer. 4549 Female sex hormones, specifically, estrogens, have also been implicated in prostate cancer.5052 Furthermore, previous studies demonstrated that hormonal cancers and CRC share common genetic factors. For instance, prostate cancer and CRC share a common polymorphism on chromosome 8q24,53 and a region of 8q24 has been associated with prostate, colorectal, and ovarian cancer risk.54 It has also been demonstrated that breast cancer and CRC share a common polymorphism in the ERBB4 gene.55 However, given the limited knowledge regarding a common hormonal pathway between breast, endometrial, ovarian, prostate, and colorectal cancers, a common etiological hormonal factor remains speculative.

HNPCC cancers

Familial aggregation of endometrial cancer and ovarian cancer with CRC may be due to the fact that these three types of cancers are HNPCC cancers. Thus, we excluded cases with known high-risk CRC mutations (MMR and MYH polymorphisms) from our analyses, and we also tested for effect modification by familial risk. As previously discussed, familial risk was not a statistically significant effect modifier in our study. Hence, familial aggregations of cancers in our results are unlikely to be due to the presence of HNPCC families in our datasets.

Validity of self-reported family history

In our dataset, there were 482, 114, 83, and 351 first-degree kin with breast, endometrial, ovarian, and prostate cancers, respectively, as reported by our cases and controls. Of these reported kin cases, we were able to obtain pathology confirmation on only 8.3%, 21.9%, 14.5%, and 7.1% of breast, endometrial, ovarian, and prostate cancer reports, respectively (this was passively done as part of the OFCCR). Hence, our major method of ascertaining family histories was through participants’ self-reports. Although this method of ascertainment is admittedly imperfect, previous studies demonstrate that self-reports of first-degree family history can be highly accurate. The accuracy of a proband’s reported family history of cancer in first-degree kin has been reported to be as high as 95.4%,37 and is generally higher than the accuracies of reporting for second- or third-degree kin.3539 Reporting accuracy also differs between cancer sites, with female breast cancer and prostate cancer having been reported with high accuracies,37, 5658 ovarian cancer generally being reported with modest accuracy,5658 and endometrial cancer being reported with low accuracy.37, 56, 59 The accuracy of participants’ reporting of kin’s age of diagnosis is less well-known. However, one study reports that the age of breast cancer diagnosis in first-degree relatives was accurately reported by 92% of their breast cancer cases,38 while another breast cancer case-control study found that the average errors of the reported age of breast cancer diagnosis in first-degree relatives among cases and controls were similar at 2.0 and 1.7 years, respectively.60 Thus, while inaccurate reporting of family history is a concern as this can lead to misclassification, previous studies suggest that misclassifications for family histories of breast and prostate cancer in the present study are unlikely, although misclassification for family history of endometrial and ovarian cancer is plausible.

As with other case-control studies, recall bias is a potential limitation of this study. However, a previous study suggests the self-reported accuracy of family history does not differ between cases and controls.56 Furthermore, within our dataset, a significant association between family history of CRC and CRC risk (age-adjusted OR = 2.4, 95% CI = 2.0, 2.9) was observed and is similar to previous literature,10,16,41 suggesting that the family history information collected is valid.

Limitations

The limitations of our study are as follows. First, since an inclusion criterion for cases was being alive at the time of diagnosis, survival bias is a possibility if fatal CRC cases differed from live cases in their reporting of, or actual, family histories of cancers. Second, the OFCCR has a moderate control response rate of 71% for both this current and study and that previously reported,41 which suggests that our control participants may not necessarily be representative of the general population. Third, since genotype information was available for only 68.9% of our high-risk participants, we may have failed to exclude some participants with high-risk HNPCC mutations, which would mean that our risk estimates are inflated. Similarly, a small number of HNPCC cases may have inadvertently been included in this study, as although MMR mutation screening was thorough based on current knowledge, false negative tests are always possible in genetic testing. Fourth, given the small frequencies and wide CIs for some of our OR estimates, our sample size was relatively small and our statistical power limited for some calculations. Finally, similar to that reported in a previous OFCCR paper61 the majority of our sample (95%) was Caucasian. While ethnic differences in the association between family histories of hormonal cancer and CRC risk are unclear, at least one previous report8 found that the odds for breast cancer among first-degree relatives of CRC patients was higher among their Japanese participants compared to all their participants combined. Thus, the applicability of our results to non-Caucasians is uncertain.

Strengths

Our study has several notable strengths. First, as the sampling frames for cases and controls were population-based, our study is not influenced by biases associated with hospital- or clinic-based samples. Second, cases were incident, pathology confirmed CRC cases, eliminating the possibility of misclassification on CRC status. Third, we collected extensive information on family histories of first-degree family members, including type of cancer diagnosed and age at diagnosis, allowing for the investigation of family history of cancer at different sites, and possible differences in the association between family history of cancer and CRC risk by kin’s age of cancer diagnosis. Finally, by testing for 19 potential confounders (number of first-degree relatives, number of first-degree female relatives, number of first-degree male relatives, familial risk of CRC, family history of CRC, CRC screening, body mass index, use of non-steroidal anti-inflammatory drugs, alcohol consumption, cigarette smoking, fiber intake, total fat intake, red meat consumption, fruit consumption, vegetable consumption, physical activity in 20s, physical activity in 30/40s, use of post-menopausal hormone replacement therapy, and use of oral contraceptives) and adjusting for actual confounders applicable to our datasets, our results are unlikely to be influenced by confounding.

Conclusion

Our results suggest that a family history of certain hormonal cancers – in particular breast cancer – may put persons in the general population at increased risk for CRC. Authors19 of a previous study investigating family history of cancers and CRC risk have noted that familial aggregation of cancers can be due to common genetic and environmental factors. Compared to previous studies, we investigated family history of hormonal cancers in greater detail; namely, evaluating the number of first-degree kin diagnosed with cancer and the age of the kin’s cancer diagnosis. Hence, our study adds to the exiting literature by providing intriguing findings regarding the possible association between hormonal cancers and CRC risk. However, given the paucity of existing studies in this area of research, further research is necessary to confirm or refute our findings. Authors of future studies should obtain larger sample sizes in order to have sufficient statistical power for evaluating family history of relatively rare hormonal cancers such as ovarian and endometrial cancer. Explorations into the roles of family history of hormonal cancers in second- and third-degree kin are also warranted.

In conclusion, family history of hormonal cancers – particularly breast cancer – may be a risk factor for CRC. Furthermore, the details of such family history such as the number of kin diagnosed and the age of the kin’s breast cancer diagnosis may be important, and thus worth considering when evaluating individual CRC risk. These findings may be of importance to clinicians and genetic counsellors advising persons on cancer prevention and screening recommendations.

Supplementary Material

sup 2

sup1

ACKNOWLEDGMENTS

This work was supported by the National Cancer Institute, National Institutes of Health under RFA # CA-95-011 (grant no. U01-CA74783). The authors would like to thank all the Ontario Familial Colon Cancer Registry (OFCCR) study staff and genetic counselors for their efforts and dedication to the study. In particular, we would like to thank Ellen Shi for her assistance with dataset preparations and variable derivations. This work was made possible through collaboration and cooperative agreements with the Colon Cancer Family Registry (CFR) and Principal Investigators. The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating institutions or investigators in the Colon CFR, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government or the Colon CFR.

Abbreviations

95% CI
95% confidence interval
CRC
Colorectal cancer
DQ
Diet Questionnaire
FHQ
Family History Questionnaire
HNPCC
Hereditary nonpolyposis colorectal cancer
ICD-9
International Classification of Diseases, 9th revision
MMR
Mismatch repair
MYH
MutY human homologue
OR
Odds ratio
OFCCR
Ontario Familial Colorectal Cancer Registry
PHQ
Personal History Questionnaire

Footnotes

2 statements regarding the novelty and impact of the paper:

  1. First study to investigate the association between colorectal cancer risk and family history of several cancers (breast, endometrial, ovarian, and prostate) in a Canadian population.
  2. This study collected extensive information regarding details of family history (e.g., age of cancer diagnosis in kin) and lifestyle factors, allowing for the investigation of family history of cancer in detail and also allowing for the adjustment of numerous potential confounders.

REFERENCES

1. Canadian Cancer Society/National Cancer Institute of Canada. Canadian Cancer Statistics 2008. Toronto, Canada: 2008.
2. Lynch HT, de la Chapelle A. Hereditary colorectal cancer. N Engl J Med. 2003;348:919–932. [PubMed]
3. de la Chapelle A. Genetic predisposition to colorectal cancer. Nat Rev Cancer. 2004;4:769–780. [PubMed]
4. Houlston RS, Tomlinson IP. Polymorphisms and colorectal tumor risk. Gastroenterology. 2001;121:282–301. [PubMed]
5. Colditz GA, Atwood KA, Emmons K, Monson RR, Willett WC, Trichopoulos D, Hunter DJ. Harvard Report on Cancer Prevention Volume 4: Harvard Cancer Risk Index. Risk Index Working Group, Harvard Center for Cancer Prevention. Cancer Causes Control. 2000;11:477–488. [PubMed]
6. Kerber RA, Slattery ML, Potter JD, Caan BJ, Edwards SL. Risk of colon cancer associated with a family history of cancer or colorectal polyps: the diet, activity, and reproduction in colon cancer study. Int J Cancer. 1998;78:157–160. [PubMed]
7. Slattery ML, Levin TR, Ma K, Goldgar D, Holubkov R, Edwards S. Family history and colorectal cancer: predictors of risk. Cancer Causes Control. 2003;14:879–887. [PubMed]
8. Le Marchand L, Zhao LP, Quiaoit F, Wilkens LR, Kolonel LN. Family history and risk of colorectal cancer in the multiethnic population of Hawaii. Am J Epidemiol. 1996;144:1122–1128. [PubMed]
9. Fuchs CS, Giovannucci EL, Colditz GA, Hunter DJ, Speizer FE, Willett WC. A prospective study of family history and the risk of colorectal cancer. N Engl J Med. 1994;331:1669–1674. [PubMed]
10. Kotake K, Koyama Y, Nasu J, Fukutomi T, Yamaguchi N. Relation of family history of cancer and environmental factors to the risk of colorectal cancer: a case-control study. Jpn J Clin Oncol. 1995;25:195–202. [PubMed]
11. Zeegers MP, Schouten LJ, Goldbohm RA, van den Brandt PA. A compendium of familial relative risks of cancer among first degree relatives: A population-based study. Int J Cancer. 2008;123:1664–1673. [PubMed]
12. Fernandez E, Gallus S, La Vecchia C, Talamini R, Negri E, Franceschi S. Family history and environmental risk factors for colon cancer. Cancer Epidemiol Biomarkers Prev. 2004;13:658–661. [PubMed]
13. Nelson CL, Sellers TA, Rich SS, Potter JD, McGovern PG, Kushi LH. Familial clustering of colon, breast, uterine, and ovarian cancers as assessed by family history. Genet Epidemiol. 1993;10:235–244. [PubMed]
14. Niell BL, Rennert G, Bonner JD, Almog R, Tomsho LP, Gruber SB. BRCA1 and BRCA2 Founder Mutations and the Risk of Colorectal Cancer. J Natl Cancer Inst. 2004;96:15–21. [PubMed]
15. La Vecchia C, Negri E, Franceschi S, Gentile A. Family history and the risk of stomach and colorectal cancer. Cancer. 1992;70:50–55. [PubMed]
16. Goldgar DE, Easton DF, Cannon-Albright LA, Skolnick MH. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst. 1994;86:1600–1608. [PubMed]
17. Phipps RF, Perry PM. Familial breast cancer and the association with colonic carcinoma. Eur J Surg Oncol. 1989;15:109–111. [PubMed]
18. Negri E, Braga C, La Vecchia C, Franceschi S, Filiberti R, Montella M, Falcini F, Conti E, Talamini R. Family history of cancer and risk of colorectal cancer in Italy. Br J Cancer. 1998;77:174–179. [PMC free article] [PubMed]
19. Slattery ML, Kerber RA. Family history of cancer and colon cancer risk: the Utah Population Database. J Natl Cancer Inst. 1994;86:1618–1626. [PubMed]
20. Vaittinen P, Hemminki K. Familial cancer risks in offspring from discordant parental cancers. Int J Cancer. 1999;81:12–19. [PubMed]
21. Dong C, Hemminki K. Modification of cancer risks in offspring by sibling and parental cancers from 2,112,616 nuclear families. Int J Cancer. 2001;92:144–150. [PubMed]
22. Campbell PT, Newcomb P, Gallinger S, Cotterchio M, McLaughlin JR. Exogenous hormones and colorectal cancer risk in Canada: associations stratified by clinically defined familial risk of cancer. Cancer Causes Control. 2007;18:723–733. [PubMed]
23. Slattery ML, Sweeney C, Murtaugh M, Ma KN, Wolff RK, Potter JD, Caan BJ, Samowitz W. Associations between ERalpha, ERbeta, and AR genotypes and colon and rectal cancer. Cancer Epidemiol Biomarkers Prev. 2005;14:2936–2942. [PubMed]
24. Kabat GC, Miller AB, Rohan TE. Oral contraceptive use, hormone replacement therapy, reproductive history and risk of colorectal cancer in women. Int J Cancer. 2008;122:643–646. [PubMed]
25. Hannaford P, Elliott A. Use of exogenous hormones by women and colorectal cancer: evidence from the Royal College of General Practitioners’ Oral Contraception Study. Contraception. 2005;71:95–98. [PubMed]
26. Chlebowski RT, Wactawski-Wende J, Ritenbaugh C, Hubbell FA, Ascensao J, Rodabough RJ, Rosenberg CA, Taylor VM, Harris R, Chen C, Adams-Campbell LL, White E, et al. Estrogen plus Progestin and Colorectal Cancer in Postmenopausal Women. N Engl J Med. 2004;350:991–1004. [PubMed]
27. Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, Kooperberg C, Stefanick ML, Jackson RD, Beresford SA, Howard BV, Johnson KC, Kotchen JM, Ockene J, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women’s Health Initiative Randomized Controlled Trial. JAMA. 2002;288:321–333. [PubMed]
28. Grodstein F, Martinez ME, Platz EA, Giovannucci E, Colditz GA, Kautzky M, Fuchs C, Stampfer MJ. Postmenopausal hormone use and risk for colorectal cancer and adenomas. Ann Intern Med. 1998;128:705–712. [PubMed]
29. Fernandez E, La Vecchia C, D’Avanzo B, Franceschi S, Negri E, Parazzini F. Oral contraceptives, hormone replacement therapy and the risk of colorectal cancer. Br J Cancer. 1996;73:1431–1435. [PMC free article] [PubMed]
30. Gerhardsson de Verdier M, London S. Reproductive factors, exogenous female hormones, and colorectal cancer by subsite. Cancer Causes Control. 1992;3:355–360. [PubMed]
31. Cotterchio M, McKeown-Eyssen G, Sutherland H, Buchan G, Aronson M, Easson AM, Macey J, Holowaty E, Gallinger S. Ontario familial colon cancer registry: methods and first-year response rates. Chronic Dis Can. 2000;21:81–86. [PubMed]
32. Newcomb PA, Baron J, Cotterchio M, Gallinger S, Grove J, Haile R, Hall D, Hopper JL, Jass J, Le Marchand L, Limburg P, Lindor N, et al. Colon Cancer Family Registry: an international resource for studies of the genetic epidemiology of colon cancer. Cancer Epidemiol Biomarkers Prev. 2007;16:2331–2343. [PubMed]
33. World Health Organization. Manual of the international statistical classification of disease, injuries, and causes of death, 9th revision conference. Geneva: World Health Organization; 1977.
34. Vasen HF, Mecklin JP, Khan PM, Lynch HT. The International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer (ICG-HNPCC) Dis Colon Rectum. 1991;34:424–425. [PubMed]
35. Love RR, Evans AM, Josten DM. The accuracy of patient reports of a family history of cancer. J Chron Dis. 1985;38:289–293. [PubMed]
36. Douglas FS, O’Dair LC, Robinson M, Evans DG, Lynch SA. The accuracy of diagnoses as reported in families with cancer: a retrospective study. J Med Genet. 1999;36:309–312. [PMC free article] [PubMed]
37. Ziogas A, Anton-Culver H. Validation of family history data in cancer family registries. Am J Prev Med. 2003;24:190–198. [PubMed]
38. Theis B, Boyd N, Lockwood G, Tritchler D. Accuracy of family cancer history in breast cancer patients. Eur J Cancer Prev. 1994;3:321–327. [PubMed]
39. Bondy ML, Strom SS, Colopy MW, Brown BW, Strong LC. Accuracy of family history of cancer obtained through interviews with relatives of patients with childhood sarcoma. J Clin Epidemiol. 1994;47:89–96. [PubMed]
40. Greenland S. Modeling and variable selection in epidemiologic analysis. Am J Public Health. 1989;79:340–349. [PubMed]
41. Cotterchio M, Manno M, Klar N, McLaughlin J, Gallinger S. Colorectal screening is associated with reduced colorectal cancer risk: a case-control study within the population-based Ontario Familial Colorectal Cancer Registry. Cancer Causes Control. 2005;16:865–875. [PubMed]
42. Ponz de Leon M, Sassatelli R, Sachetti C, Zanghieri G, Scalmati A, Roncucci L. Familial aggregation of tumors in the three-year experience of a population-based colorectal cancer registry. Cancer Res. 1989;49:4344–4348. [PubMed]
43. Mahavni V, Sood AK. Hormone replacement therapy and cancer risk. Curr Opin Oncol. 2001;13:384–389. [PubMed]
44. Gambacciani M, Monteleone P, Sacco A, Genazzani AR. Hormone replacement therapy and endometrial, ovarian and colorectal cancer. Best Pract Res Clin Endocrinol Metab. 2003;17:139–147. [PubMed]
45. Bernstein L. The Risk of Breast, Endometrial and Ovarian Cancer in Users of Hormonal Preparations. Basic Clin Pharmacol Toxicol. 2006;98:288–296. [PubMed]
46. Purdie DM, Green AC. Epidemiology of endometrial cancer. Best Pract Res Clin Obstet Gynaecol. 2001;15:341–354. [PubMed]
47. Burkman RT, Collins JA, Shulman LP, Williams JK. Current perspectives on oral contraceptive use. Am J Obstet Gynecol. 2001;185:S4–S12. [PubMed]
48. La Vecchia C. Oral contraceptives and ovarian cancer: an update, 1998–2004. Eur J Cancer Prev. 2006;15:117–124. [PubMed]
49. La Vecchia C, Bosetti C. Benefits and risks of oral contraceptives on cancer. Eur J Cancer Prev. 2004;13:503–508. [PubMed]
50. Carruba G. Estrogen and Prostate Cancer: An Eclipsed Truth in an Androgen-Dominated Scenario. J Cell Biochem. 2007;102:899–911. [PubMed]
51. Ricke WA, Wang Y, Cunha GR. Steroid hormones and carcinogenesis of the prostate: the role of estrogens. Differentiation. 2007;75:871–882. [PubMed]
52. Härkönen PL, Mäkelä SI. Role of estrogens in development of prostate cancer. J Steroid Biochem Mol Biol. 2004;92:297–305. [PubMed]
53. Haiman CA, Le Marchand L, Yamamato J, Stram DO, Sheng X, Kolonel LN, Wu AH, Reich D, Henderson BE. A common genetic risk factor for colorectal and prostate cancer. Nat Genet. 2007;39:954–956. [PMC free article] [PubMed]
54. Ghoussaini M, Song H, Koessler T, Al Olama AA, Kote-Jarai Z, Driver KE, Pooley KA, Ramus SJ, Kjaer SK, Hogdall E, DiCioccio RA, Whittemore AS, et al. Multiple loci with different cancer specificities within the 8q24 gene desert. J Natl Cancer Inst. 2008;100:962–966. [PMC free article] [PubMed]
55. Rokavec M, Justenhoven C, Schroth W, Istrate MA, Haas S, Fischer HP, Vollmert C, Illig T, Hamann U, Ko YD, Glavac D, Brauch H. A novel polymorphism in the promoter region of ERBB4 is associated with breast and colorectal cancer risk. Clin Cancer Res. 2007;13:7506–7514. [PubMed]
56. Kerber RA, Slattery ML. Comparison of self-reported and database-linked family history of cancer data in a case-control study. Am J Epidemiol. 1997;146:244–248. [PubMed]
57. Sijmons RH, Boonstra AE, Reefhuis J, Hordijk-Hos JM, de Walle HE, Oosterwijk JC, Cornel MC. Accuracy of family history of cancer: clinical genetic implications. Eur J Hum Genet. 2000;8:181–186. [PubMed]
58. Verkooijen HM, Fioretta G, Chappuis PO, Vlastos G, Sappino AP, Benhamou S, Bouchardy C. Set-up of a population-based familial breast cancer registry in Geneva, Switzerland: validation of first results. Ann Oncol. 2004;15:350–353. [PubMed]
59. King TM, Tong L, Pack RJ. Spencer C, Amos CI. Accuracy of family history of cancer as reported by men with prostate cancer. Urology. 2002;59:546–550. [PubMed]
60. Parent ME, Ghadirian P, Lacroix A, Perret C. Accuracy of reports of familial breast cancer in a case-control series. Epidemiology. 1995;6:184–186. [PubMed]
61. Ramji F, Cotterchio M, Manno M, Rabeneck L, Gallinger S. Association between subject factors and colorectal cancer screening participation in Ontario, Canada. Cancer Detect Prev. 2005;29:221–226. [PubMed]