BRCA1 and BRCA2 Gene Mutations and Colorectal Cancer Risk

Systematic Review and Meta-analysis

Mok Oh; Ali McBride; Seongseok Yun; Sandipan Bhattacharjee; Marion Slack; Jennifer R. Martin; Joanne Jeter; Ivo Abraham

Disclosures

J Natl Cancer Inst. 2018;110(11):1178-1189. 

In This Article

Systematic Review

Study Characteristics

Of the 18 studies in the systematic review, five were cohort studies with ascertained BRCA mutation carriers, six were cohort studies involving pedigree analysis, five were case–control studies, and two were kin–cohort studies. These studies are summarized in Table 1.

Cohort Studies With Ascertained BRCA Mutation Carriers (n = 5)

Mersch et al.[21] included 613 BRCA1 and 459 BRCA2 mutation carriers from patients receiving genetic counseling at the University of Texas MD Anderson Cancer Center Genetics clinics between 1997 and 2013. The number of colorectal cancer cases observed in the sample was compared with the expected number in the general population, which was calculated from the number of participants in the study sample multiplied by age-specific (5-year intervals) colorectal cancer incidence rates in the population-based incidence rates (United States Cancer Statistics, 1999–2010). The majority of the carriers were women (n = 1010, 94.2%) and white (n = 803, 74.9%). The SIRs were 1.58 (95% CI = 0.58 to 3.44) for BRCA1 and 0.53 (95% CI = 0.06 to 1.90) for BRCA2, indicating no statistically significant risk of colorectal cancer between mutation carriers and population-based incidence rates.

Phelan et al.[22] followed 7015 women with a BRCA1 or BRCA2 mutation from 50 centers in the United States, Canada, Poland, France, and Norway enrolled from 1992 to 2010 and followed for an average of 5.5 years. Age-standardized incidence rates for each participating country were obtained from GLOBOCAN 2008. In total, 21 incident colorectal cancer cases were observed among all mutation carriers, which was less than the 23.5 cases expected from the population-specific incidence rates. The age-adjusted SIR was 0.92 (95% CI = 0.54 to 1.40) for BRCA1 and 0.82 (95% CI = 0.30 to 1.80) for BRCA2 (Table 1), indicating no increased risk of colorectal cancer in the mutation carriers. However, in an adequately powered subanalysis of women younger than age 50 years carrying the BRCA1 mutation (n = 2638), there was a 4.76 (95% CI = 2.21 to 9.00) statistically significant increase in the risk of colorectal cancer.

Kadouri et al.[23] conducted a historical cohort study of 1098 Ashkenazi Jewish women with breast and/or ovarian cancer seen in the cancer genetics clinics of Hadassah Medical Center (Jerusalem, Israel) between 1995 and 2003. Of the 329 BRCA-positive subjects, 229 were BRCA1, and 100 were BRCA2 mutation carriers; leaving a comparison group of 769 noncarriers (Table 1). Six colorectal cancers were identified among BRCA1 and two among BRCA2 mutation carriers compared to 12 among noncarriers; corresponding to statistically nonsignificant ORs of 1.70 (95% CI = 0.63 to 4.57) and 1.29 (95% CI = 0.29 to 5.84), respectively. The study incorporated time to colorectal cancer by using a Cox proportional hazard analysis with the time of onset of colorectal cancer as the endpoint. This yielded HRs of 3.90 (95% CI = 1.30 to 12.10) for BRCA1 carriers and 2.31 for BRCA2 carriers (95% CI = 0.50 to 11.30), indicating that BRCA1 but not BRCA2 mutation carriers had a statistically significant higher colorectal cancer risk than noncarriers.

Thompson and Easton[11] reviewed data from 699 families with at least one known BRCA1 mutation carrier from 30 centers across Europe and North America. Of the 11 847 subjects, 2245 were BRCA1 mutation carriers, 1106 were noncarriers, and 8496 individuals had not been tested for mutations. The colorectal cancer incidence rate in the 2245 BRCA1 mutation carriers was used to estimate relative risk. Fourteen cases of colon cancer were screened among BRCA1 mutation carriers, compared with 7.36 expected cases from the SEER incidence rates for Caucasians (1983–1987), yielding a statistically significant relative risk for colon cancer in BRCA1 carriers of 2.03 (95% CI = 1.45 to 2.85). In addition, in the 8496 family members not tested for BRCAmutations, 76 cases of colon cancer were found compared to 48.77 expected cancers, resulting in a statistically significant SIR of 1.46 (95% CI = 1.24 to 1.94).

In a cohort study, Lin et al.[24] compared the lifetime colorectal cancer risk in BRCA mutation carriers relative to the general population. The Creighton Hereditary Cancer Institute Registry (1972–1997) and the Creighton University Tumor Registry (1965–1996) identified 164 BRCA1 and 88 BRCA2 gene mutation carriers and compared their colorectal cancer risk with the lifetime risk from the SEER Cancer Statistics Review (1993–1995). The risk in female and male BRCA mutation carriers was 3.2% and 5.6%, respectively; which was statistically not significantly different from the corresponding 5.9% and 6.0% risks in the general female and male populations.

Summarized, in this set of five cohort studies with ascertained BRCA mutation carriers, two studies reported an increased risk of colon and colorectal cancer among BRCA1 mutation carriers. The remaining studies, all of which included both BRCA1and BRCA2 carriers and focused on colorectal cancer, found no increased cancer risk relative to population-specific incidence rates.

Cohort Studies Involving Pedigree Analysis (n = 6)

Moran et al.[25] assessed 268 BRCA1 families and 222 BRCA2 families including 3341 individuals who had tested positive for BRCA1 or BRCA2 mutations (n = 1148) or who had not been tested, although a relative had tested positive (n = 2193) (Table 1). In total, 1815 possible BRCA1 carriers and 1526 possible BRCA2 carriers were included. Person-years from 1975 to 2005 were used for risk analysis, and an assumption was made that 50% of those not tested carried the BRCA mutation. SIR was estimated from incidence rates for the northwest of England for the same time frame. For BRCA1 mutation carriers, 11 incident colorectal cancer cases were observed, which was less than the 11.4 cases expected, yielding a statistically nonsignificant SIR of 1.00 (95% CI = 0.50 to 1.70). For BRCA2 mutation carriers, 10.4 incident colorectal cancer cases were observed, compared to 10.3 expected cases, yielding a statistically nonsignificant SIR of 1.00 (95% CI = 0.50 to 1.80) for BRCA2 mutation carriers.

Van Asperen et al.[26] conducted a retrospective cohort study in eight clinical genetics centers in the Netherlands of the members of 139 families in which someone had breast and/or ovarian cancer (Table 1). Pedigrees were drawn for all members, and the youngest typed carrier and his/her first-degree relatives were selected for the cohort. These presumed BRCA1 or BRCA2 mutation carriers comprised 1811 individuals, including 1008 women (55.6%) and 803 men (44.3%). The observed number of colon cancers in the carrier group was 20, while 16.5 were expected based on the Netherlands Cancer Statistics (1990–2005), yielding a statistically significant relative risk of 1.50 (95% CI = 0.90 to 2.30) and indicating an increased risk of colorectal cancer in presumed BRCA1 or BRCA2 mutation carriers.

Brohet et al.[27] reported on part of a nationwide study in the Netherlands of 517 families with the BRCA1 mutation (Table 1). The risk of other cancers was estimated through an assumption of 50% cohort family members being a carrier. These presumed BRCA1 mutation carriers comprised 6585 individuals, including 2843 women (43.2%) and 3742 men (56.8%). The observed number of colon cancers in the BRCA1 mutation carrier group was 93, whereas 53.1 were expected based on the Netherlands Cancer Statistics (1990–2005), yielding a statistically significant relative risk of 2.51 (95% CI = 2.02 to 3.07) and indicating an increased risk of colon cancer in presumed BRCA1 mutation carriers.

Brose et al.[28] examined the colon cancer risk in 147 families of patients from a breast cancer risk clinic in the United States (Table 1). The sample of 483 participants included 316 subjects who tested positive for BRCA1 and 167 presumed carriers, which referred to siblings or parents of the identified BRCA1 mutation carriers. The cumulative lifetime colon cancer risk in the 483 BRCA1 mutation carriers was compared with the population colon cancer risk of 5.6% in the SEER data. Nineteen cases of colon cancer (11.0%) were identified among carriers, resulting in a relative risk of 2.00 for colorectal cancer in BRCA1 mutation carriers, but the study did not report confidence intervals and P values.

The Breast Cancer Linkage Consortium (BCLC)[29] examined colon cancer risks in 173 families with BRCA2 mutations. The final cohort comprised 3728 individuals, including 681 individuals with breast or ovarian cancer, 471 unaffected BRCA2 mutation carriers, 390 tested noncarriers, and 1186 individuals with unknown BRCAmutation status (Table 1). Eight colon cancer cases were found in the BRCA2 mutation group compared with 6.56 colon cancer cases expected from SEER data, resulting in a statistically nonsignificant relative risk of 1.43 (95% CI = 0.79 to 2.58) and indicating no increased risk of colorectal cancer in the BRCA2 mutation carriers.

Ford et al.[14] examined data from a cohort of 33 families contributed by research groups in North America and Western Europe (Table 1). Each family included at least four people with ovarian or breast cancer diagnosed before age 60 years with evidence of linkage to BRCA1 mutation. The incidence of colon cancer in BRCA1 mutation carriers was compared with the expected numbers of colon cancers from the general population using age- and sex-specific incidence rates from SEER for the United States, and England and Wales incidence rates for Europe. There were seven colon cancer cases among the 464 BRCA1 mutation carriers compared with 2.22 expected cases for the general population, yielding a statistically significant relative risk for colon cancer among BRCA1 mutation carriers of 4.11 (95% CI = 2.36 to 7.15).

In summary, in this set of six cohort studies involving pedigree analysis, four studies reported no increased risk of colon cancer among BRCA1 or BRCA2 mutation carriers and one study reported no increased risk of colorectal cancer among BRCA1 or BRCA2 mutation carriers. The remaining one study showed an increased risk of colon cancer among BRCA1 mutation carriers.

Case–control Studies (n = 5)

Suchy et al.[30] compared 2398 unselected colorectal cancer patients newly diagnosed between 1998 and 2008 in nine centers in Poland to 4570 individuals without colorectal cancer from 10 hospitals throughout Poland (Table 1). Ten BRCA1mutations were identified in the case group and 22 in the control group for an OR of 0.80 (95% CI = 0.40 to 1.70), indicating no increased risk of colorectal cancer in BRCA1 mutation carriers. In a subanalysis of subjects diagnosed before the age of 60 years, the OR was 1.70 (95% CI = 0.70 to 4.00), also statistically nonsignificant. Subanalyses of associations between specific BRCA1 mutations and colorectal cancer risk yielded no statistically significant ORs (C61G: OR = 2.50, 95% CI = 0.60 to 11.40; 5382insC: OR = 0.60, 95% CI = 0.20 to 1.60)

Kirchhoff et al.[31] compared 586 unselected Ashkenazi Jewish colorectal cancer patients to 5012 unaffected Ashkenazi Jews as a control group (Table 1). In the case group, six BRCA1 or BRCA2 mutation carriers were identified compared with 118 BRCA1 or BRCA2 mutation carriers among control subjects. Among these six carriers, two were 185delAG carriers, one was a 5382insC carrier, and three were 6147delT carriers. This yielded an unadjusted OR of 0.50 (95% CI = 0.22 to 1.14), which was statistically nonsignificant and indicates no increased risk of colorectal cancer in the BRCA1 or BRCA2 mutation carriers.

Niell et al.[32] tested 1422 unselected colorectal cancer patients matched to 1566 control subjects without colorectal cancer (Table 1). Matching was done through logistic regression analysis adjusting for age at diagnosis and sex. Twenty-four BRCA1 or BRCA2 carriers were identified in the case group and 20 in the control group, for an OR of 1.24 (95% CI = 0.68 to 2.26), which was statistically nonsignificant and indicates no increased risk of colorectal cancer in the BRCA1 or BRCA2 mutation carriers. In a subgroup analysis stratified by BRCA type, BRCA1yielded an OR of 1.26 (95% CI = 0.52 to 3.06) and BRCA2 yielded an OR of 1.22 (95% CI = 0.54 to 2.73). Another subgroup analysis of participants younger than 65 years yielded a statistically nonsignificant OR of 3.14 (95% CI = 0.64 to 15.43). Subanalyses of associations between specific BRCA1 and BRCA2 mutations and colorectal cancer risk yielded no statistically significant ORs (187delAG: OR = 1.03, 95% CI = 0.39 to 2.77; 5385insC: OR = 3.11, 95% CI = 0.32 to 29.33; 187delAG or 5385insC: OR = 1.26, 95% CI = 0.52 to 3.06; 6147delT: OR = 1.22, 95% CI = 0.54 to 2.73).

Chen-Shtoyerman et al.[33] screened 225 unselected Ashkenazi Jewish colorectal cancer patients for BRCA1 or BRCA2 mutations and compared them with the mutation frequency in the general Ashkenazi Jewish population of 5318 without colorectal cancer (Table 1). Four carriers were identified (1.8%), which was not statistically significantly different from the 120 carriers from the general Ashkenazi Jewish population (2.3%). Among these four carriers, one was a 185delAG carrier, one was a 5382insC carrier, and two were 6147delT carriers. As the study did not report ORs, confidence intervals, and P values, the unadjusted OR was calculated to be 0.78 (95% CI = 0.29 to 2.14), which was not statistically significant and indicates no increased risk of colorectal cancer in the BRCA1 or BRCA2 mutation carriers.

In a preliminary study, Drucker et al.[34] screened 136 Israeli Jewish colorectal cancer patients for BRCA1 or BRCA2 mutations and compared Ashkenazi to non-Ashkenazi, but did not report statistical results beyond counts and rates (Table 1). Among Ashkenazi Jews, 3 of 87 (3.5%) carriers had colorectal cancer, whereas no carriers were found among 49 non-Ashkenazi Jews, indicating a possible elevated risk of BRCA mutation in Ashkenazi compared to non-Ashkenazi colorectal cancer patients.

Summarized, in this set of five case–control studies, three studies reported no increased risk of colorectal cancer among BRCA1 or BRCA2 mutation carriers and one study reported no increased risk of colorectal cancer among BRCA1 mutation carriers. The remaining study was a preliminary descriptive study showing a possible increased colorectal cancer risk in Ashkenazi Jews. The two studies reporting a subgroup analysis for early-onset colorectal cancer did not infer an age-related risk.

Cohort Studies Involving Kindred (n = 2)

Risch et al.[35] screened 649 unselected incident cases of ovarian cancer diagnosed in Ontario, Canada, during 1995–1996 for mutations in BRCA1 and BRCA2. Of the 515 women with invasive ovarian cancer, 60 had a BRCA mutation, including 39 with BRCA1 and 21 with BRCA2. Family histories of colorectal cancer were compared in mutation carriers and noncarriers. The incidence rate of colorectal cancer among first-degree relatives was compared with the incidence rate for 455 noncarriers from unselected ovarian cancer cases. A total of 2.9% incidence were observed among BRCA1 and 10.3% among BRCA2 carriers compared to 4.2% expected in the comparison group. The relative risk of colorectal cancer for BRCA1 mutation carriers was 0.70 (95% CI = 0.17 to 2.80), which was statistically nonsignificant, indicating no increased risk of colorectal cancer among first-degree relatives of the BRCA1 mutation carriers. However, for BRCA2 mutation carriers, the relative risk was 2.50 (95% CI = 1.00 to 6.30), which was statistically significant (Table 1).

Struewing et al.[36] screened 5318 Ashkenazi Jewish volunteers for BRCA1 and BRCA2mutations and compared family histories of colon cancer in 114 BRCA mutation carriers and 4759 noncarriers (Table 1). Nine (7.8%) colon cancer family history cases were reported in the BRCA carriers and 509 (10.6%) in the noncarriers group. As relative risk, confidence interval, and P value were not reported, we estimated the relative risk to be 0.71 (95% CI = 0.36 to 1.42), which was not statistically significant and indicated no increased risk of colorectal cancer family histories in the BRCA mutation carriers. Summarized, the two kin–cohort studies reported no increased risk of colorectal or colon cancer family histories among BRCA1 mutation carriers. One study showed an increased risk of colorectal cancer family history among BRCA2 mutation carriers.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....