The importance of a family history of breast cancer in predicting the presence of a BRCA mutation.

To the Editor:Hartge et al. (1999xThe prevalence of common BRCA1 and BRCA2 mutations among Ashkenazi Jews. Hartge, P, Struewing, JP, Wacholder, S, Brody, LC, and Tucker, MA. Am J Hum Genet. 1999; 64: 963–970Abstract | Full Text | Full Text PDF | PubMed | Scopus (122)See all References1999) describe the prevalence of the three founder Ashkenazi Jewish (AJ) mutations in BRCA1 (MIM 113705) and BRCA2 (MIM 600185) in 5,290 AJ volunteers from the Washington, DC, area. They report an overall mutation frequency of 2.3%, ranging from 1.2%, in those with no personal or first-degree-relative history of breast or ovarian cancer, to 50%, in women diagnosed with breast or ovarian cancer at age <40 years who had at least one first-degree relative with breast cancer diagnosed at age <50 years. The authors demonstrate, as we and others (Karp et al. 1997xInfluence of BRCA1 mutations on nuclear grade and estrogen receptor status on breast carcinoma in Ashkenazi Jewish women. Karp, SE, Tonin, PN, Begin, LR, Martinez, JJ, Zhang, JC, Pollak, MN, and Foulkes, WD. Cancer. 1997; 80: 435–441Crossref | PubMed | Scopus (142)See all References1997; Shattuck-Eidens et al. 1997xBRCA1 sequence analysis in women at high risk for susceptibility mutations: risk factor analysis and implications for genetic testing. Shattuck-Eidens, D, Oliphant, A, McClure, M, McBride, C, Gupte, J, Rubano, T, Pruss, D et al. JAMA. 1997; 278: 1242–1250Crossref | PubMedSee all References1997; Fodor et al. 1998xFrequency and carrier risk associated with common BRCA1 and BRCA2 mutations in Ashkenazi Jewish breast cancer patients. Fodor, FH, Weston, A, Bleiweiss, IJ, McCurdy, LD, Walsh, MM, Tartter, PI, Brower, ST et al. Am J Hum Genet. 1998; 63: 45–51Abstract | Full Text | Full Text PDF | PubMed | Scopus (154)See all References1998) have done, that, for the 297 women in their study with breast or ovarian cancer, the probability of carrying a BRCA1 or BRCA2 (BRCA) mutation decreases as age at diagnosis increases. Hartge et al. (1999xThe prevalence of common BRCA1 and BRCA2 mutations among Ashkenazi Jews. Hartge, P, Struewing, JP, Wacholder, S, Brody, LC, and Tucker, MA. Am J Hum Genet. 1999; 64: 963–970Abstract | Full Text | Full Text PDF | PubMed | Scopus (122)See all References1999, p. 965) state that, given age-at-onset information, “family history discriminated relatively little if the participant herself developed breast cancer, whereas, among other participants, family history best discriminated carriers from non-carriers.” The age of the proband is clearly a powerful predictor of carrier probability, but our experience is that family history is an important determinant of the probability of a mutation, in both unaffected and affected women. Therefore, we reanalyzed Hartge et al.'s data, estimating relative risks of carrying a BRCA mutation for each age-at-diagnosis group (stratified by decade), in association with a first-degree-relative family history of breast cancer at any age (“positive family history”) and in association with a first-degree-relative family history of at least one case of breast cancer diagnosed at age <50 years (“positive early-onset family history”). We analyzed affected and unaffected women separately. In affected women, the Mantel-Haenszel (M-H) odds ratio (OR), stratified by age at onset, for the association between a positive family history and the presence of a founder BRCA mutation, compared with a negative family history, was 2.6 (95% confidence interval [CI] 1.2–6.0, P=.022; table 1table 1, first “OR” column). For unaffected women, the M-H OR for the presence of a mutation in women with a positive family history was 3.1 (95% CI 1.9–5.1, P<.001; table 1table 1, second “OR” column). For affected women, the M-H OR for the presence of a BRCA mutation in association with a positive early-onset family history was 4.4 (95% CI 1.7–11.4, P=.003; table 2table 2, first column). This OR is greater than that observed in unaffected women with a positive early-onset family history (M-H OR 3.6, P<.001; table 2table 2, second “OR” column). Thus, a family history of breast cancer is as predictive of the presence of a BRCA mutation in affected women as it is in unaffected women. The M-H OR for the unaffected and affected subgroups is similar and has overlapping CIs. The significance levels do differ, but this reflects the much larger size of the subgroup of unaffected women (n=4,993; 94.4%) compared with those with breast or ovarian cancer (n=297; 5.6%). In addition, a comparison of the strata-specific ORs in unaffected and affected women does not reveal a consistent pattern: none of the within-strata ORs differ statistically—the smallest P value is .38 (table 1table 1, “OR” column 1 vs. column 2; table 2table 2, “OR” column 1 vs. column 2).Table 1M-H OR for the Presence of a BRCA Mutation in Association with a Positive Family History of Breast Cancer, Stratified by Age or Age at OnsetHartge et al. (1999)Warner et al. (1999): Affected WomenAffected WomenUnaffected WomenNo. ofNo. ofNo. ofNoncarriersCarriersbbOR [95%CI]ccNoncarriersCarriersbbOR [95%CI]ccNoncarriersCarriersbbOR [95%CI]ccAgeaa <40 years FH× −2161.0557 91.014 71.0 FH× + 432.6 [.5–15.0]114105.4 [2.4–12.5] 3 64.0 [.8–20.6] 40–49 years FH× −7161.0874141.080131.0 FH× +2752.2 [.6–7.6]215 92.6 [1.2–5.9]31102.0 [.8–5.0] 50–59 years FH× −5721.0628 81.081 41.0 FH× +1946.0 [1.2–30.2]169 62.8 [.9–7.8]21 54.8 [1.3–17.8] ≥60 years FH× −4611.0611 41.095 21.0 FH× +250.6 [.0–15.5]189 21.6 [.3–8.8]39 11.2 [.1–13.9]M-H2.6 [1.2–6.0]3.1 [1.9–5.1]2.6 [1.4–5.0]PPPHomogeneity.50.53.63Unity.022<.001.004aFH×= family history of breast cancer in any first-degree relative. The minus sign (−) indicates negative; the plus sign (+) indicates positive.bOf a founder AJ BRCA1 or BRCA2 mutation.cThe logit estimate of the odds ratio was used when there was a zero cell.Table 2M-H OR for the Presence of a BRCA Mutation in Association with a Positive Family History of Early-Onset Breast Cancer, Stratified by Age or Age at OnsetHartge et al. (1999)Warner et al. (1999): Affected WomenAffected WomenUnaffected WomenNo. ofNo. ofNo. ofNoncarriersCarriersbbOR [95%CI]ccNoncarriersCarriersbbOR [95%CI]ccNoncarriersCarriersbbOR [95%CI]ccAgeaa <40 years FH×50 −2371.0 614141.0 17111.0 FH×50 + 223.3 [.4–26.4] 57 53.9 [1.4–10.3] 0 27.6 [.3–173] 40–49 years FH×50 −9281.01014171.0 97171.0 FH×50 + 635.8 [1.4–23.9] 75 64.8 [2.0–11.4] 14 62.5 [.8–7.1] 50–59 years FH×50 −7041.0 743121.0 90 41.0 FH×50 + 625.8 [1.0–32.6] 54 22.3 [.5–10.1] 12 59.4 [2.7–33.1] ≥60 years FH×50 −6111.0 73451.0122 21.0 FH×50 +1002.0 [.1–51.2] 6612.2 [.3–18.3] 12 15.1 [.548.0]M-H4.4 [1.7–11.4]3.6 [2.0–6.4]4.4 [2.1–9.2]PPPHomogeneity.81.83.40Unity.003<.001<.001aFH×50 = family history of breast cancer at age <50 years in any first-degree relative. The minus sign (−) indicates negative; the plus sign (+) indicates positive.bOf a founder AJ BRCA1 or BRCA2 mutation.cThe logit estimate of the odds ratio was used when there was a zero cell.To assess whether our reinterpretation of the Washington, DC, data set is valid, we performed the same analysis in 412 prevalent cases of breast cancer diagnosed in AJ women, ascertained between November 1, 1996, and May 31, 1998, in Toronto and Montreal (Warner et al. 1999xPrevalence and penetrance of BRCA1 and BRCA2 mutations in unselected Ashkenazi Jewish women with breast cancer. Warner, E, Foulkes, W, Goodwin, P, Meschino, W, Blondal, J, Patterson, C, Ozcelik, H et al. J Natl Cancer Inst. 1999; 91: 1241–1247Crossref | PubMedSee all References1999). To compare exactly with the Washington, DC, study, we included only first-degree relatives with breast cancer in the analyses here. The definition of early-onset breast cancer was age at diagnosis of <50 years. The results are shown in tables 11 and 22, “OR” column 3. Notably, the M-H ORs seen in the Canadian study are identical to that observed in affected women in the Washington, DC, study: M-H OR 2.6 for the presence of a BRCA mutation in association with any first-degree-relative history of breast cancer and 4.4 for a positive early-onset family history. Thus, the findings from the Canadian study support our interpretation of the data published by Hartge et al. (1999xThe prevalence of common BRCA1 and BRCA2 mutations among Ashkenazi Jews. Hartge, P, Struewing, JP, Wacholder, S, Brody, LC, and Tucker, MA. Am J Hum Genet. 1999; 64: 963–970Abstract | Full Text | Full Text PDF | PubMed | Scopus (122)See all References1999) and lead us to question those authors' conclusion that the knowledge gained from knowing the family history of an affected person is “relatively small.” The weight of evidence from clinical experience, from previously published work, and from their own study supports the conclusion that family history and age at diagnosis of breast cancer are both important factors in indicating the likely presence of a mutation in BRCA1 or BRCA2.