The incidence of leukaemia in women with BRCA1 and BRCA2 mutations: an International Prospective Cohort Study.

BACKGROUND: Germline mutations in BRCA1 and BRCA2 increase the susceptibility to develop breast and ovarian cancers as well as increase the risk of some other cancers. Primary objective was to estimate the risk of leukaemia in BRCA1 and BRCA2 mutation carriers.
METHODS: We followed 7243 women with a BRCA1 or a BRCA2 mutation for incident cases of leukaemia. We used the standardised incidence ratio (SIR) to estimate the relative risk of leukaemia, according to mutation and history of breast cancer.
RESULTS: We identified five incident cases of leukaemia (two BRCA1, three BRCA2). All five women had a prior history of breast cancer and four had received chemotherapy. The mean time from breast cancer diagnosis to the development of leukaemia was 10.2 years (range 3-18 years). The SIR for BRCA1 carriers was 0.66 (95% CI: 0.11-2.19, P=0.61) and the SIR for BRCA2 carriers was 2.42 (95% CI: 0.61-6.58, P=0.17). The SIR was significantly higher than expected for women with a BRCA2 mutation and breast cancer (SIR=4.76, 95% CI:1.21-12.96, P=0.03), in particular for women who received chemotherapy (SIR=8.11, 2.06-22.07, P=0.007).
CONCLUSIONS: We observed an increased risk of leukaemia in women with a BRCA2 mutation who receive chemotherapy for breast cancer.

Germline mutations in the BRCA1 and the BRCA2 genes increase the susceptibility for a woman to develop breast and ovarian cancer (Miki ; Wooster ). Women with an inherited mutation in the BRCA1 or the BRCA2 gene have a life-time risk of ∼70% of developing breast cancer (Claus ; Ford ; Antoniou ; Chen and Parmigiani, 2007; Mavaddat ). Reports from the Breast Cancer Linkage Consortium suggested that women with BRCA1 and BRCA2 mutations have elevated risk of cancers other than breast and ovarian cancers (The Breast Cancer Linkage Consortium, 1999; Thompson ). However, it is not clear to what extent a mutation in BRCA genes affects the risk of leukaemia.

Cytotoxic chemotherapy for breast cancer is known to increase the risk of therapy-related myeloid neoplasms (Swerdlow ). Chemotherapy-related myeloid neoplasms constitute up to 20% of all acute myeloid leukaemias and myelodysplastic syndromes (Morton ). The National Surgical Adjuvant Breast and Bowel Project trials suggest that the use of an intense dose cytotoxic chemotherapy (such as adriamycin and cyclophosphamide), and radiation therapy for breast cancer is associated with a higher incidence of acute myeloid leukaemias, and myelodysplastic syndrome (Smith, 2003).

Most women who develop a BRCA1- or a BRCA2-associated breast cancer are likely to receive chemotherapy and radiation therapy. It is hypothesised that faulty mechanisms, such as altered cellular signalling, enhanced proliferation, diminished apoptosis, and abnormal differentiation increase the risk of chemotherapy-related leukaemias and myelodysplastic syndromes in patients with the BRCA-associated breast cancer (Friedenson, 2007). Some case reports have suggested that mutations in BRCA1 or the BRCA2 gene might add to the risk of developing leukaemia in patients who receive chemotherapy and radiation therapy for their breast cancer (Fruscalzo ; Cole and Strair, 2010). However, there are no prospective data on the subject.

We have previously reported on the incidence of pancreatic cancer, endometrial cancer, and colorectal cancer in BRCA mutation carriers (Iqbal ; Segev ; Phelan ). Here we report the incidence of leukaemia in a prospective cohort of BRCA1 and the BRCA2 mutation carriers, and also estimate the incidence of leukaemia in women with a BRCA mutation who were treated with chemotherapy for breast cancer.

Materials and Methods

Study cohort

We identified a total of 12 310 women with a BRCA1 or BRCA2 mutation through a registry of mutation carriers at the Women's College Research Institute, Toronto. The registry collects data from women who have a known pathogenic BRCA1 or BRCA2 mutation at 50 different centres in 11 countries in North America and Europe. The study protocol was approved by Ethics Committees/Human Subjects Review Boards of all participating centres. We obtained informed consent from all women before genetic testing. Mutation detection was confirmed by the direct sequencing of DNA.

Women were eligible for study entry if they were 25–74 years of age, and had a mutation in either the BRCA1 or the BRCA2 gene. All BRCA carrier women had completed a risk assessment questionnaire at study entry and a follow-up questionnaire every 2 years thereafter. The questionnaires collected information on all new cancers diagnosed including site of cancer and age of diagnosis. At some centres, the questionnaires were completed by telephone interview. The eligible women completed at least one follow-up questionnaire. The diagnosis of leukaemia was confirmed by reference to a pathology report or medical record in all cases.

We excluded women who were <25 years (n=454; 3.0%) and >75 years of age (n=107; 0.71%) at baseline, who had no date of birth recorded (n=5; 0.03%), who had not completed a follow-up questionnaire (n=3,434; 22.9%), who were lost to follow-up (n=986; 6.6%), and who had a missing date of death (n=81; 0.54), leaving 7243 women who were confirmed carriers eligible for study entry. We followed all women from the date of study entry (baseline) until either: 1) the date of completion of the last follow-up questionnaire, 2) diagnosis of leukaemia, and 3) death from any cause. Table 1 describes characteristics of study cohort.

Table 1

Baseline characteristics of study cohort

Characteristic	All women N=7243	BRCA1 N=5397	BRCA2 N=1790	Othera N=56
Age, years
Mean (min, max)	47.0 (25–74.9)	46.2 (25–74.9)	49.4 (25–74.9)	48.3 (28–73)
Breast cancer
Yes	3200	2353	813	34
No	4040	3042	976	22
Not known	3	2	1	0
Chemotherapyb
Yes	2346	1776	540	30
No	788	544	241	3
Not known	66	33	32	1
Country
Canada	2255	1229	994	32
Italy	103	84	19	0
USA	2185	1414	749	22
Poland	2700	2670	28	2

Other category includes women who were confirmed carriers of either BRCA mutation, but the type of BRACA mutation (BRCA1 or BRCA2) was not known.

Analysis is restricted to women who developed breast cancer (n=3200).

Incidence and standardised incidence ratios

We estimated the incidence of leukaemia by dividing the total number of new leukaemia cases by the total person-years of follow-up. The incidence was calculated as rate per 100 000 women per year. We calculated the risk of developing leukaemia in the BRCA1 and the BRCA2 mutation carriers relative to general population by calculating standardised incidence ratios (SIR). We calculated the SIRs in the following steps. First, we obtained incidence rates of leukaemia for each participating country using GLOBCAN 2012 (Ferlay ). The incidence rates were calculated for each 5-year age bins (<40, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74). Second, we calculated the number of person-years for each age bin in our carrier database. We then calculated the expected number of leukaemia cases for each age bin by multiplying incidence rate of leukaemia with person-years divided by 100 000. The SIRs were calculated as the ratio of observed cancers to expected cancers. The SIRs were calculated according to the type of BRCA mutation (BRCA1, BRCA2, combined), country of residence (Canada, Italy, Poland, and the United States), age groups (<50 years, ⩾50 years), breast cancer status (yes/no), and chemotherapy (yes/no) for breast cancer.

Results

After a mean follow-up time of 6.1 years (range 1.0–19.5 years), five women (two BRCA1, three BRCA2) were diagnosed with leukaemia. Table 2 presents characteristics of incident cases of leukaemia. Three women (one BRCA1, two BRCA2) were diagnosed with acute myeloid leukaemia and one woman (BRCA2) developed chronic lymphocytic leukaemia; one women (BRCA1) developed acute leukaemia but the subtype of leukaemia could not be determined. The mean age at diagnosis of leukaemia was 53.4 years. The mean age was 43 years for BRCA1 mutation carriers and was 60.3 years for BRCA2 mutation carriers.

Table 2

Characteristics of incidence cases of leukaemia in BRCA carriers

Case no.	Age at breast cancer diagnosis	Treatment of breast cancer	Age at leukaemia diagnosis	Subtype of leukaemia	BRCA mutation	Mutation description	Status	Country
1	35	No treatment	49	Acute leukaemiaa	BRCA1	Exon 11 3913delA	Dead	Italy
2	34	Surgery Chemotherapy	37	Acute myeloid leukaemia (M3)	BRCA1	Exon 12 Q1395X	Alive	Canada
3	52	Surgery Chemotherapy Tamoxifen	70	Acute myeloid leukaemia	BRCA2	Exon 11 6024delTA	Dead	Italy
4	50	Surgery Chemotherapy Radiation Tamoxifen Letrozole	60	Acute myeloid leukaemia	BRCA2	Exon 11 3772delTT	Alive	Canada
5	45	Surgery Chemotherapy Tamoxifen	51	Chronic lymphocytic leukaemia	BRCA2	IVS24–18C>A	Alive	USA

Acute myeloid leukaemia, M3 is the acute promyelocytic leukaemia.

The subtype of acute leukaemia is not known for this patient.

All five leukaemia patients had previous history of breast cancer. The mean age at the diagnosis of breast cancer was 43.2 years (BRCA1=34.5 years, BRCA2=49.0 years). Four women (one BRCA1, three BRCA2) received chemotherapy for their breast cancer. The BRCA1 mutation carrier who received chemotherapy for breast cancer developed acute myeloid leukaemia 3 years after initial treatment for breast cancer. Of three BRCA2 mutation carriers, two developed acute myeloid leukaemia after 18 and 10 years, and one patient developed chronic lymphocytic leukaemia 6 years after the treatment for breast cancer.

Overall, five cases of leukaemia were observed vs 4.3 cases expected (SIR=1.16, 95% CI: 0.43–2.58, P=0.69). Among the entire cohort, the incidence of leukaemia was 11.3 per 100 000 women per year. Two incident leukaemia cases were observed among the BRCA1 carriers vs 3.02 cases expected (SIR=0.66, 95% CI: 0.11–2.19, P= 0.61) (Table 3). For BRCA2 carriers, three leukaemia cases were observed vs 1.24 cases expected (SIR=2.42, 95% CI: 0.61–6.58, P=0.17). Among BRCA2 mutation carriers the incidence rate of leukaemia was 27.1 per 100 000 women per year. Compared to the general population, the risk of leukaemia was higher for women <50 years (SIR=2.30, 95% CI: 0.38–7.59, P=0.27); the risk was lower for women 50 years and above (SIR=0.87, 95% CI: 0.22–2.38, P=0.88). However, the absolute difference in the annual risk of leukaemia was statistically not significant between the two age groups (P=0.77).

Table 3

Observed and expected number of leukaemia cases in BRCA1 and BRCA2 carriers according to country of origin

Country	BRCA status	N	Person-years	Observed cancers	Expected cancers	Incidence (per 100 000 women)	SIR (95% CI)	P-value
All combined	All	7243	44255.18	5	4.3	11.3	1.16 (0.43–2.58)	0.69
	BRCA1	5397	32864.14	2	3.02	6.1	0.66 (0.11–2.19)	0.61
	BRCA2	1790	11074.09	3	1.24	27.1	2.42 (0.61–6.58)	0.17
Canada	All	2255	16431.59	2	2.02	12.2	0.99 (0.17–3.27)	0.98
	BRCA1	1229	9211.06	1	1.05	10.8	0.95 (0.02–5.31)	0.56
	BRCA2	994	7010.24	1	0.95	14.3	1.05 (0.05–5.19)	0.86
Italy	All	103	463.32	2	0.04	431.7	50 (8.38–165.2)	0.0007
	BRCA1	84	348.18	1	0.03	287.2	33.3 (1.67–164.4)	0.03
	BRCA2	19	115.14	1	0.01	868.5	100 (5.0–493.2)	0.01
USA	All	2185	12343.86	1	1.30	8.1	0.77 (0.04–3.79)	0.9
	BRCA1	1414	8379.13	0	0.87	0	—	—
	BRCA2	749	3862.16	1	0.43	25.9	2.33 (0.12–11.47)	0.42
Poland	All	2700	15016.41	0	0.84	—	—	—
	BRCA1	2670	14925.77	0	0.84	—	—	—
	BRCA2	28	86.56	0	0.006	—	—	—

Abbreviations: CI, confidence interval; SIR, standardised incidence ratio.

Compared to general population, the risk of leukaemia was five-fold higher for women who had a prior BRCA2-associated breast cancer (SIR=4.76, 95% CI: 1.21–12.96, P=0.03), but was not significantly higher for women with BRCA1-associated breast cancer (Table 4). Among all women with BRCA-associated breast cancer who received chemotherapy, four cases of leukaemia were observed vs 1.4 cases expected (SIR=2.86, 95% CI: 0.91–6.89, P=0.07). Among women who did not receive chemotherapy, one case of leukaemia was observed vs 0.83 cases expected (SIR=1.20, 95% CI: 0.06–5.94, P=0.76) (Table 4).

Table 4

Observed and expected number of leukaemia in BRCA carriers according breast cancer status

BRCA status	Breast cancer	N	Person-years	Observed cancers	Expected cancers	Incidence rate per 100 000 per year	SIR 95% CI	P-value
All	No	4040	24628.12	0	2.03	—	—	—
	Yes	3200	19617.74	5	2.26	25.5	2.21 (0.81–4.90)	0.11
BRCA1	No	3042	18483.11	0	1.42	—	—	—
	Yes	2353	14376.04	2	1.60	13.9	1.25 (0.21–4.13)	0.69
BRCA2	No	976	6027.48	0	0.60	—	—	—
	Yes	813	5042.28	3	0.64	59.5	4.69 (1.19–12.76)	0.03

Abbreviations: CI, confidence interval; SIR, standardised incidence ratio.

The risk of leukaemia was eight-fold higher for women who received chemotherapy for a BRCA2-associated breast cancer (SIR=8.11, 95% CI: 2.06–22.07, P=0.007) (Table 5). For BRCA1 mutation carriers who received chemotherapy for breast cancer, the risk of leukaemia was similar to the risk in the general population (SIR=0.97, 95% CI: 0.05–4.79, P=1.01). The absolute risk of leukaemia for women who received chemotherapy for breast cancer was 0.06% (1 of 1776 women) for the BRCA1 carriers and was 0.55% (3 of 540 women) for the BRCA2 carriers.

Table 5

Observed and expected number of leukaemia in BRCA-associated breast cancer patients according to chemotherapy

BRCA status	Chemotherapy	N	Person-years	Observed cancers	Expected cancers	Incidence rate per 100 000 per year	SIR 95% CI	P-value
All	No	788	5257.98	1	0.83	19.0	1.20 (0.06–5.94)	0.76
	Yes	2,346	14107.09	4	1.40	28.3	2.86 (0.91–6.89)	0.07
BRCA1	No	544	3582.22	1	0.56	27.9	1.79 (0.09–8.81)	0.54
	Yes	1,776	10674.21	1	1.03	9.4	0.97 (0.02–5.41)	0.55
BRCA2	No	241	1662.88	0	0.27	—	—	—
	Yes	540	3350.50	3	0.37	89.5	8.11 (2.06–22.07)	0.007

Abbreviations: CI, confidence interval; SIR, standardised incidence ratio.

Discussion

Women who carry a mutation in BRCA genes are concerned about their risk of developing cancers other than breast and ovarian cancers. In this prospective study of the BRCA mutation carriers, we observed a two-fold increase in the risk of leukaemia among carriers of a BRCA2 mutation, compared to the risk in general population. The increased risk of leukaemia was statistically significant if a BRCA2 mutation carrier had breast cancer (SIR=4.69, P=0.03), and had received chemotherapy (SIR=8.11, P=0.007).

The risk of leukaemia in the BRCA1 and the BRCA2 mutation carriers was first reported in the earlier studies of the Breast Cancer Linkage Consortium (The Breast Cancer Linkage Consortium, 1999; Thompson ). In their first study, the Breast Cancer Linkage Consortium estimated the risk of leukaemia among the probable carriers of a BRCA2 mutation, relative to non-carriers, and subjects with an unknown mutation status (The Breast Cancer Linkage Consortium, 1999). The relative risk of leukaemia in probable carriers of the BRCA2 mutation was 1.12 (95% CI: 0.30–4.25). The risk of leukaemia was relatively lower in the BRCA1 mutation carriers, compared to non-carriers and untested women (relative risk=0.88, 95% CI: 0.37–2.14) (Thompson ). Relative risks reported in studies of the Breast Cancer Linkage Consortium were based on the review of family histories, and diagnosis of leukaemia was based on the information provided by a family member. In comparison, our estimates are based on a prospective follow-up of women with a known BRCA mutation status.

Two case reports suggest that carriers of a BRCA2 mutation who receive treatment for breast cancer might be at a higher risk of developing haematologic malignancies, in particular, leukaemias (Friedenson, 2007; Cole and Strair, 2010). Fruscalzo et al (2006) reported of a patient who developed chronic lymphocytic leukaemia 2 years after receiving chemotherapy for the BRCA2-associated breast cancer. In another case-series, Cole and Strair (2010) reported on six patients who developed either acute myelogenous leukaemia or myelodysplasia after receiving treatment for breast cancer. Three of six (50%) patients in this case-series had a mutation of one of the BRCA genes (BRCA1 or BRCA2), and all of them received adriamycin and cyclophosphamide-based chemotherapy for their breast cancer. Hall et al (2006) reported acute myelogenous leukaemia in three Hispanic BRCA2 carriers who did not receive chemotherapy or radiation therapy for early-stage breast cancers.

The actual risk of leukaemia in our BRCA cohort was <1%. One potential reason why the impact of BRCA mutations on leukaemia incidence might be less than that for breast and ovarian cancer susceptibility could be that the self-renewal potential of leukaemic cells is dependent on genome stability, which is compromised when there is a BRCA mutation (Santos ).

Our results are based on the prospective follow-up of the largest cohort of women who are confirmed carriers of a BRCA mutation, and with two-yearly follow-up of all women.

Our results should be interpreted with caution. Our BRCA cohort included only women and therefore we cannot estimate the incidence of leukaemia in male carriers of a BRCA mutation. Moreover, the number of incidence cases was relatively small in our study which limits the precision of risk estimates for subgroups. Furthermore, missing follow-up information for nearly 30% women could potentially bias our estimates.

In conclusion, our findings suggest that women with a BRCA2-associated breast cancer might be at a higher risk of developing leukaemia; in particular if these women receive chemotherapy for breast cancer. However, given a very low actual risk, we suggest that the risk of leukaemia in the BRCA mutation carriers should not influence the choice of chemotherapy to treat their breast cancer.