Long-term heart-specific mortality among 347 476 breast cancer patients treated with radiotherapy or chemotherapy: a registry-based cohort study

Long-term heart-specific mortality among 347 476 breast cancer patients treated with... Abstract Aims Breast cancer survival has improved throughout the last decades, but treatment-induced cardiotoxicity remains a major concern. This study aimed to investigate competing causes of death and prognostic factors within a large cohort of breast cancer patients and to describe the heart-specific mortality in relation to the general population. Methods and results In this registry-based cohort study, women diagnosed with breast cancer between 2000 and 2011, who were treated with radiotherapy or chemotherapy and followed until 2014, were identified from the Surveillance, Epidemiology, and End Results-18 (SEER-18) database. Cumulative mortality functions were computed. To investigate heart-specific mortality relative to the general population, long-term (≥10 years) standardized mortality ratios (SMRs) were calculated. Prognostic factors for heart-specific mortality were assessed by calculating cause-specific hazard ratios (HRcs) with corresponding 95% confidence intervals using the Cox proportional hazards regression. Subgroup analysis on intermediate-term mortality according to molecular subtypes, for which information was available since 2010, was performed. In total, 347 476 breast cancer patients were eligible to be included in the study. Among all possible competing causes of death, breast cancer accounted for the highest cumulative mortality. Compared with the general population, heart-specific mortality of breast cancer patients treated with radiotherapy or chemotherapy was lower [SMRoverall 0.84 (0.79–0.90)]. In subgroup analysis, human epidermal growth factor receptor 2 (HER2)-positive subtype was not associated with increased heart-specific mortality relative to HER2-negative patients [HRcs 0.96 (0.70–1.32)]. Conclusion Heart-specific mortality among breast cancer survivors is not increased compared with the general population. Human epidermal growth factor receptor 2-positive patients do not have increased heart-specific mortality compared to HER2-negative patients. Cardio-oncology, Breast cancer, Cardiovascular, Competing risk, Cancer registry Introduction With 1.67 million cases in 2012, breast cancer is the second most common cancer worldwide.1 Whereas improvement in treatment options including radiation and human epidermal growth factor receptor 2 (HER2)-targeted chemotherapy led to progress in terms of breast cancer survival, long-term mortality due to persistent decline in left ventricular ejection fraction (LVEF) and incidental cardiac diseases have remained one of the major concerns in breast cancer survivors.2 Recent studies found that especially among older breast cancer patients and those with a concomitant cardiovascular comorbidity at time of breast cancer diagnosis, the cumulative mortality due to a cardiovascular disease was higher than the cumulative mortality due to the underlying breast cancer disease.3,4 However, most of these findings were analysed within breast cancer cohorts only, and not much is known about cardiovascular mortality compared with the general population.5 As increasing age is a natural driver for the incidence of cardiovascular diseases and cardiovascular mortality among the general population, this comparison is crucial in order to disclose associations potentially attributable to breast cancer therapy such as radiotherapy or chemotherapy. Moreover, it is of utmost importance that these analyses are performed with careful consideration of competing risks.3,6 In the light of possible treatment-related cardiotoxic effects, we aimed to investigate long-term heart-specific mortality within a large breast cancer population including 347 476 US female breast cancer patients treated with radiotherapy or chemotherapy both in the presence of competing risks and in comparison with the general US female population. Methods Data sources For this registry-based retrospective cohort study, the Surveillance, Epidemiology, and End Results-18 (SEER-18) database was used, which covers data from 18 regional cancer registries throughout the USA.7 Access to the data was provided through a standard request to SEER, which was accepted after filling the required data usage agreement. In total, the SEER cancer registries draw on a base population of approximately 28% of the US population and capture comprehensive demographic and cancer-specific information including the International Classification of Diseases-10 (ICD-10) codes for the underlying cause of death. The reference cohort representing the general US population was selected from the underlying cause of death (WONDER) online database of the Centers for Disease Control and Prevention (CDC).8 These county-level national mortality and population data are based on death certificates for US residents. Study population Breast cancer cases were defined as women at age 15 years or older with a first primary malignant breast cancer diagnosis (ICD-10: C50.x) between 1 January 2000 and 31 December 2011. This calendar year interval was chosen to provide a sufficient minimum follow-up time for all included patients. Follow-up time was defined as the time from diagnosis to the date of last contact, death or end of the study period (31 December 2014), whichever occurred first. Patients without or with unknown status of surgery (8.2%), missing information on cause of death (0.5%), stage 0 (0.1%), or unknown stage (3.6%) were excluded in order to perform a complete case analysis. To check if results hold when the above mentioned variables were imputed, all main analyses were repeated and estimates from 30 multiple imputed datasets were combined according to Rubin’s rule.9 Finally, the analysis cohort was restricted to patients who received either radiotherapy or chemotherapy or both. Outcome and covariate assessment The primary focus of this study was given to mortality due to diseases of the heart as a potential long-term side-effect of cardiotoxic treatment such as radiation and chemotherapy. Hence, heart-specific causes of death were defined according to the SEER recode 50060 (ICD-10 codes are presented in Supplementary material online, Table S1). Sensitivity analyses were also conducted including patients dying from primary hypertension and hypertensive renal disease (ICD-10 codes I10 and I12, respectively). Table 1 Basic characteristics of all included female breast cancer patients at time of diagnosis*   All patients  Treatment (chemotherapy or radiotherapy)   Yes  No/unknown  N  501 547  347 476  154 071  Age, mean (SD)  59.94 (13.71)  57.51 (12.61)  65.41 (14.48)  Age groups   <50  125 328 (25.0)  99 795 (28.7)  25 533 (16.6)   50–64  190 536 (38.0)  144 409 (41.6)  46 127 (29.9)   65–74  100 059 (20.0)  66 184 (19.0)  33 875 (22.0)   75+  85 624 (17.1)  37 088 (10.7)  48 536 (31.5)  Ethnicity, n (%)   White  411 241 (82.0)  283 319 (81.5)  127 922 (83.0)   Black  49 410 (9.9)  35 577 (10.2)  13 833 (9.0)   American Indian/Alaska Native  2524 (0.5)  1818 (0.5)  706 (0.5)   Asian/Pacific Islander  36 424 (7.3)  25 620 (7.4)  10 804 (7.0)   Other/unknown  1948 (0.4)  1142 (0.3)  806 (0.5)  Year of diagnosis, n (%)   2000–2005  242 584 (48.4)  165 135 (47.5)  77 449 (50.3)   2006–2011  258 963 (51.6)  182 341 (52.5)  76 622 (49.7)  Stage, n (%)   I  246 040 (49.1)  156 504 (45.0)  89 536 (58.1)   II  179 950 (35.9)  128 560 (37.0)  51 390 (33.4)   III  65 773 (13.1)  55 096 (15.9)  10 677 (6.9)   IV  9784 (2.0)  7316 (2.1)  2468 (1.6)  Grading, n (%)a   I  102 147 (21.6)  65 133 (19.7)  37 014 (26.1)   II  199 202 (42.2)  135 484 (41.0)  63 718 (44.9)   III  164 513 (34.9)  125 255 (38.0)  39 258 (27.7)   IV  6156 (1.3)  4178 (1.3)  1978 (1.4)  ICD-10 site, n (%)   C50.0 (Nipple and areola)  2533 (0.5)  1324 (0.4)  1209 (0.8)   C50.1 (Central portion of breast)  27 092 (5.4)  17 207 (5.0)  9885 (6.4)   C50.2 (Upper-inner quadrant of breast)  55 226 (11.0)  39 697 (11.4)  15 529 (10.1)   C50.3 (Lower-inner quadrant of breast)  28 205 (5.6)  19 644 (5.7)  8561 (5.6)   C50.4 (Upper-outer quadrant of breast)  174 951 (34.9)  126 602 (36.4)  48 349 (31.4)   C50.5 (Lower-outer quadrant of breast)  35 136 (7.0)  24 914 (7.2)  10 222 (6.6)   C50.6 (Axillary tail of breast)  3191 (0.6)  2362 (0.7)  829 (0.5)   C50.8 (Overlapping lesion of breast)  106 527 (21.2)  74 179 (21.3)  32 348 (21.0)   C50.9 (Breast, unspecified)  68 686 (13.7)  41 547 (12.0)  27 139 (17.6)  Cause of death, n (%)   Alive  381 201 (76.0)  276 198 (79.5)  105 003 (68.2)   Breast cancer  54 424 (10.9)  39 802 (11.5)  14 622 (9.5)   Heart-specific disease  17 278 (3.4)  7401 (2.1)  9877 (6.4)   Other  48 644 (9.7)  24 075 (6.9)  24 569 (15.9)  Radiotherapy, n (%)b  260 144 (53.0)  260 144 (76.2)  0 (0.0)  Chemotherapy, n (%)  212 600 (42.4)  212 600 (61.2)  0 (0.0)  HER2+, n (%)c  11 921 (14.5)  9733 (16.4)  2188 (9.6)  ER+, n (%)d  367 708 (79.5)  254 177 (77.4)  113 531 (84.7)  PR+, n (%)e  310 036 (68.1)  214 590 (66.3)  95 446 (72.7)    All patients  Treatment (chemotherapy or radiotherapy)   Yes  No/unknown  N  501 547  347 476  154 071  Age, mean (SD)  59.94 (13.71)  57.51 (12.61)  65.41 (14.48)  Age groups   <50  125 328 (25.0)  99 795 (28.7)  25 533 (16.6)   50–64  190 536 (38.0)  144 409 (41.6)  46 127 (29.9)   65–74  100 059 (20.0)  66 184 (19.0)  33 875 (22.0)   75+  85 624 (17.1)  37 088 (10.7)  48 536 (31.5)  Ethnicity, n (%)   White  411 241 (82.0)  283 319 (81.5)  127 922 (83.0)   Black  49 410 (9.9)  35 577 (10.2)  13 833 (9.0)   American Indian/Alaska Native  2524 (0.5)  1818 (0.5)  706 (0.5)   Asian/Pacific Islander  36 424 (7.3)  25 620 (7.4)  10 804 (7.0)   Other/unknown  1948 (0.4)  1142 (0.3)  806 (0.5)  Year of diagnosis, n (%)   2000–2005  242 584 (48.4)  165 135 (47.5)  77 449 (50.3)   2006–2011  258 963 (51.6)  182 341 (52.5)  76 622 (49.7)  Stage, n (%)   I  246 040 (49.1)  156 504 (45.0)  89 536 (58.1)   II  179 950 (35.9)  128 560 (37.0)  51 390 (33.4)   III  65 773 (13.1)  55 096 (15.9)  10 677 (6.9)   IV  9784 (2.0)  7316 (2.1)  2468 (1.6)  Grading, n (%)a   I  102 147 (21.6)  65 133 (19.7)  37 014 (26.1)   II  199 202 (42.2)  135 484 (41.0)  63 718 (44.9)   III  164 513 (34.9)  125 255 (38.0)  39 258 (27.7)   IV  6156 (1.3)  4178 (1.3)  1978 (1.4)  ICD-10 site, n (%)   C50.0 (Nipple and areola)  2533 (0.5)  1324 (0.4)  1209 (0.8)   C50.1 (Central portion of breast)  27 092 (5.4)  17 207 (5.0)  9885 (6.4)   C50.2 (Upper-inner quadrant of breast)  55 226 (11.0)  39 697 (11.4)  15 529 (10.1)   C50.3 (Lower-inner quadrant of breast)  28 205 (5.6)  19 644 (5.7)  8561 (5.6)   C50.4 (Upper-outer quadrant of breast)  174 951 (34.9)  126 602 (36.4)  48 349 (31.4)   C50.5 (Lower-outer quadrant of breast)  35 136 (7.0)  24 914 (7.2)  10 222 (6.6)   C50.6 (Axillary tail of breast)  3191 (0.6)  2362 (0.7)  829 (0.5)   C50.8 (Overlapping lesion of breast)  106 527 (21.2)  74 179 (21.3)  32 348 (21.0)   C50.9 (Breast, unspecified)  68 686 (13.7)  41 547 (12.0)  27 139 (17.6)  Cause of death, n (%)   Alive  381 201 (76.0)  276 198 (79.5)  105 003 (68.2)   Breast cancer  54 424 (10.9)  39 802 (11.5)  14 622 (9.5)   Heart-specific disease  17 278 (3.4)  7401 (2.1)  9877 (6.4)   Other  48 644 (9.7)  24 075 (6.9)  24 569 (15.9)  Radiotherapy, n (%)b  260 144 (53.0)  260 144 (76.2)  0 (0.0)  Chemotherapy, n (%)  212 600 (42.4)  212 600 (61.2)  0 (0.0)  HER2+, n (%)c  11 921 (14.5)  9733 (16.4)  2188 (9.6)  ER+, n (%)d  367 708 (79.5)  254 177 (77.4)  113 531 (84.7)  PR+, n (%)e  310 036 (68.1)  214 590 (66.3)  95 446 (72.7)  ER, estrogen receptor; HER, human epidermal growth factor receptor; ICD, international classification of diseases; PR, progesterone receptor. * P-values for differences in categorical variables between treated and untreated patients are all <0.001. a Information on grading missing for 29 529 patients. b Information missing/unknown for 10 771 patients. c Not available for 419 384 patients, since HER2 status is recorded in SEER from 2010+ on. d ER status not available for 38 930 patients. e PR status not available for 46 578 patients. Table 1 Basic characteristics of all included female breast cancer patients at time of diagnosis*   All patients  Treatment (chemotherapy or radiotherapy)   Yes  No/unknown  N  501 547  347 476  154 071  Age, mean (SD)  59.94 (13.71)  57.51 (12.61)  65.41 (14.48)  Age groups   <50  125 328 (25.0)  99 795 (28.7)  25 533 (16.6)   50–64  190 536 (38.0)  144 409 (41.6)  46 127 (29.9)   65–74  100 059 (20.0)  66 184 (19.0)  33 875 (22.0)   75+  85 624 (17.1)  37 088 (10.7)  48 536 (31.5)  Ethnicity, n (%)   White  411 241 (82.0)  283 319 (81.5)  127 922 (83.0)   Black  49 410 (9.9)  35 577 (10.2)  13 833 (9.0)   American Indian/Alaska Native  2524 (0.5)  1818 (0.5)  706 (0.5)   Asian/Pacific Islander  36 424 (7.3)  25 620 (7.4)  10 804 (7.0)   Other/unknown  1948 (0.4)  1142 (0.3)  806 (0.5)  Year of diagnosis, n (%)   2000–2005  242 584 (48.4)  165 135 (47.5)  77 449 (50.3)   2006–2011  258 963 (51.6)  182 341 (52.5)  76 622 (49.7)  Stage, n (%)   I  246 040 (49.1)  156 504 (45.0)  89 536 (58.1)   II  179 950 (35.9)  128 560 (37.0)  51 390 (33.4)   III  65 773 (13.1)  55 096 (15.9)  10 677 (6.9)   IV  9784 (2.0)  7316 (2.1)  2468 (1.6)  Grading, n (%)a   I  102 147 (21.6)  65 133 (19.7)  37 014 (26.1)   II  199 202 (42.2)  135 484 (41.0)  63 718 (44.9)   III  164 513 (34.9)  125 255 (38.0)  39 258 (27.7)   IV  6156 (1.3)  4178 (1.3)  1978 (1.4)  ICD-10 site, n (%)   C50.0 (Nipple and areola)  2533 (0.5)  1324 (0.4)  1209 (0.8)   C50.1 (Central portion of breast)  27 092 (5.4)  17 207 (5.0)  9885 (6.4)   C50.2 (Upper-inner quadrant of breast)  55 226 (11.0)  39 697 (11.4)  15 529 (10.1)   C50.3 (Lower-inner quadrant of breast)  28 205 (5.6)  19 644 (5.7)  8561 (5.6)   C50.4 (Upper-outer quadrant of breast)  174 951 (34.9)  126 602 (36.4)  48 349 (31.4)   C50.5 (Lower-outer quadrant of breast)  35 136 (7.0)  24 914 (7.2)  10 222 (6.6)   C50.6 (Axillary tail of breast)  3191 (0.6)  2362 (0.7)  829 (0.5)   C50.8 (Overlapping lesion of breast)  106 527 (21.2)  74 179 (21.3)  32 348 (21.0)   C50.9 (Breast, unspecified)  68 686 (13.7)  41 547 (12.0)  27 139 (17.6)  Cause of death, n (%)   Alive  381 201 (76.0)  276 198 (79.5)  105 003 (68.2)   Breast cancer  54 424 (10.9)  39 802 (11.5)  14 622 (9.5)   Heart-specific disease  17 278 (3.4)  7401 (2.1)  9877 (6.4)   Other  48 644 (9.7)  24 075 (6.9)  24 569 (15.9)  Radiotherapy, n (%)b  260 144 (53.0)  260 144 (76.2)  0 (0.0)  Chemotherapy, n (%)  212 600 (42.4)  212 600 (61.2)  0 (0.0)  HER2+, n (%)c  11 921 (14.5)  9733 (16.4)  2188 (9.6)  ER+, n (%)d  367 708 (79.5)  254 177 (77.4)  113 531 (84.7)  PR+, n (%)e  310 036 (68.1)  214 590 (66.3)  95 446 (72.7)    All patients  Treatment (chemotherapy or radiotherapy)   Yes  No/unknown  N  501 547  347 476  154 071  Age, mean (SD)  59.94 (13.71)  57.51 (12.61)  65.41 (14.48)  Age groups   <50  125 328 (25.0)  99 795 (28.7)  25 533 (16.6)   50–64  190 536 (38.0)  144 409 (41.6)  46 127 (29.9)   65–74  100 059 (20.0)  66 184 (19.0)  33 875 (22.0)   75+  85 624 (17.1)  37 088 (10.7)  48 536 (31.5)  Ethnicity, n (%)   White  411 241 (82.0)  283 319 (81.5)  127 922 (83.0)   Black  49 410 (9.9)  35 577 (10.2)  13 833 (9.0)   American Indian/Alaska Native  2524 (0.5)  1818 (0.5)  706 (0.5)   Asian/Pacific Islander  36 424 (7.3)  25 620 (7.4)  10 804 (7.0)   Other/unknown  1948 (0.4)  1142 (0.3)  806 (0.5)  Year of diagnosis, n (%)   2000–2005  242 584 (48.4)  165 135 (47.5)  77 449 (50.3)   2006–2011  258 963 (51.6)  182 341 (52.5)  76 622 (49.7)  Stage, n (%)   I  246 040 (49.1)  156 504 (45.0)  89 536 (58.1)   II  179 950 (35.9)  128 560 (37.0)  51 390 (33.4)   III  65 773 (13.1)  55 096 (15.9)  10 677 (6.9)   IV  9784 (2.0)  7316 (2.1)  2468 (1.6)  Grading, n (%)a   I  102 147 (21.6)  65 133 (19.7)  37 014 (26.1)   II  199 202 (42.2)  135 484 (41.0)  63 718 (44.9)   III  164 513 (34.9)  125 255 (38.0)  39 258 (27.7)   IV  6156 (1.3)  4178 (1.3)  1978 (1.4)  ICD-10 site, n (%)   C50.0 (Nipple and areola)  2533 (0.5)  1324 (0.4)  1209 (0.8)   C50.1 (Central portion of breast)  27 092 (5.4)  17 207 (5.0)  9885 (6.4)   C50.2 (Upper-inner quadrant of breast)  55 226 (11.0)  39 697 (11.4)  15 529 (10.1)   C50.3 (Lower-inner quadrant of breast)  28 205 (5.6)  19 644 (5.7)  8561 (5.6)   C50.4 (Upper-outer quadrant of breast)  174 951 (34.9)  126 602 (36.4)  48 349 (31.4)   C50.5 (Lower-outer quadrant of breast)  35 136 (7.0)  24 914 (7.2)  10 222 (6.6)   C50.6 (Axillary tail of breast)  3191 (0.6)  2362 (0.7)  829 (0.5)   C50.8 (Overlapping lesion of breast)  106 527 (21.2)  74 179 (21.3)  32 348 (21.0)   C50.9 (Breast, unspecified)  68 686 (13.7)  41 547 (12.0)  27 139 (17.6)  Cause of death, n (%)   Alive  381 201 (76.0)  276 198 (79.5)  105 003 (68.2)   Breast cancer  54 424 (10.9)  39 802 (11.5)  14 622 (9.5)   Heart-specific disease  17 278 (3.4)  7401 (2.1)  9877 (6.4)   Other  48 644 (9.7)  24 075 (6.9)  24 569 (15.9)  Radiotherapy, n (%)b  260 144 (53.0)  260 144 (76.2)  0 (0.0)  Chemotherapy, n (%)  212 600 (42.4)  212 600 (61.2)  0 (0.0)  HER2+, n (%)c  11 921 (14.5)  9733 (16.4)  2188 (9.6)  ER+, n (%)d  367 708 (79.5)  254 177 (77.4)  113 531 (84.7)  PR+, n (%)e  310 036 (68.1)  214 590 (66.3)  95 446 (72.7)  ER, estrogen receptor; HER, human epidermal growth factor receptor; ICD, international classification of diseases; PR, progesterone receptor. * P-values for differences in categorical variables between treated and untreated patients are all <0.001. a Information on grading missing for 29 529 patients. b Information missing/unknown for 10 771 patients. c Not available for 419 384 patients, since HER2 status is recorded in SEER from 2010+ on. d ER status not available for 38 930 patients. e PR status not available for 46 578 patients. As potential predictors of mortality, age (categorized as age <35, 35–49, 50–54, 55–64, 65–74, and ≥75 years), ethnicity (white, black, American Indian/Alaska native, Asian/Pacific islander, and other), year of diagnosis, stage [coded according to the breast cancer adjusted staging of the American Joint Committee on Cancer (AJCC) 6th edition], grade (I–IV), and tumour site were considered. Statistical analysis The distributions of all baseline characteristics were summarized by calculating the mean and standard deviation for continuous variables and frequencies for categorical variables, respectively. The median follow-up time was assessed using the reverse Kaplan–Meier method.10 Crude cumulative mortality functions were calculated and plotted for cause-specific deaths overall and stratified by age to describe the probability of experiencing a specific endpoint in the presence of competing risks among patients treated with radiotherapy or chemotherapy. As standard Kaplan–Meier analyses treat failures from competing events as censored, this approach would lead to an overestimation of the absolute risk of the event of interest because competing events would then violate the assumption of non-informative censoring.6 Furthermore, the Kaplan–Meier estimate would reflect mortality from the event of interest in a hypothetical world without competing events, which is less clinically relevant. Hence, to compute the probability to die from a specific cause of death (cumulative mortality), breast cancer patients who died due to competing causes of death were retained in the underlying risk set instead of being censored. To investigate if risk factors such as age might be only associated with a potentially higher heart-specific mortality by the naturally increasing mortality due to aging, overall and age-specific standardized mortality ratios (SMRs) were calculated with the US female standard population from the CDC database as reference.8 As potential long-term cardiotoxicity was of major interest, conditional SMRs (cSMRs) were calculated for the calendar period 2012–2014 restricting the breast cancer cohort to patients who were diagnosed between 2000 and 2002 to ensure a conditional survival of at least ten years. Conditional SMRs were calculated by dividing the observed numbers of heart-specific deaths divided by the expected numbers. Latter were determined by multiplying the cumulative person-time across breast cancer patients within 5-year age-groups (55–59, 60–64, 65–69, 70–74, 75–79, 80–84, and ≥85) and the calendar period 2012–2014 with the age- and calendar period-specific mortality rates from heart diseases in the general female population. All age-specific and overall cSMRs were calculated for patients aged 55 years or older who received radiotherapy or chemotherapy. The 5-year age groups described above were chosen for age classification for two reasons: Firstly, for the aggregated data for the US general female population only categorizations in 1-year, 5-, and 10-year steps were available, of which the 5-year categorizations were found best to operate for the cSMR analyses. Secondly, official mortality statistics of the US population are categorized in exactly the same 5-year steps, which we adopted to make comparisons with cSMRs easier. Additionally, stratified analyses were performed by Stage I–III and for all-cause mortality. Stage-specific analyses for Stage IV patients were not conducted due to the very low numbers of 10-year survivors in this group. 95% confidence interval (95% CI) and P-values were estimated according to the method by Vandenbroucke et al.11 and Altman et al.,12 respectively. Additionally, adjusted cause-specific hazard ratios (HRcs) were computed to estimate the relative association of individual risk and prognostic factors on both heart- and breast cancer related mortality, using the Cox proportional hazards regression. Potential confounding was addressed by adjusting all models for the above listed variables. Stratified analyses by molecular subtype were conducted as a separate analysis on intermediate-term heart-specific mortality including all eligible breast cancer patients for whom this information was available since 2010.13 Table 1 was created using the R package ‘tableone’ (R Foundation for Statistical Computing, Vienna, Austria).14 All other analyses were performed with SAS software, version 9.4 (SAS Institute Inc., Cary, NC, USA) according to an apriori defined study protocol. Statistical significance was defined by a two-sided P < 0.05. Results Out of 572 341 identified breast cancer cases between 2000 and 2011, 501 547 (87.6%) female breast cancer patients were eligible to be included in the study, of whom 347 476 (69.3%) received either radiotherapy or chemotherapy or both (Figure 1). Figure 1 View largeDownload slide Selection of eligible patients for competing risk analysis. Figure 1 View largeDownload slide Selection of eligible patients for competing risk analysis. Compared with patients who did not receive radiotherapy or chemotherapy, treated patients were on average younger at diagnosis (57.5 years vs. 65.4), rather diagnosed in later stages (Stage III or IV 18.0% vs. 8.5%) and with higher gradings (Grade III or IV 39.3% vs. 29.1%) (Table 1). The majority of patients were white (81.5% and 83.0%, respectively) and the most common tumour site in both groups was the upper-outer quadrant of the breast (36.4% and 31.4%, respectively). In general, patients treated with radiotherapy or chemotherapy were more likely to die due to breast cancer (11.5% vs. 9.5%), and less likely to die due to a heart-specific disease (2.1% vs. 6.4%). Among all patients for whom the information was available, treated patients were more frequently diagnosed with a HER2-positive (HER2+) molecular subtype (16.4% vs. 9.6%), but less often diagnosed with oestrogen receptor positive (ER+) and progesterone receptor positive (PR+) molecular subtypes (77.4% vs. 84.7 and 66.3 vs. 72.7%, respectively). The median follow-up time for patients treated with radiotherapy or chemotherapy comprised 8.4 years [interquartile range (IQR) 5.5–11.6] and was slightly higher for untreated patients [8.6 (IQR 5.6–11.2)]. Cumulative mortality The cumulative mortalities for all causes of death among breast cancer patients treated with radiotherapy or chemotherapy are illustrated in Figure 2. By far the highest cumulative mortality of death was caused by breast cancer, followed by other non-cancer causes of death, which were not related to heart diseases. Heart-specific cumulative mortality was marginally lower than cumulative mortality due to cancers other than breast cancer. These patterns also did not substantially change when primary hypertension and hypertensive renal disease were included among the heart-specific causes of death or when using the imputed datasets (see Supplementary material online, Figure S1 and S2, respectively). Figure 2 View largeDownload slide Cumulative cause-specific mortality among treated breast cancer patients. Figure 2 View largeDownload slide Cumulative cause-specific mortality among treated breast cancer patients. Stratified by age groups, women aged 50 years or younger had the highest cumulative breast cancer mortality compared with heart-specific deaths (Take home figure). Nevertheless, it was observed that the cumulative heart-specific mortality steadily increased with age at diagnosis and length of follow-up. In the oldest age group (≥75 years), the cumulative mortality from heart-specific diseases exceeded the cumulative mortality from breast cancer around 12 years after diagnosis. When using the multiple imputed datasets, however, the cumulative mortality from heart-specific diseases was close, but did not exceed the cumulative mortality from breast cancer within the first 15 years after diagnosis (see Supplementary material online, Figure S3). Take home figure View largeDownload slide Cumulative heart-specific and breast cancer-specific mortality by age at diagnosis. Take home figure View largeDownload slide Cumulative heart-specific and breast cancer-specific mortality by age at diagnosis. Conditional standardized mortality ratios The comparison of heart-specific mortality between breast cancer patients and the general population is displayed in Table 2. Table 2 Age-specific and overall standardized mortality ratios for the years 2012–2014 among breast cancer patients who survived at least 10 years and were treated with radio- or chemotherapy diagnosed between 2000 and 2002 relative to the USA female standard population Group  cSMR heart-specific  P-value  cSMR overall  P-value  Agea           55–59  0.67 (0.35–1.08)  0.1553  2.25 (1.99–2.53)  <0.0001   60–64  0.99 (0.69–1.34)  0.9541  1.75 (1.58–1.94)  <0.0001   65–69  0.77 (0.56–1.01)  0.0830  1.48 (1.35–1.61)  <0.0001   70–74  0.66 (0.50–0.85)  0.0028  1.22 (1.12–1.33)  <0.0001   75–79  0.87 (0.71–1.05)  0.1648  1.13 (1.05–1.22)  0.0017   80–84  0.89 (0.75–1.04)  0.1601  1.09 (1.02–1.17)  0.0120   85+  0.86 (0.78–0.94)  0.0011  0.88 (0.84–0.93)  <0.0001  Stage           I  0.75 (0.68–0.82)  <0.0001  0.88 (0.84–0.91)  <0.0001   II  0.89 (0.79–1.00)  0.0598  1.23 (1.17–1.29)  <0.0001   III  0.99 (0.79–1.21)  0.9188  1.90 (1.76–2.04)  <0.0001  Overall  0.84 (0.79–0.90)  <0.0001  1.12 (1.09–1.15)  <0.0001  Group  cSMR heart-specific  P-value  cSMR overall  P-value  Agea           55–59  0.67 (0.35–1.08)  0.1553  2.25 (1.99–2.53)  <0.0001   60–64  0.99 (0.69–1.34)  0.9541  1.75 (1.58–1.94)  <0.0001   65–69  0.77 (0.56–1.01)  0.0830  1.48 (1.35–1.61)  <0.0001   70–74  0.66 (0.50–0.85)  0.0028  1.22 (1.12–1.33)  <0.0001   75–79  0.87 (0.71–1.05)  0.1648  1.13 (1.05–1.22)  0.0017   80–84  0.89 (0.75–1.04)  0.1601  1.09 (1.02–1.17)  0.0120   85+  0.86 (0.78–0.94)  0.0011  0.88 (0.84–0.93)  <0.0001  Stage           I  0.75 (0.68–0.82)  <0.0001  0.88 (0.84–0.91)  <0.0001   II  0.89 (0.79–1.00)  0.0598  1.23 (1.17–1.29)  <0.0001   III  0.99 (0.79–1.21)  0.9188  1.90 (1.76–2.04)  <0.0001  Overall  0.84 (0.79–0.90)  <0.0001  1.12 (1.09–1.15)  <0.0001  cSMR, standardized mortality ratio conditional on at least 10 year survival among eligible breast cancer patients. a Age of the breast cancer patients/population in the respective follow-up year (2012–2014). Table 2 Age-specific and overall standardized mortality ratios for the years 2012–2014 among breast cancer patients who survived at least 10 years and were treated with radio- or chemotherapy diagnosed between 2000 and 2002 relative to the USA female standard population Group  cSMR heart-specific  P-value  cSMR overall  P-value  Agea           55–59  0.67 (0.35–1.08)  0.1553  2.25 (1.99–2.53)  <0.0001   60–64  0.99 (0.69–1.34)  0.9541  1.75 (1.58–1.94)  <0.0001   65–69  0.77 (0.56–1.01)  0.0830  1.48 (1.35–1.61)  <0.0001   70–74  0.66 (0.50–0.85)  0.0028  1.22 (1.12–1.33)  <0.0001   75–79  0.87 (0.71–1.05)  0.1648  1.13 (1.05–1.22)  0.0017   80–84  0.89 (0.75–1.04)  0.1601  1.09 (1.02–1.17)  0.0120   85+  0.86 (0.78–0.94)  0.0011  0.88 (0.84–0.93)  <0.0001  Stage           I  0.75 (0.68–0.82)  <0.0001  0.88 (0.84–0.91)  <0.0001   II  0.89 (0.79–1.00)  0.0598  1.23 (1.17–1.29)  <0.0001   III  0.99 (0.79–1.21)  0.9188  1.90 (1.76–2.04)  <0.0001  Overall  0.84 (0.79–0.90)  <0.0001  1.12 (1.09–1.15)  <0.0001  Group  cSMR heart-specific  P-value  cSMR overall  P-value  Agea           55–59  0.67 (0.35–1.08)  0.1553  2.25 (1.99–2.53)  <0.0001   60–64  0.99 (0.69–1.34)  0.9541  1.75 (1.58–1.94)  <0.0001   65–69  0.77 (0.56–1.01)  0.0830  1.48 (1.35–1.61)  <0.0001   70–74  0.66 (0.50–0.85)  0.0028  1.22 (1.12–1.33)  <0.0001   75–79  0.87 (0.71–1.05)  0.1648  1.13 (1.05–1.22)  0.0017   80–84  0.89 (0.75–1.04)  0.1601  1.09 (1.02–1.17)  0.0120   85+  0.86 (0.78–0.94)  0.0011  0.88 (0.84–0.93)  <0.0001  Stage           I  0.75 (0.68–0.82)  <0.0001  0.88 (0.84–0.91)  <0.0001   II  0.89 (0.79–1.00)  0.0598  1.23 (1.17–1.29)  <0.0001   III  0.99 (0.79–1.21)  0.9188  1.90 (1.76–2.04)  <0.0001  Overall  0.84 (0.79–0.90)  <0.0001  1.12 (1.09–1.15)  <0.0001  cSMR, standardized mortality ratio conditional on at least 10 year survival among eligible breast cancer patients. a Age of the breast cancer patients/population in the respective follow-up year (2012–2014). When heart-specific mortality among breast cancer patients treated with radiotherapy or chemotherapy was compared to the general female US population, it was observed that the cSMRs tended to be slightly decreased for all age groups with an overall decrease by 16% (95% CI 21–10%) in the age standardized analysis. When this analysis was repeated using the multiple imputed datasets, the cSMRs varied marginally but still showed the same patterns (see Supplementary material online, Table S2). In terms of all-cause mortality, breast cancer patients treated with radiotherapy or chemotherapy showed an over two-fold increased mortality among younger breast cancer patients [cSMR 2.25 (1.99–2.53)] in the age group 55–59 years), which then steadily decreased by increasing age. The overall cSMRs suggested a 12% [1.12 (1.09–1.15)] increased mortality. When stratified by stage, it was observed that both heart-specific and overall cSMRs increased by stage. Cause-specific hazard ratios The associations between individual prognostic factors and heart-specific mortality and breast cancer mortality among patients receiving radiotherapy or chemotherapy are presented in Table 3. Table 3 Cause-specific hazard and 95% confidence intervals for heart-related and breast-cancer related mortality among breast cancer patients who received either chemotherapy or radiotherapy or both* Strata/model  N  Cause-specific hazards ratiosa   Heart  P-value  Breast cancer  P-value  Age at diagnosis (categorical)             <35  8837  0.29 (0.18–0.48)  <0.0001  1.24 (1.18–1.31)  <0.0001   35–49  90 958  0.59 (0.50–0.69)  <0.0001  0.99 (0.96–1.02)  0.5131   50–54  49 435  1.00 (ref.)    1.00 (ref.)     55–64  94 974  1.89 (1.66–2.15)  <0.0001  1.05 (1.01–1.08)  0.0076   65–74  66 184  5.98 (5.28–6.76)  <0.0001  1.20 (1.15–1.24)  <0.0001   ≥75  37 088  22.78 (20.20–25.70)  <0.0001  1.78 (1.71–1.86)  <0.0001  Year of diagnosis             2000–2005  165 135  1.00 (ref.)    1.00 (ref.)     2006–2011  182 341  0.83 (0.78–0.88)  <0.0001  0.82 (0.81–0.84)  <0.0001  Stage             I  156 504  1.00 (ref.)    1.00 (ref.)     II  128 560  1.26 (1.19–1.33)  <0.0001  3.01 (2.91–3.12)  <0.0001   III  55 096  1.82 (1.70–1.95)  <0.0001  9.45 (9.13–9.79)  <0.0001   IV  7316  2.90 (2.43–3.46)  <0.0001  34.52 (33.07–36.04)  <0.0001  Gradingb             I  65 133  1.00 (ref.)    1.00 (ref.)     II  135 484  1.02 (0.96–1.08)  0.5608  2.06 (1.95–2.16)  <0.0001   III  125 255  1.14 (1.07–1.22)  0.0001  3.88 (3.69–4.07)  <0.0001   IV  4178  1.16 (0.94–1.43)  0.1729  3.96 (3.66–4.29)  <0.0001  Ethnicity             White  283 319  1.00 (ref.)    1.00 (ref.)     Black  35 577  1.85 (1.72–1.99)  <0.0001  1.49 (1.45–1.53)  <0.0001   American Indian/Alaska Native  1818  0.86 (0.54–1.37)  0.5243  1.14 (1.00–1.29)  0.0451   Asian/Pacific Islander  25 620  0.74 (0.65–0.84)  <.0001  0.85 (0.82–0.89)  <0.0001  Strata/model  N  Cause-specific hazards ratiosa   Heart  P-value  Breast cancer  P-value  Age at diagnosis (categorical)             <35  8837  0.29 (0.18–0.48)  <0.0001  1.24 (1.18–1.31)  <0.0001   35–49  90 958  0.59 (0.50–0.69)  <0.0001  0.99 (0.96–1.02)  0.5131   50–54  49 435  1.00 (ref.)    1.00 (ref.)     55–64  94 974  1.89 (1.66–2.15)  <0.0001  1.05 (1.01–1.08)  0.0076   65–74  66 184  5.98 (5.28–6.76)  <0.0001  1.20 (1.15–1.24)  <0.0001   ≥75  37 088  22.78 (20.20–25.70)  <0.0001  1.78 (1.71–1.86)  <0.0001  Year of diagnosis             2000–2005  165 135  1.00 (ref.)    1.00 (ref.)     2006–2011  182 341  0.83 (0.78–0.88)  <0.0001  0.82 (0.81–0.84)  <0.0001  Stage             I  156 504  1.00 (ref.)    1.00 (ref.)     II  128 560  1.26 (1.19–1.33)  <0.0001  3.01 (2.91–3.12)  <0.0001   III  55 096  1.82 (1.70–1.95)  <0.0001  9.45 (9.13–9.79)  <0.0001   IV  7316  2.90 (2.43–3.46)  <0.0001  34.52 (33.07–36.04)  <0.0001  Gradingb             I  65 133  1.00 (ref.)    1.00 (ref.)     II  135 484  1.02 (0.96–1.08)  0.5608  2.06 (1.95–2.16)  <0.0001   III  125 255  1.14 (1.07–1.22)  0.0001  3.88 (3.69–4.07)  <0.0001   IV  4178  1.16 (0.94–1.43)  0.1729  3.96 (3.66–4.29)  <0.0001  Ethnicity             White  283 319  1.00 (ref.)    1.00 (ref.)     Black  35 577  1.85 (1.72–1.99)  <0.0001  1.49 (1.45–1.53)  <0.0001   American Indian/Alaska Native  1818  0.86 (0.54–1.37)  0.5243  1.14 (1.00–1.29)  0.0451   Asian/Pacific Islander  25 620  0.74 (0.65–0.84)  <.0001  0.85 (0.82–0.89)  <0.0001  * For 6294 patients the radiotherapy status was missing/unknown. a Adjusted for age at diagnosis (categorical), period of diagnosis, stage, grading, race, and topography (ICD-10). b A total of 17 426 patients with missing information on grading were excluded from analysis. Table 3 Cause-specific hazard and 95% confidence intervals for heart-related and breast-cancer related mortality among breast cancer patients who received either chemotherapy or radiotherapy or both* Strata/model  N  Cause-specific hazards ratiosa   Heart  P-value  Breast cancer  P-value  Age at diagnosis (categorical)             <35  8837  0.29 (0.18–0.48)  <0.0001  1.24 (1.18–1.31)  <0.0001   35–49  90 958  0.59 (0.50–0.69)  <0.0001  0.99 (0.96–1.02)  0.5131   50–54  49 435  1.00 (ref.)    1.00 (ref.)     55–64  94 974  1.89 (1.66–2.15)  <0.0001  1.05 (1.01–1.08)  0.0076   65–74  66 184  5.98 (5.28–6.76)  <0.0001  1.20 (1.15–1.24)  <0.0001   ≥75  37 088  22.78 (20.20–25.70)  <0.0001  1.78 (1.71–1.86)  <0.0001  Year of diagnosis             2000–2005  165 135  1.00 (ref.)    1.00 (ref.)     2006–2011  182 341  0.83 (0.78–0.88)  <0.0001  0.82 (0.81–0.84)  <0.0001  Stage             I  156 504  1.00 (ref.)    1.00 (ref.)     II  128 560  1.26 (1.19–1.33)  <0.0001  3.01 (2.91–3.12)  <0.0001   III  55 096  1.82 (1.70–1.95)  <0.0001  9.45 (9.13–9.79)  <0.0001   IV  7316  2.90 (2.43–3.46)  <0.0001  34.52 (33.07–36.04)  <0.0001  Gradingb             I  65 133  1.00 (ref.)    1.00 (ref.)     II  135 484  1.02 (0.96–1.08)  0.5608  2.06 (1.95–2.16)  <0.0001   III  125 255  1.14 (1.07–1.22)  0.0001  3.88 (3.69–4.07)  <0.0001   IV  4178  1.16 (0.94–1.43)  0.1729  3.96 (3.66–4.29)  <0.0001  Ethnicity             White  283 319  1.00 (ref.)    1.00 (ref.)     Black  35 577  1.85 (1.72–1.99)  <0.0001  1.49 (1.45–1.53)  <0.0001   American Indian/Alaska Native  1818  0.86 (0.54–1.37)  0.5243  1.14 (1.00–1.29)  0.0451   Asian/Pacific Islander  25 620  0.74 (0.65–0.84)  <.0001  0.85 (0.82–0.89)  <0.0001  Strata/model  N  Cause-specific hazards ratiosa   Heart  P-value  Breast cancer  P-value  Age at diagnosis (categorical)             <35  8837  0.29 (0.18–0.48)  <0.0001  1.24 (1.18–1.31)  <0.0001   35–49  90 958  0.59 (0.50–0.69)  <0.0001  0.99 (0.96–1.02)  0.5131   50–54  49 435  1.00 (ref.)    1.00 (ref.)     55–64  94 974  1.89 (1.66–2.15)  <0.0001  1.05 (1.01–1.08)  0.0076   65–74  66 184  5.98 (5.28–6.76)  <0.0001  1.20 (1.15–1.24)  <0.0001   ≥75  37 088  22.78 (20.20–25.70)  <0.0001  1.78 (1.71–1.86)  <0.0001  Year of diagnosis             2000–2005  165 135  1.00 (ref.)    1.00 (ref.)     2006–2011  182 341  0.83 (0.78–0.88)  <0.0001  0.82 (0.81–0.84)  <0.0001  Stage             I  156 504  1.00 (ref.)    1.00 (ref.)     II  128 560  1.26 (1.19–1.33)  <0.0001  3.01 (2.91–3.12)  <0.0001   III  55 096  1.82 (1.70–1.95)  <0.0001  9.45 (9.13–9.79)  <0.0001   IV  7316  2.90 (2.43–3.46)  <0.0001  34.52 (33.07–36.04)  <0.0001  Gradingb             I  65 133  1.00 (ref.)    1.00 (ref.)     II  135 484  1.02 (0.96–1.08)  0.5608  2.06 (1.95–2.16)  <0.0001   III  125 255  1.14 (1.07–1.22)  0.0001  3.88 (3.69–4.07)  <0.0001   IV  4178  1.16 (0.94–1.43)  0.1729  3.96 (3.66–4.29)  <0.0001  Ethnicity             White  283 319  1.00 (ref.)    1.00 (ref.)     Black  35 577  1.85 (1.72–1.99)  <0.0001  1.49 (1.45–1.53)  <0.0001   American Indian/Alaska Native  1818  0.86 (0.54–1.37)  0.5243  1.14 (1.00–1.29)  0.0451   Asian/Pacific Islander  25 620  0.74 (0.65–0.84)  <.0001  0.85 (0.82–0.89)  <0.0001  * For 6294 patients the radiotherapy status was missing/unknown. a Adjusted for age at diagnosis (categorical), period of diagnosis, stage, grading, race, and topography (ICD-10). b A total of 17 426 patients with missing information on grading were excluded from analysis. Heart-specific mortality most strongly increased with age, with hazard ratios (HRs) (95% CIs) as high as 1.89 (1.66–2.15), 5.98 (5.28–6.76), and 22.78 (20.20–25.70) for patients aged 55–64, 65–74, and ≥75 relative to patients aged 50–54, respectively. The association with age was much stronger for heart-related mortality than for breast cancer related mortality. Moreover, women who were diagnosed between 2006 and 2011 tended to have a 17% decreased heart-specific mortality [HRcs 0.83 (0.78–0.88)] relative to those diagnosed between 2000 and 2005. Heart-specific mortality also increased with stage, even though the association was less pronounced than for breast cancer mortality [HRcs 2.90 (2.43–3.46)] for Stage IV relative to Stage I). Compared with white ethnicities, black breast cancer patients were at higher risk for heart-specific death [HRcs 1.85 (1.72–1.99)], whereas patients with an Asian/Pacific island ethnicity were at lower risk [HRcs 0.74 (0.65–0.84)]. None of the results did significantly change when the multiple imputed datasets were used (see Supplementary material online, Table S3). When comparing intermediate-term heart-specific mortality according to their HER2/neu status in subgroup analysis, it was observed that being diagnosed with a HER2-positive receptor status was not associated with heart-specific mortality [HRcs 0.96 (0.70–1.32)] relative to HER2-negative receptor status (see Supplementary material online, Table S4). Discussion In this large registry-based cohort study, it was found that within a cohort of female breast cancer patients receiving radiotherapy or chemotherapy, heart-specific mortality was an important competing risk, which was observed to steadily increase by age. However, when comparing long-term heart-specific mortality to the general population, mortality was not increased. Moreover, relative to patients with a HER2-negative subtype, also HER2-positive patients did not show increased mortality. To our knowledge, this is the largest registry-based study, which has been conducted so far on long-term heart-specific mortality among breast cancer patients treated with radiotherapy or chemotherapy. In line with previously published literature, also within this large US-based cohort an improvement of breast cancer-specific survival on the population level over time was observed, leaving potentially treatment-induced heart-specific diseases as a major competing risk for mortality among breast cancer survivors.15,16 Potentially relevant systemic treatments include anthracyclines, taxanes, alkylating agents, tyrosine-kinase inhibitors, and several new targeted therapeutic options, which are known to hold cardiotoxic profiles.17 The underlying mechanisms have not been well understood yet, but are hypothesized to be caused by impaired stress-protective signalling mechanisms in cardiomyocytes and damages of myocardial cells due to reactive oxygen species.18,19 Also radiotherapy is known as a major risk factor for treatment-induced heart-specific diseases, particularly if applied to patients affected by left-sided breast cancer.19 Especially the combination of radiation-induced macrosvascular damages, accelerating age-related atherosclerosis, and microvascular damages, reducing capillary density, is assumed to lead dose-dependently to long-term myocardial ischaemias in breast cancer survivors.20 Nevertheless, modern imaging and radiation techniques might be promising developments that are likely to reduce radiation-induced heart-diseases in the future.19 To investigate potential excess heart-specific mortality, which might be attributable to cardiotoxic treatment, it is crucial to also take into account the steadily increasing cumulative heart-specific mortality by age in the general population. In this study, the cSMRs showed either a slightly lower or comparable heart-specific mortality for breast cancer patients who received radiotherapy or chemotherapy relative to the standard US female population. However, these results have to be carefully interpreted as patients who received radiotherapy or chemotherapy significantly differed from patients who did not receive these treatments in terms of their baseline characteristics such as younger age at diagnosis, higher stage, higher grading, and molecular subtype status. This is consistent with existing literature, as it was found that breast cancer patients who receive chemotherapy have in general less comorbid conditions, rather no cardiovascular diseases and are younger at baseline.21 Hence, we refrained from performing additional analyses on patients who did not receive radiotherapy or chemotherapy as Table 1 showed that these were on average much older and most probably more frail, which makes the interpretation of potential analyses much more challenging. Moreover, additional analyses on this patient subgroup would have been beyond the scope of this study. In subgroup analyses, it was found that especially Stage III breast cancer patients might have a high cardiotoxic burden as they are most likely to receive both radiotherapy and chemotherapy (55.2%) (see Supplementary material online, Table S5). Hence, the interpretation of the presented cSMRs is not straightforward but needs to take into account that patients at higher stages, who receive potentially cardiotoxic treatment, are also more likely to die earlier due to their underlying breast cancer disease before they might develop a heart-specific disease in the long term. Furthermore, clinicians might already have established risk stratification strategies and pre-treatment screenings in routine care, which might exclude patients at highest cardiovascular risk from potentially cardiotoxic treatment. It might also be likely that breast cancer patients with radiotherapy or chemotherapy treatment might have still developed different adverse heart-specific diseases but did not necessarily die from these diseases. While dose-dependent anthracycline administration might cause irreversible cardiomyopathy, trastuzumab-induced cardiomyopathy appears to be reversible after the end of treatment.17 Considering that targeted therapies for HER2-positive patients were also found to induce cardiotoxic effects, our results on intermediate-term survival according to molecular subtypes are encouraging as they show a decreased breast cancer-specific mortality while heart-specific mortality did not differ compared with HER2-negative patients, which was also suggested by another recently published study.22 However, not every HER2-positive diagnosed patient might also receive a HER2-targeted therapy and potential high-risk patients could be less likely to receive HER2-targeted therapy in the first place. Furthermore, follow-up time for this restricted subgroup was shorter than for the general cohort. The results of our study are quite consistent to results from other studies. For example, in a recently published Canadian study, comparing breast cancer patients and non-cancer population controls, cardiovascular death was not significantly associated among breast cancer patients who received anthracyclines only [HR (95% CI) 0.94 (0.41–2.20)], trastuzumab only [1.14 (0.85–1.52)], or a sequential therapy [0.81 (0.25–2.66)].23 In another study, comparing deaths due to cardiovascular diseases in a sample of 1413 breast cancer patients who were diagnosed in 1996–1997 and age-matched to women without breast cancer,5 it was found that an increase in cardovascular deaths among breast cancer patients was only evident 7 years after diagnosis [HRcs 1.8 (1.3–2.5)] and mainly present among women who received chemotherapy [HRcs 1.7 (1.1–2.6)]. A comparable study to ours from the Netherlands, which investigated in great detail the association of different treatment approaches and cardiovascular causes of death, reported that especially 5-year breast cancer survivors who were diagnosed in Stage I–III and had left-sided radiotherapy after mastectomy or chemotherapy after 1997 were at increased risk to die from ischaemic heart diseases and congestive heart failure when accounting for other competing causes of death.24 When comparing this cohort to the general Dutch population, the overall results for all cardiovascular causes of death [SMR 0.92 (0.88–0.97)] and cardiovascular disease subgroups tended to be rather below the null with the exception of valvular dysfunction [SMR 1.28 (1.08–1.52)]. These analyses are quite conclusive to our findings from the SMR analysis. This study has some strengths and limitations. Firstly, the SEER database contains information about chemotherapy and radiation therapy with a high specificity. However, as cancer treatments are increasingly administered outside the hospital there might be the possibility that the cancer registries might have missed cases of patients having received treatment and hence coded them as ‘no/unknown’. Nevertheless, it was found that the positive predictive value of the treatment information captured in SEER is reasonably high (>85%), reflecting sufficient specificity of treatment information for our analyses as especially just patients with radiotherapy or chemotherapy treatment were examined.25 Furthermore, details of the type, dosage and duration of administered chemotherapy or radiotherapy were not available. Also adjuvant endocrine therapy in postmenopausal women would have been of interest, since it was found that compared with tamoxifen, aromatase inhibitors are associated with a higher risk of myocardial infarction.26 Nevertheless, it should be noted that despite a decreasing trend of anthracycline use, studies that investigated the patterns of chemotherapy use among breast cancer patients, found that 88.6% (SEER/Texas Cancer Registry-Medicare-linked database) and 92.8% (Kaiser Permanente Northern California electronic medical records database) of all patients who received chemotherapy received anthracyclines and 88.8% of HER2-positive chemotherapy-treated patients received trastuzumab, respectively.21,27 Hence, use of chemotherapy, which was used as exposure variable in our analysis, should approximate the relevant exposures reasonably well, which is supported by the concordance to other smaller studies, that found an increase in heart-specific events (e.g. hospitalization or major cardiac events) but not in heart-specific mortality.23,24 Secondly, cSMRs are not adjusted for further confounders as in multivariable regression analyses and also do not address the time-to-event temporality as in Cox proportional hazards regressions. Hence, risk factors other than sex, age and stage might still have still confounded these results and future multivariable analyses comparing breast cancer patients and non-breast cancer cohorts are highly desirable. Moreover, mortality data on the general population is just available on aggregated level in the WONDER database and hence does not provide the option to exclude women with a diagnosis of breast cancer. However, due to the good breast cancer prognosis cSMRs might not substantially change. Thirdly, the SEER coverage population was found to have a lower socio-economic status and a greater diversity in the distribution of ethnic minorities and hence might be not representative for the whole United States (US) population.28 Lastly, due to the very long follow-up time, HRcs derived from regression analysis should be interpreted as average effects over time. In a sensitivity analysis to assess time-varying confounding, calculating HRcs for each quartile of follow-up time, associations mostly became stronger with longer follow-up time, but still showed the same directions and patterns (data not shown). Nevertheless, unique strengths of this study are the large sample size, the stratified analyses by molecular receptor subtype and the long follow-up time. Randomized clinical trails are often limited in sample size and follow-up time, which leads to challenges in examining long-term side effects on mortality and hence surrogates such as LVEF are used which might, due to the above mentioned reasons, not necessarily reflect long-term heart-specific mortality outcomes among breast cancer patients. Conclusion In conclusion, this study identified some potential prognostic factors for heart-specific and breast cancer-specific mortality, some of which were quite consistent with previously published literature. More importantly, however, it was observed that the heart-specific mortality among breast cancer patients treated with radiotherapy or chemotherapy is not increased and rather comparable with the general female population. Further studies with more detailed data on treatment are strongly desirable and needed to both validate the results presented in this study and further examine potential prognostic factors. This might add knowledge for further development of cardio-oncological risk stratification among breast cancer patients and help clinicians to make informed decisions about which patient strata are at highest risk and who would benefit most from preventive actions. Supplementary material Supplementary material is available at European Heart Journal online. Acknowledgements The authors would like to thank Silvia Calderazzo from the Division of Biostatistics, German Cancer Research Center, Heidelberg, Germany for her statistical support. Conflict of interest: none declared. References 1 Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer  2015; 136: E359– E386. Google Scholar CrossRef Search ADS PubMed  2 Zamorano JL, Lancellotti P, Rodriguez Munoz D, Aboyans V, Asteggiano R, Galderisi M, Habib G, Lenihan DJ, Lip GY, Lyon AR, Lopez Fernandez T, Mohty D, Piepoli MF, Tamargo J, Torbicki A, Suter TM. 2016 ESC Position Paper on cancer treatments and cardiovascular toxicity developed under the auspices of the ESC Committee for Practice Guidelines: the Task Force for cancer treatments and cardiovascular toxicity of the European Society of Cardiology (ESC). Eur Heart J  2016; 37: 2768– 2801. Google Scholar CrossRef Search ADS PubMed  3 Abdel-Qadir H, Austin PC, Lee DS, Amir E, Tu JV, Thavendiranathan P, Fung K, Anderson GM. A population-based study of cardiovascular mortality following early-stage breast cancer. JAMA Cardiol  2017; 2: 88– 93. Google Scholar CrossRef Search ADS PubMed  4 Patnaik JL, Byers T, DiGuiseppi C, Dabelea D, Denberg TD. Cardiovascular disease competes with breast cancer as the leading cause of death for older females diagnosed with breast cancer: a retrospective cohort study. Breast Cancer Res  2011; 13: R64. Google Scholar CrossRef Search ADS PubMed  5 Bradshaw PT, Stevens J, Khankari N, Teitelbaum SL, Neugut AI, Gammon MD. Cardiovascular disease mortality among breast cancer survivors. Epidemiology  2016; 27: 6– 13. Google Scholar CrossRef Search ADS PubMed  6 Austin PC, Lee DS, Fine JP. Introduction to the analysis of survival data in the presence of competing risks. Circulation  2016; 133: 601– 609. Google Scholar CrossRef Search ADS PubMed  7 Surveillance, Epidemiology, and End Results (SEER) Program Populations (1969–2014). National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch. http://www.seer.cancer.gov/popdata (4 May 2017). 8 Centers for Disease Control and Prevention. CDC Wonder. http://wonder.cdc.gov/ (18 May 2017). 9 Rubin DB. Multiple Imputation for Nonresponse in Surveys . New York, Chichester, Brisbane, Toronto, Singapore: John Wiley & Sons; 2004. 10 Altman DG, De Stavola BL, Love SB, Stepniewska KA. Review of survival analyses published in cancer journals. Br J Cancer  1995; 72: 511– 518. Google Scholar CrossRef Search ADS PubMed  11 Vandenbroucke JP. A shortcut method for the calculation of the 95 per cent confidence interval of the standarized mortality ratio. Am J Epidemiol  1982; 115: 303– 304. Google Scholar CrossRef Search ADS   12 Altman DG, Bland JM. How to obtain the P value from a confidence interval. BMJ  2011; 343: d2304. Google Scholar CrossRef Search ADS PubMed  13 Howlader N, Altekruse SF, Li CI, Chen VW, Clarke CA, Ries LAG, Cronin KA. US incidence of breast cancer subtypes defined by joint hormone receptor and HER2 status. J Natl Cancer Inst  2014; 106: dju055. Google Scholar CrossRef Search ADS PubMed  14 Yoshida K, Bohn J. R Package ‘tableone’. R Foundation for Statistical Computing, Vienna, Austria (30 November 2016). 15 Plana JC, Barac A. Good news, bad news, but not fake news. Circulation  2017; 135: 1413– 1416. Google Scholar CrossRef Search ADS PubMed  16 Swain SM, Baselga J, Kim S-B, Ro J, Semiglazov V, Campone M, Ciruelos E, Ferrero J-M, Schneeweiss A, Heeson S, Clark E, Ross G, Benyunes MC, Cortés J. Pertuzumab, trastuzumab, and docetaxel in HER2-positive metastatic breast cancer. New Engl J Med  2015; 372: 724– 734. Google Scholar CrossRef Search ADS PubMed  17 Schlitt A, Jordan K, Vordermark D, Schwamborn J, Langer T, Thomssen C. Cardiotoxicity and oncological treatments. Dtsch Arztebl Int  2014; 111: 161– 168. Google Scholar PubMed  18 Wysocki PJ, Wysocki H. Cardiovascular complications associated with biological therapies for breast cancer. Expert Opin Biol Ther  2008; 8: 1551– 1559. Google Scholar CrossRef Search ADS PubMed  19 Meattini I, Guenzi M, Fozza A, Vidali C, Rovea P, Meacci F, Livi L. Overview on cardiac, pulmonary and cutaneous toxicity in patients treated with adjuvant radiotherapy for breast cancer. Breast Cancer  2017; 24: 52– 62. Google Scholar CrossRef Search ADS PubMed  20 Darby SC, Cutter DJ, Boerma M, Constine LS, Fajardo LF, Kodama K, Mabuchi K, Marks LB, Mettler FA, Pierce LJ, Trott KR, Yeh ET, Shore RE. Radiation-related heart disease: current knowledge and future prospects. Int J Radiat Oncol Biol Phys  2010; 76: 656– 665. Google Scholar CrossRef Search ADS PubMed  21 Kurian AW, Lichtensztajn DY, Keegan THM, Leung RW, Shema SJ, Hershman DL, Kushi LH, Habel LA, Kolevska T, Caan BJ, Gomez SL. Patterns and predictors of breast cancer chemotherapy use in Kaiser Permanente Northern California, 2004-2007. Breast Cancer Res Treat  2013; 137: 247– 260. Google Scholar CrossRef Search ADS PubMed  22 Dall P, Koch T, Gohler T, Selbach J, Ammon A, Eggert J, Gazawi N, Rezek D, Wischnik A, Hielscher C, Keitel S, Cirrincione U, Hinke A, Feisel-Schwickardi G. Trastuzumab in human epidermal growth factor receptor 2-positive early breast cancer: results of a prospective, noninterventional study on routine treatment between 2006 and 2012 in Germany. Oncologist  2017; 22: 131– 138. Google Scholar CrossRef Search ADS PubMed  23 Thavendiranathan P, Abdel-Qadir H, Fischer HD, Camacho X, Amir E, Austin PC, Lee DS. Breast cancer therapy-related cardiac dysfunction in adult women Treated in routine clinical practice: a population-based cohort study. J Clin Oncol  2016; 34: 2239– 2246. Google Scholar CrossRef Search ADS PubMed  24 Boekel NB, Schaapveld M, Gietema JA, Russell NS, Poortmans P, Theuws JC, Schinagl DA, Rietveld DH, Versteegh MI, Visser O, Rutgers EJ, Aleman BM, van Leeuwen FE. Cardiovascular disease risk in a large, population-based cohort of breast cancer survivors. Int J Radiat Oncol Biol Phys  2016; 94: 1061– 1072. Google Scholar CrossRef Search ADS PubMed  25 Noone AM, Lund JL, Mariotto A, Cronin K, McNeel T, Deapen D, Warren JL. Comparison of SEER treatment data with medicare claims. Med Care  2016; 54: e55– e64. Google Scholar CrossRef Search ADS PubMed  26 Abdel-Qadir H, Amir E, Fischer HD, Fu L, Austin PC, Harvey PJ, Rochon PA, Lee DS, Anderson GM. The risk of myocardial infarction with aromatase inhibitors relative to tamoxifen in post-menopausal women with early stage breast cancer. Eur J Cancer  2016; 68: 11– 21. Google Scholar CrossRef Search ADS PubMed  27 Barcenas CH, Niu J, Zhang N, Zhang Y, Buchholz TA, Elting LS, Hortobagyi GN, Smith BD, Giordano SH. Risk of hospitalization according to chemotherapy regimen in early-stage breast cancer. J Clin Oncol  2014; 32: 2010– 2017. Google Scholar CrossRef Search ADS PubMed  28 Kuo T-M, Mobley LR. How generalizable are the SEER registries to the cancer populations of the USA? Cancer Causes Control  2016; 27: 1117– 1126. Google Scholar CrossRef Search ADS PubMed  Published on behalf of the European Society of Cardiology. All rights reserved. © The Author(s) 2018. For permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Heart Journal Oxford University Press

Long-term heart-specific mortality among 347 476 breast cancer patients treated with radiotherapy or chemotherapy: a registry-based cohort study

Loading next page...
 
/lp/ou_press/long-term-heart-specific-mortality-among-347-476-breast-cancer-KtIssB5FjK
Publisher
Oxford University Press
Copyright
Published on behalf of the European Society of Cardiology. All rights reserved. © The Author(s) 2018. For permissions, please email: journals.permissions@oup.com.
ISSN
0195-668X
eISSN
1522-9645
D.O.I.
10.1093/eurheartj/ehy167
Publisher site
See Article on Publisher Site

Abstract

Abstract Aims Breast cancer survival has improved throughout the last decades, but treatment-induced cardiotoxicity remains a major concern. This study aimed to investigate competing causes of death and prognostic factors within a large cohort of breast cancer patients and to describe the heart-specific mortality in relation to the general population. Methods and results In this registry-based cohort study, women diagnosed with breast cancer between 2000 and 2011, who were treated with radiotherapy or chemotherapy and followed until 2014, were identified from the Surveillance, Epidemiology, and End Results-18 (SEER-18) database. Cumulative mortality functions were computed. To investigate heart-specific mortality relative to the general population, long-term (≥10 years) standardized mortality ratios (SMRs) were calculated. Prognostic factors for heart-specific mortality were assessed by calculating cause-specific hazard ratios (HRcs) with corresponding 95% confidence intervals using the Cox proportional hazards regression. Subgroup analysis on intermediate-term mortality according to molecular subtypes, for which information was available since 2010, was performed. In total, 347 476 breast cancer patients were eligible to be included in the study. Among all possible competing causes of death, breast cancer accounted for the highest cumulative mortality. Compared with the general population, heart-specific mortality of breast cancer patients treated with radiotherapy or chemotherapy was lower [SMRoverall 0.84 (0.79–0.90)]. In subgroup analysis, human epidermal growth factor receptor 2 (HER2)-positive subtype was not associated with increased heart-specific mortality relative to HER2-negative patients [HRcs 0.96 (0.70–1.32)]. Conclusion Heart-specific mortality among breast cancer survivors is not increased compared with the general population. Human epidermal growth factor receptor 2-positive patients do not have increased heart-specific mortality compared to HER2-negative patients. Cardio-oncology, Breast cancer, Cardiovascular, Competing risk, Cancer registry Introduction With 1.67 million cases in 2012, breast cancer is the second most common cancer worldwide.1 Whereas improvement in treatment options including radiation and human epidermal growth factor receptor 2 (HER2)-targeted chemotherapy led to progress in terms of breast cancer survival, long-term mortality due to persistent decline in left ventricular ejection fraction (LVEF) and incidental cardiac diseases have remained one of the major concerns in breast cancer survivors.2 Recent studies found that especially among older breast cancer patients and those with a concomitant cardiovascular comorbidity at time of breast cancer diagnosis, the cumulative mortality due to a cardiovascular disease was higher than the cumulative mortality due to the underlying breast cancer disease.3,4 However, most of these findings were analysed within breast cancer cohorts only, and not much is known about cardiovascular mortality compared with the general population.5 As increasing age is a natural driver for the incidence of cardiovascular diseases and cardiovascular mortality among the general population, this comparison is crucial in order to disclose associations potentially attributable to breast cancer therapy such as radiotherapy or chemotherapy. Moreover, it is of utmost importance that these analyses are performed with careful consideration of competing risks.3,6 In the light of possible treatment-related cardiotoxic effects, we aimed to investigate long-term heart-specific mortality within a large breast cancer population including 347 476 US female breast cancer patients treated with radiotherapy or chemotherapy both in the presence of competing risks and in comparison with the general US female population. Methods Data sources For this registry-based retrospective cohort study, the Surveillance, Epidemiology, and End Results-18 (SEER-18) database was used, which covers data from 18 regional cancer registries throughout the USA.7 Access to the data was provided through a standard request to SEER, which was accepted after filling the required data usage agreement. In total, the SEER cancer registries draw on a base population of approximately 28% of the US population and capture comprehensive demographic and cancer-specific information including the International Classification of Diseases-10 (ICD-10) codes for the underlying cause of death. The reference cohort representing the general US population was selected from the underlying cause of death (WONDER) online database of the Centers for Disease Control and Prevention (CDC).8 These county-level national mortality and population data are based on death certificates for US residents. Study population Breast cancer cases were defined as women at age 15 years or older with a first primary malignant breast cancer diagnosis (ICD-10: C50.x) between 1 January 2000 and 31 December 2011. This calendar year interval was chosen to provide a sufficient minimum follow-up time for all included patients. Follow-up time was defined as the time from diagnosis to the date of last contact, death or end of the study period (31 December 2014), whichever occurred first. Patients without or with unknown status of surgery (8.2%), missing information on cause of death (0.5%), stage 0 (0.1%), or unknown stage (3.6%) were excluded in order to perform a complete case analysis. To check if results hold when the above mentioned variables were imputed, all main analyses were repeated and estimates from 30 multiple imputed datasets were combined according to Rubin’s rule.9 Finally, the analysis cohort was restricted to patients who received either radiotherapy or chemotherapy or both. Outcome and covariate assessment The primary focus of this study was given to mortality due to diseases of the heart as a potential long-term side-effect of cardiotoxic treatment such as radiation and chemotherapy. Hence, heart-specific causes of death were defined according to the SEER recode 50060 (ICD-10 codes are presented in Supplementary material online, Table S1). Sensitivity analyses were also conducted including patients dying from primary hypertension and hypertensive renal disease (ICD-10 codes I10 and I12, respectively). Table 1 Basic characteristics of all included female breast cancer patients at time of diagnosis*   All patients  Treatment (chemotherapy or radiotherapy)   Yes  No/unknown  N  501 547  347 476  154 071  Age, mean (SD)  59.94 (13.71)  57.51 (12.61)  65.41 (14.48)  Age groups   <50  125 328 (25.0)  99 795 (28.7)  25 533 (16.6)   50–64  190 536 (38.0)  144 409 (41.6)  46 127 (29.9)   65–74  100 059 (20.0)  66 184 (19.0)  33 875 (22.0)   75+  85 624 (17.1)  37 088 (10.7)  48 536 (31.5)  Ethnicity, n (%)   White  411 241 (82.0)  283 319 (81.5)  127 922 (83.0)   Black  49 410 (9.9)  35 577 (10.2)  13 833 (9.0)   American Indian/Alaska Native  2524 (0.5)  1818 (0.5)  706 (0.5)   Asian/Pacific Islander  36 424 (7.3)  25 620 (7.4)  10 804 (7.0)   Other/unknown  1948 (0.4)  1142 (0.3)  806 (0.5)  Year of diagnosis, n (%)   2000–2005  242 584 (48.4)  165 135 (47.5)  77 449 (50.3)   2006–2011  258 963 (51.6)  182 341 (52.5)  76 622 (49.7)  Stage, n (%)   I  246 040 (49.1)  156 504 (45.0)  89 536 (58.1)   II  179 950 (35.9)  128 560 (37.0)  51 390 (33.4)   III  65 773 (13.1)  55 096 (15.9)  10 677 (6.9)   IV  9784 (2.0)  7316 (2.1)  2468 (1.6)  Grading, n (%)a   I  102 147 (21.6)  65 133 (19.7)  37 014 (26.1)   II  199 202 (42.2)  135 484 (41.0)  63 718 (44.9)   III  164 513 (34.9)  125 255 (38.0)  39 258 (27.7)   IV  6156 (1.3)  4178 (1.3)  1978 (1.4)  ICD-10 site, n (%)   C50.0 (Nipple and areola)  2533 (0.5)  1324 (0.4)  1209 (0.8)   C50.1 (Central portion of breast)  27 092 (5.4)  17 207 (5.0)  9885 (6.4)   C50.2 (Upper-inner quadrant of breast)  55 226 (11.0)  39 697 (11.4)  15 529 (10.1)   C50.3 (Lower-inner quadrant of breast)  28 205 (5.6)  19 644 (5.7)  8561 (5.6)   C50.4 (Upper-outer quadrant of breast)  174 951 (34.9)  126 602 (36.4)  48 349 (31.4)   C50.5 (Lower-outer quadrant of breast)  35 136 (7.0)  24 914 (7.2)  10 222 (6.6)   C50.6 (Axillary tail of breast)  3191 (0.6)  2362 (0.7)  829 (0.5)   C50.8 (Overlapping lesion of breast)  106 527 (21.2)  74 179 (21.3)  32 348 (21.0)   C50.9 (Breast, unspecified)  68 686 (13.7)  41 547 (12.0)  27 139 (17.6)  Cause of death, n (%)   Alive  381 201 (76.0)  276 198 (79.5)  105 003 (68.2)   Breast cancer  54 424 (10.9)  39 802 (11.5)  14 622 (9.5)   Heart-specific disease  17 278 (3.4)  7401 (2.1)  9877 (6.4)   Other  48 644 (9.7)  24 075 (6.9)  24 569 (15.9)  Radiotherapy, n (%)b  260 144 (53.0)  260 144 (76.2)  0 (0.0)  Chemotherapy, n (%)  212 600 (42.4)  212 600 (61.2)  0 (0.0)  HER2+, n (%)c  11 921 (14.5)  9733 (16.4)  2188 (9.6)  ER+, n (%)d  367 708 (79.5)  254 177 (77.4)  113 531 (84.7)  PR+, n (%)e  310 036 (68.1)  214 590 (66.3)  95 446 (72.7)    All patients  Treatment (chemotherapy or radiotherapy)   Yes  No/unknown  N  501 547  347 476  154 071  Age, mean (SD)  59.94 (13.71)  57.51 (12.61)  65.41 (14.48)  Age groups   <50  125 328 (25.0)  99 795 (28.7)  25 533 (16.6)   50–64  190 536 (38.0)  144 409 (41.6)  46 127 (29.9)   65–74  100 059 (20.0)  66 184 (19.0)  33 875 (22.0)   75+  85 624 (17.1)  37 088 (10.7)  48 536 (31.5)  Ethnicity, n (%)   White  411 241 (82.0)  283 319 (81.5)  127 922 (83.0)   Black  49 410 (9.9)  35 577 (10.2)  13 833 (9.0)   American Indian/Alaska Native  2524 (0.5)  1818 (0.5)  706 (0.5)   Asian/Pacific Islander  36 424 (7.3)  25 620 (7.4)  10 804 (7.0)   Other/unknown  1948 (0.4)  1142 (0.3)  806 (0.5)  Year of diagnosis, n (%)   2000–2005  242 584 (48.4)  165 135 (47.5)  77 449 (50.3)   2006–2011  258 963 (51.6)  182 341 (52.5)  76 622 (49.7)  Stage, n (%)   I  246 040 (49.1)  156 504 (45.0)  89 536 (58.1)   II  179 950 (35.9)  128 560 (37.0)  51 390 (33.4)   III  65 773 (13.1)  55 096 (15.9)  10 677 (6.9)   IV  9784 (2.0)  7316 (2.1)  2468 (1.6)  Grading, n (%)a   I  102 147 (21.6)  65 133 (19.7)  37 014 (26.1)   II  199 202 (42.2)  135 484 (41.0)  63 718 (44.9)   III  164 513 (34.9)  125 255 (38.0)  39 258 (27.7)   IV  6156 (1.3)  4178 (1.3)  1978 (1.4)  ICD-10 site, n (%)   C50.0 (Nipple and areola)  2533 (0.5)  1324 (0.4)  1209 (0.8)   C50.1 (Central portion of breast)  27 092 (5.4)  17 207 (5.0)  9885 (6.4)   C50.2 (Upper-inner quadrant of breast)  55 226 (11.0)  39 697 (11.4)  15 529 (10.1)   C50.3 (Lower-inner quadrant of breast)  28 205 (5.6)  19 644 (5.7)  8561 (5.6)   C50.4 (Upper-outer quadrant of breast)  174 951 (34.9)  126 602 (36.4)  48 349 (31.4)   C50.5 (Lower-outer quadrant of breast)  35 136 (7.0)  24 914 (7.2)  10 222 (6.6)   C50.6 (Axillary tail of breast)  3191 (0.6)  2362 (0.7)  829 (0.5)   C50.8 (Overlapping lesion of breast)  106 527 (21.2)  74 179 (21.3)  32 348 (21.0)   C50.9 (Breast, unspecified)  68 686 (13.7)  41 547 (12.0)  27 139 (17.6)  Cause of death, n (%)   Alive  381 201 (76.0)  276 198 (79.5)  105 003 (68.2)   Breast cancer  54 424 (10.9)  39 802 (11.5)  14 622 (9.5)   Heart-specific disease  17 278 (3.4)  7401 (2.1)  9877 (6.4)   Other  48 644 (9.7)  24 075 (6.9)  24 569 (15.9)  Radiotherapy, n (%)b  260 144 (53.0)  260 144 (76.2)  0 (0.0)  Chemotherapy, n (%)  212 600 (42.4)  212 600 (61.2)  0 (0.0)  HER2+, n (%)c  11 921 (14.5)  9733 (16.4)  2188 (9.6)  ER+, n (%)d  367 708 (79.5)  254 177 (77.4)  113 531 (84.7)  PR+, n (%)e  310 036 (68.1)  214 590 (66.3)  95 446 (72.7)  ER, estrogen receptor; HER, human epidermal growth factor receptor; ICD, international classification of diseases; PR, progesterone receptor. * P-values for differences in categorical variables between treated and untreated patients are all <0.001. a Information on grading missing for 29 529 patients. b Information missing/unknown for 10 771 patients. c Not available for 419 384 patients, since HER2 status is recorded in SEER from 2010+ on. d ER status not available for 38 930 patients. e PR status not available for 46 578 patients. Table 1 Basic characteristics of all included female breast cancer patients at time of diagnosis*   All patients  Treatment (chemotherapy or radiotherapy)   Yes  No/unknown  N  501 547  347 476  154 071  Age, mean (SD)  59.94 (13.71)  57.51 (12.61)  65.41 (14.48)  Age groups   <50  125 328 (25.0)  99 795 (28.7)  25 533 (16.6)   50–64  190 536 (38.0)  144 409 (41.6)  46 127 (29.9)   65–74  100 059 (20.0)  66 184 (19.0)  33 875 (22.0)   75+  85 624 (17.1)  37 088 (10.7)  48 536 (31.5)  Ethnicity, n (%)   White  411 241 (82.0)  283 319 (81.5)  127 922 (83.0)   Black  49 410 (9.9)  35 577 (10.2)  13 833 (9.0)   American Indian/Alaska Native  2524 (0.5)  1818 (0.5)  706 (0.5)   Asian/Pacific Islander  36 424 (7.3)  25 620 (7.4)  10 804 (7.0)   Other/unknown  1948 (0.4)  1142 (0.3)  806 (0.5)  Year of diagnosis, n (%)   2000–2005  242 584 (48.4)  165 135 (47.5)  77 449 (50.3)   2006–2011  258 963 (51.6)  182 341 (52.5)  76 622 (49.7)  Stage, n (%)   I  246 040 (49.1)  156 504 (45.0)  89 536 (58.1)   II  179 950 (35.9)  128 560 (37.0)  51 390 (33.4)   III  65 773 (13.1)  55 096 (15.9)  10 677 (6.9)   IV  9784 (2.0)  7316 (2.1)  2468 (1.6)  Grading, n (%)a   I  102 147 (21.6)  65 133 (19.7)  37 014 (26.1)   II  199 202 (42.2)  135 484 (41.0)  63 718 (44.9)   III  164 513 (34.9)  125 255 (38.0)  39 258 (27.7)   IV  6156 (1.3)  4178 (1.3)  1978 (1.4)  ICD-10 site, n (%)   C50.0 (Nipple and areola)  2533 (0.5)  1324 (0.4)  1209 (0.8)   C50.1 (Central portion of breast)  27 092 (5.4)  17 207 (5.0)  9885 (6.4)   C50.2 (Upper-inner quadrant of breast)  55 226 (11.0)  39 697 (11.4)  15 529 (10.1)   C50.3 (Lower-inner quadrant of breast)  28 205 (5.6)  19 644 (5.7)  8561 (5.6)   C50.4 (Upper-outer quadrant of breast)  174 951 (34.9)  126 602 (36.4)  48 349 (31.4)   C50.5 (Lower-outer quadrant of breast)  35 136 (7.0)  24 914 (7.2)  10 222 (6.6)   C50.6 (Axillary tail of breast)  3191 (0.6)  2362 (0.7)  829 (0.5)   C50.8 (Overlapping lesion of breast)  106 527 (21.2)  74 179 (21.3)  32 348 (21.0)   C50.9 (Breast, unspecified)  68 686 (13.7)  41 547 (12.0)  27 139 (17.6)  Cause of death, n (%)   Alive  381 201 (76.0)  276 198 (79.5)  105 003 (68.2)   Breast cancer  54 424 (10.9)  39 802 (11.5)  14 622 (9.5)   Heart-specific disease  17 278 (3.4)  7401 (2.1)  9877 (6.4)   Other  48 644 (9.7)  24 075 (6.9)  24 569 (15.9)  Radiotherapy, n (%)b  260 144 (53.0)  260 144 (76.2)  0 (0.0)  Chemotherapy, n (%)  212 600 (42.4)  212 600 (61.2)  0 (0.0)  HER2+, n (%)c  11 921 (14.5)  9733 (16.4)  2188 (9.6)  ER+, n (%)d  367 708 (79.5)  254 177 (77.4)  113 531 (84.7)  PR+, n (%)e  310 036 (68.1)  214 590 (66.3)  95 446 (72.7)    All patients  Treatment (chemotherapy or radiotherapy)   Yes  No/unknown  N  501 547  347 476  154 071  Age, mean (SD)  59.94 (13.71)  57.51 (12.61)  65.41 (14.48)  Age groups   <50  125 328 (25.0)  99 795 (28.7)  25 533 (16.6)   50–64  190 536 (38.0)  144 409 (41.6)  46 127 (29.9)   65–74  100 059 (20.0)  66 184 (19.0)  33 875 (22.0)   75+  85 624 (17.1)  37 088 (10.7)  48 536 (31.5)  Ethnicity, n (%)   White  411 241 (82.0)  283 319 (81.5)  127 922 (83.0)   Black  49 410 (9.9)  35 577 (10.2)  13 833 (9.0)   American Indian/Alaska Native  2524 (0.5)  1818 (0.5)  706 (0.5)   Asian/Pacific Islander  36 424 (7.3)  25 620 (7.4)  10 804 (7.0)   Other/unknown  1948 (0.4)  1142 (0.3)  806 (0.5)  Year of diagnosis, n (%)   2000–2005  242 584 (48.4)  165 135 (47.5)  77 449 (50.3)   2006–2011  258 963 (51.6)  182 341 (52.5)  76 622 (49.7)  Stage, n (%)   I  246 040 (49.1)  156 504 (45.0)  89 536 (58.1)   II  179 950 (35.9)  128 560 (37.0)  51 390 (33.4)   III  65 773 (13.1)  55 096 (15.9)  10 677 (6.9)   IV  9784 (2.0)  7316 (2.1)  2468 (1.6)  Grading, n (%)a   I  102 147 (21.6)  65 133 (19.7)  37 014 (26.1)   II  199 202 (42.2)  135 484 (41.0)  63 718 (44.9)   III  164 513 (34.9)  125 255 (38.0)  39 258 (27.7)   IV  6156 (1.3)  4178 (1.3)  1978 (1.4)  ICD-10 site, n (%)   C50.0 (Nipple and areola)  2533 (0.5)  1324 (0.4)  1209 (0.8)   C50.1 (Central portion of breast)  27 092 (5.4)  17 207 (5.0)  9885 (6.4)   C50.2 (Upper-inner quadrant of breast)  55 226 (11.0)  39 697 (11.4)  15 529 (10.1)   C50.3 (Lower-inner quadrant of breast)  28 205 (5.6)  19 644 (5.7)  8561 (5.6)   C50.4 (Upper-outer quadrant of breast)  174 951 (34.9)  126 602 (36.4)  48 349 (31.4)   C50.5 (Lower-outer quadrant of breast)  35 136 (7.0)  24 914 (7.2)  10 222 (6.6)   C50.6 (Axillary tail of breast)  3191 (0.6)  2362 (0.7)  829 (0.5)   C50.8 (Overlapping lesion of breast)  106 527 (21.2)  74 179 (21.3)  32 348 (21.0)   C50.9 (Breast, unspecified)  68 686 (13.7)  41 547 (12.0)  27 139 (17.6)  Cause of death, n (%)   Alive  381 201 (76.0)  276 198 (79.5)  105 003 (68.2)   Breast cancer  54 424 (10.9)  39 802 (11.5)  14 622 (9.5)   Heart-specific disease  17 278 (3.4)  7401 (2.1)  9877 (6.4)   Other  48 644 (9.7)  24 075 (6.9)  24 569 (15.9)  Radiotherapy, n (%)b  260 144 (53.0)  260 144 (76.2)  0 (0.0)  Chemotherapy, n (%)  212 600 (42.4)  212 600 (61.2)  0 (0.0)  HER2+, n (%)c  11 921 (14.5)  9733 (16.4)  2188 (9.6)  ER+, n (%)d  367 708 (79.5)  254 177 (77.4)  113 531 (84.7)  PR+, n (%)e  310 036 (68.1)  214 590 (66.3)  95 446 (72.7)  ER, estrogen receptor; HER, human epidermal growth factor receptor; ICD, international classification of diseases; PR, progesterone receptor. * P-values for differences in categorical variables between treated and untreated patients are all <0.001. a Information on grading missing for 29 529 patients. b Information missing/unknown for 10 771 patients. c Not available for 419 384 patients, since HER2 status is recorded in SEER from 2010+ on. d ER status not available for 38 930 patients. e PR status not available for 46 578 patients. As potential predictors of mortality, age (categorized as age <35, 35–49, 50–54, 55–64, 65–74, and ≥75 years), ethnicity (white, black, American Indian/Alaska native, Asian/Pacific islander, and other), year of diagnosis, stage [coded according to the breast cancer adjusted staging of the American Joint Committee on Cancer (AJCC) 6th edition], grade (I–IV), and tumour site were considered. Statistical analysis The distributions of all baseline characteristics were summarized by calculating the mean and standard deviation for continuous variables and frequencies for categorical variables, respectively. The median follow-up time was assessed using the reverse Kaplan–Meier method.10 Crude cumulative mortality functions were calculated and plotted for cause-specific deaths overall and stratified by age to describe the probability of experiencing a specific endpoint in the presence of competing risks among patients treated with radiotherapy or chemotherapy. As standard Kaplan–Meier analyses treat failures from competing events as censored, this approach would lead to an overestimation of the absolute risk of the event of interest because competing events would then violate the assumption of non-informative censoring.6 Furthermore, the Kaplan–Meier estimate would reflect mortality from the event of interest in a hypothetical world without competing events, which is less clinically relevant. Hence, to compute the probability to die from a specific cause of death (cumulative mortality), breast cancer patients who died due to competing causes of death were retained in the underlying risk set instead of being censored. To investigate if risk factors such as age might be only associated with a potentially higher heart-specific mortality by the naturally increasing mortality due to aging, overall and age-specific standardized mortality ratios (SMRs) were calculated with the US female standard population from the CDC database as reference.8 As potential long-term cardiotoxicity was of major interest, conditional SMRs (cSMRs) were calculated for the calendar period 2012–2014 restricting the breast cancer cohort to patients who were diagnosed between 2000 and 2002 to ensure a conditional survival of at least ten years. Conditional SMRs were calculated by dividing the observed numbers of heart-specific deaths divided by the expected numbers. Latter were determined by multiplying the cumulative person-time across breast cancer patients within 5-year age-groups (55–59, 60–64, 65–69, 70–74, 75–79, 80–84, and ≥85) and the calendar period 2012–2014 with the age- and calendar period-specific mortality rates from heart diseases in the general female population. All age-specific and overall cSMRs were calculated for patients aged 55 years or older who received radiotherapy or chemotherapy. The 5-year age groups described above were chosen for age classification for two reasons: Firstly, for the aggregated data for the US general female population only categorizations in 1-year, 5-, and 10-year steps were available, of which the 5-year categorizations were found best to operate for the cSMR analyses. Secondly, official mortality statistics of the US population are categorized in exactly the same 5-year steps, which we adopted to make comparisons with cSMRs easier. Additionally, stratified analyses were performed by Stage I–III and for all-cause mortality. Stage-specific analyses for Stage IV patients were not conducted due to the very low numbers of 10-year survivors in this group. 95% confidence interval (95% CI) and P-values were estimated according to the method by Vandenbroucke et al.11 and Altman et al.,12 respectively. Additionally, adjusted cause-specific hazard ratios (HRcs) were computed to estimate the relative association of individual risk and prognostic factors on both heart- and breast cancer related mortality, using the Cox proportional hazards regression. Potential confounding was addressed by adjusting all models for the above listed variables. Stratified analyses by molecular subtype were conducted as a separate analysis on intermediate-term heart-specific mortality including all eligible breast cancer patients for whom this information was available since 2010.13 Table 1 was created using the R package ‘tableone’ (R Foundation for Statistical Computing, Vienna, Austria).14 All other analyses were performed with SAS software, version 9.4 (SAS Institute Inc., Cary, NC, USA) according to an apriori defined study protocol. Statistical significance was defined by a two-sided P < 0.05. Results Out of 572 341 identified breast cancer cases between 2000 and 2011, 501 547 (87.6%) female breast cancer patients were eligible to be included in the study, of whom 347 476 (69.3%) received either radiotherapy or chemotherapy or both (Figure 1). Figure 1 View largeDownload slide Selection of eligible patients for competing risk analysis. Figure 1 View largeDownload slide Selection of eligible patients for competing risk analysis. Compared with patients who did not receive radiotherapy or chemotherapy, treated patients were on average younger at diagnosis (57.5 years vs. 65.4), rather diagnosed in later stages (Stage III or IV 18.0% vs. 8.5%) and with higher gradings (Grade III or IV 39.3% vs. 29.1%) (Table 1). The majority of patients were white (81.5% and 83.0%, respectively) and the most common tumour site in both groups was the upper-outer quadrant of the breast (36.4% and 31.4%, respectively). In general, patients treated with radiotherapy or chemotherapy were more likely to die due to breast cancer (11.5% vs. 9.5%), and less likely to die due to a heart-specific disease (2.1% vs. 6.4%). Among all patients for whom the information was available, treated patients were more frequently diagnosed with a HER2-positive (HER2+) molecular subtype (16.4% vs. 9.6%), but less often diagnosed with oestrogen receptor positive (ER+) and progesterone receptor positive (PR+) molecular subtypes (77.4% vs. 84.7 and 66.3 vs. 72.7%, respectively). The median follow-up time for patients treated with radiotherapy or chemotherapy comprised 8.4 years [interquartile range (IQR) 5.5–11.6] and was slightly higher for untreated patients [8.6 (IQR 5.6–11.2)]. Cumulative mortality The cumulative mortalities for all causes of death among breast cancer patients treated with radiotherapy or chemotherapy are illustrated in Figure 2. By far the highest cumulative mortality of death was caused by breast cancer, followed by other non-cancer causes of death, which were not related to heart diseases. Heart-specific cumulative mortality was marginally lower than cumulative mortality due to cancers other than breast cancer. These patterns also did not substantially change when primary hypertension and hypertensive renal disease were included among the heart-specific causes of death or when using the imputed datasets (see Supplementary material online, Figure S1 and S2, respectively). Figure 2 View largeDownload slide Cumulative cause-specific mortality among treated breast cancer patients. Figure 2 View largeDownload slide Cumulative cause-specific mortality among treated breast cancer patients. Stratified by age groups, women aged 50 years or younger had the highest cumulative breast cancer mortality compared with heart-specific deaths (Take home figure). Nevertheless, it was observed that the cumulative heart-specific mortality steadily increased with age at diagnosis and length of follow-up. In the oldest age group (≥75 years), the cumulative mortality from heart-specific diseases exceeded the cumulative mortality from breast cancer around 12 years after diagnosis. When using the multiple imputed datasets, however, the cumulative mortality from heart-specific diseases was close, but did not exceed the cumulative mortality from breast cancer within the first 15 years after diagnosis (see Supplementary material online, Figure S3). Take home figure View largeDownload slide Cumulative heart-specific and breast cancer-specific mortality by age at diagnosis. Take home figure View largeDownload slide Cumulative heart-specific and breast cancer-specific mortality by age at diagnosis. Conditional standardized mortality ratios The comparison of heart-specific mortality between breast cancer patients and the general population is displayed in Table 2. Table 2 Age-specific and overall standardized mortality ratios for the years 2012–2014 among breast cancer patients who survived at least 10 years and were treated with radio- or chemotherapy diagnosed between 2000 and 2002 relative to the USA female standard population Group  cSMR heart-specific  P-value  cSMR overall  P-value  Agea           55–59  0.67 (0.35–1.08)  0.1553  2.25 (1.99–2.53)  <0.0001   60–64  0.99 (0.69–1.34)  0.9541  1.75 (1.58–1.94)  <0.0001   65–69  0.77 (0.56–1.01)  0.0830  1.48 (1.35–1.61)  <0.0001   70–74  0.66 (0.50–0.85)  0.0028  1.22 (1.12–1.33)  <0.0001   75–79  0.87 (0.71–1.05)  0.1648  1.13 (1.05–1.22)  0.0017   80–84  0.89 (0.75–1.04)  0.1601  1.09 (1.02–1.17)  0.0120   85+  0.86 (0.78–0.94)  0.0011  0.88 (0.84–0.93)  <0.0001  Stage           I  0.75 (0.68–0.82)  <0.0001  0.88 (0.84–0.91)  <0.0001   II  0.89 (0.79–1.00)  0.0598  1.23 (1.17–1.29)  <0.0001   III  0.99 (0.79–1.21)  0.9188  1.90 (1.76–2.04)  <0.0001  Overall  0.84 (0.79–0.90)  <0.0001  1.12 (1.09–1.15)  <0.0001  Group  cSMR heart-specific  P-value  cSMR overall  P-value  Agea           55–59  0.67 (0.35–1.08)  0.1553  2.25 (1.99–2.53)  <0.0001   60–64  0.99 (0.69–1.34)  0.9541  1.75 (1.58–1.94)  <0.0001   65–69  0.77 (0.56–1.01)  0.0830  1.48 (1.35–1.61)  <0.0001   70–74  0.66 (0.50–0.85)  0.0028  1.22 (1.12–1.33)  <0.0001   75–79  0.87 (0.71–1.05)  0.1648  1.13 (1.05–1.22)  0.0017   80–84  0.89 (0.75–1.04)  0.1601  1.09 (1.02–1.17)  0.0120   85+  0.86 (0.78–0.94)  0.0011  0.88 (0.84–0.93)  <0.0001  Stage           I  0.75 (0.68–0.82)  <0.0001  0.88 (0.84–0.91)  <0.0001   II  0.89 (0.79–1.00)  0.0598  1.23 (1.17–1.29)  <0.0001   III  0.99 (0.79–1.21)  0.9188  1.90 (1.76–2.04)  <0.0001  Overall  0.84 (0.79–0.90)  <0.0001  1.12 (1.09–1.15)  <0.0001  cSMR, standardized mortality ratio conditional on at least 10 year survival among eligible breast cancer patients. a Age of the breast cancer patients/population in the respective follow-up year (2012–2014). Table 2 Age-specific and overall standardized mortality ratios for the years 2012–2014 among breast cancer patients who survived at least 10 years and were treated with radio- or chemotherapy diagnosed between 2000 and 2002 relative to the USA female standard population Group  cSMR heart-specific  P-value  cSMR overall  P-value  Agea           55–59  0.67 (0.35–1.08)  0.1553  2.25 (1.99–2.53)  <0.0001   60–64  0.99 (0.69–1.34)  0.9541  1.75 (1.58–1.94)  <0.0001   65–69  0.77 (0.56–1.01)  0.0830  1.48 (1.35–1.61)  <0.0001   70–74  0.66 (0.50–0.85)  0.0028  1.22 (1.12–1.33)  <0.0001   75–79  0.87 (0.71–1.05)  0.1648  1.13 (1.05–1.22)  0.0017   80–84  0.89 (0.75–1.04)  0.1601  1.09 (1.02–1.17)  0.0120   85+  0.86 (0.78–0.94)  0.0011  0.88 (0.84–0.93)  <0.0001  Stage           I  0.75 (0.68–0.82)  <0.0001  0.88 (0.84–0.91)  <0.0001   II  0.89 (0.79–1.00)  0.0598  1.23 (1.17–1.29)  <0.0001   III  0.99 (0.79–1.21)  0.9188  1.90 (1.76–2.04)  <0.0001  Overall  0.84 (0.79–0.90)  <0.0001  1.12 (1.09–1.15)  <0.0001  Group  cSMR heart-specific  P-value  cSMR overall  P-value  Agea           55–59  0.67 (0.35–1.08)  0.1553  2.25 (1.99–2.53)  <0.0001   60–64  0.99 (0.69–1.34)  0.9541  1.75 (1.58–1.94)  <0.0001   65–69  0.77 (0.56–1.01)  0.0830  1.48 (1.35–1.61)  <0.0001   70–74  0.66 (0.50–0.85)  0.0028  1.22 (1.12–1.33)  <0.0001   75–79  0.87 (0.71–1.05)  0.1648  1.13 (1.05–1.22)  0.0017   80–84  0.89 (0.75–1.04)  0.1601  1.09 (1.02–1.17)  0.0120   85+  0.86 (0.78–0.94)  0.0011  0.88 (0.84–0.93)  <0.0001  Stage           I  0.75 (0.68–0.82)  <0.0001  0.88 (0.84–0.91)  <0.0001   II  0.89 (0.79–1.00)  0.0598  1.23 (1.17–1.29)  <0.0001   III  0.99 (0.79–1.21)  0.9188  1.90 (1.76–2.04)  <0.0001  Overall  0.84 (0.79–0.90)  <0.0001  1.12 (1.09–1.15)  <0.0001  cSMR, standardized mortality ratio conditional on at least 10 year survival among eligible breast cancer patients. a Age of the breast cancer patients/population in the respective follow-up year (2012–2014). When heart-specific mortality among breast cancer patients treated with radiotherapy or chemotherapy was compared to the general female US population, it was observed that the cSMRs tended to be slightly decreased for all age groups with an overall decrease by 16% (95% CI 21–10%) in the age standardized analysis. When this analysis was repeated using the multiple imputed datasets, the cSMRs varied marginally but still showed the same patterns (see Supplementary material online, Table S2). In terms of all-cause mortality, breast cancer patients treated with radiotherapy or chemotherapy showed an over two-fold increased mortality among younger breast cancer patients [cSMR 2.25 (1.99–2.53)] in the age group 55–59 years), which then steadily decreased by increasing age. The overall cSMRs suggested a 12% [1.12 (1.09–1.15)] increased mortality. When stratified by stage, it was observed that both heart-specific and overall cSMRs increased by stage. Cause-specific hazard ratios The associations between individual prognostic factors and heart-specific mortality and breast cancer mortality among patients receiving radiotherapy or chemotherapy are presented in Table 3. Table 3 Cause-specific hazard and 95% confidence intervals for heart-related and breast-cancer related mortality among breast cancer patients who received either chemotherapy or radiotherapy or both* Strata/model  N  Cause-specific hazards ratiosa   Heart  P-value  Breast cancer  P-value  Age at diagnosis (categorical)             <35  8837  0.29 (0.18–0.48)  <0.0001  1.24 (1.18–1.31)  <0.0001   35–49  90 958  0.59 (0.50–0.69)  <0.0001  0.99 (0.96–1.02)  0.5131   50–54  49 435  1.00 (ref.)    1.00 (ref.)     55–64  94 974  1.89 (1.66–2.15)  <0.0001  1.05 (1.01–1.08)  0.0076   65–74  66 184  5.98 (5.28–6.76)  <0.0001  1.20 (1.15–1.24)  <0.0001   ≥75  37 088  22.78 (20.20–25.70)  <0.0001  1.78 (1.71–1.86)  <0.0001  Year of diagnosis             2000–2005  165 135  1.00 (ref.)    1.00 (ref.)     2006–2011  182 341  0.83 (0.78–0.88)  <0.0001  0.82 (0.81–0.84)  <0.0001  Stage             I  156 504  1.00 (ref.)    1.00 (ref.)     II  128 560  1.26 (1.19–1.33)  <0.0001  3.01 (2.91–3.12)  <0.0001   III  55 096  1.82 (1.70–1.95)  <0.0001  9.45 (9.13–9.79)  <0.0001   IV  7316  2.90 (2.43–3.46)  <0.0001  34.52 (33.07–36.04)  <0.0001  Gradingb             I  65 133  1.00 (ref.)    1.00 (ref.)     II  135 484  1.02 (0.96–1.08)  0.5608  2.06 (1.95–2.16)  <0.0001   III  125 255  1.14 (1.07–1.22)  0.0001  3.88 (3.69–4.07)  <0.0001   IV  4178  1.16 (0.94–1.43)  0.1729  3.96 (3.66–4.29)  <0.0001  Ethnicity             White  283 319  1.00 (ref.)    1.00 (ref.)     Black  35 577  1.85 (1.72–1.99)  <0.0001  1.49 (1.45–1.53)  <0.0001   American Indian/Alaska Native  1818  0.86 (0.54–1.37)  0.5243  1.14 (1.00–1.29)  0.0451   Asian/Pacific Islander  25 620  0.74 (0.65–0.84)  <.0001  0.85 (0.82–0.89)  <0.0001  Strata/model  N  Cause-specific hazards ratiosa   Heart  P-value  Breast cancer  P-value  Age at diagnosis (categorical)             <35  8837  0.29 (0.18–0.48)  <0.0001  1.24 (1.18–1.31)  <0.0001   35–49  90 958  0.59 (0.50–0.69)  <0.0001  0.99 (0.96–1.02)  0.5131   50–54  49 435  1.00 (ref.)    1.00 (ref.)     55–64  94 974  1.89 (1.66–2.15)  <0.0001  1.05 (1.01–1.08)  0.0076   65–74  66 184  5.98 (5.28–6.76)  <0.0001  1.20 (1.15–1.24)  <0.0001   ≥75  37 088  22.78 (20.20–25.70)  <0.0001  1.78 (1.71–1.86)  <0.0001  Year of diagnosis             2000–2005  165 135  1.00 (ref.)    1.00 (ref.)     2006–2011  182 341  0.83 (0.78–0.88)  <0.0001  0.82 (0.81–0.84)  <0.0001  Stage             I  156 504  1.00 (ref.)    1.00 (ref.)     II  128 560  1.26 (1.19–1.33)  <0.0001  3.01 (2.91–3.12)  <0.0001   III  55 096  1.82 (1.70–1.95)  <0.0001  9.45 (9.13–9.79)  <0.0001   IV  7316  2.90 (2.43–3.46)  <0.0001  34.52 (33.07–36.04)  <0.0001  Gradingb             I  65 133  1.00 (ref.)    1.00 (ref.)     II  135 484  1.02 (0.96–1.08)  0.5608  2.06 (1.95–2.16)  <0.0001   III  125 255  1.14 (1.07–1.22)  0.0001  3.88 (3.69–4.07)  <0.0001   IV  4178  1.16 (0.94–1.43)  0.1729  3.96 (3.66–4.29)  <0.0001  Ethnicity             White  283 319  1.00 (ref.)    1.00 (ref.)     Black  35 577  1.85 (1.72–1.99)  <0.0001  1.49 (1.45–1.53)  <0.0001   American Indian/Alaska Native  1818  0.86 (0.54–1.37)  0.5243  1.14 (1.00–1.29)  0.0451   Asian/Pacific Islander  25 620  0.74 (0.65–0.84)  <.0001  0.85 (0.82–0.89)  <0.0001  * For 6294 patients the radiotherapy status was missing/unknown. a Adjusted for age at diagnosis (categorical), period of diagnosis, stage, grading, race, and topography (ICD-10). b A total of 17 426 patients with missing information on grading were excluded from analysis. Table 3 Cause-specific hazard and 95% confidence intervals for heart-related and breast-cancer related mortality among breast cancer patients who received either chemotherapy or radiotherapy or both* Strata/model  N  Cause-specific hazards ratiosa   Heart  P-value  Breast cancer  P-value  Age at diagnosis (categorical)             <35  8837  0.29 (0.18–0.48)  <0.0001  1.24 (1.18–1.31)  <0.0001   35–49  90 958  0.59 (0.50–0.69)  <0.0001  0.99 (0.96–1.02)  0.5131   50–54  49 435  1.00 (ref.)    1.00 (ref.)     55–64  94 974  1.89 (1.66–2.15)  <0.0001  1.05 (1.01–1.08)  0.0076   65–74  66 184  5.98 (5.28–6.76)  <0.0001  1.20 (1.15–1.24)  <0.0001   ≥75  37 088  22.78 (20.20–25.70)  <0.0001  1.78 (1.71–1.86)  <0.0001  Year of diagnosis             2000–2005  165 135  1.00 (ref.)    1.00 (ref.)     2006–2011  182 341  0.83 (0.78–0.88)  <0.0001  0.82 (0.81–0.84)  <0.0001  Stage             I  156 504  1.00 (ref.)    1.00 (ref.)     II  128 560  1.26 (1.19–1.33)  <0.0001  3.01 (2.91–3.12)  <0.0001   III  55 096  1.82 (1.70–1.95)  <0.0001  9.45 (9.13–9.79)  <0.0001   IV  7316  2.90 (2.43–3.46)  <0.0001  34.52 (33.07–36.04)  <0.0001  Gradingb             I  65 133  1.00 (ref.)    1.00 (ref.)     II  135 484  1.02 (0.96–1.08)  0.5608  2.06 (1.95–2.16)  <0.0001   III  125 255  1.14 (1.07–1.22)  0.0001  3.88 (3.69–4.07)  <0.0001   IV  4178  1.16 (0.94–1.43)  0.1729  3.96 (3.66–4.29)  <0.0001  Ethnicity             White  283 319  1.00 (ref.)    1.00 (ref.)     Black  35 577  1.85 (1.72–1.99)  <0.0001  1.49 (1.45–1.53)  <0.0001   American Indian/Alaska Native  1818  0.86 (0.54–1.37)  0.5243  1.14 (1.00–1.29)  0.0451   Asian/Pacific Islander  25 620  0.74 (0.65–0.84)  <.0001  0.85 (0.82–0.89)  <0.0001  Strata/model  N  Cause-specific hazards ratiosa   Heart  P-value  Breast cancer  P-value  Age at diagnosis (categorical)             <35  8837  0.29 (0.18–0.48)  <0.0001  1.24 (1.18–1.31)  <0.0001   35–49  90 958  0.59 (0.50–0.69)  <0.0001  0.99 (0.96–1.02)  0.5131   50–54  49 435  1.00 (ref.)    1.00 (ref.)     55–64  94 974  1.89 (1.66–2.15)  <0.0001  1.05 (1.01–1.08)  0.0076   65–74  66 184  5.98 (5.28–6.76)  <0.0001  1.20 (1.15–1.24)  <0.0001   ≥75  37 088  22.78 (20.20–25.70)  <0.0001  1.78 (1.71–1.86)  <0.0001  Year of diagnosis             2000–2005  165 135  1.00 (ref.)    1.00 (ref.)     2006–2011  182 341  0.83 (0.78–0.88)  <0.0001  0.82 (0.81–0.84)  <0.0001  Stage             I  156 504  1.00 (ref.)    1.00 (ref.)     II  128 560  1.26 (1.19–1.33)  <0.0001  3.01 (2.91–3.12)  <0.0001   III  55 096  1.82 (1.70–1.95)  <0.0001  9.45 (9.13–9.79)  <0.0001   IV  7316  2.90 (2.43–3.46)  <0.0001  34.52 (33.07–36.04)  <0.0001  Gradingb             I  65 133  1.00 (ref.)    1.00 (ref.)     II  135 484  1.02 (0.96–1.08)  0.5608  2.06 (1.95–2.16)  <0.0001   III  125 255  1.14 (1.07–1.22)  0.0001  3.88 (3.69–4.07)  <0.0001   IV  4178  1.16 (0.94–1.43)  0.1729  3.96 (3.66–4.29)  <0.0001  Ethnicity             White  283 319  1.00 (ref.)    1.00 (ref.)     Black  35 577  1.85 (1.72–1.99)  <0.0001  1.49 (1.45–1.53)  <0.0001   American Indian/Alaska Native  1818  0.86 (0.54–1.37)  0.5243  1.14 (1.00–1.29)  0.0451   Asian/Pacific Islander  25 620  0.74 (0.65–0.84)  <.0001  0.85 (0.82–0.89)  <0.0001  * For 6294 patients the radiotherapy status was missing/unknown. a Adjusted for age at diagnosis (categorical), period of diagnosis, stage, grading, race, and topography (ICD-10). b A total of 17 426 patients with missing information on grading were excluded from analysis. Heart-specific mortality most strongly increased with age, with hazard ratios (HRs) (95% CIs) as high as 1.89 (1.66–2.15), 5.98 (5.28–6.76), and 22.78 (20.20–25.70) for patients aged 55–64, 65–74, and ≥75 relative to patients aged 50–54, respectively. The association with age was much stronger for heart-related mortality than for breast cancer related mortality. Moreover, women who were diagnosed between 2006 and 2011 tended to have a 17% decreased heart-specific mortality [HRcs 0.83 (0.78–0.88)] relative to those diagnosed between 2000 and 2005. Heart-specific mortality also increased with stage, even though the association was less pronounced than for breast cancer mortality [HRcs 2.90 (2.43–3.46)] for Stage IV relative to Stage I). Compared with white ethnicities, black breast cancer patients were at higher risk for heart-specific death [HRcs 1.85 (1.72–1.99)], whereas patients with an Asian/Pacific island ethnicity were at lower risk [HRcs 0.74 (0.65–0.84)]. None of the results did significantly change when the multiple imputed datasets were used (see Supplementary material online, Table S3). When comparing intermediate-term heart-specific mortality according to their HER2/neu status in subgroup analysis, it was observed that being diagnosed with a HER2-positive receptor status was not associated with heart-specific mortality [HRcs 0.96 (0.70–1.32)] relative to HER2-negative receptor status (see Supplementary material online, Table S4). Discussion In this large registry-based cohort study, it was found that within a cohort of female breast cancer patients receiving radiotherapy or chemotherapy, heart-specific mortality was an important competing risk, which was observed to steadily increase by age. However, when comparing long-term heart-specific mortality to the general population, mortality was not increased. Moreover, relative to patients with a HER2-negative subtype, also HER2-positive patients did not show increased mortality. To our knowledge, this is the largest registry-based study, which has been conducted so far on long-term heart-specific mortality among breast cancer patients treated with radiotherapy or chemotherapy. In line with previously published literature, also within this large US-based cohort an improvement of breast cancer-specific survival on the population level over time was observed, leaving potentially treatment-induced heart-specific diseases as a major competing risk for mortality among breast cancer survivors.15,16 Potentially relevant systemic treatments include anthracyclines, taxanes, alkylating agents, tyrosine-kinase inhibitors, and several new targeted therapeutic options, which are known to hold cardiotoxic profiles.17 The underlying mechanisms have not been well understood yet, but are hypothesized to be caused by impaired stress-protective signalling mechanisms in cardiomyocytes and damages of myocardial cells due to reactive oxygen species.18,19 Also radiotherapy is known as a major risk factor for treatment-induced heart-specific diseases, particularly if applied to patients affected by left-sided breast cancer.19 Especially the combination of radiation-induced macrosvascular damages, accelerating age-related atherosclerosis, and microvascular damages, reducing capillary density, is assumed to lead dose-dependently to long-term myocardial ischaemias in breast cancer survivors.20 Nevertheless, modern imaging and radiation techniques might be promising developments that are likely to reduce radiation-induced heart-diseases in the future.19 To investigate potential excess heart-specific mortality, which might be attributable to cardiotoxic treatment, it is crucial to also take into account the steadily increasing cumulative heart-specific mortality by age in the general population. In this study, the cSMRs showed either a slightly lower or comparable heart-specific mortality for breast cancer patients who received radiotherapy or chemotherapy relative to the standard US female population. However, these results have to be carefully interpreted as patients who received radiotherapy or chemotherapy significantly differed from patients who did not receive these treatments in terms of their baseline characteristics such as younger age at diagnosis, higher stage, higher grading, and molecular subtype status. This is consistent with existing literature, as it was found that breast cancer patients who receive chemotherapy have in general less comorbid conditions, rather no cardiovascular diseases and are younger at baseline.21 Hence, we refrained from performing additional analyses on patients who did not receive radiotherapy or chemotherapy as Table 1 showed that these were on average much older and most probably more frail, which makes the interpretation of potential analyses much more challenging. Moreover, additional analyses on this patient subgroup would have been beyond the scope of this study. In subgroup analyses, it was found that especially Stage III breast cancer patients might have a high cardiotoxic burden as they are most likely to receive both radiotherapy and chemotherapy (55.2%) (see Supplementary material online, Table S5). Hence, the interpretation of the presented cSMRs is not straightforward but needs to take into account that patients at higher stages, who receive potentially cardiotoxic treatment, are also more likely to die earlier due to their underlying breast cancer disease before they might develop a heart-specific disease in the long term. Furthermore, clinicians might already have established risk stratification strategies and pre-treatment screenings in routine care, which might exclude patients at highest cardiovascular risk from potentially cardiotoxic treatment. It might also be likely that breast cancer patients with radiotherapy or chemotherapy treatment might have still developed different adverse heart-specific diseases but did not necessarily die from these diseases. While dose-dependent anthracycline administration might cause irreversible cardiomyopathy, trastuzumab-induced cardiomyopathy appears to be reversible after the end of treatment.17 Considering that targeted therapies for HER2-positive patients were also found to induce cardiotoxic effects, our results on intermediate-term survival according to molecular subtypes are encouraging as they show a decreased breast cancer-specific mortality while heart-specific mortality did not differ compared with HER2-negative patients, which was also suggested by another recently published study.22 However, not every HER2-positive diagnosed patient might also receive a HER2-targeted therapy and potential high-risk patients could be less likely to receive HER2-targeted therapy in the first place. Furthermore, follow-up time for this restricted subgroup was shorter than for the general cohort. The results of our study are quite consistent to results from other studies. For example, in a recently published Canadian study, comparing breast cancer patients and non-cancer population controls, cardiovascular death was not significantly associated among breast cancer patients who received anthracyclines only [HR (95% CI) 0.94 (0.41–2.20)], trastuzumab only [1.14 (0.85–1.52)], or a sequential therapy [0.81 (0.25–2.66)].23 In another study, comparing deaths due to cardiovascular diseases in a sample of 1413 breast cancer patients who were diagnosed in 1996–1997 and age-matched to women without breast cancer,5 it was found that an increase in cardovascular deaths among breast cancer patients was only evident 7 years after diagnosis [HRcs 1.8 (1.3–2.5)] and mainly present among women who received chemotherapy [HRcs 1.7 (1.1–2.6)]. A comparable study to ours from the Netherlands, which investigated in great detail the association of different treatment approaches and cardiovascular causes of death, reported that especially 5-year breast cancer survivors who were diagnosed in Stage I–III and had left-sided radiotherapy after mastectomy or chemotherapy after 1997 were at increased risk to die from ischaemic heart diseases and congestive heart failure when accounting for other competing causes of death.24 When comparing this cohort to the general Dutch population, the overall results for all cardiovascular causes of death [SMR 0.92 (0.88–0.97)] and cardiovascular disease subgroups tended to be rather below the null with the exception of valvular dysfunction [SMR 1.28 (1.08–1.52)]. These analyses are quite conclusive to our findings from the SMR analysis. This study has some strengths and limitations. Firstly, the SEER database contains information about chemotherapy and radiation therapy with a high specificity. However, as cancer treatments are increasingly administered outside the hospital there might be the possibility that the cancer registries might have missed cases of patients having received treatment and hence coded them as ‘no/unknown’. Nevertheless, it was found that the positive predictive value of the treatment information captured in SEER is reasonably high (>85%), reflecting sufficient specificity of treatment information for our analyses as especially just patients with radiotherapy or chemotherapy treatment were examined.25 Furthermore, details of the type, dosage and duration of administered chemotherapy or radiotherapy were not available. Also adjuvant endocrine therapy in postmenopausal women would have been of interest, since it was found that compared with tamoxifen, aromatase inhibitors are associated with a higher risk of myocardial infarction.26 Nevertheless, it should be noted that despite a decreasing trend of anthracycline use, studies that investigated the patterns of chemotherapy use among breast cancer patients, found that 88.6% (SEER/Texas Cancer Registry-Medicare-linked database) and 92.8% (Kaiser Permanente Northern California electronic medical records database) of all patients who received chemotherapy received anthracyclines and 88.8% of HER2-positive chemotherapy-treated patients received trastuzumab, respectively.21,27 Hence, use of chemotherapy, which was used as exposure variable in our analysis, should approximate the relevant exposures reasonably well, which is supported by the concordance to other smaller studies, that found an increase in heart-specific events (e.g. hospitalization or major cardiac events) but not in heart-specific mortality.23,24 Secondly, cSMRs are not adjusted for further confounders as in multivariable regression analyses and also do not address the time-to-event temporality as in Cox proportional hazards regressions. Hence, risk factors other than sex, age and stage might still have still confounded these results and future multivariable analyses comparing breast cancer patients and non-breast cancer cohorts are highly desirable. Moreover, mortality data on the general population is just available on aggregated level in the WONDER database and hence does not provide the option to exclude women with a diagnosis of breast cancer. However, due to the good breast cancer prognosis cSMRs might not substantially change. Thirdly, the SEER coverage population was found to have a lower socio-economic status and a greater diversity in the distribution of ethnic minorities and hence might be not representative for the whole United States (US) population.28 Lastly, due to the very long follow-up time, HRcs derived from regression analysis should be interpreted as average effects over time. In a sensitivity analysis to assess time-varying confounding, calculating HRcs for each quartile of follow-up time, associations mostly became stronger with longer follow-up time, but still showed the same directions and patterns (data not shown). Nevertheless, unique strengths of this study are the large sample size, the stratified analyses by molecular receptor subtype and the long follow-up time. Randomized clinical trails are often limited in sample size and follow-up time, which leads to challenges in examining long-term side effects on mortality and hence surrogates such as LVEF are used which might, due to the above mentioned reasons, not necessarily reflect long-term heart-specific mortality outcomes among breast cancer patients. Conclusion In conclusion, this study identified some potential prognostic factors for heart-specific and breast cancer-specific mortality, some of which were quite consistent with previously published literature. More importantly, however, it was observed that the heart-specific mortality among breast cancer patients treated with radiotherapy or chemotherapy is not increased and rather comparable with the general female population. Further studies with more detailed data on treatment are strongly desirable and needed to both validate the results presented in this study and further examine potential prognostic factors. This might add knowledge for further development of cardio-oncological risk stratification among breast cancer patients and help clinicians to make informed decisions about which patient strata are at highest risk and who would benefit most from preventive actions. Supplementary material Supplementary material is available at European Heart Journal online. Acknowledgements The authors would like to thank Silvia Calderazzo from the Division of Biostatistics, German Cancer Research Center, Heidelberg, Germany for her statistical support. Conflict of interest: none declared. References 1 Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer  2015; 136: E359– E386. Google Scholar CrossRef Search ADS PubMed  2 Zamorano JL, Lancellotti P, Rodriguez Munoz D, Aboyans V, Asteggiano R, Galderisi M, Habib G, Lenihan DJ, Lip GY, Lyon AR, Lopez Fernandez T, Mohty D, Piepoli MF, Tamargo J, Torbicki A, Suter TM. 2016 ESC Position Paper on cancer treatments and cardiovascular toxicity developed under the auspices of the ESC Committee for Practice Guidelines: the Task Force for cancer treatments and cardiovascular toxicity of the European Society of Cardiology (ESC). Eur Heart J  2016; 37: 2768– 2801. Google Scholar CrossRef Search ADS PubMed  3 Abdel-Qadir H, Austin PC, Lee DS, Amir E, Tu JV, Thavendiranathan P, Fung K, Anderson GM. A population-based study of cardiovascular mortality following early-stage breast cancer. JAMA Cardiol  2017; 2: 88– 93. Google Scholar CrossRef Search ADS PubMed  4 Patnaik JL, Byers T, DiGuiseppi C, Dabelea D, Denberg TD. Cardiovascular disease competes with breast cancer as the leading cause of death for older females diagnosed with breast cancer: a retrospective cohort study. Breast Cancer Res  2011; 13: R64. Google Scholar CrossRef Search ADS PubMed  5 Bradshaw PT, Stevens J, Khankari N, Teitelbaum SL, Neugut AI, Gammon MD. Cardiovascular disease mortality among breast cancer survivors. Epidemiology  2016; 27: 6– 13. Google Scholar CrossRef Search ADS PubMed  6 Austin PC, Lee DS, Fine JP. Introduction to the analysis of survival data in the presence of competing risks. Circulation  2016; 133: 601– 609. Google Scholar CrossRef Search ADS PubMed  7 Surveillance, Epidemiology, and End Results (SEER) Program Populations (1969–2014). National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch. http://www.seer.cancer.gov/popdata (4 May 2017). 8 Centers for Disease Control and Prevention. CDC Wonder. http://wonder.cdc.gov/ (18 May 2017). 9 Rubin DB. Multiple Imputation for Nonresponse in Surveys . New York, Chichester, Brisbane, Toronto, Singapore: John Wiley & Sons; 2004. 10 Altman DG, De Stavola BL, Love SB, Stepniewska KA. Review of survival analyses published in cancer journals. Br J Cancer  1995; 72: 511– 518. Google Scholar CrossRef Search ADS PubMed  11 Vandenbroucke JP. A shortcut method for the calculation of the 95 per cent confidence interval of the standarized mortality ratio. Am J Epidemiol  1982; 115: 303– 304. Google Scholar CrossRef Search ADS   12 Altman DG, Bland JM. How to obtain the P value from a confidence interval. BMJ  2011; 343: d2304. Google Scholar CrossRef Search ADS PubMed  13 Howlader N, Altekruse SF, Li CI, Chen VW, Clarke CA, Ries LAG, Cronin KA. US incidence of breast cancer subtypes defined by joint hormone receptor and HER2 status. J Natl Cancer Inst  2014; 106: dju055. Google Scholar CrossRef Search ADS PubMed  14 Yoshida K, Bohn J. R Package ‘tableone’. R Foundation for Statistical Computing, Vienna, Austria (30 November 2016). 15 Plana JC, Barac A. Good news, bad news, but not fake news. Circulation  2017; 135: 1413– 1416. Google Scholar CrossRef Search ADS PubMed  16 Swain SM, Baselga J, Kim S-B, Ro J, Semiglazov V, Campone M, Ciruelos E, Ferrero J-M, Schneeweiss A, Heeson S, Clark E, Ross G, Benyunes MC, Cortés J. Pertuzumab, trastuzumab, and docetaxel in HER2-positive metastatic breast cancer. New Engl J Med  2015; 372: 724– 734. Google Scholar CrossRef Search ADS PubMed  17 Schlitt A, Jordan K, Vordermark D, Schwamborn J, Langer T, Thomssen C. Cardiotoxicity and oncological treatments. Dtsch Arztebl Int  2014; 111: 161– 168. Google Scholar PubMed  18 Wysocki PJ, Wysocki H. Cardiovascular complications associated with biological therapies for breast cancer. Expert Opin Biol Ther  2008; 8: 1551– 1559. Google Scholar CrossRef Search ADS PubMed  19 Meattini I, Guenzi M, Fozza A, Vidali C, Rovea P, Meacci F, Livi L. Overview on cardiac, pulmonary and cutaneous toxicity in patients treated with adjuvant radiotherapy for breast cancer. Breast Cancer  2017; 24: 52– 62. Google Scholar CrossRef Search ADS PubMed  20 Darby SC, Cutter DJ, Boerma M, Constine LS, Fajardo LF, Kodama K, Mabuchi K, Marks LB, Mettler FA, Pierce LJ, Trott KR, Yeh ET, Shore RE. Radiation-related heart disease: current knowledge and future prospects. Int J Radiat Oncol Biol Phys  2010; 76: 656– 665. Google Scholar CrossRef Search ADS PubMed  21 Kurian AW, Lichtensztajn DY, Keegan THM, Leung RW, Shema SJ, Hershman DL, Kushi LH, Habel LA, Kolevska T, Caan BJ, Gomez SL. Patterns and predictors of breast cancer chemotherapy use in Kaiser Permanente Northern California, 2004-2007. Breast Cancer Res Treat  2013; 137: 247– 260. Google Scholar CrossRef Search ADS PubMed  22 Dall P, Koch T, Gohler T, Selbach J, Ammon A, Eggert J, Gazawi N, Rezek D, Wischnik A, Hielscher C, Keitel S, Cirrincione U, Hinke A, Feisel-Schwickardi G. Trastuzumab in human epidermal growth factor receptor 2-positive early breast cancer: results of a prospective, noninterventional study on routine treatment between 2006 and 2012 in Germany. Oncologist  2017; 22: 131– 138. Google Scholar CrossRef Search ADS PubMed  23 Thavendiranathan P, Abdel-Qadir H, Fischer HD, Camacho X, Amir E, Austin PC, Lee DS. Breast cancer therapy-related cardiac dysfunction in adult women Treated in routine clinical practice: a population-based cohort study. J Clin Oncol  2016; 34: 2239– 2246. Google Scholar CrossRef Search ADS PubMed  24 Boekel NB, Schaapveld M, Gietema JA, Russell NS, Poortmans P, Theuws JC, Schinagl DA, Rietveld DH, Versteegh MI, Visser O, Rutgers EJ, Aleman BM, van Leeuwen FE. Cardiovascular disease risk in a large, population-based cohort of breast cancer survivors. Int J Radiat Oncol Biol Phys  2016; 94: 1061– 1072. Google Scholar CrossRef Search ADS PubMed  25 Noone AM, Lund JL, Mariotto A, Cronin K, McNeel T, Deapen D, Warren JL. Comparison of SEER treatment data with medicare claims. Med Care  2016; 54: e55– e64. Google Scholar CrossRef Search ADS PubMed  26 Abdel-Qadir H, Amir E, Fischer HD, Fu L, Austin PC, Harvey PJ, Rochon PA, Lee DS, Anderson GM. The risk of myocardial infarction with aromatase inhibitors relative to tamoxifen in post-menopausal women with early stage breast cancer. Eur J Cancer  2016; 68: 11– 21. Google Scholar CrossRef Search ADS PubMed  27 Barcenas CH, Niu J, Zhang N, Zhang Y, Buchholz TA, Elting LS, Hortobagyi GN, Smith BD, Giordano SH. Risk of hospitalization according to chemotherapy regimen in early-stage breast cancer. J Clin Oncol  2014; 32: 2010– 2017. Google Scholar CrossRef Search ADS PubMed  28 Kuo T-M, Mobley LR. How generalizable are the SEER registries to the cancer populations of the USA? Cancer Causes Control  2016; 27: 1117– 1126. Google Scholar CrossRef Search ADS PubMed  Published on behalf of the European Society of Cardiology. All rights reserved. © The Author(s) 2018. For permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

European Heart JournalOxford University Press

Published: Apr 9, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off