Breast Cancer Mortality in Older and Younger Patients in California

Breast cancer is the most common cancer among women worldwide, accounting for a fifth of overall cancer mortality (1). In the United States, less than 20% of all breast cancer cases occur before the age of 50 years (2). Results of some studies have shown that younger compared with older breast cancer patients have poorer survival, with studies focusing on age groups less than 40 years (3–5). For breast cancer mortality endpoints, two studies based on the Surveillance, Epidemiology, and End Results (SEER) Program reported that younger compared with older breast cancer patients had higher breast cancer mortality, with HRs of 1.095 [95% confidence interval (CI), 1.101–1.183] comparing patients less than 35 with those 50 to 55 years of age in one study (6) and 1.39 (CI, 1.34 to 1.45) in the second study comparing women <40 with those 40+ years of age (7). Variations in estimates of risk could be due to the inconsistent use of referent and comparison age groups and to covariates included in multivariate models. Proposed reasons for higher mortality in younger versus older patients include later stage disease, more aggressive tumors, and less favorable tumor receptor status in younger than older patients (5, 6). However, biological, undertreatment, and socioeconomic status (SES) factors may potentially result in higher mortality among older compared with younger patients (8–11).

To our knowledge, there are no published reports regarding differences in breast cancer mortality for the age cut off of 50 years, a marker for menopausal status and for recommended initiation of screening mammography (12, 13). Furthermore, although breast cancer survival has been shown to vary according to tumor subtype (14), comparison of prognostic factors between younger and older patients by tumor subtype is poorly understood, especially in population-based settings.

Using data from the population-based California Cancer Registry (CCR), our study takes advantage of the completeness of tumor subtype information in the registry in the mid-2000s to assess breast cancer mortality differences between breast cancer patients who were younger (age 18–49) and older (age 50 and above) at diagnosis. We further assessed the moderating effects of tumor biology and clinical factors by examining whether mortality differences vary by tumor subtype, receiving guideline-appropriate care, and receiving care at an NCI-Designated Cancer Center (NCICC).

We obtained from CCR information about all female California residents age 18 years and older at diagnosis who were diagnosed with a first, primary invasive breast cancer [International Classification of Disease for Oncology, 3rd Edition, (ICD-O-3) site codes C50.0–50.9] during January 1, 2005, through December 31, 2014 (n = 196,628). As the criteria for guideline treatment were limited to patients diagnosed with American Joint Committee on Cancer (AJCC) stage I to III breast cancer, our analysis did not include patients with stage unknown or stage IV breast cancer (n = 19,842). Patients were additionally excluded from analysis hierarchically as follows: diagnosis by death certificate or autopsy only (n = 19) or diagnosis not microscopically confirmed (n = 187); ICD-O-3 histologic type other than 8000, 8001, 8010, 8020, 8022, 8050, 8140, 8201, 8211, 8230, 8255, 8260, 8401, 8453, 8480, 8481, 8500–8525, or 8575 (n = 2,217); tumor size missing because unknown (n = 667), no tumor noted (n = 236), microscopic (n = 2,009), diffuse (n = 280), or mammographic diagnosis only (n = 54); young patient insured by Medicare (n = 403); no follow-up (n = 210); second primary breast tumor diagnosed within 60 days of initial tumor (n = 5,068); bilateral tumors at initial diagnosis (n = 7); residential address that was uncertain or not geocodable (n = 5,347). Analyses thereby included 160,082 patients, of which 38,509 were younger (age 18–49) and 121,573 were older (age 50 and above, up to age 103) patients.

We obtained information from the CCR, which is derived from the patient's medical record, on age at diagnosis, race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, Asian/Pacific Islander, or other/unknown), marital status, residential address at diagnosis, stage at diagnosis, tumor size (in cm), lymph node involvement, histology, grade (I, II, III/IV, or unknown), primary source of payment (private only, any Medicaid/military/Other public, Medicare only/Medicare + private, no insurance, and unknown), hormone receptor [estrogen receptor (ER) and progesterone receptor (PR) together referred as hormone receptor (HR), and HER2] status, as well as initial treatment modalities [surgery, radiotherapy, and chemotherapy (endocrine therapy is undercaptured in cancer registry data)]. We followed patients for vital status from linkage with vital records as of December 31, 2014.

We used a multicomponent measure of neighborhood SES (nSES), based on patients' residential census block group at diagnosis. This measure incorporated the 2000 U.S. Census (for cases diagnosed in 2005) and the 2006–2010 American Community Survey data (for cases diagnosed in 2006 and forward) on education, occupation, unemployment, household income, poverty, rent, and house values (15, 16). Each patient was assigned an nSES quintile, based on the distribution of SES across census block groups in California.

We used the breast cancer subtype definition as previously defined (17). Briefly, the CCR has collected information on the expression of ER and PR since 1990 and of HER2 since 1999 (18), with HER2 data completeness increasing greatly after 2005. We classified breast cancers into four mutually exclusive subtype categories: HR⁺/HER2^– (defined as ER and/or PR positive and HER2 negative), HR⁺/HER2⁺ (ER and/or PR positive and HER2 positive), HR^–/HER2⁺ (ER and PR negative and HER2 positive), and triple-negative breast cancer (TNBC, ER, PR, and HER2 negative; refs. 14, 18–21). Of the 160,082 cancers in this analysis, 16,373 (10.2%) did not have information needed to assign to one of these subtypes, including 11,012 cancers (6.9%) for whom only HER2 status was unknown, 631 cancers (0.4%) for whom only HR status was unknown, and 4,730 cancers (3.0%) for whom both HR and HER2 statuses were unknown.

Receipt of guideline-concordant care was based on whether women reported receiving treatment that aligned with the National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (22) and the American Society of Clinical Oncology Quality Oncology Practice Initiative (23, 24). Cancer registry first course of treatment data on receipt of surgery (lumpectomy, mastectomy, axillary, or sentinel node dissection), radiotherapy, and/or chemotherapy were obtained. Each woman was considered to be in one or more patient subsets based on her age and tumor characteristics (subtype, lymph node involvement, and tumor size). Women with any nonconcordant care were categorized as not receiving guideline treatment. As described in a prior SEER study (25), the subsets were used to define appropriate treatment options (Table 1). Women who did not fall into either subset were classified in regard to guideline concordant treatment as “Not applicable,” and those who were in one or more subset but were missing the treatment data needed to determine concordance were classified as “Unknown.”

Receiving care at an NCICC was based on diagnosis and/or treatment occurring at such centers. In a population-based setting, because patients may be seen and have received care at multiple facilities, and also due to how the CCR data on multiple reporting facilities are coded, it is not possible to determine the treating facilities.

Follow-up time was calculated as the number of days between the date of diagnosis and the earliest of: the date of death from breast cancer (ICD 9/10 = 174/C50), the date of death from another cause, the date of last follow-up (i.e., last known contact), or the study end date (December 31, 2014). The 545 deceased patients with an unknown cause of death were excluded from all models. Cox proportional hazards regression was used to estimate the breast cancer–specific mortality rate ratio (MRR) and corresponding associated 95% CIs for the two age groups with fully adjusted models adjusted for year of diagnosis (continuous), marital status, race/ethnicity, insurance status, nSES, lymph nodes involvement, tumor subtype, tumor size, tumor grade, tumor histology, receipt of guideline concordant treatment, and whether the patient was seen at one or more of the NCICC in California for her breast cancer. Fully adjusted models were additionally adjusted for clustering by block group, using a sandwich estimator of the covariance structure that accounts for intracluster dependence. The proportional hazards assumption was examined by statistical testing of the correlation between weighted Schoenfeld residuals and logarithmically transformed survival time. No violations of the assumption were observed, except for AJCC stage at diagnosis. Thus, stage was included as an underlying stratifying variable in the fully adjusted Cox regression models, allowing the baseline hazard to vary by stage. Wald Type 3 tests for interaction between age group (18–49, 50+) and tumor subtype, NCICC, or guideline concordant treatment (excluding Unknown) were computed using cross-product terms, in models adjusted for all statistically significant (P < 0.05) interactions with age group (year, insurance status, tumor subtype, NCICC, and guideline concordant treatment). All statistical tests were carried out using SAS software version 9.3 (SAS Institute).

In this population-based study in California (n = 160,082), 38,509 (24.1%) female patients under the age 50 years at diagnosis presented with stage I to III breast cancer; 121,573 (75.9%) were aged 50 and above. As shown in Table 2, compared with older patients, younger women were more likely to be Hispanic (27% vs. 16%), be married (64% vs. 54%), be covered by private insurance only (76% vs. 54%), have TNBC (13% vs. 10%), be diagnosed with tumors over 2 cm in size (49% vs. 36%), and have ductal tumors (83% vs. 78%). Younger patients were less likely to be diagnosed with stage I cancer (39% vs. 54%), grade I tumors (16% vs. 25%), and negative lymph node involvement (58% vs. 71%) than older patients. No substantial differences were shown for nSES, NCICC status, or guideline concordant treatment between younger and older patients.

Multivariable-adjusted risk of breast cancer mortality for older versus younger patients categorized by 10-year age groups shows that compared with patients ages 40 to 49 years, a progressively higher risk of mortality is shown for older age groups with the highest risk shown in women 80 years and older (MRR = 3.25; 95% CI, 2.98–3.55; Table 3). Results using an age cutoff at 50 years show that older patients had a higher risk of mortality than younger patients (MRR = 1.17; 95% CI, 1.11–1.23).

Breast cancer mortality according to tumor subtype, guideline-appropriate treatment, and receiving care at an NCICC, stratified by age group, is presented in Table 4. Significant interactions between age group and tumor subtype (P = 0.0008), NCICC (P = 0.003), and guideline-appropriate treatment (P = 0.015) were observed. Among younger women, patients with HR⁺/HER2⁺ disease had a lower risk of dying (MRR = 0.80; 95% CI, 0.69–0.93) than those with HR⁺/HER2^– tumors, whereas a higher risk was shown for patients with HR^–/HER2⁺ tumors (MRR = 1.37; 95% CI, 1.15–1.62) and those with TNBC (MRR = 2.50; 95% CI, 2.19–2.86). In older patients, higher mortality was observed for patients with HR^–/HER2⁺ (MRR = 1.28; 95% CI, 1.14–1.43) and TNBC (MRR = 2.35; 95% CI, 2.17–2.53) but not for HR⁺/HER2⁺ tumors (MRR = 1.05; 95% CI, 0.95–1.15) when compared with women with HR⁺/HER2^–. Older women who received care at an NCICC had a lower risk of dying from breast cancer than those who did not (MRR = 0.84; 95% CI, 0.76–0.93); no difference was seen in younger patients (MRR = 1.00; 95% CI, 0.89–1.13). In both younger (MRR = 1.20; 95% CI, 1.03–1.40) and older patients (MRR = 1.49; 95% CI, 1.36–1.63), a higher risk of dying was shown for women who did not receive guideline-appropriate treatment, compared with those who did.

Table 5 shows the multivariable-adjusted breast cancer MRRs for older compared with younger patients. Stratified multivariable analyses by tumor subtype showed higher mortality for older compared with younger patients who were diagnosed with HR⁺/HER2^– (MRR = 1.24; 95% CI, 1.14–1.35) and with HR⁺/HER2⁺ tumors (MRR = 1.38; 95% CI, 1.17–1.62) but not for those with HR^–/HER2⁺ (MRR = 0.94; 95% CI, 0.79–1.12) or TNBC (MRR = 1.01; 95% CI, 0.92–1.11). Older women who were not cared for at an NCICC had a higher risk of dying than younger patients (MRR = 1.21; 95% CI, 1.15–1.28), but the opposite was seen among women cared for at an NCICC (MRR = 0.86, 95% CI, 0.73–1.01). Older as compared with younger patients who did not receive guideline-appropriate treatment had a higher risk of dying (MRR = 1.20; 95% CI, 1.02–1.41) but those who had guideline-concordant treatment were not at higher risk (MRR = 1.06; 95% CI, 0.99–1.14).

In this large and representative series of women diagnosed with invasive stage I to III breast cancer in California, we found that after taking into account clinical and sociodemographic factors, older patients at diagnosis experience 17% higher breast cancer mortality than younger patients. However, variation in risk was shown according to tumor subtype, receipt of care at an NCICC, and receiving guideline-concordant treatment. The difference in breast cancer mortality between older and younger patients was evident for patients with HR⁺ tumors regardless of HER2 status, whereas no difference was observed for women with HR^– disease (HR^–/HER2⁺ and TNBC). The higher mortality among older versus younger women was diminished in patients receiving guideline treatment and reversed among those seen at an NCICC, suggesting that appropriate treatment improves survival among older women. Although age differences regarding breast cancer aggressiveness and mortality outcomes have been published (6, 7, 26, 27), to our knowledge no comprehensive reports exist on differences by age for associations between tumor subtype and clinical prognostic factors and breast cancer mortality. In these analyses, we were able to account for a number of sociodemographic and clinical factors as covariates, which provided a comprehensive assessment of age-specific differences in breast cancer mortality.

Presence of aggressive breast tumor subtypes was higher in younger than older women (19.2% HER2⁺ in younger vs. 13.1% in older patients, and 13.4% TNBC in younger vs. 9.5% in older patients), which is consistent with findings from other studies (28–30). TNBC has been difficult to study before 2005, especially in population settings, since routine HER2 testing for breast cancers was not implemented at large until after Trastuzumab was approved for the adjuvant treatment of early-stage breast cancer in 2005. Comprising less than 20% of breast cancers, TNBC is associated with worse survival than other subtypes, in part due to the lack of targeted therapeutic agents (30, 31). Our study shows that patients diagnosed with TNBC have a greater than 2-fold increased risk of dying compared with those with HR⁺/HER2^– breast cancer regardless of age group, underscoring the aggressive nature of TNBC subtype.

Stratified analyses by age group showed that among younger patients, patients with HR⁺/HER2⁺ tumors had lower risk of dying as compared with those with HR⁺/HER2^– tumors; older women with HR⁺/HER2⁺ tumors had a mortality rate similar to older women with HR⁺/HER2^– tumors. These findings imply a greater benefit of HER2-targeted treatment (32, 33) on survival in the younger population, who are more likely to be HER2-positive and receive targeted treatment (34). In fact, the most pronounced difference in mortality by age was shown for patients with HR⁺/HER2⁺ breast cancer, where older patients had approximately 40% increased risk of breast cancer death relative to younger patients. It is possible that HR⁺/HER2⁺ older patients are more likely to forego chemotherapy given the emerging, but understudied, use of dual antiestrogen/anti-HER2 therapy. Conversely, a higher risk of mortality regardless of age group was found for patients with HR^–/HER2⁺ tumors compared with HR⁺/HER2^– subtype. The results suggest that older women might be sacrificing some potential gain in breast cancer survival to take into account factors such as treatment-related toxicity, functional status, and other quality of life measures. Due to the population-based registry nature of our study, we are unable to assess to what degree these types of trade-offs are being made by the patient or provider.

It has been reported in the literature that older patients with breast cancer receive less guideline-appropriate treatment than their younger counterparts (8, 35, 36). Therefore, our finding of a higher mortality in older than younger patients in women who do not receive guideline-appropriate treatment but not in those who receive guideline-concordant treatment is noteworthy. Our results also show that the higher mortality associated with older compared with younger patients is present among women who were not ever seen at an NCICC and not in patients who received care at an NCICC. As improved breast cancer treatment guideline concordance and surgical outcomes at an NCICC were reported previously (37–39), our findings imply that for older breast cancer patients, which represent the vast majority of the patient population (∼80%), receiving care at an NCICC and ensuring that guideline-appropriate treatment is provided will decrease breast cancer mortality in this older age group. Although we could not completely characterize these effects nor do our data allow us to definitively attribute treatment to specific facilities, the better survival outcome for older patients may be due to improved multidisciplinary care coordination, in addition to access to tumor boards, patient-centered care programs, and clinical trials for special geriatric cancer care that may be more achievable in NCICC than in other types of facilities (40, 41). With limited evidence from clinical trials and research studies on older patients due to their comorbid conditions or belief from providers that older patients are incapable of tolerating treatment or have limited long-term benefit, it is difficult to formulate evidence-based treatment and guideline-compliance recommendations.

Our study used CCR data from the most recent decade to examine variation in breast cancer survival in the younger and older groups. Few previous studies have looked concurrently at the age cohorts or have included in the analysis tumor subtypes and receipt of guideline treatment. As the ER, PR, and HER2 designations are becoming increasingly useful in guiding clinical treatment and in breast cancer research (14, 42), however, our conclusions need further validation, as subtypes determined by receptor status serve only as a proxy for full genetic profiling. Also, our survival analyses were adjusted for sociodemographic, clinical characteristics, and first course of cancer-directed treatment, which are available in the cancer registry. However, our study is limited by the lack of data on genetic profile, unmeasured treatment information such as dosing or specific regimens, as well as comorbidities. Consequently, our findings could be subject to residual confounding from incomplete treatment and comorbidity data in the cancer registry (43), which may be especially relevant when comparing older and younger patients. We encourage further population-based studies with more detailed treatment and clinical data and individual-level measures of socioeconomic factors to explore the mechanisms associated with age-specific mortality differences.

In summary, our results based on multivariable-adjusted models show that women age 50 years and older at diagnosis with stage I to III breast cancer have a higher risk of dying from breast cancer compared with younger women, but variation in risk by age exists according to tumor subtype. In addition, the higher mortality rate among older relative to younger women was diminished in women who received guideline-concordant treatment and reversed for patients receiving care at an NCICC, suggesting that ensuring receipt of appropriate treatment and patient-centered care provided in NCICCs may help to reduce age-related differences in breast cancer mortality.

No potential conflicts of interest were disclosed.

The ideas and opinions expressed herein are those of the author(s) and do not necessarily reflect the opinions of the State of California, Department of Public Health, the NCI, and the Centers for Disease Control and Prevention (CDC) or their Contractors and Subcontractors.

Conception and design: S.L. Gomez, M. Gago-Dominguez, M.E. Martinez

Development of methodology: M.E. Martinez

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.L. Gomez

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R.B. Schwab, A.J. Canchola, J.D. Murphy, A.A. Molinolo, M.E. Martinez

Writing, review, and/or revision of the manuscript: R.B. Schwab, Y.S. Miguel, S.L. Gomez, A.J. Canchola, I.K. Komenaka, J.D. Murphy, A.A. Molinolo, M.E. Martinez

Study supervision: S.L. Gomez, M.E. Martinez

We would like to thank Valesca Largaespada for her contribution in preparation of the article.

This work was supported by the Specialized Cancer Center Support Grant to the University of California San Diego Moores Cancer Center (CA023100-29 to R. Schwab, M. Gago-Dominguez, J. Murphy, A. Molinolo, and M.E. Martinez) and by the SDSU/UCSD Comprehensive Cancer Center Partnership (CA132379 and CA132384 to Y. San Miguel and M.E. Martinez). The collection of cancer incidence data used in this study was supported by the California Department of Public Health pursuant to California Health and Safety Code Section 103885; CDC's National Program of Cancer Registries, under cooperative agreement 5NU58DP006344; the NCI's SEER Program under contract HHSN261201800032I awarded to the University of California, San Francisco, contract HHSN261201800015I awarded to the University of Southern California, and contract HHSN261201800009I awarded to the Public Health Institute, Cancer Registry of Greater California.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.