Risk of Cardiometabolic Risk Factors in Women With and Without a History of Breast Cancer: The Pathways Heart Study

PURPOSE

The incidence of cardiometabolic risk factors in breast cancer (BC) survivors has not been well described. Thus, we compared risk of hypertension, diabetes, and dyslipidemia in women with and without BC.

METHODS

Women with invasive BC diagnosed from 2005 to 2013 at Kaiser Permanente Northern California (KPNC) were identified and matched 1:5 to noncancer controls on birth year, race, and ethnicity. Cumulative incidence rates of hypertension, diabetes, and dyslipidemia were estimated with competing risk of overall death. Subdistribution hazard ratios (sHRs) were estimated by Fine and Gray regression, adjusted for cardiovascular disease–related risk factors, and stratified by treatment and body mass index (BMI).

RESULTS

A total of 14,942 BC cases and 74,702 matched controls were identified with mean age 61.2 years and 65% non-Hispanic White. Compared with controls, BC cases had higher cumulative incidence rates of hypertension (10.9% v 8.9%) and diabetes (2.1% v 1.7%) after 2 years, with higher diabetes incidence persisting after 10 years (9.3% v 8.8%). In multivariable models, cases had higher risk of diabetes (sHR, 1.16; 95% CI, 1.07 to 1.26) versus controls. Cases treated with chemotherapy (sHR, 1.23; 95% CI, 1.11 to 1.38), left-sided radiation (sHR, 1.29; 95% CI, 1.13 to 1.48), or endocrine therapy (sHR, 1.23; 95% CI, 1.12 to 1.34) continued to have higher diabetes risk. Hypertension risk was higher for cases receiving left-sided radiation (sHR, 1.11; 95% CI, 1.02 to 1.21) or endocrine therapy (sHR, 1.10; 95% CI, 1.03 to 1.16). Normal-weight (BMI < 24.9 kg/m²) cases had higher risks overall and within treatment subgroups versus controls.

CONCLUSION

BC survivors at KPNC experienced elevated risks of diabetes and hypertension compared with women without BC depending on treatments received and BMI. Future studies should examine strategies for cardiometabolic risk factor prevention in BC survivors.

INTRODUCTION

Breast cancer (BC) survivors are at higher risk of cardiovascular disease (CVD) after their cancer diagnosis, as compared with women without a history of BC.^1-3 Recent studies suggest that for women with a history of BC, the most common cause of noncancer death is CVD, thereby making it imperative to identify BC survivors at elevated CVD risk.¹ Detrimental effects on the cardiovascular system may result from cardiotoxic anticancer exposures including anthracycline chemotherapy,^4,5 trastuzumab biologic therapy,^6,7 left-sided radiation therapy,^8,9 and endocrine therapy.^10-13

CONTEXT

• Key Objective

• Breast cancer (BC) survivors are at risk of cardiovascular disease, but their risk of cardiometabolic risk factors is largely unknown. This prospective study compared risk of developing hypertension, diabetes, and dyslipidemia in 14,942 women with BC diagnosed from 2005-2013 with that in 74,702 women without BC at Kaiser Permanente Northern California.

• Knowledge Generated

• Women with BC had higher incidence of hypertension and especially diabetes up to 10 years postdiagnosis, with 1.2 times higher risk of diabetes over study follow-up, compared with women without BC. Associations became stronger in women who received cancer treatments, specifically left-sided radiation therapy and endocrine therapy, and in normal-weight women at diagnosis.

• Relevance

• Study findings highlight patients with BC as a vulnerable population at higher risk of cardiometabolic risk factors and support targeted cardiovascular surveillance by specific patient and treatment characteristics to mitigate these risks.

In vitro and animal studies have underscored a cause-effect relationship between the receipt of anticancer therapies, lipid uptake, and insulin signaling dysfunction.^14-16 Similarly, pre- and post-treatment changes in metabolic syndrome components and insulin resistance have been demonstrated in patients with BC undergoing treatment.^17-20 Previous studies suggest that BC patients with prevalent cardiometabolic risk factors at the time of cancer treatment, including being overweight or obese, having dyslipidemia, and/or having diabetes are at higher risk of cardiotoxicity events post-treatment.^21-24 After BC diagnosis, the impact is largely unknown of long-term cardiotoxic effects of anticancer treatments on developing cardiometabolic risk factors in patients of normal weight or without pre-existing CVD.

We set out to compare risk of incident hypertension, diabetes, and dyslipidemia in 14,942 women with and 74,702 women without a history of BC using data from Kaiser Permanente Northern California (KPNC). This study is one of the largest prospective studies of cardiometabolic risk factors in BC patients with a mean follow-up of 7 years. Findings could provide evidence to support intensified monitoring for cardiovascular outcomes in BC survivors after cancer treatment and lay the groundwork for future targeted intervention studies.

METHODS

Study Population

The Pathways Heart Study is an ongoing National Cancer Institute (NCI)–funded cohort study within KPNC examining the incidence of CVD events and cardiometabolic risk factors in women with and without a history of BC treatment (R01 CA214057).

For this analysis, administrative and clinical data were extracted from KPNC electronic health records (EHRs), with information of more than 4.5 million members across 21 hospitals and 262 outpatient clinics in the Northern California region. KPNC also maintains a cancer registry that reports to the NCI's SEER cancer registries. Study data include patient enrollment, sociodemographic characteristics, tumor characteristics and treatments, health care utilization, vital signs, diagnoses, procedures, laboratory results, pharmacy, and mortality.

Women with a new diagnosis of invasive BC (all stages) from November 2005 to March 2013 (cases) were identified from the KPNC Cancer Registry²⁵ and matched 1:5 to women without a history of BC (controls) from the KPNC membership files on birth year, race (White, Black, Asian, Native American and/or Alaskan Native, Native Hawaiian and/or Pacific Islander, or others), and ethnicity (Hispanic or non-Hispanic). Inclusion criteria for cases and controls included no previous history of invasive cancer, 21 years or older, female, and continuous KPNC membership for at least 12 months before and on the diagnosis date for cases or reference date for controls, allowing for up to a 31-day gap in membership. See Figure 1 for the study CONSORT diagram.

FIG 1.

Flow diagram of breast cancer case and matched control selection, Pathways Heart Study. ^aEight cases matched to four controls; all remaining cases matched to five controls. AJCC, American Joint Committee on Cancer; KPNC, Kaiser Permanente Northern California.

Measures

Incident hypertension was defined according to criteria developed by the KPNC Preventing Heart Attacks and Strokes Everyday (PHASE) program.²⁶ These were (1) two or more hypertension diagnoses of International Classification of Diseases 9th revision, clinical modification (ICD-9-CM) 401.xx (or equivalent ICD-10-CM codes) during primary care visits in the past 2 years, (2) one or more primary care hypertension diagnoses and one or more hospitalizations with a primary or secondary hypertension diagnosis in the past 2 years, or (3) one or more primary care hypertension diagnoses and one or more filled prescriptions for hypertension medication within the past 6 months.

Incident diabetes was identified from the KPNC Diabetes Registry.²⁷ Inclusion criteria are any one of the following: (1) at least one principal inpatient diagnosis of ICD-9-CM 250.xx or equivalent ICD-10-CM codes; (2) at least two outpatient diagnoses spanning the current and previous 5 years; (3) two or more abnormal laboratory test results on separate days spanning the current year and the past 2 years (ie, fasting blood glucose ≥ 126 mg/dL); and (4) at least one diabetes medication prescription such as insulin, oral hypoglycemics, or insulin sensitizers. Patients with only one of the indicators must have additional indicators during the subsequent 2 years of KPNC membership.

Incident dyslipidemia was defined as (1) two separate diagnosis codes of ICD-9-CM 272.0-272.4, ICD-10 E78.00, E78.01, and E78.1-R78.5; (2) diagnosis code and abnormal laboratory test results (LDL cholesterol ≥ 160 mg/dL); (3) diagnosis code and dispensed lipid-lowering medication such as statins and other antilipemic agents; or (4) dispensed lipid-lowering medication and abnormal laboratory test results (LDL cholesterol ≥ 160 mg/dL).

For the cardiometabolic risk factors where two diagnosis codes were required to identify incidence, the diagnosis date associated with the earlier (first) diagnosis code was considered the diagnosis date of the cardiometabolic risk factor.

Receipt of chemotherapy and radiation therapy and start dates were obtained from the KPNC Cancer Registry and supplemented by the EHR. For radiation therapy data, an additional variable indicating the receipt of left-sided radiation (yes or no) was created. Data on receipt of endocrine therapy were obtained from the outpatient pharmacy database.

CVD events were ascertained according to the ICD-9-CM and ICD-10-CM diagnosis codes and Current Procedural Terminology (CPT4) Codes. The complete list of codes is given in Appendix Table A1 (online only). Major CVD events included ischemic heart disease, heart failure, cardiomyopathy, and stroke. Other CVD events considered were arrhythmia, cardiac arrest, carotid disease, myocarditis or pericarditis, transient ischemic attack, valvular disease, and venous thromboembolism.

Mortality information was obtained from linkage to the KPNC mortality file, which contains data from the California State Department of Vital Statistics, US Social Security Administration, National Death Index, and KP membership and utilization data.

Statistical Analysis

The mean and standard deviation for continuous variables and frequency and percentage for categorical variables were used to summarize the baseline characteristics of the case and control populations. Cumulative incidence rates (CIRs) of hypertension, dyslipidemia, or diabetes were calculated using the cumulative incidence function (CIF) accounting for the competing risk of overall death and censoring individuals at the end of follow-up.^28-30 Differences in CIF between cases and controls were determined by Gray's test, the competing risk analog of the log-rank test.³¹

Subdistribution hazard ratios (sHRs) and 95% CIs for incident hypertension, diabetes, and dyslipidemia after BC diagnosis were estimated using Fine and Gray proportional hazards regression to model the CIF with covariates, by treating the CIF curve as a subdistribution function.^32,33 The time scale was defined as time since date of BC diagnosis or reference date to first of incident event, health plan disenrollment, death, or end of study (December 31, 2018). The proportional hazards assumption was tested using Schoenfeld residuals, and evidence of violation was found for case or control status. Thus, an interaction term of case or control status and log of time-varying survival time were included in all models. Models were adjusted for baseline body mass index (BMI), menopausal status, smoking status, neighborhood median household income, neighborhood education level, and any prevalent major or other CVD condition at baseline. To account for potential surveillance bias in the BC cases compared with their controls, models were also adjusted for number of outpatient primary care visits (in-person and virtual) in the year before diagnosis date or reference date. Prespecified subgroups of women who received any chemotherapy, left-sided radiation therapy, and/or endocrine therapy were compared with matched controls. Individual and combination treatment groups, along with standard BMI categories at BC diagnosis, were also examined. For each cardiometabolic risk factor model, women with a history of that specific risk factor within the past 12 months at baseline were excluded. Analyses were run in R using the survival package.³⁴

RESULTS

A total of 14,942 women with a new diagnosis of invasive BC were identified and matched to 74,702 controls (Table 1). On average, women were 61.2 (standard deviation = 12.8) years old and 65% were non-Hispanic White. For the total cohort, cases compared with their matched controls were more likely to be former smokers or nonsmokers (84% v 76%, P < .01) and overweight or obese (68% v 62%, P < .01). Although there were no general differences in comorbidity status at the time of BC diagnosis or reference date, more cases had a history of chronic hypertension compared with controls (44% v 42%, P < .01).

TABLE 1.

Characteristics of Cases and Matched Controls, Pathways Heart Study

Over a mean follow-up period of 7.0 (range < 1.0-13.4) years, 1,790 cases and 9,524 controls developed hypertension; 1,004 cases and 4,497 controls developed diabetes; and 1,775 cases and 9,681 controls developed dyslipidemia (Fig 2A). BC cases compared with controls had higher CIRs of hypertension at 2 years (10.9% v 8.9%) and lower rates after 10 years (23.9% v 27.2%) of follow-up. Incidence rates of diabetes were higher in cases at all timepoints including at 2 years (2.1% v 1.7%), 5 years (4.9% v 4.4%), and 10 years (9.3% v 8.8%). For dyslipidemia, cases consistently had lower CIRs compared with controls, especially by 10 years (23.7% v 27.0%). When examining CIRs by type of BC treatment (chemotherapy, left-sided radiation therapy, or endocrine therapy) for cases compared with controls, the rates were similar to the overall cohort and did not vary across treatment type (Figs 2B and 2D).

FIG 2.

Cumulative incidence curves of hypertension, diabetes, and dyslipidemia in BC cases and matched noncancer controls estimated from the cumulative incidence function treating overall death as a competing risk. P values are from Gray's test comparing the event rates between BC cases and controls: (A) total cohort, (B) among BC cases who received chemotherapy and matched controls, (C) among BC cases who received left-sided radiation therapy and matched controls, and (D) among BC cases who received endocrine therapy and matched controls. BC, breast cancer; CIR, cumulative incidence rate; CVD, cardiovascular disease.

In fully adjusted regression models, BC cases were significantly more likely to develop diabetes (sHR, 1.16; 95% CI, 1.07 to 1.26) but not hypertension (sHR, 1.02; 95% CI, 0.97 to 1.08) over the follow-up period, compared with non-BC controls (Table 2). By contrast, cases were less likely to develop dyslipidemia (sHR, 0.90; 95% CI, 0.86 to 0.95) compared with controls. For any of the cardiometabolic risk factors, there was no difference in risk between cases versus controls (sHR, 0.98; 95% CI, 0.93 to 1.03). The time-varying effect of survival time on case or control status decreased the sHRs further over the follow-up period for all main effects of the risk factors (not shown).

TABLE 2.

Relative Risk of Cardiometabolic Risk Factors Comparing BC Cases With Matched Controls, Overall and by Type of Treatment

When examining the risk factors by subgroups of cases who received any chemotherapy, any left-sided radiation therapy, or any endocrine therapy and their controls, the elevated associations remained for incident diabetes in all three groups, whereas higher risk for incident hypertension became apparent among cases receiving any left-sided radiation therapy (sHR, 1.11; 95% CI, 1.02 to 1.21) or any endocrine therapy (sHR, 1.10; 95% CI, 1.03 to 1.16) but not any chemotherapy (sHR, 0.97; 95% CI, 0.90 to 1.05). For risk of incident dyslipidemia, associations of lower risk persisted within the treatment subgroups, particularly among cases who received any chemotherapy versus controls (sHR, 0.89; 95% CI, 0.83 to 0.95).

We also examined associations between individual and combination BC treatments received and incidence of cardiometabolic risk factors (Appendix Table A2, online only). Risks of incident hypertension (sHR, 1.21; 95% CI, 1.07 to 1.37), diabetes (sHR, 1.28; 95% CI, 1.05 to 1.56), and any of the risk factors (sHR, 1.15; 95% CI, 1.01 to 1.30) were significantly higher in cases receiving both left-sided radiation therapy and endocrine therapy compared with controls. Higher risk of diabetes was also observed among cases who received all three treatments (chemotherapy, left-sided radiation therapy, and endocrine therapy; sHR, 1.42; 95% CI, 1.13 to 1.78). A lower risk of dyslipidemia was noted among cases who received chemotherapy only (sHR, 0.66; 95% CI, 0.51 to 0.86).

Finally, in subgroup analyses by BMI category at BC diagnosis, normal-weight BC cases had a statistically significant increased risk of hypertension (sHR, 1.18; 95% CI, 1.07 to 1.30) and diabetes (sHR, 1.35; 95% CI, 1.09 to 1.67) compared with controls (Table 3). Overweight and obese cases had no increased risk. There were no BMI differences among cases compared with controls for risk of dyslipidemia or for any risk factor. When examining by BC treatments received in the normal-weight cases versus controls, higher risk of hypertension and diabetes remained regardless of treatment subgroup and became attenuated when restricting to the subgroups who did not receive treatments (Appendix Table A3, online only).

TABLE 3.

Relative Risk of Cardiometabolic Risk Factors Comparing BC Cases With Matched Controls, by BMI Categories

DISCUSSION

In our prospective cohort study with a mean follow-up of 7 years in 14,942 women with invasive BC, we observed higher CIRs of hypertension (11.0%) and diabetes (2.1%) at 2 years postdiagnosis compared with a matched comparison group of 74,702 women without a history of BC. For diabetes, rates continued to be higher for BC cases compared with noncancer controls up to 10 years postdiagnosis, whereas for hypertension, they were initially higher during the first 2 years and then decreased by 10 years. By contrast, rates for dyslipidemia were lower in cases compared with controls over the entire follow-up period after BC diagnosis. In competing risk analyses adjusted for CVD-related risk factors, women with BC had 1.2 times higher risk of developing diabetes over study follow-up, but not for hypertension. Associations became stronger in women who received BC treatments, specifically left-sided radiation therapy and endocrine therapy, and in normal-weight women. These findings highlight patients with BC as a vulnerable population at higher risk of developing cardiometabolic risk factors compared with the general population without a history of BC and support targeted cardiovascular surveillance by specific patient and treatment characteristics to mitigate these risks.

The mechanisms underlying why patients with BC experience greater risk of cardiometabolic risk factors are not well understood. Chemotherapy and endocrine therapy can potentially cause alterations in blood pressure and metabolism, leading to the development of cardiometabolic risk factors.^8,35 Radiation therapy to the heart has been shown to induce a local proinflammatory state,^36,37 which we hypothesize might cause systemic changes to the cardiovascular system.

The evidence thus far has primarily focused on increased risk of diabetes, particularly steroid-induced diabetes caused by hyperglycemia,³⁸ with a scarcity of studies on hypertension and dyslipidemia. The treatments examined to date include chemotherapy and endocrine therapy, with little on radiation therapy. Consistent with our results, studies have shown that BC survivors experience a higher incidence of diabetes up to 10 years postdiagnosis compared with healthy controls,^39-43 with important differences by the type of anticancer treatment received. Among women who received chemotherapy, risk was highest with HR = 1.24 in the first 2 years after cancer diagnosis and then declined to HR = 1.08 by 10 years in a Canadian study.⁴⁰ For women who received endocrine therapy compared with nonreceivers, higher risk of diabetes ranged from HR = 1.31 in tamoxifen users in a Taiwanese study⁴³ to HR = 2.40 in all endocrine therapy users in an Israeli study,³⁹ whereas no association with incident diabetes was found in women receiving endocrine therapy in the only US study to date.⁴² However, these studies were limited by not accounting for CVD risk factors such as BMI and/or smoking^40,41,43 or history of cardiometabolic risk factors at BC diagnosis,⁴² small sample sizes of treatment subgroups,^39,42 or relatively short follow-up.⁴² Notably, extensive studies of cardiovascular health in childhood cancer survivors exist,^44-50 yet findings may not apply to adult survivors.

In our study, given that women who received left-sided radiation therapy or endocrine therapy were at higher risk for diabetes and hypertension, as well as normal-weight women, these specific patient populations may benefit from more frequent clinical monitoring and lifestyle interventions to mitigate this risk. In contrast to the persistently higher incidence of diabetes extending to at least 10 years, the elevated risk for hypertension appeared to normalize by 10 post-BC diagnosis. For hypertension, whether the early post-treatment period of higher risk will predispose to longer-term cardiac events requires further study. Reasons why normal-weight women might have a higher risk of cardiometabolic risk factors are unclear and require further study, but we hypothesize that variation in body composition and its impact on delivery of BC treatments could play a role. Variability in fat and muscle mass has been commonly observed in individuals with a similar height and weight.^51-54 Specifically, for chemotherapy, a normal BMI could mask loss of skeletal muscle and increased visceral and subcutaneous adiposity (ie, sarcopenic obesity), which may lead to variable effects of drug pharmacokinetics and clearance.^55-57 It is unclear why a lower risk of dyslipidemia was observed in cases compared with controls, but future studies on the impact of weight change over the follow-up period could provide some insight.

Our study strengths include being one of the largest prospective studies to date with long-term follow-up examining incidence of newly diagnosed cardiometabolic risk factors in a competing risk framework. The cohort is embedded within an integrated health care setting of a racially diverse BC patient population matched to noncancer controls with equal access to clinical care. Our study is also one of the first to examine the effects of radiation therapy on cardiometabolic risk factors. Previous studies have only focused on older patients in clinical trials, had relatively short follow-up periods, and did not account for important CVD patient risk factors (eg, BMI and smoking).

Limitations include lack of patient lifestyle data such as physical activity and diet, which can influence cardiometabolic health, yet this information is not routinely available in clinical data systems. Although details of cancer treatments including dose, duration, and specific drugs were not considered in this analysis, future analyses are underway. Also, given that women with BC may be more closely monitored than those without cancer, surveillance bias may exist. However, a strength of KPNC is providing its members with integrated medical care and population-based interventions to help maintain cardiovascular health.^58,59 Furthermore, to minimize possible surveillance bias, we adjusted for the number of outpatient primary care visits in the year before diagnosis date or reference date in the cases and controls. Although outside the scope of the current study, future analyses are planned to examine the potential time-varying effects of health care utilization on risk of CVD events. Our preliminary data show that there are differences between cases and controls in primary care utilization within the first several years after the index date.

In conclusion, the risk of developing new incident hypertension and diabetes is higher in women with BC, particularly for those who receive left-sided radiation therapy and/or endocrine therapy. Risk differences might also exist by BMI at BC diagnosis. Future studies should examine the impact of targeted cardiometabolic risk factor management on cardiovascular events in subgroups of women with BC on the basis of adjuvant therapy received and body size.

APPENDIX

TABLE A1.

Standard Codes for the Ascertainment of Nonfatal and Fatal Major Cardiovascular Events in the Pathways Heart Study

TABLE A2.

Relative Risk of Cardiometabolic Risk Factors Comparing BC Cases With Matched Controls and Individual and Combination Treatments

TABLE A3.

Relative Risk of Cardiometabolic Risk Factors Comparing Normal-Weight (BMI < 25 kg/m²) BC Cases With Matched Controls

AUTHOR CONTRIBUTIONS

Conception and design: Marilyn L. Kwan, Carlos Iribarren, Romain Neugebauer, Zaixing Shi, Dawn L. Hershman, Lawrence H. Kushi, Heather Greenlee

Financial support: Marilyn L. Kwan, Heather Greenlee

Administrative support: Marilyn L. Kwan, Heather Greenlee

Provision of study materials or patients: Marilyn L. Kwan

Collection and assembly of data: Marilyn L. Kwan, Cecile A. Laurent, Valerie S. Lee, Janise M. Roh, Heather Greenlee

Data analysis and interpretation: Marilyn L. Kwan, Richard K. Cheng, Carlos Iribarren, Romain Neugebauer, Jamal S. Rana, Mai Nguyen-Huynh, Zaixing Shi, Cecile A. Laurent, Janise M. Roh, Hanjie Shen, Eileen Rillamas-Sun, Margarita Santiago-Torres, Dawn L. Hershman, Lawrence H. Kushi, Heather Greenlee

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Risk of Cardiometabolic Risk Factors in Women With and Without a History of Breast Cancer: The Pathways Heart Study

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).