Inflammation and Clinical Decline After Adjuvant Chemotherapy in Older Adults With Breast Cancer: Results From the Hurria Older Patients Prospective Study

PURPOSE

Older breast cancer survivors are at increased risk of clinical decline after adjuvant chemotherapy. This study aimed to evaluate whether inflammatory markers assessed before adjuvant chemotherapy are associated with chemotherapy-induced clinical decline in a population of fit older adults with breast cancer.

METHODS

In a prospective study of women age ≥ 65 years with stage I-III breast cancer treated with chemotherapy, we measured interleukin-6 (IL-6) and C-reactive protein (CRP) prechemotherapy (T1). We assessed frailty status, using a Deficit Accumulation Index (DAI; categorized as robust, prefrail, and frail), at T1 and postchemotherapy (T2). The population of interest was robust women at T1. The primary outcome was chemotherapy-induced decline in frailty status, defined as decline in DAI from robust (T1) to prefrail or frail (T2). Multivariable logistic regression was used to examine the association between inflammatory markers and the primary outcome, adjusted for sociodemographic and clinical characteristics.

RESULTS

Of the 295 robust women at T1, 76 (26%) experienced chemotherapy-induced decline in frailty status, among whom 66% had high IL-6, 63% had high CRP, and 46% had high IL-6 and CRP at T1. After adjusting for sociodemographic and clinical characteristics, women with high IL-6 and CRP had a > three-fold (odds ratio, 3.52; 95% CI, 1.55 to 8.01; P = .003) odds of chemotherapy-induced decline in frailty status compared with women with low IL-6 and CRP.

CONCLUSION

In this cohort of older women with early breast cancer who were clinically fit before chemotherapy initiation, high IL-6 and CRP prechemotherapy were associated with chemotherapy-induced decline in frailty status independent of sociodemographic and clinical risk factors. Further research is needed to examine whether inflammatory markers can inform more personalized approaches to treating older breast cancer survivors.

INTRODUCTION

Although breast cancer treatment is highly effective, the control or cure of breast cancer does not always lead to a full restoration of health.^1,2 Evidence demonstrates that breast cancer treatment increases the risk of developing geriatric syndromes and chronic diseases—all signs of an accelerated aging state.^3,4 In adults over age 65 years, accelerated aging might have far greater consequences than in younger adults. One important sign of accelerated aging is frailty. Frailty is a geriatric syndrome linked to loss of independence, falls, and death.^5,6 Older breast cancer survivors treated with chemotherapy have a three-fold increased incidence of frailty as compared with age-matched survivors not treated with chemotherapy and persons without cancer.⁷ However, older adults are a heterogeneous population and experience different aging trajectories after cancer treatment.⁸ Significant gaps in knowledge remain, including how to identify patients at risk for frailty before initiating cancer treatment and whether differences in rates of biologic aging explain the observed heterogeneity in the health of older breast cancer survivors.

CONTEXT

• Key Objective

• To evaluate whether inflammatory markers assessed before adjuvant chemotherapy are associated with chemotherapy-induced clinical decline in a population of fit older adults with breast cancer.

• Knowledge Generated

• In this prospective cohort of clinically fit older women with breast cancer, patients with high interleukin-6 and C-reactive protein prechemotherapy had a greater than three-fold odds of experiencing chemotherapy-induced decline in frailty status than patients without high interleukin-6 and C-reactive protein. This increased risk was independent of age, race/ethnicity, education, body mass index, cancer stage and treatment, medications, or comorbidities at baseline.

• Relevance (K.D. Miller)

• These studies may help individualize treatment decisions for older patients by identifying those who are at greater risk of functional decline with chemotherapy. However, independent validation is critical before these results are incorporated into practice.*

  *Relevance section written by JCO Senior Deputy Editor Kathy D. Miller, MD.

Inflammation is a biologic aging process that has been linked to frailty. Frailty and prefrailty are associated with higher markers of inflammation, such as interleukin-6 (IL-6) and C-reactive protein (CRP).^9-11 In the general population, IL-6 and CRP have been correlated with functional status, cardiovascular health, and mortality.^12-16 In older adults with breast cancer, inflammatory markers at the time of cancer diagnosis have been linked to worse baseline physical function and postchemotherapy frailty.^17,18 However, there is limited evidence on the association between inflammation and chemotherapy-induced decline in frailty status.

To advance existing evidence and address limitations of previous studies, we leveraged a multicenter prospective cohort to evaluate the association between prechemotherapy IL-6 and CRP and chemotherapy-induced decline in frailty status in older adults with breast cancer.

METHODS

Study Design

The Hurria Older PatiEnts (HOPE) with Breast Cancer Study (ClinicalTrials.gov identifier: NCT01472094) enrolled patients from 16 institutions in the United States.¹⁹ The HOPE study gathered biologic and clinical data from patients age 65 years and older with stage I-III breast cancer treated with neoadjuvant or adjuvant chemotherapy. Between September 2011 and May 2017, 501 participants consented to participate in the study.¹⁹ The institutional review board approved the study at each participating institution, and all participants were provided written informed consent.

Clinical and Biologic Measures

Measure of health status.

We used a Deficit Accumulation Index (DAI)²⁰ as a measure of health status prechemotherapy (T1, within 14 days) and postchemotherapy (T2, within 30 days). The DAI is a 50-item scale that evaluates deficits in physical activities of daily living, instrumental activities of daily living, psychosocial status, nutrition, frequency of falls, number of medications, comorbid conditions, social support, and laboratory values (Data Supplement, online only). The DAI is calculated as a score that ranges from 0.0 to 1.0, with 0.0 representing no detected deficits. We categorized the DAI as robust (DAI 0 to > 0.2), prefrail (DAI 0.2 to < 0.35), or frail (DAI ≥ 0.35; see the Data Supplement for details). These cut points were chosen on the basis of prior evidence showing their relationship to mortality and morbidity outcomes in older adults.^21-24 Furthermore, a change in phenotype (robust, prefrail, and frail) defined using these cut points is considered a clinically meaningful change in frailty status. In the clinical practice and geriatric literature, the use of frailty status as a core measure has relied on these prespecified cut points to establish frailty prediction for adverse health outcomes in older adults.

For this analysis, the primary study outcome was chemotherapy-induced decline in frailty status—a dichotomized (yes/no) variable, with yes defined as a change in DAI category from robust (T1) to prefrail or frail (T2).

Biomarkers of inflammation.

Blood specimens (7.5 mL) were collected at T1 for measurement of IL-6 and CRP (see the Data Supplement). Plasma was stored at −80°C until assays were run. Plasma levels of IL-6 and CRP were measured in the City of Hope Analytical Pharmacology Core Facility using commercially available Luminex immunoassay kits—Invitrogen Luminex IL-6 Human Ultrasensitive Singleplex Bead Kit (cat.# LHC0063) and CRP Human ProcartaPlex Simplex Kit (cat. # EPX01A10288901). Each sample was analyzed in duplicate, and analyses were repeated if the coefficient of variation of the replicates exceeded 10%. None of the samples analyzed required reanalysis, and there were no outliers.

Covariates.

Baseline demographic characteristics, including the patient's age, race/ethnicity, education, employment, body mass index (BMI), number of comorbidities, and use of anti-inflammatory medications, were reported by participants. Disease characteristics including American Joint Committee on Cancer stage, hormone and human epidermal growth factor receptor 2 receptor status, prior breast cancer surgery, chemotherapy regimen, and primary granulocyte-colony stimulating factor (G-CSF) prophylaxis were collected.

Statistical Analysis

For this analysis, the population of interest was older women who were clinically fit (defined as robust per the DAI, 0.0 to < 0.2) at T1, before chemotherapy initiation. The primary outcome was chemotherapy-induced decline in frailty status, a dichotomized (yes/no) variable defined as a decline in DAI from robust (T1) to prefrail or frail (T2). The continuous change in DAI was also calculated as T2 DAI minus T1 DAI and summarized using descriptive statistics.

The distributions of IL-6 and CRP were summarized using mean (standard deviation) and median (interquartile, range). Given that there are no established clinically meaningful cutoffs for IL-6 and CRP, we first assessed the distribution of the cytokine values and log-transformed them to render parametric statistical analyses. We then performed empiric testing of various cutoff points (see the Data Supplement). Finally, we chose cutoffs above the median for each cytokine (2.5 pg/mL for IL-6; 3.5 mg/L for CRP), consistent with previous work as detailed in the Data Supplement.^17,18 We also calculated the Pearson correlation coefficient between log-transformed IL-6 and CRP, after controlling for BMI given the established relationship between BMI and cytokines. Given that there was little correlation between IL-6 and CRP in our cohort, we examined the combination of the two biomarkers.

The t-test and chi-square test were used to examine the relationships between continuous and categorical baseline variables and chemotherapy-induced decline in frailty status. Univariate analysis and multivariable logistic regression were used to examine the associations between inflammatory biomarkers at T1 and chemotherapy-induced decline in frailty status. Based on previous research^25-29 and findings from the univariate analysis, the final multivariable model was adjusted for age, race, education, BMI, number of comorbidities, cancer stage, primary breast cancer surgical procedure, and concurrent use of anti-inflammatory medication; other covariates were not statistically or clinically significant and were not retained. All statistical tests were two-sided, and P values < .05 were considered statistically significant. Data were analyzed using SAS 9.4 (SAS Institute, Cary, NC).

RESULTS

Patient Characteristics

Of the 501 participants in the HOPE cohort, we excluded 61 participants because of no biomarker data and 48 because of missing DAI scores. In addition, 76 were excluded because they were prefrail or frail at T1 (our population of interest was clinically fit women who were robust), and 21 were excluded because of having prior chemotherapy (n = 7), prior radiation (n = 10), or both (n = 4). The final cohort was 295 robust women at T1 (Fig 1).

FIG 1.

Flow diagram of study participants. Participants were included in the final cohort if they completed all study procedures including blood biomarker samples and DAI data. Participants with a baseline DAI score that categorized them as prefrail or frail were excluded. DAI, deficit accumulation index.

Overall, the 295 participants included in the analysis had similar baseline sociodemographic and clinical characteristics to those of the entire HOPE cohort. The median age at the time of starting chemotherapy was 69 (range 65-86) years. The majority (82.0%) were non-Hispanic White, with college or higher education (76.5%), and not employed (70.8%). At T1, the mean BMI was 28.5, the mean number of comorbidities was 1.9, and 109 (37%) reported taking anti-inflammatory medication. One hundred ten (37.3%) had stage I, and 185 (62.7%) had stage II or III disease; 22.0% had triple-negative disease, 31.2% had human epidermal growth factor receptor 2–positive+ disease, 37.3% had a regimen containing an anthracycline, and 72.5% received primary G-CSF prophylaxis. Most participants (82%) had completed breast cancer surgery before chemotherapy, among whom 48% had a mastectomy and 52% had a lumpectomy (Table 1).

TABLE 1.

Participant's Demographic, Disease, and Treatment Characteristics

Prechemotherapy Inflammatory Biomarkers

The median cutoff was 2.5 pg/mL and 3.5 mg/L for IL-6 and CRP, respectively. We examined the association between baseline BMI, use of anti-inflammatory medication, and breast cancer surgery with prechemotherapy inflammatory markers. After log transformation and controlling for BMI, Pearson's correlation coefficient between IL6 and CRP was 0.16. Use of anti-inflammatory medication was not significantly associated with IL-6 (median 2.7 pg/mL among nonusers v 2.3 pg/mL among users, P = .28) and CRP (median 3.4 mg/L among nonusers v 3.6 mg/L among users, P = .76). There was also no significant difference in inflammatory marker levels among participants who received mastectomy versus lumpectomy (IL-6: median 2.5 pg/mL v 2.6 pg/mL, P = .64; CRP: median 3.6 mg/L v 3.4 mg/L, P = .74).

Chemotherapy-Induced Decline in Frailty Status

Of the 295 robust women at T1, 76 (25.8%) experienced chemotherapy-induced decline in frailty status (Figs 2A and 2B). Compared with women who remained clinically fit after chemotherapy, women who experienced decline in frailty status did not differ significantly in age, race/ethnicity, education, employment, use of anti-inflammatory medications, tumor markers, breast cancer surgery, chemotherapy regimen, and G-CSF prophylaxis. Participants with higher BMI (P = .006) and more comorbidities (P < .001) had significantly higher odds of having chemotherapy-induced decline in frailty status compared with those with lower BMI and fewer comorbidities. Participants with stage II/III disease had increased odds of having chemotherapy-induced decline in frailty status (P = .04) compared with women with stage I disease (Table 1).

FIG 2.

Frailty status of clinically fit patients ≤ 1 month after chemotherapy. (A) Of the 295 participants in the clinically fit cohort, 74.2% remained clinically fit (maintained their robust status), whereas 25.8% had chemotherapy-induced decline in frailty status (became frail or prefrail) at the end of chemotherapy treatment, categorized as those who had chemotherapy-induced decline in frailty status. (B) The DAI distribution pre- (blue) and postchemotherapy (red) shown for the entire cohort. DAI, Deficit Accumulation Index.

We also examined the change in DAI as a continuous variable (T2 DAI minus T1 DAI). The median change in DAI of the final analysis cohort was 0.06, and, in accord with previous work, this change is considered a clinically meaningful difference.³⁰ In addition, compared with those who remained clinically fit after chemotherapy, patients with chemotherapy-induced decline in frailty status had a larger clinically meaningful change in DAI (median change in DAI 0.12 v 0.03, P < .001).³⁰ See the Data Supplement for additional details regarding our analysis of DAI as both a categorical and a continuous variable.

Prechemotherapy Inflammatory Biomarkers and Chemotherapy-Induced Decline in Frailty Status

Compared with participants who remained clinically fit, those who experienced chemotherapy-induced decline in frailty status had higher baseline levels of IL-6 (median 3.57 pg/mL v 2.15 pg/mL, P = .003) and CRP (median 4.75 mg/L v 3.06 mg/L, P = .007; Table 2). Among the 76 participants with chemotherapy-induced decline in frailty status, 65.8% had high IL-6 and 63.2% had high CRP. When examined in combination, 46.1% of the participants with chemotherapy-induced decline in frailty status had both high IL-6 and high CRP compared with only 25.6% of participants who remained clinically fit. The proportion of patients with one high biomarker but not the other did not differ significantly between the two groups.

TABLE 2.

Distribution of DAI, Changes in DAI, and Prechemotherapy Inflammation Biomarkers Overall and by Chemotherapy-Induced Decline in Frailty Status

In univariate analysis, high IL-6 was associated with 2.2-fold increased odds (odds ratio [OR], 2.21; 95% CI, 1.28 to 3.80; P = .004) and high CRP was associated with 2.1-fold increased odds (OR, 2.12; 95% CI, 1.23 to 3.62; P = .006) of chemotherapy-induced decline in frailty status (Table 3). After combining the two biomarkers, participants with either high IL-6 alone or high CRP alone did not have significantly increased odds of having chemotherapy-induced decline in frailty status compared with those with both low IL-6 and low CRP. However, those with both high IL-6 and high CRP had greater than three-fold increased odds of having chemotherapy-induced decline in frailty status (OR, 3.61; 95% CI, 1.75 to 7.44; P < .001; Table 3).

TABLE 3.

Univariate and Multivariable Associations Between Prechemotherapy Biomarkers of Inflammation and Chemotherapy-Induced Decline in Frailty Status

In multivariable logistic regression analysis, after adjusting for age, stage, race/ethnicity, education, BMI, breast cancer surgery, anti-inflammatory medications, and number of comorbidities, participants who had both high IL-6 and CRP were over three times more likely to experience chemotherapy-induced decline in frailty status (OR, 3.52; 95% CI, 1.55 to 8.01; P = .003) compared with participants with both low IL-6 and low CRP (Table 3).

DISCUSSION

In this prospective cohort of older women with early breast cancer who were clinically fit before chemotherapy initiation, we found that high prechemotherapy inflammatory biomarkers are associated with chemotherapy-induced decline in frailty status. Patients with high IL-6 and high CRP prechemotherapy were over three times more likely to experience chemotherapy-induced decline in frailty status compared with those with low IL-6 and low CRP. This association persisted after controlling for baseline sociodemographic and clinical characteristics, including established risk factors for clinical decline in this population (eg, age, BMI, and comorbidities).^26-28

These findings highlight several important clinical implications. First, biologic measures of inflammation may facilitate individualized risk assessment of chemotherapy-induced decline in frailty status at the time of diagnosis and before treatment initiation. Among fit older patients, we found that levels of prechemotherapy IL-6 and CRP were associated with clinical decline after exposure to chemotherapy. Specifically, patients with both high IL-6 and high CRP had significantly higher risk for chemotherapy-induced clinical decline. Both IL-6 and CRP have been linked to chronic low-grade systemic inflammation, referred to as inflammaging—a central pillar of aging-related syndromes.^31,32 Older adults with persistently elevated levels of IL-6 and CRP have decreased physical and cognitive status.^33,34 Hence, our findings suggest that older patients with breast cancer with high IL-6 and CRP, and especially those with both markers elevated, are vulnerable to experiencing a decline in physiologic reserve with chemotherapy treatment, even if they appear to be robust or clinically fit.

Second, inflammation may be a targetable biological pathway that can be leveraged to prevent and mitigate chemotherapy-induced clinical decline. Interventions targeting inflammation and activating anti-inflammatory defenses may improve fitness and functional recovery after chemotherapy. There is compelling evidence of behavioral interventions that can effectively alter inflammation, such as reducing stress and enhancing sleep.³⁵ Exercise training interventions have been shown to reduce adipose tissue inflammation and reduce circulating inflammatory markers.^36-39 Similarly, dietary interventions aimed at the link between obesity and cancer-related inflammation show promising results in lowering circulating inflammatory biomarkers.^40-43 Further work is needed to better understand the mechanism of these anti-inflammatory strategies to pave the way for future novel interventions to prevent, treat, and, ideally, reverse the unintended aging consequences of chemotherapy.

Third, our findings raise the possibility that inflammation and associated mechanisms of aging are key mediators in the impact of chemotherapy on frailty status. One such mechanism is cellular senescence, an upstream biological aging mechanism exacerbated by chemotherapy and linked to inflammaging. Cellular senescence is a state of terminal growth arrest, a fundamental aging process. Both natural aging and chemotherapy result in an accumulation of senescent cells.^44,45 Accumulated senescent cells are harmful, secrete multiple proinflammatory factors, including IL-6 and CRP, and contribute to tissue damage.⁴⁶ Clinical studies of patients with stage I-III breast cancer treated with chemotherapy indicate that senescent cell burden is associated with greater treatment-related fatigue and peripheral neuropathy.^47-51 However, whether senescent cells play a role in chemotherapy-induced clinical decline remains unclear. Further research is needed to understand their role in chemotherapy-induced decline in frailty status and translate this knowledge into interventions to mitigate risk before, during, and after cancer treatment.

There are several limitations to our study. First, frailty is a dynamic assessment, and the lack of follow-up time points beyond the completion of chemotherapy (T2) does not allow for conclusions about long-term health status beyond the end of chemotherapy. Without longer-term assessments, it is unclear which patients with increased frailty at the end of chemotherapy will recover (bounce back, are resilient).⁵² Second, we only examined baseline patient characteristics and did not account for events during therapy, such as grade 3-5 treatment-related toxicities and hospitalizations. Third, the cutoffs for high inflammatory markers were also based on the population median, and further work is needed to define optimal cutoff values. Fourth, our analysis focused on examining the DAI as a categorical variable although the DAI can be both a continuous and categorical variable. Little is known about the clinical relevance of change in DAI as a continuous variable. However, per previous studies, the change in DAI as a categorical variable, with cut points related to mortality outcomes, appears to be more clinically meaningful. Further research is needed to examine the clinical relevance of continuous change in DAI scores.

Despite these limitations, our study has several strengths. To our knowledge, this is the first study to examine the association between inflammation and frailty by using a DAI. DAIs are stochastic measures of frailty, previously validated in older cancer survivors.^1,8,20,53 Stochastic frailty reflects the process of deficit accumulation, which increases one's vulnerability to stress. By contrast, Fried's frailty phenotype is a phenotypic measure of frailty and represents a specific state of decline.⁵⁴ To date, previous studies on inflammation and frailty in breast cancer populations have relied mainly on Fried's frailty phenotype. Using a geriatric assessment–based DAI, we can better understand the effects of a pleiotropic stressor such as chemotherapy. The DAI also allowed us to exclude individuals who are not only phenotypically frail at baseline but also more vulnerable to stressors such as chemotherapy.

Our study underscores the link between inflammation and chemotherapy-induced clinical decline, and these findings serve as a foundation for future research. Although IL-6 is not commonly measured in the clinical setting, these data support the possible utility of this biomarker as an addition to CRP, a more widely available marker in clinic. Our findings, consistent with previous studies, reveal that IL-6 and CRP are only weakly correlated^55-57; however, the evidence regarding the relationship between these two markers is limited and inconsistent.^58,59 Further research is needed to understand the potential clinical utility of inflammatory biomarkers prechemotherapy. Studies are needed to (1) determine whether chemotherapy-induced frailty and inflammaging are associated with specific drugs or cytotoxic chemotherapy in general, (2) evaluate the relationship between these processes and key outcomes (eg, noncancer mortality and morbidity), and (3) examine whether these processes are reversible, at least to some extent.

In conclusion, our study demonstrates that high IL-6 and CRP prechemotherapy are associated with an increased risk of chemotherapy-induced decline in frailty status in clinically fit older patients with breast cancer. Our findings provide novel insight into the biologic processes of aging that predict tolerance of cancer treatment and form a foundation for further work to appraise inflammatory biomarkers as a novel tool to guide treatment decisions. Further research is needed to evaluate the clinical utility of inflammatory biomarkers in identifying older adults most at risk of clinical decline before initiation of chemotherapy.

AUTHOR CONTRIBUTIONS

Conception and design: Jingran Ji, Can‐Lan Sun, Harvey J. Cohen, Hyman Muss, Mina S. Sedrak

Provision of study materials or patients: Timothy Synold, Hyman Muss

Collection and assembly of data: Jingran Ji, Can‐Lan Sun, Timothy Synold, Mina S. Sedrak

Data analysis and interpretation: All authors

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

Inflammation and Clinical Decline After Adjuvant Chemotherapy in Older Adults With Breast Cancer: Results From the Hurria Older Patients Prospective 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).