Introduction

Despite advances in the treatment and management of early stage breast cancer (BC), ~ 20–30% of the 3 million US patients with BC will develop metastatic disease, signifying BC spread to the bone, liver, lung and other distal locations [1]. Additionally, 6–10% (13,776–22,096) of women with metastatic breast cancer (MBC) present with de novo disease, reflecting distal BC metastases at the time of initial diagnosis. Recent analyses highlight de novo disease is increasing among women aged 25–39, particularly non-Hispanic white and African-Americans (AA) [2]. Due to the integration and expansion of targeted therapies, the median survival for women with recurrent and de novo MBC has increased substantially [3,4,5,6,7,8]. In fact, the number of women living with MBC has doubled in the last decade [7]. Nevertheless, there is an on-going perception that the majority of these women struggle with weight loss, cancer cachexia and frailty, which is often inaccurate for many living with this disease. Although symptom burden is a challenge, numerous women with MBC frequently endorse complaints of weight gain and general physical deconditioning associated with on-going anti-neoplastic treatments [9].

Breast cancer treatment is associated with adverse body composition changes, specifically gains in adipose tissue and reductions in strength and lean mass (LM; also known as skeletal muscle) [10, 11]. No clear consensus exists on the etiology of these changes; however, BC treatment, lifestyle habits and menopausal status, likely in combination, are all implicated [10]. Excess adiposity contributes to inflammation and insulin-resistance, which are theorized to promote tumor progression and loss of LM [12, 13]. Low levels of LM are associated with chemotherapy toxicity, increased symptom burden, and compromised survival in women with MBC [14, 15]. For women with early-stage BC, health behavior changes targeting reductions in adiposity and/or increases in muscle strength and LM lead to improvements in QOL and symptom reports, as well as prognostic changes serum biomarkers [16,17,18,19,20], mitochondrial function [21] and potentially novel microRNA [22]. Thus, interventions promoting favorable body composition changes would likely have numerous benefits for women with MBC, yet this remains largely unexplored. Typically, women with MBC are excluded from nutrition and/or PA intervention trials under the assumptions that: (1) exercise could be unsafe or pose additional burdens, (2) changes in lifestyle habits would have little impact on quality of life (QOL) or disease trajectory, or (3) women with MBC lack interest in such trials. We designed Every Day Counts to challenge and test these assumptions, hypothesizing women would demonstrate trends for improvements in QOL and lean mass.

Methods

Study design

Every Day Counts (EDC) was a 12-week randomized, waitlist controlled trial designed primarily to evaluate the feasibility, adherence and safety of an intervention promoting the American Cancer Society (ACS) nutrition and physical activity (PA) guidelines [23] in women with clinically stable MBC. Our secondary objective was to assess its short-term impact on patient reported outcomes and physical performance. We also explored biomarkers of inflammation, mitochondrial function and conducted a nested study to examine the additive benefits of extending the intervention to 16-weeks. Informed by our developmental work [9], EDC reflected essential adaptations of our previously successful lifestyle interventions [24,25,26,27,28] in a novel population. The ethical review board of the Medical College of Wisconsin (MCW) and other participating institutions reflective of the investigative team approved this study. This trial was registered on clinicaltrials.gov (NCT03824145.)

Patient population and recruitment

To participate, adult (≥ 18 years) females with histologically confirmed MBC had to be deemed clinically stable by their medical oncologist defined as follows: anticipated life expectancy of at least ≥ 12 months, no unintentional weight loss, no new symptoms or change in performance status (ECOG 0–1) for the past 30 days, no clinical (including laboratory) or radiologic evidence of disease progression (past 30 days), no changes in anti-neoplastic therapies (previous or anticipated next 30 days), and no reports of severe pain precluding practical and safe participation (≥ Grade 3 per the NCI CTCAE)] [29]. In addition, women were required to have access to a mobile phone with unlimited text messaging, able to participate in moderate PA and strength training, and non-compliant with the ACS nutrition and PA recommendations (i.e., consuming < 5 fruits and vegetables/day, ≤ 90 min/week moderate exercise (which allows for progress to 150 min/week), and/or < twice/week of resistance exercise training (RET).) Women with < 12 months of life anticipated, receiving appetite stimulants and/or steroids were not approached. Women not medically cleared by their oncologist or already adherent to the ACS lifestyle guidelines were ineligible. Recruitment methods included letters sent to directly to women with MBC describing the program and oncologist/nurse referrals within the cancer center. To support a broader inclusion of women, members of the community advisory board were encouraged to use ‘word of mouth’ recruiting. Further, brochures and/or study advertisements were sent to medical oncologists at local community hospitals and EDC was promoted on targeted social media outlets (non-public Facebook breast cancer support groups) in the greater Milwaukee, WI area.

Randomization

Following the baseline interview and using a random digit generator, women were randomized to the immediate intervention or waitlist control group. Randomization allocation was stored and maintained in a secured password protected server. To ensure blinding, only the biostatistician and project coordinator had access; data collectors (phlebotomy, questionnaire, diet) and investigators did not.

Every Days Counts intervention

EDC was a 12-week behavioral intervention rooted in social cognitive theory (SCT) addressing adoption of the ACS nutrition and PA guidelines for cancer survivors, including: (1) consumption of whole grains and 5+ fruits and vegetables daily, and (2) engagement in a minimum of 150 min of moderate PA and 2 sessions of resistance exercise training (RET)/week. A certified, licensed athletic trainer (LAT, ATC, CES) experienced in lifestyle counseling spearheaded: (1) four supervised aerobic/RET sessions, (2) weekly individual telephone sessions based on the EDC program intervention manual; and (3) intervention support. Attendance at each scheduled (in person or telephone session) was tracked. If the participant failed to show up (in person or telephone session) and did not respond to text or voicemail to reschedule, this was deemed a “no show” and used to assess adherence.

Prior to session 1, the coach created an RET program tailored to the participant’s strength. Using the baseline physical assessment data and guided by normative scores for strength [30, 31], the health coach recommended upper and lower extremity and core exercises based on a predetermined list of activities targeting these areas (e.g. curls, squats, lunges, planks.) At the first supervised intervention session, the health coach delivered the EDC program binder, reviewed program structure, and provided nutrition education and PA/RET guidance tailored to the participant’s perceived level of conditioning. The EDC binder addressed knowledge (e.g., plant-based eating), attitudes (e.g., understanding barriers and facilitators to healthy eating and PA), and behavioral strategies (e.g., self-monitoring) (Table 1). The EDC binder also contained recipes, a plant-based eating and PA self-monitoring form, and other materials to support adherence, reinforcement and review. Telephone sessions began with information on the weekly theme, discussion of previous week’s achievements and challenges, a review of adherence using nutrition and PA tracking, and setting one SMART weekly goal for nutrition and one for PA/RET. The supervised exercise portion included: (1) guiding participant in her choice of appropriate modes of moderate PA to work toward engaging in 30 min most days of the week (e.g., walking, biking), (2) assessing proper type and intensity of RET for safe conditioning, and (3) instructing and observing participants in the correct execution of RET. RET exercise using appropriately prescribe exercise bands were progressed individually by altering load (reps/set) or band (resistance), using a combination of participant feedback (e.g., “feels too easy,” “not challenging”), consistency of program adherence via verbal feedback and data from the tracking device. Guidance was also provided on endurance activities (e.g., increasing step goals or time spent walking.)

Table 1 Baseline characteristics for women with metastatic breast cancer participating in Every Day Counts
Full size table

In support of self-monitoring and intervention adherence, participants received a Fitbit™ activity tracker (Charge or Charge HR), exercise bands, cooking utensils and twice weekly text messages (weeks 1–12) focused on self-efficacy, outcomes expectations, social support and/or resources or self-monitoring tools. These strategies were successfully applied in our previous lifestyle interventions [32, 33]. Women were also provided the opportunity to attend three interactive cooking classes focused on simple plant-based meals (1 salad, 1 side dish, 1 main course). These group classes, held at a community location, served to: increase skills and comfort with preparing plant-based meals and promote dietary adherence [34].

Every Day Counts waitlist control

Participants in the waitlist control group received a monthly text message thanking them for their participation in the study and reminding them of their intervention start. Following the 12-week waitlist control period, women in this group were offered the intervention and all supportive materials. This design was selected based on critical feedback from a community advisory board, who strongly believed all women should have access to the intervention.

Data collection

At baseline (prior to randomization) and post-intervention (12-weeks) two 60-min assessments were completed, involving: (1) an interview and (2) physical measurements (dual energy X-ray absorptiometry [DXA], strength, anthropometrics) and phlebotomy. During the interview, women completed questionnaires quantifying QOL (Functional Assessment of Cancer Therapy—Breast [FACT-B] [35]), symptom burden (fatigue [36], pain [37, 38], sleep [39], mood [40]) and lifestyle factors (Godin leisure time exercise [41], stress [42], self-efficacy [43], social support [35]). Physical measures included standardized measures of weight and height to calculate BMI (weight [kg]/height [m]2.) Body composition (adipose and lean mass) were measured by DXA (GE Lunar Medical Systems, enCORE software version 14.10.022.) Whole body scans were performed and analyzed by a trained technician blind to study group or outcomes, following calibration with the manufacturer’s phantom. Visceral fat mass was estimated from the android region; a region of interest automatically defined using the methods of Kaul et al. [44] Three 24-h diet recalls (24DR) were completed in person and/or via telephone by the MCW CTSI Registered Dietitian using the 5-pass methodology within the week of interview [45]. Dietary data were collected and entered into Nutrition Data System for Research to detect changes in target dietary intake patterns (i.e., F/V, red meat, alcohol.) [46]. Handgrip dynamometry (Jamar Plus handgrip) was completed for all participants following standard study procedures [47]. Other physical performance measures were completed in a subset of participants (e.g., sit to stand, bicep curls, calf raises, planks.) Blood biospecimens were acquired after an overnight fast to profile inflammatory biomarkers implicated in carcinogenesis (e.g., C-reactive protein (CRP), interleukin-6, tumor necrosis factor). Samples were sent immediately to assess mitochondrial function (Seahorse XF Analyzer, Agilent, CA) or processed and stored at − 80 °C until the time of batch analyses (Eve Technologies, Calgary, AB. Canada). For purposes of this project, spare respiratory capacity (a measurement of mitochondrial integrity and energetics) was considered a representative parameter of overall mitochondrial health.

Statistical analyses

Descriptive statistics were reported for all outcomes of interest at baseline and 12 weeks, including QOL, symptoms, body composition and lifestyle behaviors by group. We defined feasibility as meeting 100% of our recruitment goal, adherence as attending at least 75% of telephone coaching sessions and 2/3 exercise training sessions, and efficacy as evidenced by trends in improvements QOL for women in the intervention vs. the control groups. To assess safety, adverse events (i.e., intervention-related injuries) were monitored throughout the study period. Linear mixed models with an interaction term to account for differences in trend over time between groups and a random effect to account for correlation of repeated measures within each woman were used to assess each outcome. Compound symmetry covariance structures were assumed for the correlation from the same woman across time. Statistical significance of the difference between outcomes for the immediate and waitlist control groups were from the interaction term and the within group comparisons used the corresponding contrasts terms. No formal power calculations were conducted a priori; 40 women were recruited taking into consideration the duration of the parent study and to demonstrate feasibility. A significance level of 0.05 was used and statistical analysis was performed using SAS v9.4 (SAS Institute, Cary, NC).

Results

A total of 78 women were screened and 40 women were consented in 9 months (January–September, 2018.) Written letters or referrals from oncologists provided the greatest participant yields (n = 34/40), followed by our community engagement efforts (n = 6/40). Thirty-five women completed the 12-week study contributing evaluable data. (Fig. 1) On average, participants were: 55 years of age (± 12), white (91%), employed full-time (32%), retired or on disability (68%), had bone (63%), liver (26%) or lung metastases (26%), and were 3.6 (± 3.2) years after MBC diagnoses. Overall attendance was 86% during the 12-week intervention: 93% for in-person sessions and 84% for telephone sessions. No adverse events (injuries related to the intervention) occurred; however, three of the women (2 Intervention, 1 Waitlist control) developed new symptoms/disease progression during the study, something expected given the unpredictable nature of MBC.

Fig. 1
figure 1

Flow of participants for the Every Day Counts: a randomized pilot intervention for women with metastatic breast cancer

Full size image

In line with our secondary objective, we evaluated patient reported outcomes, with an intentional focus on QOL (Table 2). Using the FACT-General, significant increases in QOL were reported between groups (p = 0.003) and for women in the intervention group (p = 0.001), which were largely driven by improvements in physical (p = 0.003), emotional (p = 0.01) and functional well-being (p = 0.06). When we examined symptoms unique to BC and endocrine treatment, only women in the intervention arm reported significant within group improvements for BC (p = 0.001) and Endocrine symptoms (p = 0.002), whereas. women in the waitlist control did not (p = 0.356 and p = 0.450, respectively.) Fatigue improved for women in the intervention group (p = 0.037), but not between groups (p = 0.107). Findings related to mood were mixed. Significant declines in depression were noted between groups (p = 0.03) and for women in the intervention (p = 0.010), yet anxiety and perceived stress significantly declined for all women, regardless of group (all p values < 0.05).

Table 2 Comparison of quality of life and lifestyle behaviors between the intervention (n = 17) and waitlist control (n = 18) groups in the Every Day Counts intervention
Full size table

By design, this was not a weight-loss study; thus, declines in body weight or BMI were not observed. Trends in energy intake declined for women in the intervention and increased for those in the waitlist control, yet these differences were not statistically different within or between groups (Table 3). Although improvements in body composition were targeted, no significant changes between or within groups were noted for fat or lean mass. Visceral fat mass decreased between and within all groups (all p values < 0.05), however. Using applicable DXA cut points, 20% (n = 7) and 14% (n = 5) of women were classified as sarcopenic pre vs. post-intervention, respectively (p = 0.938); none of these women had obesity. Considering sarcopenia is ideally defined as compromises in lean mass and strength [48], the significant increases in handgrip strength for women in the intervention (p = 0.002) vs. waitlist control (p = 0.163) were deemed clinically meaningful. The Godin results support women in the intervention spent significantly more time engaged in PA pre vs. post-intervention (p < 0.001), which was significantly different compared to those in the waitlist control (p < 0.001).

Table 3 Comparison of body composition, biomarkers, physical measures and lifestyle behaviors between the intervention (n = 17) and waitlist control (n = 18) groups in the Every Day Counts intervention
Full size table

We leveraged this pilot study to address concerns related to the length of the EDC intervention and to explore potentially important serum and cellular changes in a novel population. Specifically, we extended the EDC intervention to 16 weeks for women in the waitlist control who were interested (n = 14) to see if they derived additional benefit compared to women who completed the 12-week intervention. QOL results were similar; however, women in the 16-week program displayed more favorable, non-significant trends for lean mass gain, fat mass loss, and biomarker changes and significantly greater increases in strength (sit to stands, planks, bicep curls.) Due to unforeseen challenges regarding the analyses of blood biospecimens, we did not have complete data on all participants pre vs. post-intervention. We also encountered issues with the timely transport of blood biospecimens and reliable machinery, and thus, we were only able to explore mitochondrial function at in a subset of women (n = 12). Spare respiratory capacity increased by a mean 69.7% (± 165.6, p = 0.24), signifying important cellular changes in muscle functioning associated with intervention participation.

Discussion

Every Day Counts was implemented following our developmental work in a small group of women with MBC who provided novel insights on the unmet needs of this population [9]. Our present pilot study results support: (1) feasibility as evidenced by participant interest (78 women contacted and screened) and 100% of our enrollment goal (n = 40 in 9 months vs. our projected 15 months); (2) adherence as evidenced by 86% of women completing all 12 coaching sessions (84% of telephone and 93% in-person sessions) and documentation of improved exercise and dietary behaviors; and (3) potential efficacy as evidenced by significant improvements in QOL. These pilot findings also reinforce the safety of a dietary and PA intervention in this population. Given the common prevalence of bone metastases in women with MBC, the lack of intervention-related injuries is clinically important. Further, our preliminary findings reinforce the impact of lifestyle behaviors on QOL and symptom burden; key somatic reports with potential to complement and support treatment tolerance.

A variety of methods are available to quantify changes in QOL; however, we specifically chose the FACT tools because they afforded us the opportunity to directly compare our findings to a similar patient population. In doing so, we discovered our participants’ baseline scores for physical, social, emotional and functional well-being were higher compared to a larger, more heterogeneous sample of women with MBC [49], but below those for physical, social and functional well-being when compared to a normative sample compromised largely of women with early stage BC [50]. Considering our participants were required to be clinically stable and not near the end of life, these findings are not surprising. Interestingly, intervention participants reported significant improvements in physical well-being post-intervention achieving scores on par with those in the normative sample (22.7 vs. 22.1, respectively.) [50]. We suspect these increases were partially driven by clinically significant improvements in fatigue (> 3 points) [51]. Studies in women with early-stage BC support the positive and sustained impact of lifestyle (nutrition and/or PA) interventions on QOL and other BC-specific outcomes [20, 52,53,54,55,56]. Only a few studies focus on women with MBC. In an early study by Headly et al., these investigators demonstrated that a seated exercise intervention in women with MBC initiating chemotherapy significantly reduced fatigue and slowed decline in total and physical well-being scores [57]. More recently, Jones [58] and Yee [59] showed that women with MBC have reduced cardiopulmonary functioning and lower aerobic fitness, reduced muscular strength and less daily PA compared to healthy controls. Combining 20 patients with bone metastases from prostate and BC, Cormie et al. showed improvements in functional ability, lean mass and QOL at 3- and 6-months following a supervised resistance exercise trial [60]. However, two recent studies conducted exclusively in women with MBC reported high dropout rates and no significant differences in QOL for interventions focused solely on aerobic exercise [61, 62]. Our data support women in the Every Day Counts intervention exhibited statistically and clinically meaningful increases in overall (FACT-G; 3–7 points) [63] and disease-specific (FACT-Breast; 2–3 points) [63] QOL. They also increased their PA levels from ‘insufficiently active’ to ‘active,’ whereas women in the waitlist control remained ‘moderately active,’ albeit on the lower end of this interpretive scale [64]. Together, these data highlight the complexities of this population and emphasize the need for carefully structured programs to achieve positive outcomes.

Body composition is gaining increasing recognition as an important tissue biomarker predictive of prognosis and treatment tolerance in women with early [65] and late stage BC [14, 15, 66]. Using archived abdominal computed tomography (CT) images conducted for clinical purposes, Prado [14] and Shachar [15] quantified body composition in women with MBC. Results support women with MBC who are classified as sarcopenic more often experience shorter time to tumor progression and treatment toxicities [14], hospitalization, dose-reductions and dose delays [15], compared to women without sarcopenia. Within EDC, the prevalence of sarcopenia was 20% (n = 7/35) at baseline without new incident cases post-intervention (using DXA.) We anticipated that increases in PA with twice weekly RET would result in decreased DXA-detected sarcopenia. However, while increases in strength were observed, we also observed overall decreases in lean mass in the intervention and waitlist control groups. These findings are in stark contrast to Cormie et al. who had a similar intervention window (3 months), twice weekly RET and reported significant gains in LM, albeit in a mixed population of metastatic breast (n = 3) and prostate cancer (n = 17) survivors [60]. During the initial design of EDC, we contemplated the potential impact of a 12- vs. 16-week intervention on body composition. Based on the work of Cormie [60], Capozzi [67] and Lonbro et al. [68], 12 weeks of RET seemed a sufficient window to observe lean mass gains. Because these studies did not target women with MBC, we conducted a nested study within this EDC pilot to extend the 12-week intervention by 4 weeks for waitlist control participants. The nested study findings support more favorable changes in body composition, physical performance and biomarker trends. Accompanied by the qualitative feedback from our participants, we advocate a 16-week lifestyle intervention to better meet the quantitative endpoints and desires of this population.

Although this study was methodologically rigorous in design and tailored to meet the perceived needs of this population, it is not without limitations. First, we lacked sufficient measures to comprehensively evaluate changes in physical performance. Due to its ease and acceptability, we collected handgrip strength for all participants; however, we failed to consider other field measures to capture changes in other muscle groups until later in the study providing us with results for only a subset of participants. Additionally, we did not include validated methods to evaluate exercise intensity or cardiorespiratory fitness. Second, our study sample is not representative of all women with MBC. By design, we targeted women capable of healthy eating and PA; thus, our sample may be ‘healthier’ and reflects an inherent selection bias in studies of this nature. Further, although we met 100% of our recruitment goal, we did not adequately engage minority participants. Additional community engagement efforts are needed. Third, our attempts to explore biomarkers of inflammation failed to demonstrate meaningful change. Given our small sample size, we suspect some of this may be due to a Type II error. Finally, feedback from our participants highlighted the need for more information on dietary change strategies and safety guidelines for RET with bone metastases. Interestingly, our exploratory analyses reveal moderate to high adherence to several of the dietary components of the ACS guidelines at baseline for all participants (data not shown); thus, it would be difficult to show significant improvements. Regardless, targeted messaging regarding increasing whole grains and reducing alcohol consumption, along with cooking resources (video links, more plant-based recipes) provide opportunities for further benefit. In addition, reassurance that RET is a safe and recommended treatment modality for treated and medically stable bone metastasis [60, 69] should have been shared earlier in the intervention to help support the need for more frequent and higher load RET to see LM gains.

Conclusions

The design and implementation of Every Day Counts was informed by women with MBC; a growing group of cancer survivors for whom we have limited evidence to guide their lifestyle choices. Results from this pilot highlight the interests, capabilities and importantly, the safety of a nutrition and PA intervention in women with clinically stable MBC, providing strong evidence to support favorable improvements in QOL and symptom burden. A larger trial targeting a more diverse group of women (racial/ethnic, rural, community) and adequately powered to explore the mechanistic pathways between QOL and lifestyle behaviors is warranted.