Feasibility and efficacy of a weight gain prevention intervention for breast cancer patients receiving neoadjuvant chemotherapy: a randomized controlled pilot study

Abstract

Purpose:

Weight gain is common among breast cancer patients and may contribute to poorer treatment outcomes. Most programs target breast cancer survivors after the completion of therapy and focus on weight reduction. This study examined the feasibility and preliminary efficacy of an intervention designed to prevent primary weight gain among women receiving neoadjuvant chemotherapy for breast cancer.

Methods:

Thirty-eight newly diagnosed stage II or III breast cancer patients were randomized to the BALANCE intervention or usual care within three weeks of starting neoadjuvant chemotherapy. The intervention used a size acceptance-based approach, and encouraged home-based resistance and moderate-intensity aerobic exercise as well as a low energy-dense diet to prevent weight gain. Assessments were conducted at baseline, mid-chemotherapy (3 months) and post-chemotherapy (6 months). Intervention feasibility, acceptability, and preliminary effects on anthropometric, quality of life, and circulating biomarker measures were evaluated.

Results:

Intervention participant retention (100%) and in-person session attendance (80%) was high during the intervention period, although attendance dropped to 43% for telephone-delivered sessions. The majority of participants reported being satisfied with the intervention during chemotherapy (88%). Participants in the intervention group had greater reductions in waist circumference (p=.03) and greater improvements in self-reported vitality scores (p= .03) than the control group at the end of chemotherapy. Significant effects on biomarkers were not observed.

Conclusions:

A size acceptance weight management program is feasible during neoadjuvant chemotherapy among breast cancer patients, and may have beneficial effects on waist circumference and patient vitality.

BACKGROUND

Weight gain is a common problem among women diagnosed with breast cancer [1], particularly for women who are normal weight or overweight at diagnosis [2,3], younger women [4–6], and those receiving chemotherapy [7,4]. Most weight gain is observed in the first year after diagnosis [8]. Weight gain takes place throughout the course of chemotherapy [3,2,9–16], continues after treatment ends [9,10,17,18,16,19,4], and is greater than that observed in healthy women as they age [7,20].

Weight gain has a negative effect on the well-being of breast cancer patients and may influence cancer-specific and non-cancer-related health outcomes. Weight gain is distressing and can result in negative body image [21–23]. Breast cancer survivors who gain weight report poorer physical functioning, lower vitality, and more role limitations due to physical symptoms [24]. Several studies indicate that weight gain after diagnosis increases risk of all-cause mortality [25]. In a large prospective US cohort, the Nurses’ Health Study, weight gain after diagnosis was associated with an increased risk of recurrence, breast cancer death, and total mortality among never smokers, those with BMI <25 at baseline, women with tumors ≤ 2 cm, and those with no nodal involvement [26].

Most weight management trials among breast cancer populations focus on weight loss after it is gained [27], as opposed to targeting weight gain prevention. Weight loss interventions have been shown to be safe and feasible for breast cancer survivors recruited after the completion of treatment [27]. However, less is known about the feasibility of weight gain prevention interventions during chemotherapy, a time when patients are likely to initiate weight gain. A recent review of weight gain prevention interventions during active breast cancer treatment noted major limitations in the current research including small numbers of studies, and infrequent use of blood biomarkers in analysis [28].

The current pilot study assessed the feasibility and acceptability of conducting a randomized behavioral weight gain prevention intervention during neoadjuvant chemotherapy, and evaluated its effects on weight, waist circumference, and quality of life. Additionally, blood biomarkers related to adiposity and known to influence breast cancer prognosis, including sex hormone-binding globulin (SHBG), fasting insulin and glucose, IGF-I, and IGFBP-3, adiponectin, and leptin were assessed. A-priori feasibility and acceptability targets were not set for this study, but rather were examined broadly in order to identify potential for a larger trial. We hypothesized that the intervention would be feasible and acceptable to patients, and would positively affect body weight, waist circumference, quality of life, and biological markers.

METHODS

This study was a randomized pilot test of the Project BALANCE (Balancing Activity, Lifestyle And Nutrition throughout the Cancer Experience) weight gain prevention intervention (www.clinicaltrials.gov, NCT00533338). The University of Texas MD Anderson Cancer Center (UTMDACC) institutional review board approved this protocol (2005–0888). Written informed consent was collected from all participants. CONSORT guidelines were followed in the reporting of this trial.

Participants

Participants were recruited at UTMDACC. Eligibility criteria included 1) newly diagnosed stage II or III breast cancer patients within 3 weeks of starting neoadjuvant chemotherapy, 2) ability to speak and read English, 3) cognitive ability to give informed consent, 4) willingness to come to UTMDACC once a week, 5) reside in Harris county or a contiguous county, and 6) have physician clearance to engage in unsupervised physical activity. Patients with absolute or relative contraindications for exercise testing, as defined by the American College of Sport Medicine [29], were excluded as were patients with uncontrolled asthma or other medical contraindications as defined by the patient’s physician. Patients with inflammatory breast cancer were also excluded. A total of 129 patients were assessed for eligibility (Figure 1). Thirty-nine patients (30%) were ineligible based on the criteria above, and 42 patients (33%) declined to participate. Thirty-eight participants were randomized and 37 completed the study.

Fig 1:

CONSORT diagram for Project BALANCE

Procedures

Eligible patients were identified by providers or study staff using UTMDACC’s electronic medical record. After identification, a research coordinator met with potential participants during previously scheduled appointments or by phone to describe the study and assess interest. Physicians were required to sign a release form to indicate the participant was medically eligible. Enrolled participants provided written informed consent and completed baseline assessments including anthropometrics, serum collection, and questionnaires related to demographics, body image (Body Image Scale) and quality of life (Medical Outcomes Study (SF-36)).

Following baseline, participants were randomized to Project BALANCE or the usual care control group. A form of adaptive randomization called minimization, which is similar to stratification in that participant characteristics are used to assign them to the treatment conditions [30], was used to ensure group balance with respect to participant characteristics. Covariates considered in the randomization of participants to study condition included baseline BMI (median), age (median), disease stage (I / II), menopausal status (yes / no), treatment regimen (short interval/long interval), physical activity (meets standard guidelines / does not meet standard guidelines), and vegetable consumption (high / low as per questionnaire).

All participants completed questionnaires and anthropometric measurements at baseline (T0), mid-chemotherapy three months later (T1) and post-chemotherapy six months after baseline (T2). Serum was collected at T0 and T2 only.

Intervention conditions

Project BALANCE

Recognizing that women undergoing breast cancer treatment are at particularly heightened risk of having body image concerns [31–33], the intervention used a size acceptance-based approach. The size acceptance approach to weight management has shown effectiveness at maintaining weight and improving cardiovascular risk factors [34,35] and resonates well with middle-aged women [34]. It focuses on improving health through lifestyle change, and de-emphasizes weight loss. The goal of Project BALANCE was to provide information, increase behavioral capability, and enhance self-efficacy in four areas stressed by the size acceptance approach: body acceptance, eating behavior, nutrition, and activity. Project BALANCE materials focus on teaching women to disentangle feelings of self-worth from weight and to develop realistic expectations about body size and shape. Intervention materials include modeling of body acceptance through stories of women with breast cancer coping with the physical changes that result from treatment.

Participants received the intervention via 20 individual counseling sessions over approximately 6 months during chemotherapy (14 in-person and 6 by telephone). Participants attended one 60-minute session per week for 13 weeks, where they performed the resistance and flexibility exercises and received corrective feedback; reviewed the previous week’s food records; discussed progress and barriers related to exercise and diet goals; and set goals for next week. In-person appointments were typically scheduled on the same day participants received chemotherapy. Following the in-person sessions, participants received one weekly telephone session for the following 6 weeks of chemotherapy. Phone calls were scheduled with research staff at the participant’s convenience. A final in-person session was conducted at the end of chemotherapy. All sessions were conducted with masters-level health educators.

The dietary aspect of Project BALANCE focused on promoting foods with low energy density, which have been shown to increase satiety more than high energy density foods with the same caloric value [36–38]. This approach was selected as low energy dense foods are low in fat and high in water and fiber, thus emphasizing a diet consistent with published recommendations for survivors [39]. The intervention counseling sessions focused on developing skills for internal regulation and mindful eating. Study participants were asked to keep diet records on at least one weekday and one weekend day to keep track of food eaten, time eating, and mood while eating during the entire intervention.

The exercise aspect of the intervention focused on resistance and flexibility training to maintain or increase lean body mass. Each participant was prescribed tailored resistance and flexibility exercises for the major muscle groups based on a baseline strength assessment. Lighter weight and more repetition rather than heavier weight and less repetition was emphasized during exercise sessions. Participants were asked to perform the exercises 3 times per week including the weekly supervised session. Exercise sessions targeted major muscle groups and generally proceeded from core muscle groups (torso), to large muscle groups (quadriceps, semitendinosis, biceps femoris, gastrocnemius, soleus, biceps, triceps) and then with smaller muscle groups (brachoradialis, deltoids, shoulder girdle muscle, etc). Flexibility exercises included all joint areas but focused mainly on areas being resistance trained and areas with less range of motion. Specific flexibility exercises were prescribed based on assessment results, resistance training progress, and individual needs.

Participants were asked to maintain levels of aerobic activity. Baseline levels of activity were established with a pedometer during a one-week baseline period. Participants were then asked to continue wearing the pedometer and recording steps, with the goal of maintaining the baseline number of steps per week throughout the six months of chemotherapy. Because fatigue or other treatment-related symptoms may cause patients to decrease usual occupational and household activities, participants were encouraged to take brief moderate intensity walks to maintain baseline activity levels and avoid transitioning into sedentary activity patterns.

Project BALANCE also included written materials (including newsletters and instructions developed specifically for this study plus the Volumetrics cookbook [40]) and exercise equipment (hand-held weights, steps, stability balls and pedometers). The written materials incorporated Social Cognitive Theory constructs to provide information and skills necessary to support behavior change. For example, the materials included photos illustrating how to properly perform exercises, charts for monitoring progress, low-energy-density recipes, recommended shopping lists, and tips on choosing foods when eating out. In addition, they included stories about other cancer survivors who had made positive diet and activity changes in their lives, to increase self-efficacy through modeling. To encourage size acceptance, the materials included illustrations and photographs of people with a range of body sizes, and stories on developing a healthy body image and coping with appearance changes after breast cancer diagnosis. The materials also provided information on behavioral skills such as overcoming barriers, rewarding yourself, and time management.

Usual care

Participants received publically available written materials on cancer survivorship only. These materials address concerns of breast cancer survivors and provide basic information about appropriate diet and exercise during chemotherapy, but do not provide individually tailored recommendations or behavior change skills.

Measures

Feasibility and Acceptability were assessed by attrition (remaining in the study), adherence to intervention (completing study sessions), and satisfaction (based on self-reported satisfaction at T2 and likelihood of recommending the intervention to a friend).

Questionnaires included a demographic questionnaire covering questions on race, medical history, age, education, employment, marital status, and household. Body Image was assessed using the self-report Body Image Scale (BIS). The BIS is a 10-item questionnaire measuring body image concerns resulting from cancer and cancer treatment, with higher scores indicating more body image distress; it has demonstrated validity and reliability among cancer patients [41]. Health-related quality of life was assessed using The Medical Outcomes Study (SF-36) US Standard Version 1.0. Internal consistency reliability is high for the SF-36’s 8 scales (physical functioning, role functioning, bodily pain, general health, vitality, social functioning, role emotional and mental health) that range from 0 (lowest level of functioning) to 100 (highest level of functioning) [42]. Questionnaires were administered to participants during in-person assessments. If participants needed additional time to complete the questionnaires they were allowed to take them home and return by mail or at a subsequent appointment.

Anthropometrics included height and weight measured in light clothing and without shoes at in-person visits using an electronic scale and stadiometer. If a participant missed an assessment, their weight was obtained from the medical record as measured at the clinic visit closest to the scheduled assessment. Waist circumference was measured at the narrowest part of the torso using a Gulick flexible tape.

Biomarkers were assessed from serum samples collected after an overnight fast and stored in a −80-degree freezer until analyzed. Quantitative enzyme-linked immunosorbent assay (ELISA) was conducted using various kits, which included Estradiol Ultrasensitive (20-ESTHUU-E01; Alpco Diagnostics, Salem NH), Sex Hormone Binding Globulin (SHBG; 11-SHBHU-E01; Alpco Diagnostics, Salem NH), IGF-1 (11-SHBHU-E01; Alpco Diagnostics, Salem NH), IGFBP-3 (DGB300; R&D Systems, Inc, Minneapolis, MN), Leptin (EZHL-80SK; EMD Millipore, St. Charles, Missouri), and Adiponectin (EZHADP-61K; EMD Millipore, St. Charles, Missouri). Results were quantified according to manufacturer’s guidelines using a Wallac VICTOR3™ Multilabel Plate Reader (PerkinElmer, Shelton, CT).

Statistical analysis

Because this was primarily a descriptive and exploratory feasibility study, a power analysis was not conducted. Given the developmental nature of the study as well as the small sample size, we did not expect to have sufficient power to detect a small or moderate effect for anthropometric or biological outcomes. Instead, findings from this study will be used for justification of further investigation on the weight gain prevention program; and the estimation of effect sizes may be helpful for planning future larger studies. Baseline characteristics of the sample were compared between participants in the intervention and usual care groups using Fisher’s exact and t-tests. Participant attrition was calculated as percentage of total randomized participants that remained in study (did not withdraw). Adherence to intervention was determined by examining the percentage of total participants that completed study sessions.

Secondary study aims were explored primarily using intent-to-treat analysis with worst observation carried forward (WOCF) imputation for all participants. As such, missing values for waist circumference or weight were substituted with a highest measurement across the time points. A sensitivity analysis using observed data only (no imputation) was also conducted. Differences between participants in the intervention and usual care groups for weight, waist circumference, quality of life, body image, and levels of biomarkers were assessed using linear mixed models. Linear mixed models with random intercepts were used, which allowed for subject-specific response trajectories and accounted for the correlations among observations arising from the same individuals over time.

Biomarker data were examined for all participants with WOCF imputation. Changes in biomarkers from baseline (T0) to end of chemotherapy (T2) were assessed using paired t-tests in each group. All tests were conducted at a two-sided significance level of .05. Statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).

RESULTS

Participants and group comparability

Table 1 provides the characteristics of the randomized sample. The sample was predominantly white or African-American, with high levels of education. The majority of participants were overweight or obese, and just over half were post-menopausal at diagnosis. The most common treatment regime included 12 weekly cycles of paclitaxel followed by 4 cycles (every 21 days) of 5-fluorouracil, Adriamycin/epirubicin, and cyclophosphamide (FAC or FEC) with or without trastuzumab. There were no significant group differences with regard to demographic or clinical characteristics, body composition, patient-reported outcomes, or biomarkers at baseline. Figure 1 depicts the study CONSORT diagram highlighting the percentage of eligible patients who agreed to participate (42%) and the percent completing each assessment.

Table 1:

Characteristics of the sample

Characteristic	Intervention (n=19) % (N)	Control (n=18) % (N)	aP-value
Race/ethnicity
African-American	26.3% (5)	16.7% (3)	0.863
Hispanic, white	0% (0)	0% (0)
White, non-Hispanic	52.6% (10)	66.7% (12)
Asian	10.5% (2)	5.6% (1)
Other	10.5% (2)	11.1% (2)
Education
Less than high school	5.3% (1)	5.6% (1)	0.903
High school degree	5.3% (1)	5.6% (1)
Some post-high school	42.1% (8)	38.9% (7)
Bachelor’s degree	31.6% (6)	44.4% (8)
Advanced degree	15.8% (3)	5.6% (1)
Marital status
Single, never married	26.3% (5)	5.6% (1)	0.154
Married/in committed relationship	47.4% (9)	72.2% (13)
Divorced / Widowed	26.3% (5)	22.3% (4)
Baseline BMI
<25	21.1% (4)	27.8% (5)	0.913
25–29.9	26.3% (5)	27.8% (5)
≥30	52.6% (10)	44.4% (8)
Menopausal status
Premenopausal	42.1% (8)	44.4% (8)	>.999
Postmenopausal	57.9% (11)	55.6% (10)
Estrogen Receptor Positive	57.9% (11)	58.8% (10)	.995
Progesterone Receptor Positive	52.6% (10)	35.3% (6)	.335
HER2 Neu Positive	52.6% (10)	35.3% (6)	.219
Triple Negative (Yes)	10.5% (2)	11.1% (2)	.481
Disease Stage
II	73.7% (14)	72.2% (13)
III	26.3% (5)	27.8% (5)	>.999
Chemotherapy Protocol
Paclitaxel + FAC/FEC	79.0% (15)	66.7% (12)	.468
Paclitaxel + FAC/FEC + trastuzumab	10.5% (2)	27.8% (5)
Other	10.5% (2)	5.6% (1)
Radiation
No	15.8% (3)	33.3% (6)	.269
Yes	84.2% (16)	66.7% (12)
Additional Cancer Dx (other than breast)
No	89.5% (17)	88.9% (16)	>.999
Yes	10.5% (2)	11.1% (2)
	Mean (SD)	Mean (SD)
Age	49.6 (13.3)	49.2 (9.2)	.925

^aFisher’s exact tests (two-sided) were used to examine associations between intervention groups and population characteristics for categorical variables. For continuous variables, t-tests were used to test for differences between the intervention groups.

Feasibility and Acceptability

Patient retention in the intervention group was 100%, with no participants withdrawing from study for any reason. However, some participants failed to complete all study measures, with 5% (n=1) and 16% (n=3) missing assessment sessions at mid-chemotherapy and post-chemotherapy time points, respectively. Intervention participants attended an average of 80% of the in-person sessions and 43% of the telephone sessions during chemotherapy (Figure 2). In the control group, 5% (n=1) withdrew from the study; 19% (n=4) and 38% (n=7) missed the mid-chemotherapy and post-chemotherapy assessments, respectively. The intervention group reported high levels of satisfaction; 88% (n=8) of participants reported at T2 that they were satisfied with Project BALANCE, and 88% (n=8) said they would be likely to recommend it to a friend. Two optional open-ended questions were included on final surveys. Responses to the question “What did you like about the study” included the study materials and information (n= 3) and the study staff (n=1). In response to the question “What changes would you recommend?” participants noted Spanish materials (n=1), improved questionnaires (n=1) and a fully equipped and staffed gym (n=1).

Fig 2:

Scheme depicting intervention session attendance and transition from weekly treatment (12 cycles paclitaxel) to treatment every 3 weeks (4 cycles 5-fluorouracil, Adriamycin/epirubicin, and cyclophosphamide (FAC or FEC) with or without trastuzumab) among Project BALANCE participants.

Preliminary Effect on Anthropometrics

Table 2 shows estimated intervention effects on anthropometric and quality of life variables. Using intent-to-treat analysis, the intervention group weighed less across T1 and T2 time points compared to control participants but between group changes due to the intervention were not significant (p = .083). Intervention participant weight at baseline (mean = 76.0kg +/− 15.5) was reduced by mid-chemotherapy (73.5kg +/− 15.4) with some weight regain seen at the end of chemotherapy (74.8kg +/− 16.9). This trend also held true for the control group with weight fluctuating slightly from baseline (76.9kg +/− 15.5) to mid-chemotherapy (75.4kg +/− 15.0) to end of chemotherapy (75.8kg +/− 15). This finding is shown in terms of percent body weight lost in Figure 3. Waist circumference was significantly lower across time points T1-T2 in the intervention group compared to control (p= .019). Estimated intervention effects were similar using observed measures only (no imputation) versus the WOCF approach for weight (−2.66 (SE = 1.26) vs −2.01 (SE = 1.15)) and waist circumference (−3.03 (SE = 1.52) vs −3.27 (SE = 1.37)).

Table 2:

Intervention effects on weight, waist circumference, and quality of life and body image

	Baseline (T0) Mean(SD)	Change T1-T0 Mean(SD)	Change T2-T1 Mean(SD)	Adjusted Mean (SE)b	Estimated Intervention Effect (SE)a	P-value
Weight (kg)					−2.01 (1.15)	0.083
Intervention	74.23 (15.54)	−0.5 (3.65)	−2.59 (2.88)	73.53 (0.83)
Usual Care	75.61 (15.26)	1.32 (1.8)	−1.54 (2.32)	75.54 (0.78)
Waist Circumference (cm)					−3.27 (1.37)	0.019*
Intervention	89.27 (13.59)	−1.4 (5.08)	−0.51 (4.64)	86.66 (1.00)
Usual Care	87.13 (12.15)	2.21 (4.88)	−0.41 (2.39)	89.94 (0.94)
Quality of Life (SF-36) c
Physical Functioning					4.46 (7.50)	0.554
Intervention	66.33 (33.51)	−5.77 (28.78)	10 (17.16)	65.63 (5.08)
Usual Care	80.76 (19.69)	−21.07 (25.73)	5.2 (22.61)	69.13 (5.35)
Mental Health					−5.32 (3.77)	0.163
Intervention	67 (18.45)	.29 (17.64)	−2 (14.39)	73.66 (2.66)
Usual Care	75.11 (17.76)	4.57 (10.74)	3.64 (6.31)	78.98 (2.56)
Vitality					14.14 (4.84)	0.005**
Intervention	41.56 (17.96)	.48 (16.74)	7.83 (19.55)	53.66 (3.24)
Usual Care	57.5 (19.72)	−22.14 (17.94)	10.91 (18.82)	39.53 (3.19)
Role Physical					9.38 (9.38)	0.321
Intervention	40 (40.97)	9.62 (42.74)	0.83 (46.22)	46.28 (6.85)
Usual Care	45.83 (46.38)	−16.07 (73.78)	4.55 (47.19)	36.91 (6.45)
Bodily Pain					5.42 (5.60)	0.337
Intervention	63.93 (34.36)	.08 (28.41)	2.5 (35.92)	64.42 (4.11)
Usual Care	63 (22.9)	−7.57 (21.87)	1.18 (19.18)	59.00 (3.83)
General Health					4.93 (5.09)	0.337
Intervention	61.83 (18.15)	−.59 (15.76)	−0.2 (17.4)	68.63 (3.56)
Usual Care	69.61 (16.39)	−11.3 (25.16)	6.93 (24.65)	63.70 (3.48)
Social Functioning					2.66 (5.51)	0.631
Intervention	55 (27.87)	4.81 (18.07)	12.5 (31.73)	70.81 (3.93)
Usual Care	70.14 (28.81)	−10.71 (21.85)	13.64 (19.73)	68.15 (3.70)
Role Emotional					0.27 (11.92)	0.598
Intervention	52.22 (39.27)	19.23 (47.06)	10 (35.31)	69.54 (8.65)
Usual Care	66.67 (41.22)	2.38 (56.18)	0 (42.16)	69.28 (8.03)
Body Image Scale (BIS) d					−3.00 (2.23)	0.184
Intervention	29.63 (6.83)	−.21 (1.47)	−1.03 (1.1)	8.34 (1.60)
Usual Care	28.62 (5.62)	.5 (.71)	−.6 (.91)	(1.56)

^a)Effect of intervention across the study period were analyzed using linear mixed models with random intercepts. The estimated effect represents average change in the outcome variable due to intervention across the study period, adjusting for the baseline level. Standard errors (SE) of the estimates are provided in parentheses.

^b)Represents the expected value of the outcome across the study period estimated by the linear mixed model, for each intervention groups adjusting for the baseline level. Standard errors (SE) of the estimates are provided in parentheses.

^c)Higher score on the SF-36 sub-scales indicates less disability.

^d)Higher score on the BIS indicates more unfavorable body attitude

^*p < .05

^**p <.01

Fig 3:

Graph depicting percent weight loss over the study period by group

Changes in Quality of Life and Body Image

Exploratory effects of the intervention on quality of life and body image are shown in Table 2. The intervention group reported significantly greater improvements on the SF-36 vitality subscale than control participants at the end of chemotherapy (M=53.66, SE=3.24 versus M=39.53, SE=3.19, p=.005). Differences in other SF-36 scores and the BIS were not significant.

Changes in Biomarkers

There were no significant differences in changes in biomarkers between the intervention and control group (Table 3). The intervention group had significant increases in SHBG (p<.001) from baseline to T2. The control participants had significant increases in insulin (p=.004) from baseline to T2. Only a small number of participants had estradiol levels that were detectable using the ELISA. Therefore, estradiol differences between groups were not explored.

Table 3.

Change in biomarkers from baseline to end of chemotherapy N=37

	Within group change			Between group change
Biomarkera	Mean (SD)	T	P-value	T	P-value
SHBG
Intervention	0.29(0.18)	6.02	<.001**	1.78	0.087
Usual care	0.1(0.39)	1.11	0.281
Estradiol b
Intervention	−0.59(1.6)	−1.23	0.248	—	—
Usual care	—	—	—	—	—
IGF-I
Intervention	0.1(0.41)	0.93	0.367	−0.24	0.810
Usual care	0.14(0.35)	1.68	0.111
IGFBP-3
Intervention	0.08(0.33)	0.88	0.393	1.04	0.306
Usual care	−0.05(0.34)	−0.58	0.571
Insulin c
Intervention	0.14(1.05)	0.50	0.623	−1.23	0.233
Usual care	0.54(0.69)	3.32	0.004**
Glucose c
Intervention	0.07(0.24)	1.08	0.299	0.59	0.557
Usual care	0(0.37)	0.05	0.961
Adiponectin
Intervention	−0.17(0.31)	−1.76	0.109	0.73	0.496
Usual care	−0.33(0.46)	−1.6	0.186
Leptin
Intervention	0(0.18)	0.05	0.961	−1.18	0.266
Usual care	0.1(0.15)	1.59	0.187

^a)all biomarkers studied were natural log transformed

^b)Estradiol values were all zero for the usual care group for all time points. Only one participant in the usual care group reported still having menstrual cycles.

^c)Glucose and Insulin were measured in mmol/L and pmol/L

^*significant at p< .05

^**significant at p< .01

DISCUSSION

This randomized pilot study indicates that a size acceptance-based weight management program delivered during neoadjuvant chemotherapy merits further investigation in a larger, fully-powered trial. Preliminary analysis shows the intervention is acceptable to participants, can prevent weight gain and even promote modest weight loss, and has a positive effect on patient vitality and waist circumference among breast cancer patients. This pilot study adds to the small but growing literature indicating weight control interventions are feasible and beneficial for breast cancer patients during treatment [28].

Project BALANCE participants had high attendance at in-person sessions during chemotherapy, but lower adherence to telephone sessions, potentially due to their treatment schedule. The majority of the patients were receiving 12 weekly cycles of paclitaxel followed by 4 cycles (every 21 days) of FAC or FEC. Most of the in-person sessions were conducted when participants were coming into the hospital weekly, with Project BALANCE appointments typically scheduled on the same day they received chemotherapy. Once the patients transitioned to FEC/FAC, they came to the hospital less frequently and also reported more side effects, potentially making adherence more difficult (Figure 2). A review of weight gain prevention studies among breast cancer patients noted successful interventions were more likely to use frequent physical contact with participants [28]. Because attendance was greater for in-person counseling sessions delivered at the time of chemotherapy appointments, future iterations of Project BALANCE should consider concentrating intervention sessions during the in-clinic phase of treatment to optimize adherence. Although adherence to telephone sessions was lower than in-person sessions, more digitized intervention delivery methods, such as text messaging, mobile apps, and social media groups, may support participant engagement during the transition to home-based care [43].

Although not statistically significant, intervention participants weighed less than control group participants at the end of chemotherapy and lost a larger percentage of body weight over the intervention period. Beyond weight, our study demonstrated a significant effect on waist circumference. An elevated waist circumference, which may indicate greater visceral adiposity, is associated with more adverse health outcomes, such as cardiovascular disease, diabetes, and colorectal and post-menopausal breast cancer [44–46], and central adiposity has been shown to be related to poorer overall and disease-specific survival in breast cancer survivors [47].

These anthropometric results are comparable to other post-treatment weight loss interventions for breast cancer survivors. One systematic review [27] identified 10 randomized trials and four single-arm trials of weight loss interventions for breast cancer survivors. Thirteen out of 14 trials showed statistically significant weight loss, with losses ranging from 1.1 kg to 11.6 kg. Another smaller review of weight gain prevention interventions during active breast cancer therapy, noted similar results to the current study. In that review, five completed interventions and five ongoing trials were included. Of the seven intervention arms examined, two achieved weight losses, and two reported successful weight maintenance [28]. The authors concluded weight gain prevention interventions during chemotherapy were feasible and generally effective at preventing weight gain among breast cancer patients. Given the difficulty most people have with maintaining weight loss over time [48], prevention of primary weight gain may be a more effective strategy for this population [28].

The size-acceptance approach was a unique aspect of this study. A recent systematic review examined the impact of ‘health, not weight loss, focused’ weight loss interventions compared to conventional weight loss programs on cardiovascular disease risk factors and body weight among overweight and obese adults. The meta-analysis of eight studies found that although mean weight loss was greater in the conventional weight loss groups at the end of treatment, those in the ‘health, not weight loss, focused’ group had more favorable improvements in weight after a longer follow up period (53 to 104 weeks) [35]. Future iterations of Project BALANCE should consider longer follow-up periods in order to better understand the maintenance of effects.

Project BALANCE had a positive impact on fatigue, as indicated by the vitality subscale of the SF-36, a validated measure of cancer-related fatigue [49]. This finding is in line with previous studies. Reeves et al demonstrated a significant improvement in fatigue after participation in a telephone-based diet and exercise study among women that recently completed breast cancer therapy [50]. Fatigue is a highly prevalent side effect of chemotherapy that can significantly impair quality of life and functioning, and has been associated with higher BMI among breast cancer survivors [51–53]. The integration of other measures of fatigue and related conditions, such as sleep quality, into a future randomized trial may help further elucidate the impact of weight loss interventions on this important side effect of cancer therapy.

This study has several strengths including the use of a body size- acceptance model and integration of multiple outcome measures of adiposity, biological markers and quality of life. Project BALANCE focused on healthy behavior change and emphasizing positive body image instead of highlighting weight loss. It is worth noting that the intervention appeared to have a small positive impact on body image, and this outcome warrants further exploration in future, fully powered trials.

This study had several limitations. The small sample was overall highly educated, and therefore may not be representative of the general breast cancer population. Additional research in a community hospital serving a more diverse population may be warranted. This study was not sufficiently powered to determine effects on biological, anthropometric or self-reported participant outcomes, although many areas for potential exploration in future studies were identified. Waist circumference may be subject to measurement error. These limitations are consistent with the extant literature on weight gain prevention among breast cancer patients on active therapy, which is generally limited to pilot studies with small sample sizes and little post-intervention follow-up.[28] Fully powered, long term trials are needed to test the preliminary findings of this emerging area of research.

CONCLUSIONS

Our findings indicate that implementing a weight gain prevention intervention among breast cancer patients during treatment is feasible and that using a size acceptance-based approach was well-received. Preliminary efficacy findings further suggest positive effects on anthropometric and quality of life outcomes that could prevent or limit future weight gain and help patients manage treatment-related symptoms. Future research is warranted to test similar interventions, particularly those that leverage the frequency of physical patient interactions occurring during treatment, to address weight control in this population.