Effect of Combined Training on Body Image in Women with Breast Cancer: Controlled Clinical Trial

Background: Improvement in body image in women with breast cancer may be linked to altered functional capacity, regardless of body composition. The aim of this study was to evaluate the effect of combined training (CT) on body image in patients with breast cancer. Methods: A Clinical Trial study, this was a controlled study including 26 patients with breast cancer (30 to 59 years). The patients were allotted to either the Training Group (TG)(n = 13), which underwent 12 weeks of CT (aerobic, resistance, and exibility), or the Control Group (CG) (n = 13), which received only the standard hospital treatment. Participants were evaluated at baseline and after 12 weeks. Results: The patients in the TG showed a signicant reduction in the limitation dimension (p = 0.036), and an increase in VO2max (p = < 0.001) and strength in the right (p = 0.005) and left arms (p = 0.033). However, the participants also presented an increase in waist circumference (p = 0.034). Conclusion: Women with breast cancer presented improved body image and functional ability through CT.


Functional Capacity
To verify the functional capacity, strength and maximum oxygen volume (VO2max) were considered.
The static force was evaluated using a manual dynamometer (Jamar Sammons Preston) scale from 0 to 100 kilograms. Guidance was provided to press the equipment with maximum force, without exing the elbow or changing the posture. Three attempts were allowed on both sides (alternately) [29].

Co-variables
Anthropometric, demographic and clinical aspects Anthropometric measurements were recorded for body mass (kg), height (cm), and age (years) [29]. The level of education, family income (minimum wage value = R$ 788.00), employment status, marital status, period without physical training practice (eligibility criteria), treatment and follow-up phase, type of neoplasia, and patient surgery were veri ed by means of anamnesis.

Intervention
The CT program consisted of aerobic, resistance, and exibility exercises lasting 12 weeks, with 3 sessions per week of aerobic and resistance training in the same session (supervised by trainers specialized in physical exercise) and an additional 2 sessions per week of exibility training.
The exibility sessions lasted for 20 minutes and were intercalated between sessions of aerobic and resistance training in the daytime period. Familiarization with exibility occurred over two weeks, with three sessions per week, where each patient was followed to ensure the correct execution of all stretching exercises.
Each aerobic and resistance training session lasted 60 minutes, following the order: 30 minutes on cycle ergometer, hip exion and extension, shoulder development, Swiss ball squatting, French triceps, and curved paddling.

Familiarization
Familiarization to the aerobic and resistance training was performed over one week in three sessions, with a cycloergometer load of 15 watts, and the individual's body weight and an elastic band in the execution of 8 to 12 repetitions in the resistance exercises with one minute of recovery between one exercise and the next.

Aerobic training
The aerobic training was controlled by the training heart rate (THR) using the formula: THR= (Y x (A-B) + B) [30], if Y:% of the desired effort, A: Expected maximum heart rate (measured by cardiorespiratory test) and B: Resting heart rate (measured at rest).
In the cardiorespiratory test, the ramp protocol adapted by Neil et al. [31] was used on a cycle ergometer (ERGO FIT brand, model ERGO 167-FITC CYCLE), with an initial load of 15 watts during 5 minutes of warm-up, followed by incremental stages of 60 seconds with the addition of 15 watts each stage [31]. After the maximum stage reached, an active recovery of 3 minutes with the initial load was performed, the stages had from 70 to 90 rotations per minute. Blood pressure was measured with conventional mercury column apparatus (BD®), heart rate (Polar FT2) and subjective perception of exertion using the Borg scale (PSE) (Infor sic Mark) in the nal 15 seconds of the stages. Before and after the cardiorespiratory test, the patients remained seated at rest to verify the Borg subjective perception of exertion, PA, and HR. The test was performed after a 72-hour interval of familiarization.

Resistance
The resistance training protocol included 3 sets for each exercise with 12 repetitions and a one-minute interval between sets and repetitions. The speed of execution of each movement was three seconds in the concentric phase and three seconds in the eccentric phase [32]. The exercises were alternated by segment, prioritizing the large muscle groups. The loads were by means of shin guards, dumbbells, elastic bands, and the weight of the body itself.
The resistance training load was veri ed by means of the maximal repetition test, with 12 repetitions and a 72-hour interval of familiarization [33]. Patients who exceeded 12 repetitions were given a 5-minute interval before performing the 12 repetitions with a new load.

Flexibility
The exibility training was active (greater range of motion performed in a contraction of the agonists and relaxation of the antagonists), without pain, where each exercise lasted 20 seconds in 3 series [34]. The exibility exercises were: 1) Adduction of the shoulder with extension of the elbow, bilateral; 2) Shoulder and elbow exion with palm on the back, bilateral; 3) Wrist exion; 4) Wrist Extension; 5) Hip abduction with exed knees; 6) Sitting, hip exion with shoulder adduction and elbow exion; 7) Sitting, legs stretched touching the feet; 8) Sitting, legs extended and crossed touching the feet, bilateral; 9) Shoulder exion and adduction with hands joined in front of the trunk; 10) Back bending at the foot of the wall.

Weight progression
The load progressions were performed every 4 weeks, respecting the biological individuality in the cardiorespiratory capacity test and maximal repetitions to predict the initial load [34]. The initial intensity of the aerobic training was 50 to 60% of the THR, ending with 80 to 90% of the THR. The load of the resistance training started with the weight of the body itself or 1 kg in dumbbells and shin guards, and moderate intensity in the elastic band. In the fth week there was an increase to 1 kg and strong intensity in the elastic band, remaining until the twelfth week (Table 1).

Blind study
The evaluations for both the TG and CG were performed at baseline and after 12 weeks, corresponding to the intervention period. The research team was trained to apply each instrument and perform the tests, and the physical evaluations were performed blindly by the evaluator, who was only informed of the day and time of evaluations.

Calculation of the sample size
The sample size was calculated using statistical software (G-power 3.1 Düsseldorf, Germany), which showed that twelve participants were needed, through a medium effect size of 0.50 [35], α error probability of 0.05, power (1-β error probability) of 0.8, 0.5 correlation between repeated measure analysis (repeated measure ANOVA within-between interaction), with a nonsphericity correction of one for two groups, and two analyses throughout the study.

Statistical Analysis
Data normality was checked using the Shapiro-Wilk and Kolmogorov-Smirnov tests. The comparisons of the baseline variables between groups were analyzed using the Student t test for independent samples when parametric statistics were observed. If the data presented non-parametric distribution, the Mann-Whitney and dichotomous variables χ 2 and Fisher's Exact test were used.
The differences between the combined training and control groups were analyzed by two-way repeated measures of ANOVA (group x time). When a signi cant interaction was observed, a Bonferroni post hoc test was conducted. For all measured variables the estimated sphericity was veri ed according to Mauchly's W test and the Greenhouse-Geisser correction when necessary. The partial eta-squared was classi ed according to Cohen [35]. Statistical signi cance was set at p < 0.05. The data were analyzed using the Biostatistic (Statsoft, version 10) and Stata 14.0.

Page 8/27
The original patient records contained 300 patients who could be contacted, of whom only 31 were available and participated in the sample. However, one patient in the TG was excluded due to a diagnosis of a mental disorder and another because she did not complete the physical evaluation. In the CG, one patient died and two did not complete the physical evaluation. Thus, twenty-six patients (13 TG and 13 CG) completed the study (Fig. 1).
The TG and CG patients did not present statistical differences in the anthropometric, social, demographic, and clinical variables, which presented homogeneity between the groups (Table 2). Regarding the clinical aspects, only one patient in the TG had a bilateral mastectomy and patients of neither group performed breast reconstructions. The body image of the patients who performed CT presented a reduction after 12 weeks for the limitations dimension (f = 4.931, p = 0.036), in addition to an interaction between time and group (f = 10.289, p = 0.003). In the concern with the body dimension, there was also a signi cant interaction (f = 6.306, p = 0.019), however, with only a tendency for TG reduction (Δ = 5.23, f = 1.876, p = 0.183). In the transparency dimension, the groups presented signi cant differences (f = 8.015, p = 0.009). The other dimensions of body image did not present signi cant differences (Table 3). Referring to the interaction between time and group.
In The other variables of body composition and functional capacity did not present signi cant differences (Table 4).  and moderate for the CA(0.019) and T dimensions (p = 0.035)( Table 5). The concern with the arm dimension in the CG presented different correlations with the other dimensions, such as: a positive and strong correlation with the waist-hip ratio (p = 0.011) and fat-free mass (p = 0.011), and negative correlation with the volume of oxygen (p = 0.018) in the basal period ( Table 5).
The TG did not present any correlations between variables of functional capacity and body image. On the other hand, body composition variables, speci cally the body mass index (p = 0.027), weight (0.021), waist-hip ratio (p = 0.018), waist circumference (p = 0.026), and fat-free mass (p = 0.011) were negatively correlated with vulnerability, in the pre-training, but gained statistical signi cance only after the intervention.

Discussion
The main ndings of this research can be summarized as follows: 1) CT promoted improvements in functional capacity, but not in body composition in TG women; 2) CT favored positive changes in body image; 3) Variables related to appearance were correlated with the vulnerability dimension in TG; 4) Variables related to body appearance and function were directly related to body image in the CG. The latter was our most surprising nding, particularly the waist circumference, which appears to be an important marker in the body image of women with breast cancer not submitted to physical activity.
The effect of CT was con rmed by increased arm muscle strength and improvements in VO2max. These results corroborate studies in the literature which submitted women with breast cancer to protocols composed of aerobic and strength exercises and found similar results [20][21][22]. CT prevents the physical deconditioning inherent in cancer treatment [22]. These variables enhance the perception of this population of improvements in their quality of life [18]. De Luca et al. [21] further point out that CT promotes greater adherence to physical activity, due to the diversi cation of exercises.
However, in the current study, there were no positive changes in body mass, percentage of fat, BMI, or other anthropometric indices in either group. It is worth noting that, despite this, the maintenance of these variables already indicates a good follow-up, since the disease and its treatment promote negative changes in body composition [7]. Similar results were found in women with breast cancer submitted to strength training [18], aerobic [17] and combined (strength and aerobic) protocols [20]. On the other hand, De Luca et al. [21] observed a reduction in fat percentage after intervention with CT in the same population. This difference may be attributed to the superiority in the intervention time (24 weeks) and method of analysis (bioimpedance), performed by the researchers. However, the multi-frequency electrical bioimpedance analysis method indicates greater precision when performed in a segmental way, due to the morphological variation in the tissues [36].
On the other hand, there was an interaction between time and group for the right thigh and changes in fatfree mass and waist circumference after 12 weeks. Despite the maintenance of these variables, promoting health through physical training, there was an increase in the waist circumference (WC) of both groups, which shows a negative trend and can be attributed to the cancer treatment. According to Oliveira et al. [37] and Figueiredo et al. [38] there is a large incidence of high WC in patients with breast cancer, which is linked to cardiovascular risk; these authors suggest that this population requires adherence to a nutritional program. However, the study by Kim et al. [39] demonstrated a reduction in WC after a 12-week CT intervention, this difference may be due to the stage of the treatment, since all the patients were survivors of breast cancer. Further studies are needed to de ne the type of intervention effective in the reduction of visceral fat in patients in the treatment of breast cancer.
Another central axis of this research is the investigation of body image changes. This was analyzed from the theoretical model with six dimensions [23] and validated for Brazilian women with breast cancer [5].
Three dimensions demonstrated sensitivity to CT: limitation, concern for the body, and transparency. After the intervention, the TG demonstrated a signi cant reduction in the perception of functional limitations of the body, such as movement restrictions and oncologic fatigue. Arab et al. [16] in one of the few studies in this scope performed in Brazil, presented similar results, although the protocol applied by them was different, with only resistance training over 12 weeks. The authors attribute the improvement to the higher physical competence acquired for the performance of motor tasks. Unlike our data, Arab et al. [16] did not nd training effects in the other dimensions [23]. It is possible that this difference is associated with the speci city of the intervention given.
Concern with the body is a striking feature in women with breast cancer, which in Brazilian women (who culturally already have this concern) [40], is accentuated either through the chemotherapy or mastectomy process [41]. As we hypothesized, the TG showed a tendency to reduce concerns with the body, that is, with their general appearance, including concern about the gain or loss of weight. Previous studies have demonstrated similar results after both strength training [19] and aerobic training [17], in which improvements in the perception of body appearance and lower concern with weight were detected, respectively.
Concern with appearance is related to the alterations promoted by the disease and treatment, which may be less or more visible [23]. Issues relating to concern with how obvious the disease is (expressed by changes in appearance) were denominated transparency by the authors. This variable was signi cantly different between the groups, so that women who did not receive the intervention with the CT presented higher scores in this dimension. We did not nd similar studies that addressed this point, which limits our discussion. However, a qualitative research with Latina women with breast cancer, identi ed that the acceptance of changes in appearance is considered a central axis in body image [11]. The authors encourage the development of intervention strategies that favor the acceptance of appearance during and after treatment. Our results suggest that CT may be of potential assistance. This becomes more consistent when we observe the gross scores of all analyzed dimensions, and nd that, although there is no statistical signi cance, there is an increase in the CG and reduction in the TG, indicating a tendency to reduce body image impairment with the practice of CT [5].
In the same direction, the next question was to identify whether improvements in body image could be attributed to changes in appearance and/or functionality as a result of CT. This hypothesis was partially rejected because the TG did not present a signi cant correlation between body image and functional capacity. On the other hand, the body composition, BMI, weight, WHR, WC, and fat-free mass variables, although not presenting signi cant changes in our sample, were negatively correlated with body image, speci cally with the vulnerability dimension, in the basal period, assuming statistical signi cance after the intervention. This fact leads us to re ect that the body experience with the CT may have directed the attention of these women to their body measurements, however, differently to women without breast cancer, since the literature indicates that BMI, WHR, WC, and fat-free mass are predictors of negative changes in body image, such as body dissatisfaction [42,40]. In our study, the opposite occurred, the higher these scores, the lower the feeling of invasion of the body by the disease, which may have consequently caused a lower sensation of vulnerability.
Unlike our results, Speck et al. [19]and Pinto et al. [17], after interventions with strength and aerobic training, respectively, in women with breast cancer, did not identify any variable of body composition and/or function capable of mediating the effect of training on the improvements found in body image. On the other hand, the positive effect on functional capacity, identi ed here and in the studies above, is pointed out by the authors as a factors that in uences body image, although indirectly. Speck et al. [19] explain that muscular strength provides bene ts to the general quality of life and this, in turn, mediates the intervention in the perception of the body. Pinto et al. [17] concluded that the improvement found in patients' self-assessment of their physical condition (energy, strength, and agility) is consistent with the increase in VO2max, thus indicating a re nement of the patient about her physical condition.
In contrast to the TG, the body image of the women who were not submitted to the intervention was in uenced, over time, as much by the variables related to appearance as by body functionality. Waist circumference, a variable commonly associated positively with female body dissatisfaction [40], precisely because it delineates female body forms, manifested itself in an opposite way for all dimensions of body image. Thus, the smaller this variable, the greater the perception of functional limitations, the concern with the arm and with the body, accentuating the feelings of vulnerability, visibility of the disease (transparency), and body stigma.
The percentage of fat, also considered a predictor of body dissatisfaction in women [40], especially for the lean body ideal [42], maintained this characteristic for the CG, in proportion to the Body Stigma, body concerns, transparency, and vulnerability. It is possible that this is due to the gradual and complex process of acceptance of changes in appearance from disease and treatment, requiring women to learn and deal with these changes [11].
Functional ability also presented an in uence on CG body image. The CG showed a negative correlation between arm muscle strength and the limitations and transparency dimensions (despite the absence of signi cance in the basal period) in all phases. Concerns with the arm were positively related (although not signi cant at baseline) to muscle strength, assuming statistical signi cance, but negative, after 12 weeks.
The opposite occurred between concerns about the arm and VO2max. Ohira et al. [43], when developing research similar to ours, hypothesized that women with breast cancer feel empowered psychologically as they become more physically effective. Although we cannot state that the bene ts of CT positively and directly impacted body image in the TG, the authors' idea applies in our results, since we observed that impairments in functional capacity were negatively associated with CG body image.
However, we believe that, unlike the CG, the TG may have bene ted from body experiences in the intervention, thus impacting dimensions which, although not evaluated herein, are indicated in the speci c literature as linked to this process: cognitive, affective, and behavioral [10]. Interventions with physical exercise can provide the sensation of regaining control of the body itself, which may translate into a greater sense of self-e cacy in other areas of life [43]. Thus, it is possible to infer that CT promotes subjective experiences that go beyond body appearance and function, although indirectly in uencing it.
Finally, we agree with the point of view of Speck et al. [19], in which the authors point out that the mechanism of impact of training on the concern with the body of women with breast cancer remains unclear. Thus, we encourage researchers to engage in research of this scope that seeks to identify whether CT can be considered a protective element for negative changes in body image in this population. In addition to this gap, we also recognize some methodological limitations of this research.
Initially, the sample size is restricted, which hampers generalizations of the results. The high exclusion of participants may negatively impact the results of randomized clinical trials, biasing the research. In addition, the history of physical exercise of non-eligible patients was not investigated, information that may be useful for understanding some results. Finally, the reduced number of articles with the same theme in Brazilian women affected by breast cancer made the discussion di cult, since cultural factors are extremely relevant to the body image constructs [10].
Although we analyzed important variables in the elaboration of the body image of this population, such as the type of surgery [13,4,44], a relevant point in this context is breast reconstruction, since it is known that women undergoing reconstruction are less dissatis ed with their bodies [41]. Accordingly, we recommend new studies that compare women with and without breast reconstruction and analyze the relationship of body image with body composition and functional capacity.

Conclusion
Combined training was shown to be a useful strategy capable of promoting improvements in the functional capacity and body image of women with breast cancer. We recommend its use together with conventional treatment. Thus, the data obtained here have theoretical and practical implications. The rst refers to the need to broaden the understanding of body image adaptations as a function of speci c physical changes in breast cancer, using a speci c tool and theoretical axis that considers body image as a multidimensional and independent construct.
We believe that this information could help in the delineation of facilitating factors, mediators, and protectors of body image in the treatment process, allowing the elaboration of adequate interventions.
Practical implications involve the use of this information in interventions that deal directly with the body, making them more assertive and e cient.