Association between Nutritional Status and Health-Related Quality of Life (HRQoL) in Women with Breast Cancer undergoing Chemotherapy: A Prospective Study

Purpose This study aimed to evaluate the association between chemotherapy (CT) time, nutritional status and the effect of the interaction between these variables on the symptoms and domains of Health-Related Quality of Life (HRQoL). Despite the adverse effects CT, it’s common the gain weight and unfavorable changes in the body composition of women during and after the treatment. And an inadequate nutritional status can negatively impact the HRQoL and can inuence the prognosis of breast cancer (BC). Methods This study was carried out with 55 women with BC over three periods of CT, T0, before the rst cycle; T1, intermediate period; and T2, after the end of CT. HRQoLwas assessed using the 30-Item Quality of Life Questionnaire (QLQ-C30) and its BC module, the Quality of Life Questionnaire Breast Cancer − 23 (QLQ-Br23). Results Overweight, waist circumference (WC) indicative of increased risk and substantially increased risk of metabolic complications, waist-hip ratio (WHR) indicative of risk of developing chronic diseases, waist-to-height ratio (WHtR) indicative of excess abdominal fat and longer CT times (T1 and T2), each separately, as well as the interaction of these anthropometric variables with CT time, were associated with worse symptom and HRQoL domains scores. Conclusion The course of treatment and a worse nutritional status, have negative effects on HRQoL, and it is relevant to encourage women with BC undergoing CT to adopt a healthy diet and to practice physical activity in order to maintain an adequate nutritional status, that could contribute to a better HRQoL. well-differentiated tumor (low grade); G2, moderately differentiated tumor (intermediate grade); G3, poorly differentiated tumor (high degree); ER, estrogen receptor; PR, progesterone growth factor cyclophosphamide; 5-uoracil, adriamycin cyclophosphamide; cyclophosphamide, methotrexate and 5-uorouracil.


Introduction
Breast cancer (BC) is the leading cause of cancer death in women [1] Among the treatment modalities for cancer is chemotherapy (CT) [2]. Recent data from our research group identi ed that CT negatively interferes with diet and perceptions related to food intake, as well as with the nutritional status and Health-Related Quality of Life (HRQoL) of women with BC [3][4][5][6]. HRQoL includes aspects directly related to illnesses,, health interventions and the impact of symptoms, disabilities or limitations on functioning and perception of well-being [7].
Despite the adverse effects of CT and dietary inadequacy, between 50 and 96% of women gain weight during the treatment, with a progressive increase in the months and years after diagnosis [8]. In addition, there may be a signi cant increase in the waist-hip ratio (WHR), a predictor of increased risk for metabolic syndrome [9]. Furthermore, there is the possibility of unfavorable changes in the body composition of these women throughout CT, with sarcopenia (loss of muscle mass), accompanied by increases of adipose tissue, representing a signi cant risk for the development of comorbidities and worse survival in the long term [8].
It is noteworthy that the body changes presented during CT are not easily reversed after the end of the treatment. On the contrary, many continue to gain weight [10] which is unfavorable, since excess weight, whether gained before or after cancer diagnosis, is a factor that negatively impacts quality of life, prognosis and survival of women with BC [11].
When compared to the population without BC, these women have worse HRQoL, with worse scores for global health status, functional and social capacity, in addition to more symptoms of fatigue and insomnia [12].
Studies indicate that quality of life during CT can be a predictor of the disease's prognosis [13,14]. Better physical capacity, emotional condition, global health status and lower incidence of nausea and vomiting are predictive factors for survival in women with BC [13]. Additionally, Lee et al. (2010) found that physical well-being and appetite were able to predict overall survival after adjustments, while worse physical well-being was associated with less objective tumor response, greater risk of weight loss and to non-hematological adverse effects in women with advanced metastatic BC. Evidence points, therefore, that quality of life may be related to the response to CT and the patient's survival [14].
Considering that CT can negatively impact nutritional status and HRQoL [15], the present prospective study aimed to assess the association between CT time, nutritional status and the effect of the interaction of these variables in the HRQoL domains. We hypothesize that women have a worse HRQoL when they have a worse nutritional status at a more advanced stage of treatment.

Study Design and Ethical Aspects
This is a prospective study conducted between August of 2014 and October of 2015, developed with women with BC undergoing CT at the Clinical Hospital of the Federal University of Uberlandia (HC/UFU). The follow-up time varied according to the CT regimen, which ranged between 4 and 6 months. The evaluations were performed at three moments during CT -T0, before the rst CT cycle; T1, CT intermediate period; and T2, after the end of CT. It is emphasized that some data were collected 21 days after each study time. This study was approved by the Ethics Committee on Research with Human Beings (no. 721,977/14, adhering no. 1,111,998/15) and complies with the principles and ethical standards of the Declaration of Helsinki and CNS Resolution 466/12.

Sample Size and Sample Eligibility
The sample size was statistically determined using the G*Power software, version 3.1 [16]. A total of 43 women were required, considering Anova F-Test of repeated measurements with intermediate effect size of 0.25, alpha level of 0.05, test power of 95%, being a group of individuals and three measurements. Considering a 20% adjustment for possible losses, a minimum of 52 women would be necessary at the beginning of the study (T0) [4]. The inclusion criteria were women aged 18 years or older, diagnosed with primary BC and who were in the rst CT cycle at T0. Women with a primary non-breast tumor site, under antineoplastic treatment that did not include CT, with change in treatment due to toxicity or recurrence of the disease, or who were ongoing CT were excluded.

Data Collection
Socioeconomic, therapeutic and clinical data were extracted from questionnaires completed from interviews and consultation of medical records to characterize the sample. The anthropometric parameters current weight and height were evaluated, following a protocol established by the WHO [17] and using a calibrated mechanical scale with a sensitivity of 100 g and a vertical statometer with an accuracy of 1 mm (model P-150C; Leader Scales, São Paulo, Brazil). The data obtained were used to calculate body mass index (BMI) in Kg/m 2 . In addition, waist circumference (WC) and hip circumference (HC) were collected using exible and inelastic tape with 1 mm precision to establish waisthip ratio (WHR) and waist-height ratio (WHtR), using a Lohman, Roche and Martorell protocol [18]. As overweight is known to have a negative in uence on quality of life [19], we chose to dichotomize the variable "overweight", in order to enable more robust and easier to interpret analyses. The categorization of BMI into overweight or not took into account the classi cation by the WHO [20] for adults and Lipschitz [21] for the elderly women. Thus, the variable "nonoverweight", including those classi ed as underweight and eutrophic (BMI ≤ 24.9 kg/m² or BMI ≤ 27 kg/m², cut-off points respectively for adults and the elderly); and the variable "overweight", including those overweight and obese (BMI ≥ 25 kg/m² or > 27 kg/m², respectively for adults and the elderly) were created. WC was categorized as "below the risk level for metabolic complications" (WC < 80cm); "high risk for metabolic complications" (WC ≥ 80cm); and "very high risk for metabolic complications" (WC ≥ 88cm) which is according to the classi cation by the WHO [20]. For WHR and WHtR, cutoff points > 0.85 [20] and ≥ 0.5 [22] were adopted as indicative of risk for developing chronic diseases and indicative of excess abdominal fat, respectively. To evaluate HRQoL, the instruments developed by the European Organization for Research and Treatment of Cancer (EORTC), the generic 30-Item Quality of Life Questionnaire (QLQ-C30) and its module for BC, the Quality of Life Questionnaire Breast Cancer − 23 (QLQ-Br23), both in the Portuguese version for Brazil [23], by permission of the EORTC for use in scienti c research, were used. The application of these questionnaires occurred 21 days after each time of evaluation (T0 + 21 days, T1 + 21 days and T2 + 21 days), always before CT. We chose to investigate only the domains and symptoms that could be more related to anthropometric variables, having been evaluated: Global Health, Global Function, Physical Functioning, Role Functioning, Emotional Functioning, Cognitive Functioning, Social Functioning, Symptoms, Fatigue, Pain, Dyspnea, Insomnia, Body Image, Sexual Function, Sexual Satisfaction.

Statistical analysis
The IBM SPSS Statistics version 21.0 software was used to perform statistical analyses. Descriptive statistics were performed to determine the mean, standard deviation, median and interquartile interval. This study aimed to verify the effect of anthropometric variables, CT time and the interaction between these variables on HRQoL scores and HRQoL symptoms. For this, General Mixed Models (GMM) were used. For those domains and symptoms that did not have the effect of the anthropometric variable and, or interaction, were tested only for the in uence of CT time, considering that the treatment alone is enough to impact HRQoL. All analyses were adjusted for age, schooling, income and menopausal status. The estimated marginal means and 95% con dence intervals (CI) were compared to the pairs applying Sidak for multiple tests.

Results
This study included 55 women with a mean age of 51.5 ± 10.1 years. The clinical, therapeutic and sociodemographic characteristics, as well as the nutritional status are shown in Table 1. Anthropometric measurements have already been detailed in previous studies by our group [24,4]. The mean weight of these women at the beginning and at the end of treatment was 70.9 ± 16.4 kg and 71.8 ± 16.8 kg, respectively, presenting a statistically signi cant change (p = 0.008). During treatment, these women also had an increase in BMI (p = 0.009) and WC (p = 0.030), emphasizing that they had already started treatment with excess weight (T0, 28.4 ± 6.4 kg / m²) and WC, indicative of increased risk for metabolic complications (T0, 90.8 ± 15.7 cm). The WHtR and WHR did not have statistically signi cant changes during the treatment, however, since the beginning of CT these measures were above the recommended (WHtR > 0.5 and WHR > 0.85), indicating excess abdominal fat and risk for the development of chronic diseases, respectively. In relation to HRQoL, we observed that women who were overweight (BMI ≥ 25 kg/m² or > 27 kg/m², respectively for adults and elderly), had a lower score on Physical Functioning (p = 0.023) and a higher score on fatigue (p = 0.017). At the end of CT (T2), lower scores for the Global Function (p < 0.001) and Physical Functioning (p < 0.001) were identi ed, and a higher Fatigue score (p = 0.003), in relation to the beginning of treatment (T0). In addition, statistically signi cant effects of the interaction between BMI and CT time on the Global Function (p = 0.040) and Physical Functioning (p = 0.011) were identi ed at T2; on insomnia (p = 0.003), at T1; and dyspnea (p = 0.035), however, for this symptom, it was not identi ed by the post-hoc at which time of treatment the signi cance occurred (Table 2). Women with substantially increased WC (WC ≥ 88 cm), that is, at risk of metabolic complications, had a lower Physical Functioning score compared to women with WC indicative of no risk for such complications (WC < 80 cm) (p = 0.004).
As for fatigue, the opposite was found, that is, those with substantially increased WC had a higher score compared to those with lower WC (WC < 80cm) (p = 0.006). In this General Mixed Model, women at the end of CT (T2) had lower scores on the Global Function (p < 0.001) and Physical Functioning (p = 0.004), and a higher Fatigue score (p = 0.010) compared to the beginning of treatment (T0). The interaction between WC and CT time had statistically signi cant effects on Global Health (p = 0.045) and Sexual Function (p = 0.035), and this difference was identi ed at T2, between those women with lower WC and those with increased WC (> 80cm); Global Function (p = 0.046) and Physical Functioning (p = 0.028), with a difference also at T2, but between women with lower WC and the others (WC > 80cm and WC > 88cm); and on Insomnia (p = 0.024), but no statistically signi cant difference was identi ed when comparing the pairs (Table 3).   (Table 4). signi cance in any of the General Mixed Models tested, these domains and symptoms were tested only for the in uence of CT time. At the end of CT (T2), women had a lower Social Functioning score (p = 0.043) and a higher Pain score (p = 0.010), compared to the beginning of treatment (T0) ( Table 6).

Discussion
Through this present study, we identi ed the effect of chemotherapy time, nutritional status and also the interaction between the chemotherapy time and the nutritional status of women with BC on the domains and symptoms of HRQoL.
Overweight women; WC indicative of increased risk and substantially increased risk of metabolic complications; WHR indicative of risk of developing chronic diseases; WHtR indicative of excess abdominal fat; that is, women with a worse nutritional status or on a more advanced time of treatment (T1 and T2) presented worse scores on important domains of HRQoL. It is important to highlight that the nutritional status, in isolation, had a statistically signi cant effect, demonstrating the importance of seeking a healthy nutritional status to minimize the negative impact on HRQoL during the treatment. We also observed that overweight, higher WC and WHtR and the interaction with CT time negatively impacted the global function, that is, they presented a worse quality of life and physical condition at the end of the treatment (T2 ≠ T0). .
Women with BC, when compared with the population free of the disease, show a decrease in HRQoL [12]. Added to that, antineoplastic treatments are also associated with a HRQoL decrease [25]. .
Besides that, we identi ed a decrease in the scores of physical function, which was associated with higher levels of BMI, WC, WHR, WHtR and longer CT times. Factors related to the disease and treatment such as cardiotoxicity, neurotoxicity, lymphedema, precocious menopause, sexual disfunction, infertility, secondary leukemia, weight gain, di culty sleeping and fatigue can justify the negative impact on HRQoL and physical function [26]. Added to that, there is the fact that a worse nutritional status, characterized by BMI, WC, WHR, WHtR measures, can be associated with a worsening physical function, as identi ed by Mosher et al, 2009, where an association with BMI was found. Mosher et al (2009) also identi ed that better diet quality and exercise are associated with better physical function (better vitality and physical function p ≤ 0.05) [27] .
We also observed a signi cant impact of the interaction of CT time and higher levels of BMI, WC and WHR on the insomnia score.. Among the factors that can be related to this problem, it is the presence of menopause symptoms caused by CT or hormonal therapy such as hot ashes [28] and also biological modi cations related to the cancer and treatment, which includes changes in pro-in ammatory cytokines and the hypothalamic-pituitary-adrenal axis (HPA) [29].
Sexual function was also signi cantly impacted by the interaction between CT time and higher WC and WHR.  [30].
Higher BMI and CT time, as well as the interaction of WHtR with CT time, impacted negatively on the fatigue score. Among the factors that can be related to this problem it is that, BMI and CT are associated with the increase of receptor two of tumoral necrosis factor (TNF-RII), which has been related to higher fatigue scores [31]. The TNF-RII is a receptor of the alfa tumoral necrosis factor (TNF-α), an in ammatory mediator [32]. A study performed on mice found that the adipose tissue is related with the production of TNF-α [33] and obese humans have a higher production of this cytokine than normal weight individuals [34].
We also identi ed a decrease in the physical function score, associated with higher BMI and WC, longer CT time and the interaction between BMI, WC, WHR and WHtR with CT time. Nutritional status in uences the HRQoL not just on BC, but also on colorectal cancer; an elevated WC is also associated with a worse physical function and higher scores of fatigue in men and women [35]. CT time also in uences the HRQoL of these women, worsening the social function domain..
Among the factors that could be related to this issue, is the possible occurrence of nausea and vomiting due to CT [36]..
Pain was also affected by CT time, where the score worsened throughout the treatment. The mechanisms involved in the pain process are not very clear, but it is known that CT acts on nociceptors and glia modulators, activating the microglia and astrocytes of the spinal medulla, and these release neuromodulators related to pain, causing chronic pain [37].
In the face of the shown facts, it becomes clear the need to reinforce the improvement of changeable factors that can increase the survival of these women, through a healthy diet, regular physical activity, and consequently an adequate nutritional status, with the purpose of achieving a better HRQoL, because CT itself already causes an 0.55 odds ratio of mortality in 10 years [38].
A systematic review that included 63 studies randomized 5761 women in two groups, one intervention (physical activity, n = 3239) and the other control (n = 2524), revealed that the physical activity intervention resulted in small to moderated improvements in HRQoL, emotional function, self reported and measured physical function, anxiety, cardiorespiratory capability and fatigue. However, these results must be interpreted carefully because the quality of evidence is small to moderated, due to heterogeneity of the interventions and measure of results [26].
The study of Phillips et al (2015) [39], also demonstrated that maintaining or increasing physical activity during the post diagnosis time was signi cantly and independently associated with lower fatigue, depression and stress scores and higher values for physical well-being, physical, social, emotional and functional scores, speci c to BC and global HRQoL (effect size = 0.23 a 0.60). Maintaining or losing weight had an independent association (p < 0.05) with lower fatigue, higher physical well-being speci c for BC and general HRQoL (effect size = 0.28 a 0.87). [39].
With regard to food intake, the dietary pattern is a factor of in uence on HRQoL of BC survivors, and the ones with a healthier diet, have a better HRQoL, presenting better role function, emotional function, cognitive function, social function and global health status [10]. Assaf et al (2016) veri ed that a change in diet, with a reduction of 20% of daily calories in fat, eating fruits and vegetables 5 times per day and grains 6 times per day, caused an improvement in HRQoL [40] and a better HRQoL is associated with a better overall survival after 1 year of diagnosis [41] .
It becomes evident that modi cations in changeable lifestyle factors directly affect the nutritional status, and consequently, can achieve signi cant improvements in HRQoL. Furthermore, it has the potential to contribute to a better prognosis [42] and a lower mortality rate in 10 years [38].
About the possible limitations of this study, it is important to consider that BMI is a method to classify the nutritional status of individuals, but is acknowledged that it is a limited method, and there may be overestimation or underestimation of the real nutritional condition. Even though, we highlight that all the measurements, including interviews, were performed by trained nutritionist.
Considering the fact that there are only a few articles in the literature that have investigated the relationship between nutritional status and HRQoL, we highlight the contribution of these results, especially because this is the rst prospective study analyzing the effects the association between nutritional status and CT time has on HRQoL of BC women.

Conclusion
It was veri ed that a worse nutritional status (overweight, increased and substantially increased WC, high WHR and high WHtR) and longer CT time (T1 and T2) are associated with worse mean scores of symptoms and HRQoL domains. Therefore, it is relevant to encourage BC women undergoing CT to have a healthier diet and regularly practice exercise, with the purpose of maintaining an adequate nutritional status. The improvement of these modi able factors during CT can contribute to a better HRQoL, better adhesion to the treatment, better prognosis and for higher overall and disease free survival rates.

Declarations
Funding This study was supported by the Conselho Nacional do Desenvolvimento Cientí co e Tecnológico, Brasil, (CNPq); Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG); and by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). The nanciers had no role in the design, analysis or writing of this article.

Competing interest
The authors declare that they have no competing interest.

Availability of data and material
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Code availability Not applicable.