Physical activity of women treated for breast cancer in the context of sociodemographic factors

Introduction

Engaging in regular physical activity (PA) is widely recommended for all cancer survivors due to its potential to improve health outcomes and quality of life. Consistent participation in PA is linked to reduction in cancer recurrence rates, increased survival, and improved health-related quality of life [1–10].
The American Cancer Society (ACS) has developed physical activity guidelines for cancer survivors, advocating for at least 150 min of exercise weekly [6, 11]. However, only approximately 37% of breast cancer survivors adhere to these guidelines [12, 13]. Data from the Behavioral Risk Factor Surveillance System (BRFSS) indicate that 31.5% of cancer survivors did not participate in PA during their leisure time [7].
The literature identifies numerous barriers to engagement in PA. Factors related to health, e.g. fatigue and joint stiffness, and emotional and cognitive challenges, as well as environmental factors, e.g. lack of facilities or inclement weather, all play a role [7, 14]. Notably, specific cancer-related impediments, including pain, fatigue and sociodemographic factors, hinder the ability to engage in PA.
The benefits of PA are indisputable. One of the most precise methods of its measurement is accelerometry, which captures 24-hour data on both PA and sedentary behaviors of study participants [15]. Duration, intensity, and frequency are recorded, at the same time limiting errors and inherent deviations often associated with subjective measurement methods [15, 16]. Precise assessment of PA and sedentary lifestyle using accelerometry enhances our understanding of those exposures and their impact on the health of patients undergoing cancer treatment. Such insights are invaluable for developing interventions aiming at modifying these behaviors. However, the quality of data recorded with an accelerometer is influenced by several decisions made during the recording and processing phases. Those critical decisions include type and location of the device, wear time protocols, epoch length, filtering techniques, and criteria for non-use and valid wear time. Another challenge is effectively processing the data to receive metrics such as total sitting time, light physical activity, and moderate to vigorous physical activity (MVPA) [15, 17, 18].
To facilitate effective and targeted interventions aimed at improving PA among BCS, our research focuses on sociodemographic characteristics influencing objectively assessed PA in this group.

Aim of the research

This research aims to evaluate the impact of sociodemographic factors on the physical activity levels of women undergoing breast cancer treatment.

Material and methods

Study population

In total, 135 women who underwent treatment for breast cancer in Holycross Cancer Center in Kielce, Poland were included in the study. The study was approved by the Ethics Committee in Kielce on May 19th, 2017 (approval no. 19/2017). A conversation was carried out informing the participants about the examination method, concluded by signing a consent form. The study was carried out in the Department of Rehabilitation, Holycross Cancer Center.

Demographics and cancer treatment variables

A questionnaire involving sociodemographic factors (age, area of residence, marital, educational, and occupational status) and medical data (comorbidities, the side of mastectomy, lymphadenectomy, and the use of radiation therapy or chemotherapy) was used in the study.

Sedentary behavior and physical activity assessment

Physical activity (PA) was objectively assessed using the triaxial ActiGraph GT3X-BT accelerometer (Pensacola, Florida, USA) paired with ActiLife 6 software (Pensacola, Florida, USA). This well-validated device measures the frequency, duration, and intensity of sedentary behavior, light PA (LPA), and MVPA. It incorporates an inclinometer that registers the wearer’s orientation in three dimensions, which enables differentiation between sitting and standing positions. Participants were instructed to wear the accelerometer at waist level continuously for a week. After this duration, they returned the devices and were provided a comprehensive activity measurement summary.
Data were analyzed using ActiLife 6 software with a low-frequency extension and were aggregated into 60-second epochs. Every minute of device wear was categorized by intensity (counts per minute, cpm), employing Freedson cut points: < 100 cpm, sedentary lifestyle; 100–1.951 cpm, LPA; 1.952–5.724 cpm, MVPA ≥ 5.725 cpm [19, 20].
Wear time was determined using ActiLife 6 software, based on the Troiano 2007 algorithm. The minimum wear time criterion was set at 3 days. Non-wear time was defined as periods of at least 60 consecutive minutes with 0 counts, though interruptions of up to 2 min registering fewer than 100 CPM were permissible within this duration [21].
For each valid accelerometer wear day, minutes spent in sedentary lifestyle, LPA, and MVPA were tabulated. These daily estimates were then averaged across all valid days for each participant at every time point to determine the daily average minutes spent on each activity type. The number of minutes in each category was normalized by the total wear time to determine the percentage of the day allocated to each behavior. Furthermore, we assessed the daily steps using the accelerometer [21, 22].

Statistical analysis

Baseline statistics are presented as either mean (standard deviation), median (interquartile range), range (minimum-maximum) or number and proportion, depending on the nature of the specific variable. Statistical differences in PA based on the analyzed sociodemographic attributes were assessed using the Wilcoxon signed-rank test. Effect sizes are presented using Cohen’s d. Relationships between PA and sociodemographic features were analyzed using robust regression models. Univariate and multivariate models were fitted. Sociodemographic features that demonstrated significance in prior analyses were incorporated into the multivariate models. P-values of < 0.05 were considered statistically significant. All analyses were executed using R software (version 3.6.3).

Results

A total of 135 women who underwent treatment for breast cancer were included in the study. Their average age was 63.2 ±10.0 years. All participants had undergone breast surgery. Regarding the methods of cancer treatment, 29.6% (n = 40) were treated using one method, 35.6% (n = 48) with two methods, and 34.8% (n = 47) with three methods, as outlined in Table 1. In terms of demographics, more than half of the women (63%) resided in urban areas, two-thirds were in a relationship, and 85% had university degrees. 70% of the patients were not currently engaged in professional activities. Approximately 70% of the women reported having comorbidities.
Table 2 presents the characteristics of PA intensities, segmented by sociodemographic variables. We observed significant differences in the number of steps and MVPA across diverse age categories. Using Cohen’s d to assess the effect size, we found the observed differences to be of medium magnitude (Table 3). Moreover, significant differences based on the comorbidity status were noted both in the number of steps and MVPA, with the effect sizes also falling into the medium range, with Cohen’s d values ranging between > 0.3 and < 0.8.
In the univariate regression models, presented in Table 4, negative values of regression coefficients signify a relative reduction in PA, while positive values denote a relative increase. A significant correlation was observed between age and the number of steps, LPA, and MVPA. Additionally, the area of residence was significantly linked to LPA, while the comorbidity status correlated with both the number of steps and MVPA.
Age, area of residence, occupational status, education, and comorbidities were included in the multivariate analyses. Their inclusion, however, differed based on the application of specific models. In multivariate regression models (Table 5), age and comorbidity status displayed a significant relationship with the number of steps. Interestingly, among the socioeconomical variables evaluated, only age was consistently related to MVPA in all subsequent multivariate models.

Discussion

Approximately two-thirds of cancer survivors do not adhere to ACS physical exercise guidelines. Various factors – including sociodemographic, economic, health-related, and cancer-related elements – are perceived as obstacles to engagement in regular PA [7, 23]. In our study, we assessed which sociodemographic factors influence objectively measured physical activity. A relationship was identified between the number of steps taken, age, and area of residence. The average number of steps among women older than 65 years undergoing treatment for breast cancer was 5,544 steps, compared to an average of 6,820 steps for women younger than 65 years.
WHO guidelines indicate that a minimum of 10,000 steps daily is required for preservation of good health. The studied population exhibited low activity, ranging from 5000 to 7499 steps/day. Therefore, on average, the women in our study did not adhere to the guidelines. Number of steps is classified as follows: 1) sedentary lifestyle (< 5000 steps/day); 2) low activity (5000–7500 steps/day); 3) somewhat active (7500– 9999 steps/day); 4) active (≥ 10,000 steps/day); 5) highly active (12500 steps/day) [24].
Achieving 10,000 steps per day has emerged as a universal benchmark for physical fitness [25]. This target equates to an energy expenditure of approximately 300–400 kcal, contingent upon walking pace and body mass. The weekly energy expenditure associated with achieving 10,000 steps on more than 3 days is comparable to 30 min of moderate PA performed on most weekdays. This level of activity mirrors the energy expenditure (pegged at 1,000 kcal/week) tied to a marked reduction in mortality from cardiovascular disorders. According to growing body of scientific evidence, in order to preserve and improve cardiovascular fitness, as well as to maximize health benefits overall, it is recommended to engage in MVPA. Assessing PA according to number of steps is considered the correct approach to increase health-related PA. However, controversies persist regarding the exact number of steps needed to preserve health and physical fitness [24, 26, 27].
Our findings align with existing literature, in which age has consistently been identified as an important barrier to engagement in PA [7, 28, 29]. Presence of comorbidities and higher BMI were also associated with decreased PA [28]. Other authors point to demographic factors, including older age and lower education levels, as contributors to lower PA among breast cancer survivors [29].
Our research did not demonstrate a relationship between PA and education level in women treated for breast cancer. Nevertheless, other studies indicate that a lower education level is associated with decreased PA following a diagnosis [30]. Compared to participants with higher or vocational education, those with only secondary education or those who did not complete high school were 2.4 and 5.9 times more likely to remain professionally inactive after diagnosis. Patients with higher education perceived PA as a method to decrease fatigue and improve the quality of life. Interestingly, 9% of participants stated that “uncertainty about what they are allowed to do” served as a barrier to engaging in PA. Less-educated patients reported this limitation more frequently [30].
In the Life and Longevity After Cancer (LILAC) study conducted by the Women’s Health Initiative (WHI), women (n = 3710) possessing higher education levels, better self-esteem, improved physical fitness, and robust support systems, were more likely to engage in any type of physical activity [31].
In line with our study, other authors have also failed to find a significant relationship between PA and marital status [7, 27]. Nevertheless, we did observe a relationship between PA (both the number of steps and MVPA) and the presence of comorbidities. With an increase of number of comorbidities, PA correspondingly decreased.
Various epidemiological studies have explored the connection between PA and comorbidities of cancer survivors. For instance, a cross-sectional study involving women found that while total PA was not related to multimorbidity, the time spent walking was inversely proportional to the number of comorbidities [32]. However, this study was limited due to the small size of its cohort.
A retrospective cohort study involving 1526 cancer survivors revealed that moderate to high levels of physical activity were correlated with a 35% to 45% decrease in the presence of cardiovascular risk factors, including diabetes or hypertension [33]. Other comorbidity groups were not studied. Additionally, a prospective cohort study of 1,696 breast cancer survivors demonstrated that moderate PA, such as 30 min of daily walking, led to a 31% decrease in the incidence of metabolic syndrome [34]. Notably, other comorbidities were not studied. Altogether, the results of these studies demonstrate an inverse relationship between PA and several comorbidities, such as diabetes and hypertension, in cancer survivors. These studies were limited in terms of range of studied comorbidities and types of PA. Dong-Woo et al. further expanded on this, investigating the associations between various comorbidities based on type of exercise (aerobic vs. strength training) and doses (completely inactive vs. insufficiently active vs. following PA guidelines). Elevated blood glucose concentrations are related to worse prognosis in cancer patients, but this association varies depending on the location of cancer [35, 36].
Jeon et al. found that diabetic patients who survived colon cancer had 20% shorter disease-free survival compared to non-diabetic patients [37]. A similar trend was observed in breast cancer, prostate cancer, and bladder cancer patients [35]. Research by Dong-Woo et al. reported an inverse relationship between frequency of aerobic PA and fasting glycemia. Physically inactive cancer survivors were found to have higher average glucose concentrations (102.6 ±1.3 mg/dl), indicative of prediabetes. Conversely, those who adhered to PA guidelines had normal glycemia, at 95.8 ±1.7 mg/dl. Patients who adhered to aerobic PA guidelines had approximately a 35% reduced risk of diabetes compared to patients who did not exercise. This is consistent with well-documented evidence on the positive influence of PA on diabetes management and glycemic control [38]. The physiological mechanisms behind the relationship between PA and cancer prognosis remain elusive. However, PA may contribute to the systemic regulation of blood glucose and insulin concentrations, subsequently restricting glucose uptake and growth in cancer cells. It could also play a role in anti-proliferative processes, through inhibition of the direct and indirect mechanisms associated with glucose and insulin intake, and cancer growth [36, 39, 40].
Elevated blood pressure or hypertension often coexists with diabetes in cancer patients [41]. While the association between hypertension and cancer patients’ prognosis is not universal, maintaining normal blood pressure decreases the risk of death due to cardiovascular disease [42–44]. Prior usage of angiogenesis inhibitors in cancer survivors also increases the risk of both hypertension onset and cardiovascular disease-related deaths [36, 45].
The results of our research imply that area of residence is associated with LPA frequency. Individuals living in rural areas showed a higher likelihood of engaging in PA. Lynch et al. found, consistent with our study, that colon cancer survivors in Australia residing in urban areas were less likely to adopt or maintain a healthy lifestyle after diagnosis when compared to those in rural settings [46]. In contrast, Weaver et al. reported that cancer survivors from rural areas were less likely to engage in PA compared to urban residents [47]. Such divergent findings could arise from variations in the classification of residence areas, contextual differences, as well as regional or societal factors [28].
Our research holds significance for patients after the diagnosis of breast cancer, for whom PA should be an integral part of lifestyle. The indicated socioeconomic factors point to specific patient demographics that would benefit from more frequent fitness-oriented interventions.

Strengths and weaknesses of the study

This study has several limitations. First, the accelerometer cannot differentiate specific subtypes of physical activity, such as cycling or swimming, and may misclassify actions that are not truly PA. For instance, driving a car could be recorded as movement rather than sedentary behavior. While the Freedson cut points are broadly acknowledged in assessment of PA of cancer survivors, they originate from studies on healthy adults with an average age of 24 years. Consequently, the accelerometer might not detect the optimal MVPA in older populations, including cancer survivors. The studied groups are often characterized by multiple comorbidities and lingering side effects of cancer treatment, which compromises their functional fitness. Therefore, the cut points derived from a younger population might not always be pertinent for cancer survivors.
However, there are multiple notable strengths to this study. Primarily, robust measurement methods were employed. Employing accelerometers to monitor sitting duration is less susceptible to errors associated with self-assessment, such as recall biases. Furthermore, accelerometers outperform self-assessment tools in registering LPA and sedentary behavior. Ability to accurately assess LPA is particularly important in this population, because, as the data indicate, it tends to engage more in LPA than in MVPA.

Conclusions

We identified demographic factors – older age, number of comorbidities, and area of residence – as significantly related to lower activity levels of women undergoing treatment for breast cancer. This information should be taken into consideration when encouraging organized PA and a healthy lifestyle for this group of patients.

Acknowledgments

Project financed under the Minister of Education and Science program called “Regional Initiative of Excellence” in the years 2019-2023, project no. 024/RID/2018/19, amount of financing PLN 11 999 000,00.

Conflict of interest

The authors declare no conflict of interest.

References