Introduction

Globally, the cancer burden continues to rise [1], with breast cancer (BC) being the most prevalent malignancy among women [2], while colorectal cancer (CRC) ranks third among both sexes in developed countries [3]. Although the incidence rates are stable, the current increases in BC and CRC prevalence are attributed to improved survival rates, owing to advancements in early diagnosis and treatment [4]. Consequently, there is a growing need for effective survivorship care, which includes monitoring for recurrences, managing comorbidities, and addressing long-term physical and psychological effects.

Several studies have highlighted the importance of incorporating physical activity (PA) and a healthy diet into follow-up care for cancer survivors [5,6]. Lifestyle factors are known to influence both cancer incidence and prognosis [7,8]. For instance, metaanalyses have demonstrated the adverse impact of saturated fat intake on BC mortality, while PA is linked to improved survival rates [6]. Additionally, maintaining a healthy diet and engaging in regular PA can enhance patients’ quality of life by alleviating depressive symptoms and reducing treatment-related side effects [10]. In contrast, excess weight poses a risk factor for morbidity, recurrences, and BC mortality [11].

Similarly, diet and PA significantly affect CRC development and prognosis [12]. Factors such as obesity, sedentary lifestyles, red and processed meat consumption, smoking, and alcohol use are associated with poorer CRC outcomes, whereas regular PA, a diet rich in fruits and vegetables, high fibre intake, and regular fish consumption are beneficial for survival and quality of life [12,13].

Improving lifestyles is, therefore, an integral part of cancer care [14-17]. Regular consultations that cancer survivors undergo can serve as opportunities for health promotion [18,19]. Healthcare professionals (HPs), including physicians, can play a crucial role in advocating for lifestyle changes [20]. The effectiveness of multidisciplinary teams in empowering patients has been well-documented [21]. However, the implementation of lifestyle interventions remains limited [22], with only a small proportion of cancer survivors adhering to recommendations for smoking cessation, alcohol reduction, PA, and dietary guidelines [23,24]. Furthermore, HPs often lack sufficient awareness regarding appropriate PA and dietary practices during cancer care [22], and there is a dearth of guidelines for designing survivorship interventions [5,25].

This paper presents the methods and results of a lifestyle intervention conducted among a cohort of BC and CRC survivors in the Piedmont and Aosta Valley regions in Northern Italy.

Materials & Methods

Study Design

The “Follow-up and Healthy Lifestyle Study” (FUCSAM) is a prospective, multicentre investigation conducted between June 2014 and June 2017. Participants were recruited from all Oncology Multidisciplinary Groups caring for patients with BC and CRC in Piedmont and Aosta Valley, Italy. Eligible patients, who were in remission from CRC or BC and consented to participate, were followed until death, recurrence, diagnosis of other cancers, or until the study’s end in December 2018. The primary objectives of FUCSAM were to:

• Assess the feasibility of lifestyle counselling during cancer follow-ups (FUs).
• Describe the clinical, anthropometric, and serologic profiles of BC and CRC survivors, and track the evolution of these parameters and lifestyle habits over the medium and long term.

Secondary objectives included evaluating participation in lifestyle interventions and documenting the rates of recurrence and second tumors in this cohort. The study was supported by the Piedmont and Aosta Valley Cancer Network.

Participants

Eligibility criteria were:

• A histologically confirmed diagnosis of BC or CRC.
• Enrollment in follow-up care within the Piedmont and Aosta Valley Oncology Network.
• Completion of all primary treatments (surgery and any necessary adjuvant therapy).
• Absence of disease at the first follow-up visit.
• Ability to attend scheduled outpatient visits.

Intervention

HPs,  including oncologists, radiotherapists, surgeons, and nurses, received specific training on the study protocol, health promotion recommendations, and available local health promotion opportunities for cancer survivors. Informative materials providing straightforward behavioural recommendations and a list of local lifestyle interventions were distributed to all participating units. All trained HPs have had the discretion to suggest these activities and provide brief lifestyle counselling during follow-up visits. Enrolled patients have been considered all those who, at the moment of the baseline follow up, signed a specific consent for the study.

Data collection

At the first study visit, the following parameters were recorded using a standard computer-aided form:

• Socio-demographic data.
• Disease information (e.g., cancer site, tumour stage, detection modality, previous therapies).
• Current comorbidities.
• Anthropometric parameters (e.g., height, weight, body mass index [BMI], waist circumference, blood pressure).
• Serological analyses (e.g., fasting blood glucose, total cholesterol, triglycerides, HDL cholesterol).
• Study information, including the delivery of informative materials, lifestyle counselling, and suggested interventions (type and location).

Clinical information was compiled to determine the baseline presence of metabolic syndrome.

FU visits

Follow-up visits were conducted according to local protocols (approximately every 6 months). During these visits, baseline measurements were repeated and additional parameters were collected:

• Occurrence of comorbidities, recurrences, or second tumors.
• Participation in active lifestyle interventions (type and compliance).
• Variations in PA and nutrition.

Definitions of Improvement

Improvement was assessed by comparing follow-up measures to baseline values using the following criteria:

• BMI: a reduction of at least 0.5 kg/m² for participants who were overweight or obese at baseline; stable (±0.5 kg/m²) or reduced BMI for participants with normal baseline values.
• Waist circumference: a reduction of at least 1 cm for participant exceeding waist circumference recommendations at baseline; stable (±1 cm) or reduced circumference for participants with normal baseline values. A sensitivity analysis used a 2 cm cut-off.
• Metabolic syndrome: no change for participants without metabolic syndrome at baseline or a change of category for those meeting the criteria for metabolic syndrome.

A summary measure of anthropometric improvement was computed, defined as an improvement in at least one of the aforementioned parameters.

Participants completed self-reported lifestyle questionnaires, and a global score for PA and diet was computed based on items from the European Code Against Cancer. For PA, the following items were scored:

• Total weekly time spent in PA.
• Variation in time spent cycling.
• Variation in time spent walking.

Scores were attributed as follows: +1 for increase, -1 for decrease, and 0 for stable. The overall score was treated as both a continuous and categorical variable (0: score ≤ 0; 1: 1-2 points; 2: > 2 points).

For nutrition, scores were attributed as follows:

• Healthy foods (e.g., vegetables, legumes, whole cereals): +1 for increased consumption, -1 for decreased, and 0 for stable.
• Unhealthy foods (e.g., red and processed meat, sweets, alcohol, sugary drinks): +1 for reduced consumption, -1 for increased, and 0 for stable.

The overall nutritional score was categorized as 0: score ≤2 points; 1: 2-5 points; 2: 5-7 points; 3: > 7 points.

Statistical Analysis

Descriptive statistics and two-sample independent t-tests for continuous variables or χ² tests for nominal data were used. McNemar’s test was employed for pre/post analyses. At each follow-up, a statistical variable was created to define improvements, and crude and adjusted Mantel Haenszel odds ratios (ORs) were computed. Logistic regression models were fitted using variable selection based on univariate analysis (p<0.200). Sensitivity analyses were conducted on a restricted sample (patients completing all the first three follow-ups) and with different cut-offs. Due to significant participant attrition during follow-ups, an attrition analysis was performed using χ² tests and logistic regression. All analyses were conducted using STATA 15.

Results

Baseline data

Overall, 1,858 patients were registered within FUCSAM. Eleven men with BC were excluded, leaving 1,847 patients for following analyses (n=1,536; 83.2% with BC and 311; 16.8% with CRC). Table 1 shows patients’ baseline characteristics. Analyses stratifying by sex and cancer diagnosis were performed (data not in table).

Table 1. Baseline characteristics of study participants.

Overall, 695 cancers (39.1%) were screen-detected. The proportion of screen-detected lesions was higher considering BC (n=623; 42.1%) than CRC (n=72; 24.3%; p<0.001). Regarding BC, screendetected patients were in general diagnosed at an earlier stage (p<0.001), while no differences were retrieved for CRC. Overall, women were in general younger (men: 64.7±10.6 vs. women: 59.8±11.2; p<0.001), and had an earlier stage of disease (p<0.001). However, no differences were found between men and women with a previous CRC diagnosis (p>0.10).

To complete the clinical assessment, comorbidities were collected. Nearly all diagnoses were significantly more common among men (hypertension: W: 32.7% vs. M: 46.0%; diabetes: W: 5.4% vs. M: 11.2%, and cardiologic diseases: W: 12.2% vs. M: 24.0%; p<0.001). Similarly, anthropometric parameters were significantly worse among men (excess weight: M: 64.3% vs. W: 50.9% p=0.002; metabolic syndrome: M: 52.2% vs. W: 42.0% p=0.020), with the exception of waist circumference (W: 76.8% vs. M: 68.8%; p=0.025). The sex-related BMI differences were confirmed for patients with a previous CRC (p=0.007). Similar sex-differences were confirmed for biochemical analyses: glycaemia (100+ or in active therapy: M: 46.0% vs. W: 32.8% p<0.001), triglycerides (150+: M: 35.5% vs. W: 19.1%; p<0.001), and blood pressure (diastolic or systolic pressure increased: M: 81.0% vs. W: 63.5%; p= p<0.001), while no discrepancies were observed for HDL (p=0.171).

Considering FUCSAM intervention, 1,761 (95.3%) patients received information leaflets, while 1,403 patients (79.0%) performed a baseline counselling with the suggestion of active interventions. In this regard, diet programs were suggested to 61.1% participants, PA to 7.6% and combined programs to 31.3%. Overall, 160 patients (24.9%) participated to at least one program.

Most interventions took place in hospitals (n=276; 60.0%), followed by community centres (n=161; 35.0%). Diet-programs were offered predominantly in hospitals (n=216; 76.9%), while both PA and combined interventions were more frequently performed in community settings (respectively, n=23 56.7% and n=86 59.7%; p<0.001).

Both the delivery of information materials (99.3% vs. 97.5%, p=0.016) and counselling (80.0% vs. 68.6%; p=0.001) were more common for women, even considering CRC patients-only. In addition, the proportion of patients receiving some counselling greatly differed by Centres, ranging from 7.4% to 100% (p<0.001), and participants without metabolic syndrome were more likely to be offered an active counselling (82.9% vs. 76.6%; p=0.002).

Follow-up visits

FUs took place at 6-month intervals (median 184 days; range: 23-500 days). At first FU, 706 (38.2%) patients were lost, and at 3-year attrition exceeded 90% (n=1,668). Altogether, 28 recurrences (1.5%) were registered, with a higher proportion for CRC patients (respectively, 6.1% vs. 1.8%; p=0.001), while no differences were outlined by lifestyle programs participation (p>0.100). Finally, 72 non-neoplastic diseases were diagnosed (6.3%), with no differences by sex, cancer diagnosis or program participation (p>0.100).

Logistic regression for early-dropouts documented an association with sex (women: OR: 1.75 CI 1.05-2.91), diagnosis (CRC: OR: 1.61 CI: 1.10-2.91), medical counselling (OR: 0.54 CI 0.41-0.70), and metabolic syndrome (OR: 0.71 CI: 0-58-0.87).

Pre- and post- analyses at 6-month are shown in table 2. Improvements were more relevant among women (p=0.001), and patients recommended for behavioural interventions (p<0.001), with the exception of HDL-amelioration significant only among men (p=0.037).

Table 2:Results of pre-/post-analysis (McNemar test) at first follow-up visit (around 6-month since baseline) considering overall anthropometric and serological variations, and by sex and counselling.

Regarding interventions compliance, 156 participants (46.7%) attended more than half of the scheduled appointments in PA courses, with no differences by sex (p=0.220) or cancer type (p=0.203), while attendance was higher in patients without metabolic syndrome (60.1% vs. 47.2%; p=0.037). As regards the combined intervention, 136 participants (65.1%) attended more than half of scheduled lessons, with no differences by sub-groups. Finally, 326 (70.6%) patients attended more than half of the planned diet appointments, with the only disparity outlined for the absence of metabolic syndrome (25.7% vs. 35.3%; p=0.029).

Anthropometric measures variation during FUs

The following rates of BMI improvements were obtained: 54.3% (n=604) at 6-month, 52.3% (n=450) at 1-year, 50.2% (n=317) at 1.5 years, 47.4% (n=204) at 2 years, and 44.2% (n=118) at 2.5 years. Alongside, the rates of waist improvement were 58.8% (n=527) at 6 months, 51.6% (n=433) at 1-year, 49.4% (n=300) at 1.5 years, 48.9% (n=203) at 2 years, and 46.6% (n=124) at 2.5 years. The rates of metabolic syndrome improvement were 58.9% (n=581) at 6 months, 56.7% (n=438) at 1-year, 59.0% (n=338) at 1.5 years, 67.0% (n=264) at 2 years, and 54.4% (n=130) at 2.5 years. Finally, overall improvement reached 82.9% (n=946) at 6-months, 80.4% (n=715) at 1-year, 89.1% (n=541) at 1.5 years, 83.1% (n=379) at 2 years, and 77.8% (n=224) at 2.5 years.

Table 3 and Table 4 show the results of logistic regression for single anthropometric improvements and for overall improvement. At 6-month FU, the positive impact of program participation was confirmed for BMI (OR 1.43 CI 1.06-1.92), metabolic syndrome (OR 1.51 CI 1.01-2.26), and overall improvement (OR 1.71 CI 1.11-2.63).

Table 3:Odds ratio, deriving from logistic regression, for BMI and waist circumference improvement by follow-up visit.

Table 4:Odds ratios, resulting from logistic regression, for metabolic syndrome improvement and overall improvement by follow-up visit.

Conversely, for waist circumference, a substantial impact was observed for counselling (OR 1.77 CI 1.25-2.50). Although the effects of lifestyle interventions decreased over time, an association was still observed at 2.5 years. In long-term FUs, the association of anthropometric improvements with baseline metabolic syndrome, age, and cancer diagnosis was maintained. Results of sensitivity analyses did not show consistent variations.

Lifestyle habits

Overall lifestyle improvements are illustrated in figure 1, while stratified analyses are reported below.

Figure 1:a. Healthy habits increase declared by participants at first FU.b. Unhealthy habits decrease declared by cancer survivors at first FU

Considering PA, the only difference between sexes was observed for self-reported time spent in PA (≥3 hr/week: men: 53,3% vs. women: 27.4%; p=0.001). People participating in lifestyle programs reported higher modifications of PA (respectively, 43.3% vs. 16.1%; p<0.001).

Regarding diet, greater changes were seen in women, in particular for reduction in sweet consumption (reduction: M: 60.7% vs. W: 72.9%; p=0.027), sugar drinks (reduction: M: 52.7% vs. W: 58.5%; p=0.001), and refined cereals (reduction: M: 40.0% vs. W: 63.0%; p=0.009). A major reduction in alcohol consumption was documented for men (M: 51.8% vs. W: 38.4%; p<0.001). Finally, comparing participants by receiving or not a medical counselling, there were differences for overall diet modification (65.3% vs. 23.7%; p<0.001), legumes increase (62.9% vs. 50.9%; p=0.001) and whole cereals intake (61.6% vs. 49.1%; p=0.004).

Table 5:Odds ratios, resulting from logistic regression, for anthropometric improvements, including in the analysis the lifestyle score at the first FU visit.

PA and nutritional scores were computed. The association between these scores and anthropometric variations was investigated (Table 5), outlining an association with dietary score.