Feasibility and preliminary efficacy of a physical activity intervention in adults with lymphoma undergoing treatment

Abstract

Background

To determine the feasibility, acceptability, and preliminary efficacy of a 6-month tailored non-linear progressive physical activity intervention (PAI) for lymphoma patients undergoing chemotherapy.

Methods

Patients newly diagnosed with lymphoma (non-Hodgkin (NHL) or Hodgkin (HL)) were randomized into the PAI or healthy living intervention (HLI) control (2:1). Feasibility was assessed by examining accrual, adherence, and retention rates. Participants completed assessments of exercise capacity (VO₂ peak and 6-min walk distance (6MWD)), objective and self-reported levels of physical activity, MRI-derived cardiovascular functioning (Left Ventricular Ejection Fraction [LVEF], stroke volume, and cardiac output), and self-reported health-related and disease-specific quality of life and self-efficacy for exercise at baseline, 3, and 6 months.

Results

One hundred and forty-five individuals were screened with 23 of 84 eligible patients agreeing to participate (27%). Three participants withdrew before baseline testing. Out of the 20 participants randomized to the PAI (n = 13) and HLI groups (n = 7), 18 completed the intervention resulting in an overall retention rate of 78%. The adherence rates to the PAI and HLI were 85% and 87%, respectively. One non-serious adverse event was registered. VO₂ peak ranged from 15.5–28.0 ml/kg/min at baseline and participants in both groups improved by 6 months. Physical activity levels and cardiovascular function were reduced prior to treatment but did not deteriorate further.

Conclusions

Implementing a tailored PAI in adults with lymphoma during active treatment is feasible, was well received by participants and shows preliminary efficacy for limiting a decline in function during treatment. Potential Implications for Cancer Survivors: Physical activity may be beneficial for improving exercise capacity and health-related quality of life in patients undergoing chemotherapy.

Trial registration

#NCT01719562 ClinicalTrials.gov. Registered July 2, 2019—retrospectively registered.

Key messages regarding feasibility

• Uncertainties: Initial uncertainties included patient commitment during early cancer treatment, the feasibility of conducting assessments amid pandemic constraints, and concerns about potential dropout rates and adverse events.

• Feasibility findings: The study demonstrated a 27% accrual rate, strong adherence rates (85% for PAI, 87% for HLI), a 78% retention rate, and minimal adverse events.

• Implications for main study: These findings underscore the study's readiness for a larger-scale clinical trial, affirming the effectiveness of tailored interventions that fostered high adherence rates and meaningful participant engagement.

Background

Anthracycline-based chemotherapy, a curative therapy for several forms of lymphoma, is associated with left ventricular dysfunction and heart failure, fatigue, cognitive and physical function declines, and reduced health-related quality of life (HRQL) [1, 2]. Physical activity (PA) has emerged as a non-invasive approach to effectively mitigate short- and long-term cancer and treatment-related side effects [3, 4]. Despite PA’s established benefits across cancer populations, there remains a limited understanding of the feasibility of physical activity interventions specifically among adults undergoing lymphoma treatment.

To this end, the Physical Activity in Lymphoma Study (PALS) was a pilot-feasibility study designed to assess (1) the feasibility, acceptability, and safety of implementing a behavioral theory-based, tailored 6-month physical activity intervention (PAI) for adults undergoing treatment for lymphoma, and (2) the preliminary effects of PAI on objective and self-reported outcomes when compared to those receiving a healthy living education intervention (HLI).

Methods

Study design/population/recruitment

The PALS pilot study was a 6 month, 2-armed, randomized controlled (2:1) intervention study conducted at the Atrium Health Wake Forest Baptist Medical Center (AHWFBMC) and Virginia Commonwealth University Massey Comprehensive Cancer Center (VCUMCCC) from January 2019 to March 2022. The study was approved by the Wake Forest University Institutional Review Board (IRB00020968) and registered on clinicaltrials.gov (NCT01719562).

Inclusion criteria were (a) men and women aged 18 to 85 years of age, (b) diagnosis of stage I-IV non-Hodgkin’s or Hodgkin’s lymphoma, (c) receipt of anthracycline-based chemotherapy or other potentially cardiotoxic cancer therapies (e.g., immunotherapies or radiation), (d) self-reported capacity to walk at least 2 city blocks (~ 650 feet/200 m) on a flat surface, (e) ability to speak and understand English, and (f) survival beyond 6 months. Participants were excluded for (a) uncontrolled hypertension (systolic blood pressure > 190 mmHg or diastolic blood pressure > 100 mmHg, (b) recent history of alcohol or drug abuse, (c) contraindication for magnetic resonance imaging (MRI), (d) contraindications for exercise training or testing, (e) inability to exercise on stationary cycle, (f) claustrophobia, (g) unstable angina, (h) pregnant, (i) moving within 12 months of enrollment, and (j) inability to provide informed consent. Prior to study initialization, it was anticipated that 1 in 4 (25%) of approached patients would consent to the study.

Recruitment of patients occurred through the respective Comprehensive Cancer Centers and recruitment coordinators at both sites conducted eligibility assessments, gauged interest, and obtained written informed consent through in-person or telephone interviews with potential participants. Participants completed baseline assessments within 8 weeks of their initial treatment date. Randomization occurred immediately following baseline assessments. Randomization in a (2:1) was done using a variably sized permuted block randomization devised by the study biostatistician (RD) and implemented through an online clinical trials management system. The use of a variably sized permuted block randomization was to ensure small groups of participants were enrolled in each intervention arm simultaneously and to reduce the risk of selection bias. Allocation sequence was concealed from study staff and investigators by the biostatistician prior to randomization and follow-up assessments were conducted by study staff who were blinded to treatment group allocation.

Interventions

Physical activity intervention (PAI)

The PAI was a multilevel, supervised center- and/or home-based tailored physical activity intervention. The intervention was grounded in social cognitive theory, using behavioral skills training to support the successful adoption and maintenance of physical activity and exercise behaviors [5]. We aimed to build self-efficacy through the development of skills such as goal setting and self-monitoring and by addressing any individual-level barriers to accommodate each participant’s baseline functional status, treatment-related side effects, and access to exercise facilities/resources. Participants attended two weekly in-person PAI sessions either at the Wake Forest University’s Clinical Research Center (CRC) or Virginia Commonwealth University’s Department of Kinesiology and Health Sciences (KHS), where they received direct supervision from a master’s level clinical exercise physiologist. Each session followed a structured format, beginning with a 15-min slow aerobic warm-up, followed by 20 min of resistance training, 15 min of progressively intense aerobic exercise, and concluded with a 10-min cool-down involving stretching and toning exercises. The PAI design was informed by the exercise recommendations for cancer survivors outlined by the American College of Sports Medicine and the American Cancer Society [3, 4]. Each participant was also provided with a Fitbit device, primarily to support the self-monitoring of their physical activity and sedentary behavior. Participants could therefore gauge whether they were meeting their general goals or prescribed PA recommendations (e.g., steps/day or minutes of moderate to vigorous physical activity (MVPA) per week), and whether they were exercising at the appropriate intensity using heart rate.

The COVID-19 pandemic and closure of the study’s intervention facilities (CRC & KHS) prompted adjustments to facilitate online, home-based interventions. Alterations were duly submitted to the IRB and incorporated into the informed consent form. The PAI was restructured to allow for home-based physical activity sessions delivered with a software platform accessible on both PC and mobile devices that could interface with wearable devices (i.e., Fitbits) to monitor adherence to exercise prescriptions and daily activity patterns (i.e., steps). Study coordinators encouraged participants to explore various exercise options, providing guidance on walking, resistance, and aerobic exercises. As transmission rates subsided and the facility reopened, participants were given the option to attend sessions either at the center or remotely from home. Home-based exercise sessions were documented using the software platform described previously. Fitbit data, such as heart rate response and session duration, were used to gauge PA intensity and adherence, with oversight from the study investigators.

Healthy living intervention (HLI)

Participants assigned to the HLI engaged in a series of bi-weekly health workshops and interactive sessions aimed at enhancing psychosocial well-being and fostering healthy lifestyle habits. Sessions were initially conducted in-person at both sites and later adapted into an online format (via Zoom) in response to the COVID-19 pandemic. Sessions were tailored to the study's participant demographics and spanned a range of topics, including nutrition guidance for cancer survivors, stress management techniques, sleep hygiene, and transitioning to life after treatment. Each 60-min session comprised a brief review of the previous content and an introduction to the new topic (10 min), an interactive presentation on healthy living (40 min), and concluded with instructor-led gentle stretching exercises (10 min). The HLI was designed to serve as a comparison group that did not directly impact physical activity levels, while still encouraging participant interest and adherence to a program beneficial for their well-being.

Outcomes

Table 1 provides a summary of measures and timepoints for outcomes assessments across the study. Equivalent assessments were performed at baseline, and then 3 and 6-month visits. The overall goals and primary outcomes of this study were the determination of the feasibility of recruitment, our ability to complete a comprehensive battery of secondary and exploratory outcomes assessments, and the acceptability of interventions among patients undergoing active chemotherapy regimens (adherence and retention). Secondary outcomes were (a) VO₂ peak [6,7,8,9,10,11,12] and 6-min walk distance [13,14,15] and (b) device-monitored and self-reported levels of physical activity [16]. Exploratory outcomes included (c) left ventricular (LV) volumes (including LV stroke volume and heart rate to determine exercise-associated cardiac output [17,18,19,20], (d) psychosocial factors, (e) health-related quality of life [21] and fatigue [22,23,24], and (f) cognitive functioning. All assessments across both sites were conducted with adaptations to allow for the incorporation of online or home-based measures in response to COVID-19 restrictions. Following completion of baseline assessments, participants were randomized into either the PAI or HLI groups (2:1). Semi-structured, qualitative interviews were conducted with participants who either dropped out or completed either of the study intervention arms to refine a larger follow-up study. Results from the interviews are described in a separate report.

Table 1 Outcomes measures

The a priori criteria for our feasibility trial to be deemed a success and that were used to determine suitability for the conduct of larger randomized controlled trial were as follows: (1) the enrollment of 2–3 participants per month for a total of 21 participants with at least 80% in each study arm completing baseline and 6-month measures of key secondary outcomes (VO₂ peak and 6-min walk distance, physical activity levels and cardiac function); (2) a 6-month trend in improved % change in peak VO₂ after PAI versus HLI; and (3) a refined PAI intervention based on qualitative interviews.

Sample size

Our target enrollment of 21 patients was determined by the study biostatistician (RD) to estimate retention (and other binomial proportions) to within ± 23.6% with 95% confidence if 80% of participants remained in the study. In addition, this would allow a 95% confidence to be estimated for the difference between groups for the 6-month % change in peak VO₂. These data are shown in Table 3.

Statistical analyses

Descriptive statistics were used to describe the sample characteristics (means, frequencies) and all measures of study feasibility including completion rates of outcome assessments. Accrual rate was calculated by dividing the number of patients who consented to participate by the total number of eligible patients identified through clinic screenings and referral by the study’s oncologists. Adherence was assessed by comparing the number of completed sessions with the total number prescribed. PAI adherence was based on the number of completed sessions out of a maximum of 52 prescribed sessions, while HLI adherence was determined based on attendance at 13 offered sessions over 26 weeks. Make-up sessions were accounted for in cumulative adherence calculations. Retention rate was determined by monitoring participant dropout rates during the intervention, with successful retention defined as the collection of at least one 6-month measure. Adverse events were monitored and reported according to the IRB's stringent criteria, accompanied by brief explanations. The study team monitored and discussed participant data on a weekly basis. This included during assessment visits and throughout the intervention period. Testing and exercise session logs were used to monitor and report on any unexpected or adverse events. The study team also maintained contact with participants outside of the intervention sessions so study participants could report any adverse events that occurred in the home-based setting.

In addition to the feasibility of conducting this early phase pilot study, we examined the preliminary effects of the PAI compared to the HLI intervention on secondary and exploratory outcome measures of exercise capacity, cardiovascular and cognitive functioning, physical activity levels, psychosocial factors, and health-related quality of life. For these preliminary efficacy assessments, we examined and compared the means and variance (standard deviations) of each outcome in both PAI and HLI groups across the baseline, 3- and 6-month visits of the study to inform a fully powered future trial.

Results

Feasibility

Figure 1 illustrates the flow of participants through the study, detailing reasons for ineligibility, refusal to participate, and withdrawal. Between January 1, 2019, and April 1, 2022, 145 patients across both sites underwent screening for this study. A total of 84 patients met the eligibility criteria, 61 of which opted not to participate. Ultimately, 23 patients consented to join the study. However, 3 of these patients withdrew prior to baseline testing and randomization. This left 20 patients who completed baseline assessments and who were subsequently randomized into either the PAI (n = 13) or the HLI (n = 7) groups as shown in Table 2. Overall, participants had a mean age of 50 and 57 years in the PAI and HLI groups, respectively. The majority of participants were white, with a roughly equal distribution of males and females in each group. Approximately 70% of participants had a diagnosis of non-Hodgkin lymphoma, being between 9 and 13 weeks post-diagnosis at the baseline visit. Anthracycline-based treatment was received by 85% and 57% of participants in the PAI and HLI groups, respectively.

Fig. 1

CONSORT study flow diagram

Table 2 Baseline demographic and clinical characteristics

Accrual and retention

Throughout the 3-year duration of the study, the accrual rate was calculated as 27%. Out of the 23 individuals who signed informed consent, 3 withdrew before randomization, resulting in a dropout rate prior to randomization of 13%. Successful retention was defined by the completion of at least one testing measure at the 6-month visit. Notably, only 10% (n = 2) of randomized participants failed to complete any final testing, yielding an overall retention rate of 78.3% (n = 18). The 2 dropouts post-randomization (10%) was less than the a priori estimated 20% to establish feasibility for this pilot trial. Reasons for non-completion included one participant’s new cancer diagnosis and the second was lost to follow-up.

Adherence

Physical activity intervention

The average adherence rate for all PAI participants (n = 13) was 85%. Nine out of thirteen participants achieved a perfect adherence rate of 100%. Notably, only one participant exhibited an adherence rate lower than 60%. Participants were offered the choice between center-based or home-based exercise sessions. Four participants (30%) attended at least one center-based session. Among these, two participants predominantly completed sessions at the center (~ 75%), while one primarily exercised at home (93%), and another did both home- and center-based sessions. Most participants (78%) completed all their exercise sessions at home, especially once COVID-19 restrictions were put in place.

Healthy living intervention

Seven participants (n = 7) completed at least one baseline measure and were subsequently randomized into the HLI group. Adherence to the HLI was calculated by dividing the attendance to online healthy lifestyle sessions by the total number of sessions prescribed. The average adherence rate for all HLI participants (n = 7) was determined to be 87.1%. Notably, over half of the participants (4 out of 7) achieved a perfect adherence rate of 100%.

Secondary and exploratory outcomes

Table 3 depicts the secondary and exploratory outcomes assessments completed across the 3 study visits (baseline, 3- and 6-month visits). The key secondary outcome of interest in this pilot study to inform sample size calculations for a larger study was the VO₂ peak. For certain measures (CPET, MRI), COVID-19 restrictions meant we were unable to complete in-person study visits at the respective study sites resulting in between 30–60% of assessment visits incomplete. For other measures, we were either able to conduct remote assessments (6-min walk tests, accelerometry) or have participants complete quality-of-life questionnaires via online surveys (i.e., REDCap). For all measures across study visits and by treatment arm, we show the means (SD) as well as median (range) and number of participants who completed assessments. There were several participants who had VO₂ peak values below the threshold for functional independence (< 18 ml/kg/min). This was also reflected in short distances walked in the 6-min walk tests by some participants (~ 300 m). However, on average, the VO₂ peak and 6-min walk distances did increase in both PAI and HLI groups across the study. Assessments of physical activity levels and MRI-derived metrics of cardiovascular function also reflected this variability in functional capacity, with some participants having very low levels of activity (< 60 min of MVPA per week, < 2000 steps per day) and poor cardiovascular functioning (resting LVEF < 50%). Many participants reported high levels of persistent fatigue that likely impacted their health-related quality of life.

Table 3 Primary and secondary outcomes by study visit

Adverse events

One adverse event (AE; 2 participants) was registered as an unanticipated problem involving risks to subjects or others. This AE entailed a participant feeling nauseous and a headache after participating in an online kickboxing class provided by their local gym. The participant fully recovered within a span of 2 h. Another participant experienced a fall while walking their dog but they were not badly injured and continued with their participation in the intervention.

Discussion

Following successful completion of this pilot feasibility study, we met a priori criteria for determining feasibility showing we were able to randomize patients undergoing active treatment for lymphoma to a 6-month physical activity intervention or a healthy living education control arm. In addition, despite COVID-related restrictions, we were able to collect important outcomes concurrent with the receipt of active therapy regimens. Feasibility was also determined by favorable accrual and adherence rates, a moderate level of retention, and a low incidence of adverse events. Notably, participant retention emerged as the primary challenge, with potential factors such as the COVID-19 pandemic as a possible underreported reason for participant drop-out rates. However, protocol adjustments, including the introduction of home-based assessment options in response to the difficulties posed by center-based activities during the pandemic resulted in improved outcomes assessment rates and may have helped with retention to the trial given participants were able to receive feedback on their progress.

The recruitment rate of 27% achieved the predetermined objective of obtaining consent from 25% of patients referred by the oncologists involved in the study. This finding aligns with similar physical activity interventions, which have reported recruitment rates ranging from 12 to 87%, with longer interventions generally associated with lower recruitment rates [25,26,27,28]. However, the overall dropout rate observed in our study was moderately higher than the anticipated 20%, although this is consistent with findings in the existing literature. Dropout rates vary across different physical activity interventions for cancer survivors, ranging from 7 to 30% [26, 27, 29, 30]. The reasons for withdrawal reported in our study were also comparable to those documented in the literature. A meta-analysis of physical activity interventions for adults with advanced cancer revealed that the most common reasons for participant withdrawal were disease progression, health deterioration, cancer-related mortality, and loss to follow-up [29]. Overall, the moderately higher dropout rate observed in this pilot trial prior to randomization underscores the necessity of recruiting a greater number of patients than initially planned. In future studies, we plan to explore employing innovative recruitment strategies, such as social media, targeted outreach to healthcare providers, and patient advocacy groups, to increase participation. This approach may help mitigate patient loss due to reasons that may not be as relevant in similar interventions targeting patients with other chronic illnesses or cancer survivors not undergoing treatment. Furthermore, a larger sample size would increase the statistical power and improve the precision of the estimates.

The adherence rate of 85% for the PAI is noteworthy, especially given the context of patients undergoing cancer treatment, the lengthy duration of the intervention (6 months), and the added challenges presented by the COVID-19 pandemic. Adherence rates in the literature vary widely, ranging from 50 to 93%, depending on factors such as the type of cancer, duration of the intervention, timing of initiation (during or after treatment), and other intervention characteristics [25, 26, 28, 31, 32]. Notably, a high proportion of participants in the PAI group (n= 9) achieved 100% adherence, indicating their strong commitment to participating in the intervention and the effectiveness of adapting the protocol to accommodate unsupervised, home-based physical activity sessions. Recent studies have highlighted the benefits of home-based PA interventions, demonstrating their efficacy in enhancing functional capacity, reducing fatigue, and improving health-related quality of life among cancer survivors [29, 33]. The transition towards home-based options is both promising and expected, particularly in light of the challenges presented by the COVID-19 pandemic [34].

Tailored sessions and a strong theoretical framework, specifically Social Cognitive Theory, constituted key strategies in the study to accommodate patients’ treatment-related side effects and bolster their self-efficacy towards physical activity (PA). Interventionists collaborated closely with patients to design PA sessions that were realistic and attainable, thereby enhancing participants' confidence in their PA capabilities. Moreover, interventionists remained cognizant of patients' treatment schedules and encouraged them through verbal persuasion to capitalize on days when they experienced fewer treatment-related side effects (e.g., fatigue or nausea). The study's capacity to personalize PA prescriptions within the framework of a theoretical model was essential for maintaining high adherence rates to the intervention.

Given this pilot study was not powered to detect differences in the effects of the two intervention arms, we instead examined descriptive statistics (means, medians, and variances) for secondary and exploratory outcomes as an indicator of average levels of exercise capacity, cardiovascular function, health-related quality of life and physical activity levels across the study among participants receiving active treatment. For our key secondary outcome of interest, we saw a slightly higher level of exercise capacity in the PAI group at baseline and 3 months, but a higher mean VO₂ peak in the HLI group at 6 months. This is certainly due to the small sample size for this outcome given restricted testing during the trial as well as a small number of highly active participants in the control group. What was promising was the degree to which participants in the trial maintained their exercise capacity, despite receiving active treatment when typically, patients experience sharp declines. For submaximal exercise capacity, we had a higher proportion of complete data and we consistently saw participants in the PAI walk further. As with most outcomes, the range in performance and function varied widely with some participants being very low on these outcomes. Self-reported health-related quality of life seemed to be slightly better in the PAI group across the trial and these measures also had a higher completion rate than some of the objective assessments of function. While objective physical activity assessments were similar across study groups, these data were likely skewed by a small number of highly active controls; however, the participants in the PAI reported a slightly higher level of self-efficacy for walking across the study, perhaps reflecting the benefit of engagement with the study teams and the tailored activity programs. Verbal fluency scores (COWAT) were slightly higher in the PAI vs the HLI groups at 3 months, but the reverse was true at 6 months, likely reflecting differences in test completion rates at each of these timepoints (see Table 3).

Strengths and limitations

While our study showcased numerous strengths, such as showing it was feasible to recruit patients undergoing chemotherapy to an active intervention, completing a battery of novel assessments across two geographic sites (when not COVID restricted), and adapting to home-based models, it also encountered some limitations. Initially, the COVID-19 pandemic introduced significant challenges to both intervention strategies and testing protocols in the study. The transition to home-based interventions may have altered the delivery and effectiveness of the interventions, potentially impacting participant engagement and adherence. For example, despite the objectivity of the Fitbit activity monitor, participants occasionally neglected to log their physical activity sessions due to a preference for personal devices, difficulty grasping the Fitbit technology, or simply forgetting to register their sessions. Moreover, the inability to conduct certain hospital-based tests due to safety concerns limited the comprehensiveness of the assessment, potentially affecting the study’s ability to capture relevant outcomes and generalize findings. The successes and challenges of completing this pilot study suggest several potential future directions. Due to the relatively low recruitment rates in exercise oncology trials, it seems important to provide patients who are undergoing treatment with alternative intervention delivery options (online, in-person, hybrid) to further support their varying ability to change their health behaviors. A comprehensive integration of these intervention modes may offer greater opportunities for participation for those lymphoma patients who have limitations related to travel and access. In addition, future studies that utilize novel imaging techniques to determine how and whether cardiovascular function changes in response to cancer treatments and or the adoption of lifestyle behaviors (i.e., exercise training) may be enhanced by employing advanced data processing and analytical techniques such as those provided by artificial intelligence and machine learning models [35]. For example, these techniques may facilitate integrating electronic health record data with imaging data and or tracking of activity patterns to more accurately predict health outcomes or response to treatments.

Conclusion

The PALS pilot trial underscores the feasibility of implementing a tailored six-month physical activity intervention for adults undergoing active treatment for lymphoma. The study highlights patients' willingness to engage with and adhere to the intervention, particularly when provided with adequate support and accommodations tailored to their individual fitness levels and treatment-related side effects. Insights gained from this pilot trial informed the redesign of a larger-scale study, aimed to recruit 100 lymphoma patients across 2 sites. With an expanded sample size, the forthcoming study aims to evaluate the efficacy of the physical activity intervention in preserving cardiac and physical function, managing fatigue, and enhancing health-related quality of life within this population.