Tailoring a physical activity intervention to older adults receiving intensive chemotherapy for acute myeloid leukemia (AML): one size does not fit all

Introduction

Older adults are at risk for functional decline during acute myeloid leukemia (AML) therapy.¹ While some survivors demonstrate resilience,^2,3 short-term declines in physical function can be dramatic during an induction hospitalization.¹ Physical performance (measured by the Short Physical Performance Battery [SPPB]) is predictive of survival at the time of diagnosis,⁴ declines during induction therapy,¹ and is associated with shorter survival when measured at the time of remission.⁵ Interventions designed to maintain physical performance may enhance the quality of survivorship, eligibility for subsequent therapies, and could potentially improve survival.^6-10

Whether physical activity (PA) interventions targeting older adults receiving intensive chemotherapy can feasibly be administered to hospitalized patients experiencing disease and treatment-related symptoms has been sparsely studied.^6,8,11 The intervention used in our first pilot study, which tested the feasibility of intervening on older adults receiving inpatient AML chemotherapy, was developed using principles of behavioral interventions applied to healthier, outpatient cancer populations.⁸ We learned that hospitalized older adults with newly-diagnosed AML were willing to enroll in a PA study while receiving intensive chemotherapy. However, barriers to participating in the fixed thrice-weekly intervention included fluctuating symptoms, frequent changes in clinical status, and competing care needs. It was clear that a successful intervention should be adapted to both the unique functional needs of the older adult¹² as well as the realities of a symptomatic and acutely ill patient population.

Tailoring physical activity interventions to participants and settings

In designing the next study (ClinicalTrials.gov Identifier: NCT01519596), we applied lessons learned from the first pilot and combined them with insights from other settings, including interventions designed for ventilated patients ^13,14 and a landmark trial demonstrating the effectiveness of a multi-modal exercise program to prevent disability among older adults.¹⁵ In particular, studies demonstrating that even “low dose movements” for bed-bound ventilated patients positively impacted clinic outcomes provided support for including low intensity bed exercises when participants were too sick to perform standing activities, a common occurrence during the course of AML treatment. We focused our adaptations on: maximizing flexibility and opportunity to participate, tailoring exercise sessions to daily changes in symptoms and condition, applying a multi-modal training paradigm with additional focus on balance training, and integrating behavioral counseling to overcome barriers to participation. We hypothesized that the adapted PA intervention would minimize declines in physical function associated with chemotherapy, leading to decreased disability, improved quality of life and potentially improved survival (Figure 1).

Figure 1:

Conceptual model of potential benefits for a physical activity intervention during AML chemotherapy

Symptom-adapted intervention: The right activities for the right patient at the right time

Eligibility for this study was intentionally broad and included patients age ≥60 years, newly-diagnosed with AML, scheduled for intensive chemotherapy, ambulatory, with no acute medical problems precluding participation (e.g., bleeding, thrombosis, hemodynamic instability). Unlike most physical activity interventions, our design was inclusive of representative older adults with multi-morbidity and poor performance status to ensure any effect seen could be widely applicable.¹⁶ Specifically, most physical activity intervention studies exclude participants with active comorbid conditions or acute illness such as infection or anemia.¹² These conditions are common among older adults diagnosed with AML and our eligibility criteria were written to ensure the majority of patients would be eligible; the program was adapted to the patients’ symptoms and daily clinical status. Participants were randomized 1:1 post baseline to intervention versus control.

A key innovation was the development of a “symptom-adapted” tiered intervention that allowed participants more opportunity to engage in some physical activity despite experiencing fluctuating symptoms related to treatment or disease complications (Table 1). We learned from our prior pilot that offering the intervention only three times per week was insufficient in this acute setting, resulting in an average of only one completed session per week.⁸ Taking advantage of the inpatient setting, we offered the intervention five days per week with timing based on participant preference (i.e. morning or afternoon). Our goal was to enhance the likelihood of participation at least three times per week. Offering five days versus seven was pragmatic; staffing a weekend interventionist was not feasible. As this research constraint mirrors resource constraints across United States hospitals, it reasonably reflects available ancillary services that could implement this type of intervention if proven effective. Sessions were conducted by a trained exercise interventionist with oversight from a physical therapist. To ensure safety, the interventionist checked daily with the participant’s nurse to identify medical contraindications to PA.⁸ If contraindicated or the participant declined the session, the interventionist checked daily to determine appropriateness to resume.

Table 1.

Symptom Adapted Physical Activity (PA) Program

PA Session Type (offered 5x/week while hospitalized)*	Description Overview	Description Details
Standard (Ward-based)	2-3 minute warm up (walking in place) Walking (approximately 10 minutes) Strength and flexibility (approximately 20 minutes) Balance training (approximately 10 minutes) Cool down/stretching	Goal Rating of Perceived Exertion (RPE) during walking of 12-14 (light to moderate) Strength protocol included 10 exercises utilizing resistance bands: chest press, backward row, biceps curl, triceps press, shoulder raise, lat pull downs, leg raise, leg curl, hip flexion, calf point and chair stands. Balance training included: standing marching, hip side kicks, toe/heel raises, knee bends and hip backward kicks.
Intermediate (Room-based)	Strength and flexibility (approximately 20 minutes) Balance training (approximately 10 minutes) Upper extremity ergometer (approximately 10-15 minutes)	Intermediate session was used for participants who were unable or unwilling to participate in hall-based walking. Strength and flexibility using resistance bands as above. Aerobic activity was provided by use of the upper extremity ergometer (hand crank cycle). Participants were seated in an upright position with the ergometer placed on the overbed table directly opposite. Participants were started with low resistance which was adjusted to goalRPE 12-14.
Low-intensity (Bed-based)	Upper extremity ergometer (approximately 10 minutes) Progressive resistance (approximately 10 minutes)	Low intensive sessions were utilized for participants who were unable or unsafe getting out of bed. Upper extremity ergometer was positioned on the overbed table directly in front of the participant while resting in bed. Head of bed was raised to comfort. Low resistance was adjusted as needed to facilitate light/moderate intensity level as tolerated. Progressive resistance exercises were conducted in bed with our without resistance bands tailored to condition with goal of maintaining form. Modified exercises included chest press, backward row, bicps curl, triceps press, arm raise, lat pull down, short arc quads, hip flexion/straight leg raise, hamstring sets, calf points and hip abduction/adduction. Upper and lower extremity exercises were offered on alternate days. RPE goal 12-14
Behavioral counseling (offered weekly while hospitalized with phone follow-up post discharge every 2 weeks for one month then monthly up to 24 weeks)	Review participation Identify barriers Identify short and long term functional goals Relate goals to elements of intervention Trouble shoot barriers Set weekly goals for participation	Counselor was a separate team member from PA interventionists Principles of counseling included: emphasizing rationale for intervention (i.e. linking intervention to maintenance of independence and treatment tolerance); normalizing the patient situation (i.e. recognize clinical status will change with focus on accommodation), reviewing physical activity principles (i.e. frequency of participation over intensity, quality of activities over intensity, importance of multi-modality, emphasize that ANY activity mattered) and individualizing goal setting (i.e. responsive to symptoms and personal goals). Goals and strategies discussed during counseling sessions were relayed to the PA interventionists to further individualize session for maximal benefit. Post hospital phone follow-up reviewed continued PA activities, identified barriers, facilitated trouble shooting barriers to on-going PA, reviewed new individual functional goals, and re-set personal goals for on-going PA.

^*Intensity of training was regulated using the Rating of Perceived Exertion Scale (RPE) and heart rate⁸.

Participants randomized to the intervention were offered a standard, intermediate, or low-intensity session at each visit. Session intensity was a collaborative decision based on the interventionists’ assessments and the participants’ preferences.¹⁷ The relational approach we used in the hospital environment is unique, but it has strong support from a conceptual point of view in the behavior change literature.¹⁸ Importantly, participants were able to move back and forth through the tiered program as needed to adapt to symptoms and clinical condition, ensure safety, and maximize opportunities for participation. The goal was to facilitate participation, not to achieve a specific dose of PA . The collaborative approach provided the participant agency in tailoring activities to their symptoms and preferences, thereby decreasing barriers to participation.

The standard (ward-based) session included cardiovascular, strength, flexibility, and balance training.⁸ This multi-modality approach was chosen due to evidence supporting benefits of aerobic exercise in AML¹⁹ as well as our observational data showing objective declines in lower extremity strength, gait speed, and balance after induction chemotherapy measured by the SPPB.¹ Given the negative association between SPPB and survival our goal was to target each component to maximize maintenance of function during therapy.^4,20 Duration and intensity of walking segments were individualized, emphasizing mild intensity and a gradual progression in duration. This was followed by strength and flexibility exercises using resistance bands. The strength protocol included ten exercises, five targeting large muscle groups of the upper body and five targeting the lower body; these five alternated each session⁸. Range of motion exercises followed the strength phase to increase flexibility. Participants then performed in-room balance training, adapted from the Lifestyle Interventions and Independence for Elders (LIFE) study to an inpatient setting utilizing a chair or counter for support.¹⁵

The intermediate (room-based) sessions were done entirely in the participant’s room using an arm ergometer for the cardiovascular component.²¹ During our prior pilot study, many participants were unable or unwilling to perform the walking phase in the oncology ward. Rather than omit aerobic exercise, participants performed ten to fifteen minutes of cycling in the seated position. Subjects were offered strength, flexibility, and balance training as above.

The low-intensity sessions were done in bed for participants who were unable or unwilling to safely perform activities out of bed. They performed up to ten minutes of cardiovascular activity on the arm ergometer, and progressive strength exercises with or without resistance bands. The rationale for this session was based on evidence from ventilated patients in intensive care units showing that even passive exercise (passive range of motion of upper and lower extremity joints in repetitions of five) contributes to clinically significant functional improvement including earlier mobility and decreased length of hospitalization.^13,14

Overcoming barriers by aligning with participant functional goals

Participants randomized to the PA group received a behavioral counseling session weekly during hospitalization to enhance adherence and promote longer-term behavioral change. This novel intervention component is grounded in social cognitive theory and deemed particularly important to bolster self-efficacy and maintain physical activity behaviors in the face of challenging barriers faced during treatment for AML. Sessions were designed to: 1) discuss perceived benefits of PA sessions; 2) review progress; 3) highlight consequences of deconditioning and benefits of PA during treatment and post-discharge; 4) establish self-regulatory skills for effective behavior change; 5) discuss and trouble-shoot barriers to adherence; and 6) establish specific functional goals for the coming week. These sessions provided weekly motivational counseling for the participant. In addition, information about “what matters most” to the participant each week was fed back to the PA interventionist for additional “tailoring” of the up-coming PA sessions. Post-discharge telephone follow-ups were conducted semi-weekly for two months, and then monthly until week 24 to facilitate maintenance and adaptation to progressive functional goals.

Outcomes: A focus on feasibility and function

The primary endpoints for this pilot study were measures of feasibility which are critical to establish in this vulnerable population. A recent systematic review found only 4 published randomized studies that tested exercise among older adults receiving cancer therapy, none in the setting of AML.¹⁰ In a randomized exercise intervention for all aged patients receiving intensive therapy for AML, recruitment and adherence rates were 56% and 54% respectively.⁶ In our prior non-randomized pilot of older adults with AML, the recruitment rate was 44% with only 71% participating in any exercise and none completing the prescribed 12 sessions (3/week).⁸ We set more ambitious goals for our adapted intervention since our intention is to develop an intervention that can be implemented as usual care to benefit the majority of older AML patients receiving intensive therapy. Because we felt a high rate of recruitment was critical to ensure generalizability we aimed to enroll≥60% of eligible patients. Our target for adherence was to perform an average of three sessions per week. Our retention goal was for ≥85% of participants to complete scheduled follow-ups. Safety was assessed by adverse events attributable to the intervention.

Robust measures of physical function were included as secondary outcomes to fully capture the magnitude and trajectory of functional decline and recovery during AML therapy for older adults which had not been previously reported and to comprehensively evaluate the potential effect of the intervention. Physical function measures were assessed weekly while hospitalized (typically four to six weeks) and at three- and six-month follow-up visits (Supplemental Figure 1). Change in SPPB was the primary variable of interest to be used for sample size calculations for a fully powered efficacy trial. The SPPB consists of a four-meter walk, repeated chair stands, and three hierarchical standing balance tests and we have previously shown its predictive utility in this population.²² Self-reported function was assessed using the Pepper Assessment Tool for Disability , a validated tool to assess domains of physical function among older adults.²³ We also used the Mobility Assessment Tool (short form), a ten-item computer-based questionnaire that uses animated video clips to portray function.²⁴ The items range from walking slowly on level ground to climbing stairs while carrying bags. Participants assess their ability to perform the displayed activity. This correlates with objective performance testing in other settings and is well suited to assess mobility in hospitalized patients and via tele-health. We assessed grip strength with a hydraulic hand dynamometer. Hip, knee, and ankle strength were measured as an exploratory outcome using the Lafayette Manual Muscle Test System²⁵ to address the potential floor effect of the SPPB for bedbound subjects.

We also evaluated additional patient reported outcomes to better understand the potential benefits of the intervention and identify participants who were more or less likely to respond. These included assessing global health related quality-of-life (QOL) using the Functional Assessment of Cancer Therapy-Leukemia,²⁶ depressive symptoms using the Center for Epidemiologic Studies Depression Scale,²³ distress using the Distress Thermometer, ²³ and fatigue using the 13-item FACIT-fatigue scale .²⁷ The Digit Symbol Substitution Test measured attention and perceptual speed. This test has been shown in randomized trials to improve with physical activity among older adults without cancer, including those with poor physical performance, which is common among older adults with AML.^28,29

Control Group

Participants in the control group received usual care, which may have included physical therapy at the discretion of the clinical team. The study nurse interacted with control participants weekly to perform assessments and at three- and six-month follow-up assessments.

Discussion

Most older adults with AML have vulnerabilities at diagnosis when assessed with geriatric assessment,²³ are at high risk for treatment-associated functional decline, and will experience acute illnesses during the course of treatment. These patients are most in need of supportive care interventions to maintain function and QOL. Most studies designed to test interventions to prevent functional decline exclude vulnerable, acutely ill and hospitalized patients, although evidence in the intensive care unit shows that small doses of activity/movement can effect clinically meaningful outcomes.^13,14,30 Therefore, we adapted a PA intervention to the acute care setting for older adults that emphasized flexibility and frequent offers to engage in PA, and integrated our program into the clinical care structure to ensure safety. Additionally, we recognized the importance of ensuring that the intervention was aligned with the patient’s goals by engaging participants in behavioral counseling to personalize motivation, set individual goals, and identify strategies to overcome barriers to participation. While this study was designed for older adults with AML, principles applied here can be utilized to adapt interventions to older and vulnerable adults in varied clinical settings.

Moving forward, there are opportunities to build on this dual-component intervention (physical activity + behavioral counseling) by testing bundled interventions that target additional vulnerabilities associated with functional decline. In particular, our prior work has shown that depressed mood and cognitive impairment are associated with decline in physical performance during AML chemotherapy.¹ Use of geriatric assessment to guide additional supportive care interventions to address these vulnerabilities could enhance the benefits of an exercise program for older adults with AML. Finally, learning about our participants’ functional goals recorded from their behavioral counseling sessions and engaging older AML patients in the development of subsequent trials will help ensure that the study design and outcomes optimally match our patient and caregiver needs.