Randomized controlled trial of increasing physical activity on objectively measured and self-reported cognitive functioning among breast cancer survivors: The Memory & Motion study

Abstract

Background

Increasing physical activity can improve cognition in healthy and cognitively impaired adults; however, the benefits for cancer survivors are unknown. The current study examined a 12-week physical activity intervention, compared to a control condition, on objective and self-reported cognition among breast cancer survivors.

Methods

Sedentary breast cancer survivors were randomized to the Exercise arm (n=43) or Control arm (n=44). At baseline and 12-weeks, objective cognition was measured with the NIH Cognitive Toolbox and self-reported cognition with PROMIS scales. Linear mixed effects regression models tested intervention effects for changes in cognition scores.

Results

Participants’ (n=87) were on average 57 years old (SD=10.4) and 2.5 years (SD=1.3) post-surgery. Oral Symbol Digit (measure of processing speed) evidenced differential improvement in the Exercise vs. Control arms (b=2.01; p<0.05). The between group differences in improvement on self-reported cognition was not statistically significant, but suggestive of potential group differences. Time since surgery moderated the relationship with participants ≤2 years post-surgery having significantly greater improvement in Oral Symbol Digit score, Exercise vs. Control (b=4.00, p<0.01); no significant improvement was observed in patients who were >2 years post-surgery (b=−1.19, p=0.40). A significant dose response was observed with greater increased physical activity associated with objective and self-reported cognition in the Exercise arm.

Conclusions

The exercise intervention significantly improved processing speed, but only among those diagnosed with breast cancer within the past 2 years. Slowed processing speed can have substantial implications for independent functioning supporting the potential importance of early implementation of an exercise intervention among breast cancer patients.

INTRODUCTION

Problems with cognition are common among breast cancer survivors with estimates suggesting that up to 75% of patients experience deficits that can last for years after the end of cancer treatments.^1–3 Research suggests that cancer-related cognitive impairments are not solely due to chemotherapy or other treatments^2,4,5 but may result from common risk factors for both the development of cancer and age-related cognitive decline.^3,6 In addition, it may be a symptom of accelerated aging caused by cancer treatments and associated psychological stress.^2,7,8 Although the prevalence of cognitive deficits in breast cancer survivors is well documented, few evidence-based interventions exist to prevent or ameliorate this decline.⁹

One intervention that could potentially improve cognition is increasing physical activity. The Institute of Medicine’s report on Cognitive Aging recommends physical activity to reduce age-related cognitive decline.¹⁰ Many physical activity interventions have been effective in improving cognition in both healthy and cognitively impaired populations.^11–14 The general physical activity recommendation for breast cancer survivors is to engage in 150 minutes per week of moderate to vigorous physical activity (MVPA).¹⁵ However, National Health and Nutrition Examination Survey data indicate that breast cancer survivors spend only approximately 25 minutes per week in MVPA.¹⁶ Studies of the impact of physical activity on cancer-related cognitive impairments has great potential to improve the lives among the growing number of cancer survivors.¹⁷

Few intervention studies have examined the effects of physical activity on cognition in a cancer population. A recent review of the literature¹⁸ found only two trials that tested the impact of aerobic physical activity interventions on cancer-related cognitive impairments.^19,20 Focusing on aerobic physical activity is important as it is consistent with guidelines for brain health.^21,22 These two studies yielded mixed findings: a 6-week trial in 479 breast cancer survivors found improvements in self-reported cognition,²⁰ but a 12-week trial in 41 breast cancer survivors found no improvements in self-reported cognition.¹⁹ Of note, neither study used objective measures of cognition. Only one trial in cancer survivors using objective cognitive measures was identified. This recent study in 19 breast cancer survivors enrolled in a 24-week physical activity randomized controlled trial found that the exercise arm had greater improvements in a measure of processing speed compared to the control group.²³ Several cross-sectional studies with cancer survivors using objective measures of cognition have found greater physical activity to be associated with several domains of cognition including information processing speed,²⁴ memory,²⁵ executive functioning,²⁶ and attention.²⁶ These studies lend support to the hypothesis that increasing physical activity may be an effective intervention for improving cognition in breast cancer survivors.

To our knowledge, the current study is one of the first completed randomized controlled trials with cancer survivors testing the impact of increasing physical activity on both objective measures (i.e., neurocognitive testing) and self-reported measures (i.e., cognitive abilities and concerns). Since neurocognitive outcomes and self-reported cognition reflect different aspects of cognition, assessing both provides a fuller picture of how physical activity can impact cancer survivors’ quality of life. The purpose of this study was to examine the effects of a 12-week physical activity intervention, compared to a contact control condition, on objective and self-report measures of cognition among breast cancer survivors. We hypothesized that the physical activity intervention would result in greater improvements in neurocognitive tests and self-reported cognition compared to a wait-list wellness contact control. We also explored potential effect modifiers as well as dose-response associations between changes in physical activity and cognition.

METHODS

The Memory & Motion study was a randomized controlled trial of a 12-week physical activity intervention compared to a waitlist wellness contact control to test changes in objective neurocognitive tests and self-reported cognition among breast cancer survivors. Data were collected from February 2015 to July 2016 and analyzed July-November 2016. The UC San Diego institutional review board approved all study procedures and all participants provided written informed consent.

Eligible participants were female breast cancer survivors, age 21–85 years, who were diagnosed less than 5 years prior to study enrollment and had completed chemotherapy or radiation treatment. Other inclusion criteria were: (1) sedentary, defined as self-reporting less than 60 minutes of MVPA in 10 minute bouts per week, (2) able to travel to La Jolla, CA for study visits, (3) access to the internet, and (4) self-report that they are experiencing “fogginess” or worsening of their memory, thinking, or concentration. Exclusion criteria were: (1) medical condition that could make it potentially unsafe to be in an unsupervised physical activity intervention as determined by the Physical Activity Readiness Questionnaire (PAR-Q),²⁷ (2) other primary or recurrent invasive cancer within the last 10 years, and (3) unable to commit to a 3-month intervention.

Protocol

A detailed description of the protocol was previously published.²⁸ Briefly, participants were predominantly recruited via cancer registry lists. Potential participants were telephone-screened to determine eligibility. Interested and eligible women were scheduled for an in-person visit. At the baseline visit, participants provided signed informed consent, a fasting blood draw was collected, height and weight were measured, and participants completed a battery of neurocognitive tests and questionnaires assessing self-reported cognition (described below). At the end of this visit, all participants were given an ActiGraph Gt3X+ accelerometer to wear for 7 days and bring back to the randomization visit. All baseline measures were repeated at the 12-week visit.

Participants were randomly assigned to one of two groups, Exercise arm or Control arm, in a 1:1 ratio. Randomization was stratified by having had chemotherapy or not using a permuted block randomization scheme with random-sized blocks of size 6 or 8. A computerized randomization scheme was created by the Moores Cancer Center Biostatistics Shared Resource. The sample size was determined based on 80% statistical power for a between-group effect sizes of 0.32 (Cohen’s f). After randomization, participants in both groups reviewed the expectations and requirements of their group assignment with study staff.

Physical Activity Intervention (Exercise arm)

Participants randomized to the Exercise arm had a 30- to 45-minute in-person meeting where went on a 10-minute walk at moderate intensity and set physical activity goals. To promote behavior change and accountability, participants were provided a Fitbit electronic activity device as an intervention tool. Participants were informed that study staff would be able to see the physical activity data collected by the Fitbit and that they would receive feedback on the Fitbit data during the scheduled phone calls. Using motivational interviewing techniques, a starting goal with a specific plan for meeting that goal was set. Participants discussed how to gradually increase their aerobic exercise over time and meet the study goal of at least 150 minutes of MVPA per week.¹⁵ Participants received two phone calls (2 week and 6 week time point) and emails every three days throughout the 12-week intervention. The intervention was delivered by a clinical psychologist with extensive training and experience in promoting behavior change (SJH) and by a staff member (EM) who was trained by SJH. For further details on the intervention see Hartman et al.²⁸

Waitlist Wellness Contact Control Condition (Control arm)

Participants randomized to the Control arm received emails on the same schedule as those in the Exercise arm. Emails consisted of a variety of women’s health topics including healthy eating, stress reduction, and general brain health. After completion of measures at the final visit, participants in the Control arm were provided with the exercise intervention described above.

MEASURES

The primary outcome, objective neurocognitive functioning, was measured using The NIH Toolbox Cognition Domain (nihtoolbox.org) at baseline and 12-weeks. The NIH Toolbox is a series of computer-based tests that can be administered in approximately 45 minutes and has been validated and normed in individuals age 3 to 85. The testing uses multiple forms and Computer Adaptive Testing (CAT) to minimize practice effects. It also has minimal floor and ceiling effects. The tests provides age-standardized scores for a Fluid Cognition Composite Score and a Crystallized Cognition Composite Score.²⁹ In addition, we examined 7 individual test scores of the Fluid Composite Score. Five of the 7 tests have age-standardized scores, the other two only provide raw scores (Oral Symbol Digit and Auditory Verbal Learning).³⁰

The second primary outcome, self-reported cognition, was measured with two scales from the Patient Reported Outcomes Measurement Information System (PROMIS): Applied Cognition Abilities (Abilities) and General Concerns (Concerns). The cognitive Abilities and Concerns measures assess patient-perceived functional abilities and problems in the past 7 days with regard to cognitive tasks in specific areas (e.g. concentration, memory). Higher scores on the Abilities measure indicate more positive perceptions of cognition and higher scores on the Concerns measure represent worse perceptions of cognition. Both self-report measures are standardized T-scores that have demonstrated good reliability and validity with previous measures including the Functional Assessment of Cancer Therapy – Cognitive Function measure.³¹

The GT3X+ Actigraph, a research grade accelerometer that is considered the gold-standard for measuring free-living physical activity, was used to measure change in physical activity from baseline to 12-weeks. Participants in both study arms wore the ActiGraph on the hip for seven days before the randomization visit and 7 days before the final visit. The ActiGraph provides second-by-second estimates of activity that can be categorized into minutes spent in sedentary, light, moderate, and vigorous activity using calibration thresholds.³² Sufficient wear time across the 7 days was defined as having 5 days with ≥600 minutes each of wear time or 3000 minutes (50 hours) of total wear time across 4 days.

Body Mass Index (BMI) was calculated from height and weight collected at the baseline and final visits. Participants self-reported demographics including age, education, income, race/ethnicity, and marital status at baseline. All breast cancer information and treatment details were obtained from medical chart reviews. To estimate time since diagnosis we used date of surgery because the exact date of diagnosis was not consistently available from the medical charts.

Statistical Analysis

All analyses were performed using an intent-to-treat principal with missing data assumed missing at random and accounted for in the longitudinal random effects models by using a likelihood-based estimation method, which uses all available data and does not omit subjects with partially missing data. Group differences in baseline characteristics were assessed using t-tests, chi-square, or two-tailed Fisher’s exact tests (when warranted by small cell counts) for categorical variables. Differences in baseline physical activity were assessed using a mixed effects regression model of day-level activity, controlling for ActiGraph wear time, and included fixed effect terms for group. Logistic regressions assessed group differences in the percent of participants meeting the study goals (150 min/week of MVPA) at 12-weeks. Mixed effects regression models with a subject-level random intercept and an unstructured covariance structure, as determined by model Akaike Information Criterion comparisons in the main effects models, were used for all other models. Contrasts were used to calculate the difference of change based on the regression model when examining group difference. As the Crystalized Cognition Composite Score was not expected to change with the intervention, it was assessed as an outcome in all cognition models to confirm expected patterns.^29,33 Models to assess the effects of the intervention on change in physical activity – measured at the day level – and changes in cognition scores controlled for ActiGraph wear time and included fixed effect terms for group, time-point (baseline or 12-weeks), and time-by-group interaction. In the case of the Oral Symbol Digit and Auditory Verbal Learning Scores, participant age was also controlled for, as these scores are not age-adjusted.

Potential effect modifiers of the intervention (age, chemotherapy, current tamoxifen or aromatase inhibitor use, and time since surgery) were tested for the Fluid and Crystalized Cognition Composite Scores, any significant individual neurocognitive test scores, and self-reported cognitive abilities and concerns from the main effects models. Effect modification was examined by adding the potential effect modifier and a three-way interaction (time, group, and effect modifier) to the mixed effect regression models. The significance of this interaction term was set at the 0.1 level as is common practice with interactions. For significant effect modifiers, subgroup analyses stratifying the sample at the median for the effect modifier were conducted. Dose response was examined within the Exercise arm using mixed effects regression models of average physical activity per week (MVPA or total activity) on cognition scores, controlling for ActiGraph wear time.

Results

A total of 911 women were screened for eligibility; of those, 108 were eligible and 97 came to the baseline visit. Most common reasons for being ineligible included being too active (n=225), unable/unwilling to attend clinic visits (n=106), breast cancer surgery more than 5 years ago (n=81), and medical exclusion (n=36). At the baseline visit 10 women were deemed ineligible (high blood pressure (n=8); physical limitation (n=2)). Eighty-seven participants were randomized to the Exercise arm (n=43) or the Control arm (n=44). One participant from each arm was lost to follow up, resulting in a 97.7% retention rate (Exercise n= 42, Control n=43). See CONSORT diagram (Figure 1).

Figure 1.

CONSORT Flow Diagram

Baseline characteristics of study participants, stratified by randomization arm, are shown in Table 1. Overall, women in the study were an average of 57 years old (SD=10.4), predominantly white (82%), non-Hispanic (83%), with a college education or greater (71%). Participants were an average of 30 months’ post-surgery (SD=16.7); 61% had Stage 1 breast cancer, 53% had received chemotherapy, and 70% were currently taking an aromatase inhibitor or tamoxifen. Age-adjusted, standardized, composite scores were a mean of 115 points (SD=14.4) for Crystalized Cognition and 103 points (SD=14.7) for Fluid Cognition. Average scores were 45.4 (SD=5.7) for Cognitive Abilities and 43.3 (SD=6.9) for Cognitive Concerns. At baseline Cognitive Concerns were significantly lower in the exercise arm with a mean of 41.8 (SD=5.5) versus 44.9 (SD=7.9) in the wellness arm (p=0.04). At baseline, participants were engaging in an average of 14 minutes per day of MVPA (SD=14.4) and an average of 303 minutes (~5 hours) per day of total activity (light and MVPA) (SD=75.2) and wore the ActiGraph on average 842 (SD=103) minutes per day. There were no significant differences between the Exercise and Control arms in baseline characteristics (p>0.05).

Table 1.

Baseline Characteristics by Study Arm (n=87)

	Exercise Intervention (n = 43)	Wellness Control (n = 44)	p-value
	Mean (SD)	Mean (SD)
	or n/%	or n/%
Demographics
Age, years	58.2 (11.37)	56.2 (9.30)	0.354
Education			0.689
Some college or less	14/32.7	25-Nov
College graduate	18/41.9	22/50
Master’s degree or higher	11/25.6	11/25
Married/living with partner	32/76.2	31/70.5	0.679
Ethnicity			0.74
Not Hispanic/Latino	35/81.4	37/84.1
Hispanic/Latino	8/18.6	7/15.9
Race			0.62
White	36/83.7	35/79.5
Non-White	7/16.3	10/22.8
BMI, kg/m2	26.7 (6.20)	27.3 (6.40)	0.628
Breast Cancer Characteristics
Time since surgery, months	30.3 (17.41)	30.0 (16.08)	0.997
Cancer Stage			0.786
Stage I	27/62.8	26/59.1
Stage II	12/27.9	15/34.1
Stage III	4/9.3	3/6.8
Received Chemotherapy	23/53.5	23/52.3	0.91
Current aromatase inhibitor/tamoxifen	31/72.1	30/68.2	0.691
Neurocognitive Testing Score (age-adjusted, standardized)
Crystallized Composite	117.3 (12.68)	113.8 (15.95)	0.272
Fluid Composite	102.3 (14.18)	103.5 (15.36)	0.712
Auditory Verbal Learning (raw score)	24.5 (5.50)	24.6 (5.17)	0.961
Dimensional Card Sort	105.3 (10.18)	103.9 (10.16)	0.52
Flanker Inhibitory	96.3 (8.53)	96.6 (9.37)	0.867
List Sorting	106.8 (11.95)	107.0 (13.74)	0.935
Oral Symbol Digit (raw score)	73.7 (14.55)	77.7 (13.23)	0.185
Pattern Comparison	106.3 (18.01)	106.6 (17.74)	0.259
Picture Sequence	116.2 (13.97)	112.3 (15.24)	0.924
Self-reported measures (standardized)
Cognitive Abilities	46.6 (5.96)	44.3 (5.27)	0.0619
Cognitive Concerns	41.8 (5.45)	44.9 (7.89)	0.0356

The Exercise arm had greater increases in accelerometer-measured MVPA (mean min/day increase 14.2 vs. −0.7, b=7.24, p<0.001) and total activity (mean min/day increase 27.4 vs. 4.9, b=10.05, p=0.02) than the Control arm. The Exercise arm also had greater increases in the number of participants meeting the study goal of 150 min/week of MVPA as compared to the Control arm (p=0.006). BMI did not significantly change over time (b=0.01, p=0.84). See Table 2.

Table 2.

Change and Difference in ActiGraph Measured Physical Activity and BMI at Baseline and 12-weeks by Study Arm

	Intervention				Control				Difference of change between groups
	Baseline	12 week	Change	p-valuea	Baseline	12 week	Change	p-valuea	Estimate	95% CI		p-valueb
Total MVPA min/day (Mean/SD)	13.4 (12.97)	27.9 (15.13)	14.2 (13.98)	<.0001	15.4 (15.67)	14.9 (15.12)	−0.7 (9.74)	0.464	7.24	5.33	9.15	<.0001
Total Activity (Light + MVPA) min/day (Mean/SD)	290.5 (73.26)	320.4 (56.76)	27.4 (71.90)	<.0001	315.8 (75.83)	320.9 (75.38)	4.9 (52.30)	0.393	10.05	1.84	18.26	0.017
Meeting 150 min/week (n/%)c	8 (18.6%)	25 (59.5%)	–	<.0001	10 (22.7%)	9 (20.9%)	–	0.706	–	–	–	0.0003
BMI (Mean/SD)	26.7 (6.20)	26.8 (6.21)	0.1 (0.56)	0.929	27.3 (6.40)	27.5 (6.61)	0.1 (0.60)	0.895	0.01	−0.11	0.14	0.837

^a: Test of change over time

^b: Test of group difference of change over time

^c: analyzed using Chi-squared test

Changes in the objective and self-reported cognition scores from baseline to 12-weeks between the Exercise and Control arms are presented in Figure 2. The Exercise arm had significantly greater improvements in the Oral Symbol Digit Score, a measure of processing speed, than the Control arm (b=2.01; 95% CI = 0.01, 4.01; p = 0.049). All other neurocognitive measures (except List Sorting, a measure of working memory) showed significant increase in scores from baseline to 12-weeks, but there were no significant between-group differences. For self-reported cognition, there was no statistically significant difference between the arms for Cognitive Abilities (b=0.92; 95% CI = −.14, 1.98; p=0.087) or Cognitive Concerns (b=−1.38; 95% CI = −3.13, 0.37; p=0.120.)

Figure 2.

Change (12 week – Baseline) in Neurocognitive age-adjusted scale score and Self-reported Cognition, by randomization group (n=87)

Age, chemotherapy, current use of tamoxifen or aromatase inhibitors, and time since surgery were tested as potential effect modifiers of the relationship between intervention and change in the Crystalized and Fluid Composite Scores as well as the Oral Symbol Digit Score and self-reported Cognitive Abilities and Cognitive Concerns. Time since surgery was a significant effect modifier for change in Oral Symbol Digit Score (p=0.079). Stratifying the sample at the median time since surgery (2 years) showed that participants who were 2 years or less from surgery had a significant intervention effect on the change in Oral Symbol Digit Score (beta=4.01; 95% CI=1.40, 6.63; p = 0.0033); in contrast, there was no intervention effect in participants who were greater than 2 years from surgery (see Figure 3). No modification effect was observed for age, chemotherapy or hormone therapy.

Figure 3.

Change (12 week – Baseline) in Oral Symbol Digit score, by median split of time since diagnosis, per randomization group (n=87)

Within the Exercise arm, a dose response was observed for physical activity such that greater increase in MVPA was positively associated with greater improvement in Oral Symbol Digit Score (b=0.20; p=0.016), self-reported Cognitive Abilities (b=0.99, p <.0001), and self-reported Cognitive Concerns (b= −0.68, p <.0001). Specifically, a 15-min/day increase in MVPA was associated with, on average, a 3.0 point increase in the Oral Symbol Digit Score, a 14.8 point increase in the standardized self-report Cognitive Abilities score, and a 10.2 point decrease in standardized self-report Cognitive Concerns. Greater increase in total activity (MVPA + light activity) was associated with greater improvement in the Fluid Composite Score (b=0.03; p=0.038) and the Picture Sequence Score (b=0.04; p=0.039), indicating that a 30-min/day increase in total activity was associated with, on average, a 0.84 point increase in the Fluid Composite Score and a 1.3 point increase in the Picture Sequence Score. All other components of the Fluid Composite Score (except List Sorting) showed a non-significant positive association with increased total activity (data not shown).

DISCUSSION

This study is one of the first randomized controlled trials of physical activity in breast cancer survivors with comprehensive measures of cognition as the primary outcome. The intervention was successful in increasing physical activity over 12-weeks, and both groups showed improved performance in most cognitive domains. However, out of 9 examined cognitive domains, only processing speed had significantly greater improvements in the Exercise arm compared to the Control arm. These findings are consistent with a recent RCT of 19 breast cancer survivors showing that the exercise arm exhibited greater improvements only for a measure of processing speed²³ and a cross-sectional study of 136 breast cancer survivors showing that accelerometer-measured physical activity was only related to one cognitive domain, processing speed.²⁴ While the cross-sectional study was conducted by investigators for the current study as well, it was with a different sample of breast cancer survivors and used a different neurocognitive test. Although the benefits of physical activity were limited to this domain, it is an important aspect of cognition as it can impact daily life,^34,35 is sensitive to cancer-related cognitive impairments, and is commonly impaired in breast cancer survivors.^36–38 In addition, processing speed plays a central role in cognition and can impact other cognitive tasks, especially those related to learning and memory.^35,39 Therefore, improved processing speed could lead to improvements in other aspects of cognition. Future studies are needed to determine if sustained physical activity over longer periods of time could lead to more wide-spread improvements in cognition.

Between group differences for both measures of self-reported cognition were not statistically significant, but given the small sample size, it could be suggestive of potential group differences. This is consistent with the significant dose-effect for MVPA that was seen for both aspects of self-reported cognition. While there are no guidelines for what constitutes clinically meaningful improvement on these two measures, for other PROMIS quality of life measures minimally important differences for cancer survivors have ranged from 2.5–6 based on standardized T-scores.⁴⁰ Therefore, with an average 2.7 point improvement in Cognitive Abilities and 4.8 point reduction of Cognitive Concerns in the Exercise arm, many intervention participants may have experienced clinically meaningful improvements in their everyday cognitive functioning. Larger trials are needed to fully test the impact of physical activity on self-reported cognition.

In the current study, MVPA-associated improvement in processing speed was significantly greater for women who were closer to surgery but did not vary based on other cancer-related factors including having received chemotherapy or not. The physical and mental impact of cancer is generally greater closer in time to diagnosis treatments,⁴¹ thus, there may be more potential for improvements to occur when physical activity is increased closer in time to treatments. The benefits for processing speed that were seen suggest that increasing physical activity may be most advantageous when initiated closer to time of surgery. Also of note, neurocognitive testing improved for both groups on most domains over the 12-week study. This improvement may be related to the practice effects from repeated measurements. For instance, familiarity with the tests may have contributed to participants feeling less nervous, which could lead to improved scores.^42–44 These findings highlight the important role of a Control arm in ensuring that observed changes in cognition are causally related to the physical activity intervention. There is growing evidence that cognitive problems in cancer patients are not limited to those who have received chemotherapy.^2,45 Thus, the lack of effect modification for chemotherapy and for hormonal therapy in our study suggests that increasing physical activity may be helpful for improving cognitive processing speed in breast cancer survivors regardless of the treatments they received.

A significant dose response between minutes of increased physical activity and improved neurocognitive test performance and improved self-reported cognition in the Exercise arm was observed. Interestingly, for the neurocognitive tests, a dose response was observed for total activity, comprised of MVPA and light activity. This finding suggests that all types of physical activity, not just those conducted at moderate intensity, could be beneficial for cognition. This finding is consistent with Yoga and qigong interventions that report improved cognition in cancer survivors^46–48 as well as research in non-cancer populations.²² Engaging in large amounts of MVPA can be challenging for most people; a focus on increasing all intensities of activity could be a more feasible strategy for breast cancer survivors. Future research should examine the necessary minutes needed of different activity intensities to improve cognition.

There are many potential mechanisms through which physical activity could improve cognition. Self-reported cognition is often associated with psychological factors including fatigue, anxiety, and depression.^49,50 Physical activity in cancer patients has been shown to decrease fatigue,^51–54 anxiety,⁵³ and depression,^51,53,54 suggesting that physical activity could improve self-reported cognition as well. Additionally, psychological factors, including depression and anxiety, can impact neurocognitive testing.⁵⁵ There are also several biological mechanisms through which physical activity could improve cognition. One is through the increase of cerebral blood flow and oxygen transport to the brain.^56–60 Increased cerebral blood flow achieved through physical activity can enhance production of brain-derived neurotrophic factor,^61–63 a biomarker of brain health that plays a central role in survival of neurons, neurogenesis, synaptic plasticity, and cognitive function.^64–66 Physical activity may also reduce the impact of accelerated aging that is seen in cancer survivors.^2,3,6 At the cellular level, dynamic epigenetic DNA methylation profiles are associated with aging,⁶⁷ cancer,^67–69 physical activity,⁷⁰ and cognition.^71,72 Importantly, age-related DNA methylation is associated with age-related gene transcription.⁶⁷ More research is needed to elucidate the mechanisms through which physical activity improves cognition to support developing targeted interventions.

Although this was a RCT trial using objective and self-reported measures for cognition and physical activity, several limitations should be noted. This was a modestly-sized study with limited power to detect differences between groups and because of the pilot nature of the study we did not adjust for multiple comparisons. The intervention only lasted for 12-weeks and may not have been long enough to detect the impact of physical activity on cognition. At baseline about 20% of participants were engaging in 150 minutes of accelerometer measured MVPA per week, which may have limited our ability to see group differences. The small sample size also limited our ability to determine the optimal dose and intensity of physical activity. These limitations highlight the need for longer, fully-powered trials to determine whether the effectiveness of increased physical activity improves aspects of cognition beyond processing speed and what the optimal dose and intensity of activity is for improving cancer-related cognitive decline. We chose to use the NIH Cognition Toolbox rather than the three core tests recommended in 2011 by the International Cognitive and Cancer Taskforce.⁷³ The goal of both initiatives was to increase comparability between studies by common methodology. We chose the Toolbox because it enhances comparability across disease populations and is more comprehensive in the domains measured. Future studies combining the NIH Cognition Toolbox with the measures suggested by the Taskforce would be helpful in further comparing results across oncology and other population studies. Other limitations include the predominantly well-educated, high intelligence (as indicated by the Crystalized Composite Score), non-Hispanic white sample of breast cancer survivors; future research in cancer populations with greater diversity is needed.

This study provides preliminary support for the efficacy of increasing physical activity to improve processing speed and potentially improve self-reported cognition in breast cancer survivors. With the growing interest in testing the potential of physical activity to improve cognition in cancer survivors^74,75 this and other studies are likely to contribute to our ability to make recommendations to the growing number of cancer survivors on effective interventions to improve cognition.