Contribution of Sleep Disruption and Sedentary Behavior to Fatigue in Survivors of Allogeneic Hematopoietic Cell Transplant

Abstract

Background

Fatigue is a prominent quality of life concern among recipients of hematopoietic cell transplantation (HCT).

Purpose

The present study investigated whether objectively measured sleep efficiency and sedentary behavior are related to greater reports of fatigue.

Methods

Eighty-two allogeneic HCT recipients who were 1–5 years post-transplant and returning for a follow-up visit participated (age M = 56, 52% female, 56% leukemia). They wore an actigraph assessing sleep efficiency and sedentary behavior for one week and completed an electronic log assessing fatigue each evening during the same period.

Results

Twenty-six percent of patients reported clinically meaningful fatigue. On average, fatigue was mild (M = 2.5 on 0–10 scale, SD = 2.0), sleep was disturbed (sleep efficiency M = 78.9%, SD = 8.9), and patients spent the majority of time in sedentary (M = 55.4%, SD = 10.2) or light (M = 35.9%, SD = 8.6) activity. Multilevel model analysis of between-person differences indicated that patients who experienced less efficient sleep the previous evening provided greater evening reports of average fatigue, b = –0.06, 95% CI (–0.11, –0.01). Similarly, within-person analyses indicated that when patients experienced less efficient sleep the previous evening or were more sedentary as compared to their average, they provided greater evening reports of average fatigue, b = –0.02, 95% CI (–0.05, –0.004); b = 4.46, 95% CI (1.95, 6.97), respectively.

Conclusions

Findings demonstrate that poor sleep and daily sedentary behavior are related to evening reports of fatigue and should be considered modifiable targets for intervention.

Among survivors of stem cell transplant fatigue is common and is linked to poor sleep and sedentary activity patterns.

Introduction

Hematopoietic cell transplantation (HCT) is a potentially curative treatment for hematologic cancers but carries a high risk of complications. Medical advancements over the last several decades have led to longer survival, less symptom burden, and better quality of life for transplant recipients. However, fatigue remains one of the most commonly reported symptoms before, during, and after transplant [1–4]. Fatigue has been shown to be significantly worse among HCT recipients in the post-treatment period compared to healthy individuals [5, 6]. Moreover, as many as one-third of allogeneic HCT (donor transplant) recipients experience clinically significant fatigue three or more years post-transplant [4, 7, 8]. Quality of life is often profoundly impacted for the subset of individuals who go on to experience persistent fatigue [4, 5].

Existing literature suggests that precipitants of fatigue are multifactorial [9]. Female sex, chronic pain, and patient-reported severity of chronic graft-versus-host disease (cGVHD), an inflammatory condition where donor cells attack host tissue, have been identified as risk factors for fatigue in HCT survivors [7]. While a number of factors likely contribute to the daily experience of fatigue, from a treatment perspective it is important to consider potentially modifiable characteristics, such as sleep and physical activity. Increasingly, sleep and activity patterns are measured objectively in cancer patients by wrist-based actigraphy, which uses an accelerometer to record and average physical movement [10]. Actigraphy has been used to measure sleep/wake patterns in breast and gynecologic cancer patients [11–13], patients with advanced cancer [14], and autologous HCT patients [15]. It has also been used to measure activity and sedentary behavior among cancer patients [13, 16, 17].

A number of studies have yielded evidence supporting links between fatigue and objectively measured sleep and activity in cancer populations [18–22]. Liu et al. [11] found that fatigue severity over the past week was positively associated with subjective sleep scores and actigraphy measured total naptime and was negatively associated with total wake time during the day among newly diagnosed women with stages I–III breast cancer. At the level of daily experience, it has been demonstrated that actigraphy-measured sleep disturbance during chemotherapy can initiate a symptom cascade leading to increased fatigue and depressive symptoms [23]. Moreover, objectively measured increases in physical activity have been associated with decreases in self-reported fatigue among patients receiving radiation treatment [24]. In contrast, several studies have failed to find associations [25–27] or reported mixed findings [28–30]. Notably, there is little to no evidence describing relationships of fatigue and objectively measured sleep and activity among HCT patients. An important distinction not drawn in prior research are possible differences in within- versus between-person contributions of sleep and activity to fatigue. While a between-persons approach compares a person to others, a within-person approach evaluates outcomes and predictors in relation to themselves [31]. Thus, using this method, we were able to evaluate whether a person is likely to rate themselves as more fatigued after nights with greater sleep disruption, as compared to nights when their sleep is less disrupted. This has potentially important implications for treatment of fatigue, and further investigation into daily relationships between fatigue and objectively measured sleep and activity is warranted.

The present study had two primary aims. First, it sought to characterize the daily experience and variability of patient-reported fatigue and objectively measured sleep and activity patterns among allogeneic HCT survivors. Second, it used a daily analysis approach to investigate over a 7-day period hypotheses that greater sleep disruption during the previous night and sedentary behavior during the day would be related to greater evening reports of fatigue.

Method

Participants

Participants were recruited from Moffitt Cancer Center. Eligible participants: (a) were diagnosed with a hematologic malignancy, (b) had been treated with an allogeneic HCT approximately 1–5 years prior to enrollment, (c) were ≥18 years of age, (d) had no history of other cancers other than non-melanoma skin cancer, (e) had no evidence of disease progression at time of enrollment, (f) were ambulatory at time of enrollment, (g) had internet access, (h) were able to speak and read English, and (i) were able to provide informed consent. Data were collected from June 2017 to February 2018.

Procedures

Participants were part of a larger study assessing the relationship between aggregated versus daily measures of fatigue and sleep. The study was approved by the University of South Florida Institutional Review Board. Eligibility was determined through consultation with clinical staff and medical record and registry data review. Patients were recruited during routine clinic visits and were given an actigraph and instructions for completing an electronic web-based daily log. Participants were instructed to wear the actigraph on their nondominant wrist for seven consecutive 24-hr periods and complete a daily log of their sleep and fatigue during that time. As part of the larger study, participants completed a questionnaire assessing demographic and other measures on their seventh day of participation. Participants returned all study materials in a postage-paid envelope.

Measures

Demographic characteristics

Participants completed a standardized form assessing age, gender, race, ethnicity, marital status, education, income, and employment status, as well as self-reported functional status.

Medical characteristics

The following medical characteristics were collected from the Moffitt Bone Marrow Transplant Registry and through medical record review: diagnosis, donor type, ablation, time since transplant, and highest known rating of cGVHD grade.

Sleep disruption and sedentary behavior

These variables were measured objectively using the ActiGraph GT9X Link (Pensacola, FL). Data from the actigraph were downloaded and analyzed using ActiLife v6.13.3 (ActiGraph, LLC, Pensacola, FL). Sleep indices during valid wear times were calculated using the Cole–Kripke algorithm [32] in combination with daily sleep logs of bed and wake times. The primary sleep variable of interest was specified a priori as sleep efficiency (i.e., percentage of time spent sleeping in relation to time spent in bed), with efficiency scores of ≥85% representing efficient sleep [38]. Other exploratory sleep variables included: sleep onset latency (SOL, i.e., amount of time taken to fall asleep), wake after sleep onset (WASO, i.e., minutes awake after an extended period of sleep), and total sleep time (TST, i.e., time spent asleep at night) [33]. Sedentary behavior indices during valid wear times were calculated using the Freedson Adult (1998) algorithm [34], which categorizes activity as: sedentary, light, moderate, vigorous, and very vigorous. The primary sedentary behavior measure of interest was specified a priori as sedentary time (i.e., percentage of time spent engaging in sedentary activity).

Daily fatigue

Participants provided daily fatigue ratings at the end of the day for seven consecutive days through the use of an electronic interface. Participants received standard text messages daily granting access to the interface and reminding them to complete their fatigue ratings. Participants were asked to record their daily fatigue ratings each evening between 6 and 9 pm. Questions were adapted from the Fatigue Symptom Inventory [35] and included: level of fatigue right now (momentary fatigue), average fatigue during the day (average fatigue), and how much fatigue interfered with activities during the day (fatigue interference). Each item was rated on an 11-point Likert scale from 0 to 10 with higher scores indicating greater fatigue/interference. Average fatigue scores of ≥4 are considered to be indicative of moderate to severe fatigue (NCCN Guidelines), and therefore clinically meaningful.

Statistical Analyses

Data analyses were performed using SAS Version 9.4 (Cary, NC). Consistent with other studies, data from participants who contributed at least 3 days of fatigue and actigraphy data were included in analyses [36]. Study variables assessed daily were averaged over the 1-week study period to create summary scores. Descriptive statistics were used to characterize the sample and variables assessed daily over the 1-week period. Pearson correlation coefficients were examined to determine relationships among averaged scores for fatigue, sleep disruption, and sedentary behavior. Pearson correlation and analysis of variance were used to investigate relationships between fatigue outcomes and demographic and clinical factors to determine covariates for inclusion in subsequent models.

Multilevel models were created using SAS PROC MIXED to evaluate whether daily assessments of sleep and sedentary behavior were related to reports of worse fatigue. Preliminary models were used to evaluate the extent to which measures of sleep disruption, sedentary behavior, and fatigue varied between-persons (differences in average scores) and within-persons (fluctuations in scores from a person’s average). Subsequent models focused on same-day relationships (i.e., relationship between same-day sedentary behavior and daily fatigue) and temporal relationships (i.e., relationship between sleep efficiency and the next evening’s daily fatigue). Each fatigue outcome (i.e., momentary, interference, and average) was analyzed in separate models. These models included both sleep efficiency and sedentary behavior as between-person predictors of fatigue, with each centered at the sample mean, as well as sleep efficiency and sedentary behavior as within-person predictors of fatigue, with each centered at the person-level mean [37]. The person-level mean is a person’s usual level of daily sleep efficiency or sedentary behavior, as represented by each person’s mean across all seven days of the study. All models covaried for employment status based on results of preliminary analyses.

Results

Recruitment and Sample Characteristics

Supplementary Fig. 1 depicts patient flow through the study. The final sample for analytic purposes consisted of 82 patients. Compared to patients not included in the final sample (n = 35), those included (n = 82) were older (t = –2.37, p < .05). The groups did not differ on the basis of gender, ethnicity, or race (p values > .05).

Demographic and medical characteristics of the sample are presented in Table 1. The majority of participants were female, non-Hispanic White, and married or partnered. The sample was well educated with the majority having at least some college. At study entry, participants were an average of 2.5 years post-transplant. The majority were diagnosed with leukemia (56%) and received transplanted cells from an unrelated donor (66%).

Table 1.

Demographic and medical characteristics (N = 82)

Characteristic	n (%)	ANOVA or correlation with average fatigue
Age, years		r = –.04 (p = .71)
M (range)	56 (25–74)
Gender, no. (%)		p = .20
Male	39 (47.6)
Female	43 (52.4)
Ethnicity, no. (%)		p = .28
Not Hispanic	75 (91.5)
Hispanic	7 (8.5)
Race, no. (%)		p = .30
White	75 (91.5)
Nonwhite	7 (8.5)
Marital status, no. (%)		p = .33
Married or living with partner	60 (73.2)
Not married	22 (26.8)
Education, no. (%)		p = .24
Less than college grad	33 (40.2)
College grad or more	49 (59.8)
Employment, no. (%)		p = .0014
Work full-time or part-time	26 (31.7)
Retired	26 (31.7)
Disabled	23 (28.1)
Other	7 (8.5)
Income, no. (%)		p = .62
<40K	24 (29.3)
≥40K	58 (70.7)
Cancer type, no. (%)		p = .63
Leukemia	46 (56.1)
ALL	8 (17.5)
AML	34 (73.9)
CLL	2 (4.3)
CML	2 (4.3)
Non-Hodgkin lymphoma	14 (17.1)
Myelodysplastic syndrome	7 (8.5)
Multiple myeloma	6 (7.3)
Myeloproliferative syndrome	5 (6.1)
Other	4 (4.9)
Ablation, no. (%)		p = .35
Myeloablative	41 (50.0)
Nonmyeloablative	41 (50.0)
Donor type, no. (%)		p = .44
Related	28 (34.1)
Matched unrelated	54 (65.9)
Time since transplant, days		r = .05 (p = .68)
M (range)	942 (370–1889)
cGVHD grade, no. (%)		p = .19
None	24 (29.3)
Mild	18 (22.0)
Moderate	32 (39.0)
Severe	7 (8.5)
Unknown	1 (1.2)

Actigraphy-Assessed Sleep Disruption and Sedentary Behavior

A total of 87% of participants provided 6–7 days of actigraphy data, with the remaining 13% providing between three and five days of data. Table 2 presents descriptive statistics for sleep disruption and activity variables aggregated over the one-week study period. On average, participants demonstrated poor sleep efficiency (M = 78.93%, SD = 8.88) relative to a benchmark value of 85% [36]. While participants on average were able to fall asleep within 10 min (M=8.61, SD = 6.83) and slept for 6.7 hours (M = 399.64 min, SD = 63.64), they were awake an average of 99.29 min (SD = 50.41) during the night after initially falling asleep. On average, participants spent considerable time engaged in sedentary (M = 55.41%, SD = 10.19) and light activity (M = 35.86%, SD = 8.58), and less time in moderate (M = 8.73%, SD = 4.37) or more vigorous activity (0%).

Table 2.

Descriptive statistics

Variables	M (SD)	Min	Max
Sleep disruption
Sleep efficiency, %	78.93 (8.88)	51.79	93.19
SOL, min	8.61 (6.83)	0.43	28.57
WASO, min	99.29 (50.41)	24.86	288.86
TST, min	399.64 (63.64)	252.57	541.43
Sedentary behavior
Sedentary time, %	55.41 (10.19)	36.22	77.45
Light time, %	35.86 (8.58)	15.24	53.84
Moderate time, %	8.73 (4.37)	0.50	24.90
Vigorous time, %	0.00 (0.00)	0.00	0.00
Very vigorous time, %	0.00 (0.00)	0.00	0.00
Fatigue
Momentary fatigue	3.57 (1.91)	0	7.6
Fatigue interference	4.69 (2.05)	0	8.5
Average fatigue	2.51 (2.01)	0	7.83

Note. All variables were aggregated across the week and represent summary scores. SOL, sleep onset latency; WASO, wake after sleep onset; TST, total sleep time; sedentary time, percent spent in sedentary activity, light time, percent spent in light activity, moderate time, percent spent in moderate activity; vigorous time, percent spent in vigorous activity; very vigorous time, percent spent in very vigorous activity. Fatigue items were assessed as follows: (1) momentary fatigue = “rate your level of fatigue right now,” (2) fatigue interference = “rate how much did fatigue interfere with your general level of activity today,” and (3) average fatigue = “rate your average level of fatigue today.”

Daily Experience and Variability of Fatigue

Participants completed 93% (533/574) of daily fatigue ratings. Table 2 depicts fatigue scores aggregated over the 1-week study period. Although on average, patients reported mild fatigue that moderately interfered with daily life, 26% of patients reported a clinically meaningful aggregated average fatigue score of four or greater. Additionally, 74% of participants reported a clinically-meaningful average fatigue score of four or greater on at least one study day.

Relationships Among Sleep, Sedentary Behavior, and Fatigue

Table 3 presents bivariate relationships between fatigue variables and sleep disruption and sedentary behavior aggregated over the 1-week study period. Significant negative relationships were observed between average fatigue and fatigue interference and sleep efficiency, such that greater fatigue was associated with worse sleep efficiency; momentary fatigue was not associated with sleep efficiency. Significant positive relationships were observed between all fatigue variables and wake after sleep onset (WASO), such that greater fatigue was associated with greater WASO. In contrast, significant relationships were not observed between fatigue and sleep onset latency (SOL) or total sleep time (TST). Few significant relationships were evident between fatigue and activity variables; sedentary time was positively associated with average fatigue and light time was negatively associated with average fatigue. Activity variables were not associated with other fatigue indices.

Table 3.

Pearson correlation coefficients of fatigue variables with sedentary behavior and sleep

Variables	Momentary fatigue	Fatigue interference	Average fatigue
Sleep efficiency	–0.17	–0.24*	–0.37***
SOL	–0.07	0.02	0.11
WASO	0.26*	0.32**	0.44***
TST	0.10	0.05	0.00
Sedentary time	0.10	0.11	0.23*
Light time	–0.08	–0.06	–0.23*
Moderate time	–0.06	–0.13	–0.08

Note. * p < .05, ** p <.01, *** p < .001. All variables were aggregated across the week. Sedentary time, percent spent in sedentary activity; light time, percent spent in light activity; moderate time, percent spent in moderate activity; SOL, sleep onset latency; WASO, wake after sleep onset; TST, total sleep time.

Relationships between average fatigue and demographic and clinical factors are depicted in Table 1. Average fatigue was related to employment status, with patients who identified as disabled reporting greater fatigue. Age, gender, ethnicity, race, marital status, education, income, cancer type, ablation, donor type, time since transplant, and highest known cGVHD grade were not related to average fatigue. Based on these results, employment status was entered as a covariate in subsequent analyses.

Sources of Variability in Daily Measures of Sleep, Sedentary Behavior, and Fatigue

Figure 1 depicts variance estimates for fatigue variables, sleep disruption, and sedentary behavior as a function of differences across HCT recipients (between-person) and how symptoms differed day to day (within-person). As shown, the majority of the variance in fatigue, sleep disruption, and sedentary behavior was driven by between-person differences in average scores (56%–71%). There was sufficient within-person variation to proceed with planned analyses.

Fig. 1.

Variability in daily measures of fatigue, sleep disruption, and sedentary behavior.

Sleep Disruption and Sedentary Behavior Predicting Evening Reports of Fatigue

These models tested the hypothesis that sleep efficiency the previous evening and sedentary behavior during the day would predict evening reports of fatigue at the level of the group (between-person differences) and the individual (within-person differences). Results are displayed in Table 4. In all models, employment status was related to evening fatigue (ps < .05).

Table 4.

Multilevel linear regression for the association between sleep efficiency, sedentary behavior, and evening fatigue, controlling for employment status

Predictor	Effect	Momentary fatigue	Fatigue interference	Average fatigue
		b (95% CI)	b (95% CI)	b (95% CI)
Employment	–	–	–	–
Sleep efficiency	Between-person	–0.02 (–0.07, 0.03)	–0.04 (–.09, 0.02)	–0.06 (–0.12, –0.01)
	Within-person	–0.02 (–0.04, 0.01)	–0.02 (–0.04, 0.00)	–0.02 (–0.05, –0.01)
Sedentary time	Between-person	0.59 (–3.22, 4.40)	0.82 (–3.43, 5.09)	3.66 (–0.12, 7.45)
	Within-person	–0.63 (–3.14, 1.89)	4.00 (1.32, 6.68)	4.46 (1.95, 6.97)

Note. Each fatigue outcome was analyzed in separate models. All models included both sedentary time and sleep efficiency as predictors with between and within-person effects examined for each predictor. 95% confidence intervals that do not cross zero are statistically significant, with all bolded items indicating p < .05.

Average fatigue

Between-person differences in sleep efficiency but not sedentary behavior was related to evening reports of average fatigue (b = –0.06, p < .05; b = 3.66, p > .05, respectively). That is, when patients reported less efficient sleep the previous evening as compared to others, this pattern was associated with greater evening reports of average daily fatigue. Within-persons differences in both sleep efficiency and sedentary behavior were related to evening reports of average fatigue (b = –0.02, p < .05; b = 4.46, p < .001, respectively). That is, when patients experienced less efficient sleep the previous evening or were more sedentary as compared to their own average, this pattern was associated with greater evening reports of average daily fatigue.

Fatigue interference

Neither between- nor within-person differences in sleep efficiency the previous evening were related to next evening reports of fatigue interference (b = –0.04, p > .05; b = –0.02, p > .05, respectively). In contrast, within- but not between-person differences in sedentary behavior were related to evening reports of fatigue interference (b = 4.00, p < .01; b = 0.82, p > .05, respectively). That is, when patients reported they were more sedentary as compared to their own average, this pattern was associated with greater evening reports of fatigue interference.

Momentary fatigue

Between- and within-person differences in sleep efficiency and sedentary behavior were unrelated to evening reports of momentary fatigue (p values > .05).

Discussion

The present study sought to examine the daily experience of and relationships between fatigue, sleep, and sedentary behavior among patients who received allogeneic HCT 1–5 years previously. It is noteworthy as being one of the first published reports based on objectively measured sleep and sedentary behavior in HCT survivors. Although overall averages suggest that fatigue was mild, there was considerable variability across days and individuals. The high level of between- and within-person variability may be reflective of where study participants are in the post-transplant trajectory. While fatigue is nearly ubiquitous during inpatient hospitalization for transplant [2, 3], there are often dramatic reductions in fatigue after patients return home. However, a subset of patients may go on to experience persistent, clinically relevant fatigue long after transplant, with approximately one-fourth of patients reporting clinically meaningful fatigue in this sample. Results also revealed considerable within-person variability in fatigue across study days. Thus, assessing fatigue at a single time using the traditional retrospective measurement approach, which asks patients to summarize their experience of fatigue over a period of a week for example, may be insufficient for capturing this within-person variability. Descriptive analysis of actigraphy indices revealed that patients spent the majority of their time in sedentary or light activity and experienced trouble staying asleep throughout the night, as reflected in an average sleep efficiency rating less than a benchmark of 85% [38].

The central focus of the present study was to determine whether a patient would be likely to report greater fatigue after nights with greater sleep disruption or days with greater sedentary behavior, as compared to times when their sleep was less disrupted and they were more active. We controlled for employment status within this set of analyses given evidence from the current study that patients identifying as disabled were more likely to report greater fatigue, a finding consistent with prior literature and warranting additional future study [39, 40]. Results indicated that patients’ average levels of sleep efficiency was negatively related to their ratings of average fatigue; in other words, when patients had less efficient sleep compared to others, they reported greater fatigue the next evening. A similar pattern of relationships was found when we took a within-persons approach; when patients had a worse night’s sleep than was typical for themselves, they were more likely to rate their average fatigue as greater the next evening. We also explored relationships between fatigue and sedentary behavior using the same approach. Results indicated that patients’ average levels of sedentary behavior was not related to any of the fatigue outcomes. In contrast, a within-persons approach demonstrated a positive relationship between average levels of sedentary behavior and greater fatigue; when individuals were less active than was typical for themselves, they were more likely to rate their average fatigue and fatigue interference with daily activities as greater the same evening.

Across these analyses, the most consistent pattern of relationships was observed for average fatigue; corresponding relationships with fatigue interference were mixed and relationships with momentary fatigue were not significant. This pattern of results suggests that patients’ summaries of fatigue for the day may be more meaningful than a single, momentary rating when conducting a daily analysis of relationships with actigraphy-assessed behavior.

Findings from the present study are in line with a growing body of evidence demonstrating links between objectively measured behavior and self-reported fatigue among cancer patients using a daily analysis approach [13, 23] and more traditional summary approaches [18, 31, 41, 42]. Results from the present study were obtained through the use of sophisticated modeling techniques capable of identifying and elucidating relationships at the between-person level as well as at the within-person level. Moreover, results were obtained through actigraphy which is less burdensome on patients and reduces retrospective reporting biases. Thus, findings represent an important addition to previous literature and warrant replication and further study.

The present study had several limitations. First, findings are based on a limited one-week study period. Notably, the current study does not address the possible contribution of fatigue to subsequent sleep problems. In addition, though multiple dimensions of fatigue were evaluated as part of the present study, single-item assessments were used for each dimension. Finally, participants were asked each evening of the study to summarize their experience of fatigue for the day rather than completing fatigue assessments multiple times per day. Although there is evidence to suggest that fatigue fluctuates throughout the course of a single day [43], the present study is not able to make conclusions regarding diurnal fatigue patterns and possible within-day fluctuation in fatigue. Multiple assessments throughout the day would also allow computation of a daily fatigue average, which may be less susceptible than a single retrospective report to the influence of other variables; however, this was not done for the present study. Despite these limitations, the present study adds to existing literature characterizing the daily experience of fatigue among patients who have received allogeneic HCT. Moreover, the study characterized sleep and activity behavioral patterns using actigraphy, a methodology that complements use of self-report measures in providing a more complete picture of these behaviors in real time.

In sum, the present study builds on previous literature characterizing fatigue and its relationship with sleep and activity among allogeneic HCT recipients. Results suggest several future directions. First, further exploration into daily relationships among sleep, activity, and fatigue is needed. To date, most studies that have looked at these relationships have done so at the group level; results from the present study suggest that both interindividual and intraindividual variation are important for gaining a richer understanding of the complexity of these factors and how they relate. Second, results support investigation into the efficacy of interventions focused on ameliorating sleep disruption and reducing sedentary activity as means for lessening the severity and impact of persistent fatigue on daily life among HCT recipients. While efficacy of these interventions has been demonstrated with a large literature in other cancer populations [44–46], the same cannot be said among HCT survivors, emphasizing the need for large-scale, rigorous intervention trials aimed at behavioral targets including reducing sedentary behavior and improving sleep in the years following transplantation. Finally, results suggest that accelerometers represent a valuable methodology for measurement of behaviors alongside traditional self-report methods and should be incorporated into intervention research.

Compliance with Ethical Standards

Authors’ Statement of Conflict of Interest and Adherence to Ethical Standards H.S.L.J., PhD, discloses that she has consulted for RedHill BioPharma, Merck, and Janssen Scientific Affairs. All other authors declare that they have no conflict of interest.

Authors’ Contributions All authors made substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data. All were involved in drafting the article or revising it critically for important intellectual content. All provided final approval of the manuscript.

Ethics Approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study.

Disclosure The views expressed are those of the authors and do not necessarily represent those of the National Cancer Institute.