The impact of dexamethasone and prednisone on sleep in children with acute lymphoblastic leukemia

Abstract

Purpose

Corticosteroids can affect sleep patterns, mood, and behavior. Two of the most commonly prescribed corticosteroids in acute lymphoblastic leukemia (ALL), dexamethasone and prednisone, may impact sleep differently, but no research has compared these medications in children. The current study tested the hypothesis that dexamethasone and prednisone differentially affect sleep in children with ALL to understand how these medications contribute to health-related quality of life (HRQL).

Methods

Parents of 81 children 3–12 years old in maintenance therapy for ALL completed a baseline measure of child sleep (dexamethasone n = 55, prednisone n = 26), and 61 parents returned 28 days of child sleep diaries starting the first day of a 5-day steroid course (dexamethasone n = 43, prednisone n = 18). Parents also completed measures of HRQL and fatigue on the last day of steroids and the last day of the month.

Results

At baseline, parents reported more sleep disturbances in children taking dexamethasone than prednisone. Across the month, children taking dexamethasone experienced poorer sleep quality compared to children taking prednisone. During corticosteroid treatment, children taking dexame thasone also experienced more night awakenings than children taking prednisone. Sleep variables accounted for almost half of the variance in HRQL during time off steroids and also significantly contributed to fatigue during the corticosteroids course and time off corticosteroids.

Conclusions

Sleep is an essential component of HRQL in children taking corticosteroids, and the impact on sleep is more pronounced in children taking dexamethasone compared to prednisone. Screening for sleep disturbances and offering brief interventions to manage steroid-related sleep disruptions may improve HRQL.

Corticosteroids, generally recognized as disruptive to child behavior including sleep [1], are an essential component for the treatment of acute lymphoblastic leukemia (ALL). The 2-to 3-year-long ALL maintenance treatment consists of monthly 5-day corticosteroid courses with either prednisone or dexamethasone [2]; however, it is unknown whether behavioral disruptions are consistent across these medications. During maintenance therapy, children may return to school and activities they enjoyed pre-cancer, but side effects of steroid treatment such as disrupted sleep may limit engagement and health-related quality of life (HRQL) [1, 3].

In ALL treatment, dexamethasone is frequently chosen over prednisone due to reductions in central nervous system and bone marrow relapses [4]; however, some research has indicated that dexamethasone may result in more short- and long-term side effects [5–7]. A recent review comparing the acute and long-term neuropsychological effects of dexamethasone versus prednisone found mixed results regarding the impact of these steroids on behavior, mood, and HRQL [8]. The authors concluded that given the increased evidence of efficacy of dexamethasone on survival, inconsistent evidence of long-term neuropsychological risks, is outweighed by curative benefits. However, enhanced understanding of the differences between steroid groups on psychosocial functioning is important for family education, effectively treating side effects, and maximizing HRQL during ALL treatment. Sleep was notably absent in the studies reviewed by Warris and colleagues [8] despite the inclusion of “sleep” as a literature review search term.

Child sleep is an understudied area of HRQL [9] that can significantly impact cognitive functioning [10], behavior [11], mood [12], academic progress [12], and family functioning [13]. Furthermore, the relationships between sleep quality and quantity with immune functioning [14], wound healing [15], and resistance to infection [16] suggest that sleep is a logical and important target of assessment and intervention to improve physical and physiological outcomes. Sleep can be measured objectively, using polysomnography or actigraphy (wrist accelerometer used to estimate sleep); however, these methods fail to capture nuances of sleep-related behavior in the home. As such, sleep diaries can be useful to understand reasons for disrupted sleep, behaviors around sleep, and perceptions of sleep to inform behavioral interventions when needed [17].

Independently, both dexamethasone and prednisone have been reported to impact sleep, but no research has compared the two steroids in any pediatric population. Studies describing sleep in children with ALL taking prednisone have found more sleep disturbances [18, 19], more night awakenings [19], and greater fatigue [19] during steroid courses compared to time off steroids. These studies are limited in the measurement of sleep by one global item [18] and by small samples sizes [19]. Better powered, more recent studies have described the impact of dexamethasone on sleep finding longer sleep duration, fewer awakenings, and more napping by actigraphy but more difficulty in falling asleep, insufficient sleep, and more awakenings by parent report during dexamethasone courses [20, 21]. Discrepancies between parent report and actigraphy highlight the importance of further study of the impact of steroids on sleep.

The current study advances descriptive work within each steroid group by comparing children taking dexamethasone to those taking prednisone on sleep and fatigue over 1 month of maintenance treatment. Additionally, the study describes the relationship of disrupted sleep to fatigue and HRQL, between and within steroid groups. Understanding the side effects is important in ALL in the context of improved survival rates and a focus on improving HRQL [22]. Behavioral sleep interventions may minimize the impact of steroids on family functioning. Such behavioral sleep interventions are focused, effective, and durable [23], and their application in pediatric cancer is a novel approach to improving child and family HRQL.

Methods

Participants

English-speaking caregivers of children ages 3 to 12 with ALL categorized as low, standard, or high risk ALL (as documented in the electronic medical record), in the second month through the end of the maintenance phase of the treatment, and taking dexamethasone or prednisone as determined by ALL treatment protocol were included. Patients were treated according to the following COG treatment protocols (either on study or “like” protocol): AALL 0232 (n = 12), AALL0331 (n = 37), AALL07P4 (n = 3), and AALL0932 (n = 21). All patients received vincristine on the day they started corticosteroids. Children with genetic syndromes (e.g., Down syndrome) or neurological impairment (as described by the medical team or caregiver), who received cranial radiation, or with relapsed ALL were excluded. Participants were enrolled between January 2010 and October 2014 (Fig. 1). Of the 132 eligible families, 97 were approached, 12 were eligible but were not approached prior to the end of enrollment, and 21 completed maintenance therapy or aged out of eligibility prior to being approached. Of the 81 families enrolled, 61 returned sleep diaries. A total of 1620 diary days were completed (M=27, SD =3.20, range=10–28 days). There were no differences in demographic factors (age, gender, ethnicity, and socioeconomic status) or baseline reports of sleep problems between diary completers and non-completers.

Fig. 1.

Study enrollment flow chart

Procedures

After institutional review board approval, families were approached during monthly oncology clinic appointments for maintenance chemotherapy, which coincide with the beginning 5-day corticosteroid courses. Interested families provided informed consent and the child’s primary caregiver completed baseline measures during the clinic visit. Informed consent was provided by all individual participants in this study. Families were given a 28-day sleep diary to complete at home. Weekly reminder calls, e-mails, or text messages were made to families to maintain engagement and answer questions. During the initial year and a half of data collection, families were given an option to complete the diary electronically. There was generally low interest in electronic diaries (7 of 40 elected this option), thus electronic diaries were discontinued. Families who choose electronic diaries were of lower social status [F(1, 80) = 4.49, p = 0.037] but similar to paper diary completers in other demographic and baseline variables.

Measures

Baseline measures

Medical chart review

Participants’ medical charts were reviewed to confirm exclusion criteria and document chemotherapy protocols and to collect disease type and risk group.

Developmental history questionnaire

Parents completed a background information form to collect child demographics, medical, and developmental histories.

Barratt Simplified Measure of Social Status is a shortened version of the Hollingshead Four Factor Form of Social Status [24]. Parents’ and grandparents’ education history and occupation were used to estimate social class [25].

The Abbreviated Children’s Sleep Habits Questionnaire (CSHQ) is a 33-item parent report questionnaire about the child’s last week of sleep [26]. Parents indicate the frequency of sleep behaviors on the continuum of “rarely” to “usually.” Caregivers completed the CSHQ regarding their child’s sleep for the week prior to starting steroids (Table 2). Higher total scores indicated greater sleep disturbance. Reliability was acceptable (α= 0.77).

Table 2.

General linear models comparing sleep variables between steroid groups and on-vs. off-steroid weeks

		Total sleep timea	Sleep onset latencya	Sleep qualitya		Night wakeningsb	Nappingb
Prednisone vs. dexamethasone, on steroid	Mean difference	−0.39	0.12	1.21*	Rate ratio	0.45*	0.50
	95 % CI	−0.90, 0.14	−0.25, 0.50	0.25, 2.18	95 % CI	0.23, 0.85	0.20, 1.21
Prednisone vs. dexamethasone, off steroid	Mean difference	−0.19	0.02	0.86*	Rate ratio	0.70	0.70
	95 % CI	−0.66, 0.27	−0.23, 0.27	0.15, 1.57	95 % CI	0.37, 1.30	0.32, 1.50
Off- vs. on-steroid, dexamethasone	Mean difference	0.06	−0.03	0.70**	Rate ratio	0.54**	0.70**
	95 % CI	−0.09, 0.20	−0.10, 0.05	0.35, 1.05	95 % CI	0.42, 0.69	0.57, 0.85
Off- vs. on-steroid, Prednisone	Mean difference	0.25*	−0.13	0.35	Rate ratio	0.84	0.98
	95 % CI	0.01, 0.50	−0.30, 0.03	−0.24, 0.93	95 % CI	0.54, 1.32	0.72, 1.33

^*p < 0.05

^**p < 0.01

^aLinear models

^bPoisson models

Post-baseline measures

Sleep diary

Parents completed an 18-item sleep diary for 28 days based on the work of Acebo and colleagues [27]. The sleep diary assessed sleep time (calculated based on parent reports of bedtime to wake time), number of naps, sleep onset latency (time between bedtime and time child fell asleep), night awakenings, and sleep quality (a 1–10 scale where 1 was “poor sleep” and 10 was “slept well”).

Caregivers were asked to complete the Pediatric Quality of Life Inventory (PedsQL) and the Multidimen sional Fatigue Scale (MFS) on the last day of the steroid course (day 5) and the last day of the month (day 28) for the prior 5 days to allow for a comparison of HRQL and fatigue on and off steroids. The PedsQL is a 23-item parent report form used to assess the child’s physical, emotional, social, and school functioning [28]. The 18-item MFS assesses general fatigue, sleep/rest fatigue, and cognitive fatigue [28]. For both measures, the total score was used and higher scores indicate better functioning. Reliability was acceptable (PedsQL α = 0.91–0.94 α = 0.94–0.95).

Data analysis

Preliminary analysis examined differences between steroid groups on demographic (age, gender, social status, and number of siblings) and treatment-related variables (month in maintenance, risk group). Groups were also compared on baseline sleep disturbance using the CSHQ total score. CSHQ subscales were compared to normative values for the CSHQ [26], but due to limited power, significance tests were not conducted. A p value of <0.05 was considered significant for this study.

Longitudinal general linear models were used to examine differences in sleep diary variables between steroid groups, with linear models for the continuous variables including total sleep time, sleep onset latency, and sleep quality, and Poisson models for frequency count variables night awakenings and napping. In each longitudinal model, the daily measures (e.g., total sleep time on each day) were used as outcomes. The independent variables in the longitudinal model included steroid group, indicator of steroid week [0 = steroid week (week 1); 1 = off steroid (weeks 2–4)], and the group by steroid week interaction. The inclusion of the interaction term allowed us to estimate the main effect of steroid group on sleep within the steroid week and off-steroid period separately and also the main effect of on/off steroid time on sleep within each steroid group separately. We also adjusted for baseline variables that differed significantly between steroid groups by including these variables as covariates in longitudinal models. To control for differences in sleep schedules when attending school, daily school attendance was also entered into each model as a time-varying covariate. In all longitudinal models, generalized estimating equation (GEE) methods were used to account for potential correlations among the repeated sleep measurements within the same subject. To summarize the results, mean differences between groups and their 95 % confidence intervals (CIs) were estimated from linear models (for total sleep time, sleep onset latency, and sleep quality), and rate ratios between groups and their 95 % CIs were estimated from Poisson models (for night awakenings and napping). Observed power was 0.67 to detect group differences [29].

HRQL and fatigue were compared between and within steroid groups during time on steroids and time off steroids. To understand the contribution of sleep to HRQL and fatigue during steroids, four step-wise regression models were run: HRQL on steroids, fatigue on steroids, HRQL off steroids, and fatigue off steroids. Demographic and treatment-related covariates hypothesized to contribute to HRQL and fatigue were entered into the models: demographic variables (age, gender) were entered in step 1, medical variables (risk group, steroid, month in maintenance therapy) were entered in step 2, and sleep variables (total sleep time, sleep onset latency, night awakenings, sleep quality, napping) were entered in step 3.

Results

Preliminary analyses

The description of the sample is presented in Table 1. There were no significant differences between steroid groups on age, gender, month in maintenance, social status, or number of siblings. There were more children in the high risk group taking prednisone (n = 24) than dexamethasone [n = 4; χ²(2) = 56.45, p < .001]. Risk group was therefore entered into GEE models as a covariate.

Table 1.

Sample demographics

	Total sample		Dexamethasone		Prednisone
	N = 81		n = 55		n = 26
	M	SD	M	SD	M	SD
Age at diagnosis	4.51	2.30	4.41	2.10	4.72	2.70
Age on study	6.21	2.22	6.17	2.05	6.29	2.60
Month in maintenance	12.71	8.88	13.88	9.10	10.24	7.77
Barrett measure of social status	45.46	10.97	46.58	10.15	43.10	12.43
Children’s Sleep Habits Questionnaire total score*	45.39	7.23	46.56	7.29	42.90	6.56
	M	SD	M	95 % CI	M	95 % CI
PedsQL total score day 5a	73.96	17.14	69.52	63.52, 75.53	84.30	73.67, 94.94
PedsQL total score day 28a	82.11	15.32	78.77	73.30, 84.25	89.52	80.15, 98.90
MFS total score day 5a*	70.71	19.55	64.54	57.97, 71.11	85.10	73.47, 96.74
MFS total score day 28a*	83.12	15.34	79.06	73.75, 84.37	92.15	83.06, 101.24
	n	%	n	%	n	%
Female	35	43	21	38	14	54
ALL risk group**
Low	2	2	2	4	0	0
Standard	51	63	49	89	2	8
High	28	35	4	7	24	92
Treatment protocol
AALL 0232	12	16	2	4	10	38
AALL 0331	37	45	37	67	0	0
AALL07P4	3	4	0	0	3	12
AALL0932	21	26	16	29	5	19
AALL1131	8	10	0	0	8	31
Child ethnicity
Latino	5	6	3	5	2	8
Non-Latino	76	94	52	95	24	92
Child’s race
African-American	6	7	6	11	0	0
Asian	3	4	2	4	1	4
Caucasian	58	71	39	70	19	73
Others	3	4	2	4	1	4
More than one race	11	14	6	11	5	19
Caregiver completing measures
Mother (2 adoptive mothers)	65	80	44	84	19	73
Father	16	20	9	16	7	27
Number of siblings in the home
Only child	14	17	9	16	5	19
1	32	40	20	36	12	46
2	23	28	17	31	6	23
3	9	11	7	13	2	8
≥4	3	4	2	4	1	4

^*p < 0.05;

^**p < 0.01

^aSteroid group comparisons are adjusted for risk group

Baseline sleep

The average total CSHQ score for the current sample (M = 45.39, SD = 7.23) was higher than the clinical cut-off score of 41; 67 % (n = 54) of our sample scored greater than 41 [26]. Parents of children taking dexamethasone reported that their children had significantly more sleep problems at baseline compared to the prednisone group [F(1, 79) = 4.72; p = .033; Table 1]. CSHQ total score was therefore entered into GEE models as a covariate. CSHQ subscales that were higher than normative values were bedtime resistance, sleep anxiety, sleep onset delay, parasomnias, and daytime sleepiness.

Sleep comparison between steroid groups

Average daily values of total sleep time, sleep onset latency, sleep quality, night awakenings, and nap frequency by steroid group are presented in Figs. 2 and 3. During the steroid week, children taking prednisone experienced better sleep quality (mean difference 1.21, 95 % CI 0.25 to 2.18, p = 0.014) and fewer night awakenings (rate ratio 0.45, 95 % CI 0.23 to 0.85, p = 0.013) than children taking dexamethasone (Table 2). There were no differences between total sleep time, sleep onset latency, and napping between the groups during the steroid week.

Fig. 2.

Total sleep time, sleep onset latency, and sleep quality by steroid. Vertical dotted line at day 7 indicates end of steroid period

Fig. 3.

Night awakenings and daily naps by steroid. Vertical dotted line at day 7 indicates end of steroid period

For the time off steroids, children taking prednisone continued to experience significantly better sleep quality than children taking dexamethasone (mean difference 0.86, 95 % CI 0.15 to 1.57, p = 0.018). There were no differences on night awakenings, total sleep time, sleep onset latency, and napping between the groups during the weeks off steroids.

Sleep difference by off- vs. on-steroid weeks

Children taking dexamethasone demonstrated significantly better sleep quality (mean difference 0.70, 95 % CI 0.35 to 1.05, p < 0.001), fewer night awakenings (rate ratio 0.54, 95 % CI 0.42 to 0.69, p < 0.001), and napped less frequently (rate ratio 0.70, 95 % CI 0.57 to 0.85, p = 0.001) during the off-steroid weeks compared to steroid week (Table 2). Total sleep time and sleep onset latency were consistent across the month in children taking dexamethasone.

For children taking prednisone, total sleep times were longer during time off steroids compared to the steroid week (mean difference 0.25, 95 % CI 0.01 to 0.50, p = 0.047). Sleep onset latency, night awakenings, sleep quality, and napping were consistent across the month for children taking prednisone.

HRQL and fatigue by steroid group

HRQL was similar between steroid groups at day 5 [F(2, 59) = 2.46, p = 0.094] and day 28 [F(2, 57) = 1.57, p = 0.198] when controlling for risk group (Table 1). Fatigue was significantly lower indicating worse fatigue in children taking dexamethasone at day 5 [F(2, 59) = 5.15, p = 0.009] and day 28 [F(2, 57) = 4.81, p = 0.012] when controlling for risk group (Table 1).

Predictors of HRQL and fatigue

Step-wise regression models were run entering demographic variables in step 1, medical variables in step 2, and sleep variables in step 3 to predict HRQL and fatigue in the weeks on and off steroids. Model 1 predicting HRQL during the steroid week was not significant [F(10, 56) = 1.98, p = 0.057, R² = 0.30] (Table 1). Model 2 predicting HRQL during the last week of the month (off steroids) was significant [F(10, 50) = 3.05, p = 0.006, R² = 0.43]. Models 3 and 4 predicting fatigue were also significant during the steroid week [F(10, 56) = 3.67, p = 0.001, R² = 0.44] and the last week of the month (off steroids) [F(10, 50) = 4.83, p < .001, R² = 0.55]. With the exception of HRQL during steroids (model 1), the addition of the sleep variables was responsible for a significant change in the percentage of the variance explained (change of R² in step 3 is 0.32, 0.27, and 0.34 for model 2, 3, and 4, respectively; Table 3).

Table 3.

Step-wise regression models predicting health-related quality of life (HRQL) and fatigue during time on and off steroids

	(1) HRQL on steroids		(2) HRQL off steroids		(3) Fatigue on steroids		(4) Fatigue off steroids
	ΔR2	β	ΔR2	β	ΔR2	β	ΔR2	β
Step 1	0.01		0.05		0.01		0.06
Age		0.10		0.20		0.06		0.21
Gender		0.01		−0.15		−0.10		−0.17
Risk group		−0.23		−0.14		−0.15		0.01
Steroid		0.39*		0.24		0.47*		0.31
Month in maintenance		−0.21		0.19		−0.16		0.25
Step 2	0.12		0.06		0.16		0.15
Age		0.06		0.17		0.01		0.17
Gender		0.12		−0.18		0.03		−0.19
Risk group		−0.23		−0.14		−0.15		0.01
Steroid		0.39*		0.24		0.47*		0.31
Month in maintenance		−0.21		0.19		−0.16		0.25
Step 3	0.18		0.32		0.27		0.34
Age		0.10		−0.04		0.04		−0.05
Gender		0.01		−0.17		−0.03		−0.16
Risk group		−0.16		−0.19		−0.10		0.04
Steroid		0.19		0.18		0.24		0.15
Month in maintenance		−0.16		0.16		−0.09		0.18
Sleep quality		0.34*		−0.01		0.39**		0.21
Night awakenings		−0.07		−0.42*		−0.14		−0.33*
Total sleep time		−0.02		0.08		0.01		0.08
Sleep onset latency		−0.04		−0.40**		0.21		−0.31*
Nap frequency		−0.16		−0.16		−0.19		−0.18
Total R2	0.30		0.43		0.44		0.55

^*p < 0.05;

^**p < 0.01

Discussion

Sleep is an important component of HRQL that can be significantly disrupted during corticosteroid treatments. The current study indicates that sleep is differentially affected by prednisone and dexamethasone steroid courses for ALL maintenance. Caregivers of children taking dexamethasone reported that their children had more sleep disturbances at baseline, poorer sleep quality across the month, and more nighttime awakenings during steroid courses compared to children taking prednisone. Children taking dexamethasone also experienced more difficulty with fatigue on and off steroids relative to the prednisone group. Results offer a prospective comparison of the impact of steroids on sleep, HRQL, and fatigue over 1 month of maintenance treatment.

With the exception of HRQL during steroids, regression models predicting HRQL off steroids and fatigue at both time points indicated a unique, significant contribution of sleep controlling for demographic (age, gender) and medical (risk group, steroid, month in maintenance) variables. In these models, sleep accounted for a larger portion of the variance during the week off steroids compared to the week on steroids. That almost half the variance in HRQL is accounted for by sleep when children are not taking steroids supports the importance of screening for sleep problems in children with cancer, even when children are not undergoing treatment thought to impact sleep. Because of the importance of sleep for child neurocognitive functioning [30, 31], psychological well-being [32, 33], and health [14–16], attending to such sleep disruptions is essential to improving outcomes.

Results in the dexamethasone group indicating more napping and poorer sleep quality during steroid treatment are consistent with prior research [20, 21]. Both Hinds and Rosen described an extension in nighttime and daytime sleep during dexamethasone courses. The dexamethasone group of the current sample did not exhibit a significant difference in nighttime sleep duration, which may reflect differences in study methodologies. Both studies used actigraphy which may better capture small changes in sleep times. However, differences in collecting sleep data for time off steroids (i.e., using only the 5 days before a steroid course [20] or an average of non-steroid days [21]) may have also affected the results by not reporting the full variation in sleep times across the month as we present here.

Mechanisms underlying the differential effect on sleep between prednisone and dexamethasone have not been described. The longer half-life of dexamethasone relative to prednisone (200 versus 60 min) [34, 35] may be the primary factor contributing to more sleep disturbances in patients taking dexamethasone compared to prednisone. It has been suggested that steroid-induced insomnia can be best managed by once-per-day dosing of corticosteroids in the morning [36]. Previous research in non-cancer populations indicates that dexamethasone reduces slow wave sleep and rapid eye movement sleep in adults [37, 38] as well as latency to rapid eye movement sleep [39]. Such changes to sleep architecture may be responsible for the subjective experience of poor sleep quality and reports of increased awakenings. Older patient age and higher risk group have also been shown to contribute to slower dexamethasone clearance, which may contribute to some of the variabilities observed within steroid groups in the current study [40].

Results suggesting more psychosocial morbidity with dexamethasone treatment must be taken in the broader context of the reduction in relapses with dexamethasone compared to prednisone [4]. Longer-term follow-up studies have not consistently demonstrated differences in psychosocial or neurocognitive outcomes between steroid groups [8]. The impact on HRQL, sleep, and fatigue during dexamethasone treatment warrants further clinical and research attention. Because dexamethasone treatment is a necessity for many patients, the results speak to the need for increased screening and referrals for treatment of sleep disturbances in children taking dexamethasone to optimize HRQL during ALL maintenance.

This study is limited in the use of parent report measures. Prior research in children with ALL has differences in parent report of child awakenings during the night and actigraphy measurement of child sleep during steroids [20]. The current study sought to understand parental reports as these are the most salient for clinical interventions based on parent management of child sleep. Parent report of sleep, fatigue, and quality of life is limited in shared method variance. Furthermore, parents may not have been the most accurate reporters of internal experiences such as sleep quality and fatigue. The use of a 28-day sleep diary is also a limitation that may have resulted in some of the missing data observed. Children were treated on several ALL protocols, including a study that delivered vincristine and dexamethasone once every 3 months. For these children (n = 3), the reduced amount of chemotherapy may have resulted in less disrupted sleep and HRQL. It was not possible to disentangle the effects of vincristine and steroids in the current sample. All children received vincristine prior to starting steroids, which limits the generalizability of the current results to other non-cancer populations.

Although some sleep disruptions may be due to steroids, this does not preclude behavioral treatments. Elevated CSHQ scores indicated high bedtime resistance and anxiety around sleep, areas that can be addressed by behavioral treatments. Furthermore, sleep disruptions have been related to lower parental sleep quality and higher levels of maternal stress [13], and providing parents with sleep training has resulted in higher levels of parental competence [41]. Thus, treating sleep problems may also improve parental functioning. The average sleep time reported in the sample was 10 h 23 min (SD = 1:04; range of 5–14 h; 3–9 year olds; M = 10:04, SD = 1:01; 10–12 year olds; M = 9:18, SD = 1:12), indicating that some children are receiving less than optimal sleep (10–13 h for preschoolers and 9–11 h for school-aged children [42]). The wide range of sleep times reported suggests that it is important to assess and educate families about developmentally appropriate sleep needs. Children with ALL may need more sleep than recommendations for healthy children because of increased physical and psychosocial demands of cancer treatment. Sleep is extremely sensitive to stressors of chronic illnesses [43] and corticosteroids [37]. Changes in sleep that occur during steroid bursts can be maintained by the recurrent and stressful nature of steroids, thereby perpetuating poor sleep, impairing HRQL, and slowing recovery.

Results indicate the importance of assessing and treating sleep disturbances in children taking corticosteroids. Brief screening tools (e.g., BEARS [44]) can be employed during medical visits to assess steroid-related sleep disturbances. Furthermore, providing families with education regarding the potential impact of steroids on sleep, fatigue, and HRQL can empower parents to proactively manage their child’s sleep during steroids through good sleep hygiene, minimizing disruptions to the child’s sleep behaviors (e.g., changing sleeping locations and sleep times), and potentially considering supplements to promote regular sleep during steroid courses (i.e., melatonin, although melatonin has not been tested in pediatric oncology to date).

Abbreviations

ALL

acute lymphoblastic leukemia

CSHQ

Children’s Sleep Habits Questionnaire

HRQL

health-related quality of life

MFS

Multidimensional Fatigue Scale

PedsQL

Pediatric Quality of Life Inventory