Experimental sleep restriction increases somatic complaints in healthy adolescents

Abstract

Short duration sleep and somatic complaints (bodily complaints like aches/pains or GI distress) are common in adolescence and may be linked, yet no published studies have tested causation. In this study, healthy adolescents (n = 30; 14–18yrs) completed a three-week, within-subject cross-over experiment. Following a sleep stabilization week, adolescents were randomized (in counterbalanced order) to five nights of 6.5hrs in bed (Insufficient Sleep) or 9.5hrs in bed (Sufficient Sleep), each preceded by a two-night “washout.” Somatic complaints were assessed via the Children’s Somatic Symptoms Inventory (CSSI-24) and the Pain and Symptom Assessment Tool (PSAT) concluding each condition. Adherence to experimental condition was confirmed via actigraphy. Adolescents slept two fewer hours and reported significantly greater frequency and severity of somatic complaints during the Insufficient Sleep (vs. Sufficient Sleep) condition. Restricting sleep opportunity to a level common in adolescence causally increased somatic complaints in otherwise healthy adolescents. Findings support clinical and preventative efforts to address pervasively inadequate sleep in adolescence.

1. Introduction

Sleep deprivation is pervasive during adolescence, with most American adolescents sleeping <7 h on school nights [1], well below consensus recommendations of 8–10 h/night [2]. Concurrently, the transition from childhood to adolescence is also marked by an increase in somatic complaints (eg unexplained headaches, nausea, aches/pains). Similar to the effects of sleep deprivation [3], these bothersome and ambiguous body-based perceptual experiences can have profound negative impacts on the daily functioning [4] and quality of life of adolescents [5]. There is reason to suspect that sleep contributes to the somatic experience. Adult studies find that chronic [6] and experimental [7] sleep restriction alters sensory and pain thresholds. Pediatric studies also find connections between sleep and key somatic complaints, such as pain [8]. However, these pediatric studies have been limited in study design (correlational) and scope (restricted to one specific somatic complaints or patient population). In adolescents, it remains unknown whether short sleep causes somatic complaints, or if the two simply co-occur within this developmental period.

We aimed to address this question using rigorous experimental methodology. To avoid ethical risks of worsening symptoms within an already-vulnerable population (children with medical conditions), we conducted this as a small proof-of-concept study using a healthy sample with no risk factors for somatic complaints. We aimed to test whether experimentally-induced Insufficient Sleep (6.5 h/night in bed) vs. Sufficient Sleep opportunity (9.5 h/night in bed) alters the self-report of a broad range of somatic complaints in otherwise healthy adolescents. If we find changes in this healthy population, it would suggest this is an important area for future work with more vulnerable populations. We hypothesized that adolescents would report more frequent somatic complaints during Insufficient Sleep than during Sufficient Sleep. Secondarily, we also explored whether this effect was moderated by age, sex, race, family income, or randomization order.

2. Material and methods

2.1. Data source and sample

Study data were part of a larger, IRB-approved experimental study on sleep restriction and health behaviors, using a previously-reported protocol [9] in a new sample. Healthy 14–18 year-olds were recruited through emails and flyers posted in a large Mid-western children’s hospital. Exclusion criteria included conditions that could render sleep restriction unsafe (neurological, mood or psychotic disorder), use of medications affecting sleep or alertness, body mass index>30, or habitually high caffeine consumption (>1 coffee/energy drink or >2 caffeinated sodas daily).

2.2. Procedure

Adolescents participated in the three-week within-subjects cross-over experiment during the summer of 2017 (see Fig. 1). Wake-time was held constant throughout (set at a time to accommodate 8am office visits). Bedtimes were adjusted to accomplish changes in sleep duration. During an initial five-night sleep stabilization period, adolescents self-selected bedtimes and adjusted to the new constant wake-times. This period was not conceptualized as a baseline of typical sleep patterns, but instead was intended to allow adolescents on a summer sleep schedule to adjust, prior to experimental randomization, to a more school-like schedule in which sleep was bounded primarily by early wake times. Adolescents were then block-randomized to one of two sequences of experimental sleep conditions. Participants in one sequence experienced a period of Insufficient Sleep (6.5 h/night sleep opportunity) followed by Sufficient Sleep (9.5 h/night sleep opportunity), while those in the other sequence experienced the same conditions in the opposite order. Following the logic detailed in Beebe et al. [10], sleep opportunity in the Insufficient Sleep condition was set to fall below consensus recommendations for healthy adolescent sleep duration [2]; sleep opportunity in the Sufficient Sleep condition was set to increase the likelihood consensus recommendations of 8–10 h of sleep. Each experimental sleep condition was preceded by a two-night “washout” (>8 h/night sleep opportunity with similar wake times), and all adolescents who completed the full study had both Insufficient and Sufficient Sleep conditions (only the order randomly varied). Somatic complaints were assessed during office visits the morning concluding the Insufficient and Sufficient Sleep conditions. Adolescents were instructed against naps or increased caffeine consumption throughout the study.

Fig. 1.

Sleep Manipulation Design. Note: Wake times were held constant throughout the entire study procedures, set at a time that would allow adolescents to attend 8am Saturday study visits.

2.3. Measures

Adolescent age, sex, race, and family income were reported by the adolescent’s caregiver at the first study visit.

Sleep was measured by the wrist-worn Micro Motionlogger actiwatch (Ambulatory Monitoring, Inc.). Adolescents also completed sleep diaries that included in and out of bed to help screen for actigraphy artifacts (eg, failure to wear during part of the night). Trained coders reviewed actigraph data (supported by sleep diaries) with families during office visits to screen for artifacts, set rest intervals, and encourage accountability and adherence to the assigned sleep conditions. Actigraphy-estimated sleep period (onset to offset) was objectively derived using the Sadeh algorithm [11], which has been validated for child/teen sleep.

Our primary outcome, frequency and magnitude of being bothered by somatic complaints, was assessed using the Children’s Somatic Symptoms Inventory (CSSI-24) [12]. The CSSI-24 measures the severity (or bothersome-ness) of broad range of non-specific somatic complaints such as headaches, chest pain, weakness, and nausea over the past two weeks. Participants answer questions on a five-point Likert Scale (0 = not at all, 4 = a whole lot). It has been validated within a pediatric population with good psychometric properties.

Secondarily, we used the Pain and Symptom Assessment Tool (PSAT) [13] as a complementary measure of spatial pain distribution and the presence of behavioral and somatic symptoms. The self-reported PSAT assesses the presence of 18 specific pain locations, 24 cross-cutting somatic complaints, and burden of several sleep and fatigue-specific symptoms (eg feeling fatigued, struggling to return to wakefulness, and concentration or memory problems). Scores across each sub-scale are added together into a continuous score reflecting a continuum (or severity) of greater symptom burden related to widespread pain and somatic symptomology.

2.4. Analytic approach

All participants who completed the study were included in intent-to-treat analyses. Difference scores between Insufficient Sleep and Sufficient Sleep were computed for all variables (with positive difference scores reflecting higher values in Sufficient Sleep). Wilcoxon Signed Rank Test (to account for skewed distributions) was used to test if difference scores on somatic complaints significantly differed from 0. Paired-samples t-tests was used to test differences between normally-distributed sleep periods. Exploratory analyses tested whether main effects differed by sex, race, or condition order via Wilcoxon tests, family income via Kruskale—Wallis test, and age via Spearman’s Rho.

3. Results

Of the 42 adolescents who were recruited into the study, five did not meet eligibility, five dropped out before randomization (three due to poor adherence to the sleep instructions during the stabilization week, two due to personal or unknown reasons), and two dropped out after randomization (one due to not wearing the actigraph and one for personal or unknown reasons). In the final sample (n = 30), mean age was 16.2 (±1.1), 74% were female, 58% were non-Hispanic white, 32% were non-Hispanic black, and 10% were multiracial or of another race. See Supplemental Table for more detailed demographic information. The average sleep period in the stabilization week was 6.86 h (±1.07). As shown in Table 1, adolescents averaged >90 min longer nightly sleep in Sufficient vs. Insufficient Sleep. The vast majority (22 of 30) averaged a nightly difference of >120 min between Sufficient and Insufficient Sleep conditions.

Table 1.

Changes in sleep period and somatic complaints across experimental conditions.

	Sufficient Sleep Mean ± SD	Insufficient Sleep Mean ± SD	Difference Score Mean ± SD	p-value	Effect size
Actigraphy
Sleep Period (hours)	7.99 ± 1.03	6.49 ± 0.71	92.93 ± 1.30	<0.001	1.19
Somatic Complaints - CSSI
CSSI with all items	3.97 ± 4.85	6.48 ± 5.36	2.6 ± 4.90	0.0042	0.53
CSSI without low energy item	3.33 ± 4.33	5.52 ± 4.85	2.23 ± 4.09	0.004	0.55
Somatic Complaints - PSAT
PSAT with all items	2.52 ± 3.50	4.06 ± 3.64	1.55 ± 2.59	0.0009	0.60
PSAT without sleep/fatigue items	1.17 ± 3.01	2.03 ± 3.17	0.93 ± 2.03	0.0226	0.46

Note: Sufficient Sleep = 9.5 h/night sleep opportunity; Insufficient Sleep = 6.5 h/night sleep opportunity. Effect size for Sleep Period calculated from t-test results; effect sizes for difference scores of CSSI, CSSI w/out one item explicitly about low energy, Full PSAT, and PSAT w/out sleep/fatigue items calculated from Wilcoxon Signed Rank test.

While somatic complaints were generally low across conditions, adolescents reported significantly greater frequency and/or severity of somatic complaints on our primary outcome, the CSSI-24 (p = 0.004, d = 0.53) in Insufficient Sleep (vs. Sufficient Sleep). They also endorsed >60% more somatic symptomology on the PSAT (p = 0.0009, d = 0.60). Effects maintained for both the CSSI-24 (p = 0.004, d = 0.55) and PSAT (p = 0.023, d = 0.46) after removing items manifestly related to sleep or low energy (ie feeling fatigued, waking up tired; p = 0.02, d = 0.46). Exploratory analyses found that these cross-condition effects did not differ across sex, age, race, or the order of experimental sleep conditions. However, the effect of Insufficient Sleep on the PSAT was significantly larger for those coming from families with family income of >$100,000 than $50–100,000 (p = 0.006).

4. Discussion

Complementing prior adult and pediatric findings, this small experimental study confirms that, compared to when they were allowed Sufficient Sleep opportunity, adolescents experiencing Insufficient Sleep had significantly increased somatic complaints. On an item level, adolescents particularly reported having low back pain and headaches via the Pain and Symptom Assessment Tool more often during Insufficient Sleep compared to Sufficient Sleep (23% vs. 9% and 23% vs. 5%, respectively). Although increases in complaints were generally modest, they are notable, as this was a proof-of-concept study using a small sample. We found that even among a healthy adolescent sample, insufficient sleep that was similar to what many adolescents experience on school nights caused bothersome and ambiguous body sensations.

These findings open the door to replicating and expanding this line of research to other populations (eg adolescents with insomnia and/or greater risk/prevalence of somatic complaints). Future studies that extend or stabilize sleep would be particularly interesting, and could inform sleep prevention/intervention efforts. Sleep may be an underappreciated, modifiable target to non-pharmacologically promote resiliency and reduce somatic complaints in adolescence.

Findings also raise important questions about possible mediating factors linking sleep with somatic complaints. Previous work has already connected sleep with emotional states, stress-related systems, and pain sensitivity, all of which are relevant to the somatic experience. Among adolescents, restricting sleep causally lowers positive affect and heightens anxiety and negative affect [14,15]. Affective states are closely linked to the somatic experience in the adult literature [16]. Experimentally restricting sleep opportunity in adults also increases spontaneous pain complaints and blunts habituation to specific types of pain [6]. Finally, altered immune functioning has also been posed as a common underlying biology of physical complaints, pain sensitivity, depressive disorders, and sleep disturbance [17].

Results suggest a need for research that allows for temporal investigation of these potential mechanisms linking sleep with adolescents’ somatic experience. Future research is also needed to address limitations of this study, including a modestly-sized and relatively homogeneous sample, relatively short experimental sleep periods (five nights each), and exclusive reliance on self-report measures (though somatic complaints are by nature subjective). In the meantime, current findings dovetail with prior work suggesting that providers who treat adolescents with somatic complaints and Insufficient Sleep can reasonably link the two phenomena – and consider sleep as a possible intervention target in a clinical setting. On the prevention science side, these findings provide further support for the need to alleviate pervasively inadequate sleep during adolescence to improve population health and well-being.

Supplementary Material

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.sleep.2020.07.020.

Abbreviations:

IRB

Institutional Review Board

CSSI-24

Children’s Somatic Symptoms Inventory

PSAT

Pain and Symptom Assessment Tool