section epub:type=”chapter”> Maida Chen, Margaret Wardlaw and Mark A. Stein, Seattle Children’s and University of Washington, Seattle, Washington, United States Difficulties with sleep initiation and maintenance frequently occur in children with attention-deficit hyperactivity disorder (ADHD). All stimulants can produce insomnia, with little empirical data to suggest that there are substantial differences in sleep onset latency for the different stimulant formulations. While most children fall asleep with 15–20 minutes, children taking stimulant medications often take longer to fall asleep, especially during initial treatment and with dose increases. Insomnia related to stimulant medication appears to be dose-dependant, with 20%–30% of children treated in controlled trials taking more than 30 minutes to fall asleep when using low to moderate stimulant dosages. When initiating pharmacotherapy for ADHD, sleep patterns should be closely monitored. Sleep hygiene and behavioral procedures to reduce bedtime problems should be emphasized at all phases of ADHD treatment. If insomnia persists after initiating an effective ADHD treatment, alternative dosages, formulations, timing of administration or medications should be considered to produce optimal benefit during the day without compromising sleep. Presumably, reducing the variability in sleep–wake cycles due to ADHD pharmacotherapy will promote attention and alertness during the day. Stimulants; methylphenidate; nonstimulants; amphetamine; insomnia; melatonin According to Cortese, “the awareness of a link between sleep disturbances and attention deficit hyperactivity disorder (ADHD) is not novel at all” (Cortese, 2015). It is also not a recent development. Indeed, sleep problems are strongly linked with both ADHD, and its treatment (Stein, 1999). Insomnia or delayed sleep onset was first reported by Bradley on the “calming” behavioral effects of Benzedrine. In his initial report, 6 of 30 children displayed a delay in falling asleep for “the first night or two,” while one child remained awake “to a late hour for four nights” (Bradley, 1937). Despite the subsequent development of numerous stimulant and nonstimulant medications used to treat ADHD, transient sleep problems, as well as variability in their severity and duration of sleep are common side effects in children receiving pharmacotherapy (Efron, Jarman, & Barker, 1997). Indeed, since Bradley, much has been written on the association between stimulant medications and longer sleep onset latency (SOL), worse sleep efficiency (SE), shorter sleep duration, and the broader umbrella term of “insomnia” (Cortese, 2015; Stein, Weiss, & Hlavaty, 2012). In contrast to stimulants, which increase alertness acutely when administered in the morning, sleepiness is a common side effect of the alpha agonists and atomoxetine (ATX), especially when administered in the morning (Block et al., 2009) or as monotherapy (Spencer, Greenbaum, Ginsberg, & Murphy, 2009). Additionally, night-to-night variability and sleep phase disorder are also common, especially in adolescents, who vary both their medication and sleep schedule, based on either preference for school day treatment, differences in weekday–weekend schedule, or nonadherence. As there are currently numerous approved immediate and delayed release stimulant and nonstimulant formulations for ADHD, as well as several more in development, information is needed to help providers and patients select medications to manage their ADHD symptoms while minimizing adverse effects on sleep. Although all approved medications reduce ADHD symptoms, there is wide individual variability in the duration, tolerability, and withdrawal effects. In this chapter, we will review studies evaluating the effects of medication on sleep, as well as clinical trials that report on sleep as an adverse event associated with medication. We will also discuss strategies to address medication-related sleep problems and areas for future study. In evaluating the methodologically heterogeneous literature on ADHD treatment and sleep, it is important to first consider how sleep problems are defined and measured. Objective measures, such as actigraphy and polysomnography (PSG) are often used in studies directly assessing effects of medication on sleep. Although both subjective and objective sleep measures provide valuable information ranging from subjective parental impressions to clearly defined sleep stages and night-to-night variability, respectively, data are not highly correlated (Cortese, Faraone, Konofal, & Lecendreux, 2009; Hvolby, 2015). Although sleep problems are often treated as a discrete or categorical variables based on a single measure, sleep problems vary in frequency and severity, and are highly influenced by the rater, method, and timing of measurement during the course of a clinical trial. Common sleep variables reported include SOL, total sleep time (TST), SE, number of awakenings (AW), and severity of daytime sleepiness. In examining the relationship between ADHD medications and sleep problems, detailed information is needed on the formulation, dose, dosing schedule, titration procedure and if the medication was effective in reducing ADHD symptoms at time of sleep measurement, in order to determine the external validity or how generalizable findings are to clinical practice. For example, a study that examined sleep problems at low doses may result in fewer side effects but also lower efficacy of ADHD treatment (Stein et al., 1996). While fixed dose or dose–response studies provide more distinct information on the medication and dose effects (Stein et al., 2011), flexible dosing is more similar to real-world care, and mimics more closely how the medication would perform in clinical settings. When evaluating sleep problems as a side effect of stimulant medications, an assessment prior to initiating treatment is crucial (Rapport & Moffitt, 2002), since there is high baseline prevalence of sleep difficulties in ADHD children and youth regardless of medication status (Owens et al., 2013). Furthermore, children with preexisting sleep problems are at increased risk for more severe sleep problems during stimulant therapy (Faraone et al., 2009). In longer-term follow-up studies, poor responders often dropout early, resulting in a biased study sample of predominantly positive responders who tolerate the medication (Wilens et al., 2003). The longer the trial, the more likely that only responders are continuing treatment (Wilens et al., 2005). This means that detrimental effects of ADHD medication on sleep contributing to poor tolerability may be missed. On the other hand, shorter-term studies are primarily measuring acute medication effects, and do not capture the temporary nature of common but mild adverse effects on sleep that often gradually improve. Another major confounder in evaluating the potential for adverse effects on sleep is the potential for selection biases, and thus the previous treatment history should be well-described. Clinical trials utilizing enrichment designs where subjects were selected for previous exposures or following a “run in” phase may report fewer sleep difficulties as pretreated cohorts are likely to include fewer nonresponders or individuals with intolerable side effects compared to treatment naïve participants (Poulton & Nanan, 2008). In contrast, stimulant naïve patients appear more likely to display sleep problems and other adverse events (Wigal et al., 2012). In clinical trials, the most common method for monitoring adverse events is through spontaneous reporting to the investigator. Severe or worsening episodes of insomnia are reported as frequency counts or rates of occurrence. This naturalistic approach is limited by the lack of detail on the problem, subjectivity of what constitutes a sleep problem, as well as poor recall. Parental report of perceived sleep problems are highly influenced by parents’ own levels of stress, anxiety, and sleep deprivation. Indeed, parents of children with ADHD display increased rates of depression, anxiety, and substance disorder even when compared to children with other developmental disorders (Roizen et al., 1996). Another index of poor tolerability in clinical trials is premature withdrawal from a study due to severe sleep problems, which are usually listed in the CONSORT diagram (Begg et al., 1996). A more systematic approach compared to spontaneous report of adverse events is the utilization of a structured rating scales (Adler, Goodman, Weisler, Hamdani, & Roth, 2009), or the use of sleep diaries or sleep questionnaires (Owens, Spirito, & McGuinn, 2000; Stein, Mendelsohn, Obermeyer, Amromin, & Benca, 2001). These scales are also subject to rater bias, however, and baseline sleep status should also be considered. Table 6.1 includes brief descriptions of studies that utilize sleep diaries, structured rating scales, actigraphy and PSG to measure how ADHD medications effect sleep. Table 6.1 As can be seen, there is considerable heterogeneity in study design, treatment, and measures. Studies also varied considerably in sample size, ranging from a case series of seven children in which full PSG was performed before and after initiation of stimulant medication (Vigliano, Galloni, Bagnasco et al., 2016) to a multicenter clinical trial involving several hundred children (Faraone et al., 2009). Methylphenidate (MPH), in various formulations, was the most commonly studied medication, including both extended and immediate release preparations. Consistent with prior reviews (Cortese, 2015), the most common finding was delayed sleep onset and reduction in TST in stimulant treated youth. Generally, dose–response studies indicate a linear or dose-dependent relationship between dose and insomnia or delayed sleep onset for MPH and amphetamine preparations (Santisteban, Stein, Bergmame, & Gruber, 2014; Stein et al., 2003). The majority of studies suggest that with appropriate titration and over time, effects on sleep were mild and generally improved (Faraone et al., 2009), although selection and ascertainment biases should always be considered. Surprisingly, there are far fewer studies of amphetamine although amphetamine has now surpassed MPH, at least in the United States, as the most commonly prescribed stimulant for adolescents and adults (Safer, 2018). Unfortunately, there are also very few studies of clinically relevant samples utilizing an appropriate range of doses, reliable measures of sleep, with active comparators of medications frequently utilized. Stein et al. (2011) compared three dose levels of extended-release, mixed amphetamine salts (ER-MAS) with extended-release dexmethylphenidate (ER-dMPH) and a placebo, in 56 children (30% stimulant naive) in a placebo-controlled, crossover 8-week trial. Parent ratings of severe insomnia were significantly higher for ER-MAS at the 10 mg dose level, but at higher dose levels, there was no drug-related difference in percentage with severe insomnia. Although limited in statistical power, when actigraphy data were examined, both MAS and MPH were both associated with a dose-dependent effect on sleep onset and duration (Santisteban et al., 2014). In this study, the dose was confounded with time, which partially explains the lower rates of insomnia at the highest dose level, as participants receiving the highest dose level had more time to accommodate to the stimulant. Table 6.2 includes representative medication trials in which impacts on sleep were measured as a side effect. Table 6.2
ADHD Medications and Sleep
Abstract
Keywords
6.1 Introduction
6.1.1 Sleep and ADHD Treatment: Methodological Considerations
6.1.2 Studies Measuring Sleep as a Primary Outcome
Authors (year)
Study drug
Design (as pertains to sleep metrics)
Sleep-related metrics
Findings
Comments
Haig & Schoeder. (1974)
MPH: 15 mg b.i.d. or 40 q/day
PSG on/off MPH
Feinberg, Hibi, Braun, Cavness, Westerman & Small. (1974)
PSG pre/on/off medication
Ashkenasi (2011)
Transdermal MPH: 10–30 mg, 9–12 hours/day
Sleep diary, sleep quality scale between baseline/4 different patch wear sequences
Faraone et al. (2009)
OROS-MPH: 18–54 mg MTS: 10–30 mg
CSHQ at baseline vs OROS/MTS
High CSHQ scores pre-MPH predicted High CSHQ scores on MPH (either OROS or MTS)
Galland Tripp & Taylor (2010)
MPH: dose not provided, various formulations
PSG on/off MPH vs controls.
PSG findings off MPH were similar to non-ADHD controls
Giblin & Strobel (2011)
LDX: 30–70 mg
PSG, Actigraphy, CSHQ pre/on LDX
Actigraphy and PSG with similar findings
Greenhill, Puig-Antich, Goetz, Hanlon & Davies (1983)
MPH: 0.99–1.44 mg/kg/day
PSG pre/on MPH
Hollway et al. (2018)
ATX: 0.3–1.8 mg/kg/day
CSHQ, TST, WASO pre/post
Kent, Blader, Koplewicz, Abikoff & Foley (1995)
MPH: 10 or 15 mg
Ad hoc sleep adequacy measure, sleep latency by inpatient staff
SOL: no change across 3 groups Sleep adequacy: improved at 10 mg, worsened in placebo and 15 mg
Lee, Seo, Sung, Choi, Kim & Lee (2012)
Sleep diary pre/on
Morash-Conway, Gendron & Corkum (2017)
MPH-LA: dose not provided
PSG on/off MPH
Owens, Weiss, Nordbrock et al. (2016)
MPH-MLR: 15–60 mg
CSHQ or ASHQ, Child-Self Sleep Report
O’Brien et al. (2003)
MPH, MPH SR, MAS: doses not provided
Parental sleep survey, PSG
Few differences between either ADHD group were found vs controls on PSG
Rugino (2018)
GXR: 1–4 mg
PSG
Study terminated early due to significant decreases in TST
6.1.3 Studies Measuring Sleep as a Side Effect
