Hyperactivity and Polysomnographic Findings in Children Evaluated for Sleep-Disordered Breathing

Transcription

1 PEDIATRIC SLEEP Hyperactivity and Polysomnographic Findings in Children Evaluated for Sleep-Disordered Breathing Ronald D. Chervin MD, MS, and Kristen Hedger Archbold RN, PhD Sleep Disorders Center, Department of Neurology, University of Michigan, Ann Arbor, Michigan, U.S.A Study Objectives: Children with sleep-disordered breathing (SDB) or periodic leg movements during sleep (PLMS) often have hyperactive behavior that improves when the sleep disorder is treated. Some children with SDB also have PLMS. To determine what polysomnographic features of SDB might be associated with hyperactive behavior, we studied behavior, SDB, and PLMS in a series of patients. Design: Prospective and observational Setting: University-based sleep disorders laboratory Subjects: Children (n=113) aged 2 to 18 years, referred for suspected SDB Interventions: Parents completed the hyperactivity index of the Connors Parental Rating Scale, and results were converted to age-adjusted t- scores. Children underwent laboratory-based polysomnography, with esophageal pressure monitoring when requested (n=19) by referring physicians. Results: Children with SDB (n=59) showed high hyperactivity scores (mean 59.5±18.3 SD, 95% C.I. [54.7, 64.2]) but these scores were no INTRODUCTION CHILDREN WITH OBSTRUCTIVE SLEEP-DISORDERED BREATHING (SDB), IN THE FORM OF OBSTRUCTIVE SLEEP APNEA OR UPPER AIRWAY RESISTANCE SYN- DROME, are reported to have inattentive, hyperactive, and aggressive behavior. 1-3 Some pediatric patients arrive at sleep centers already diagnosed with attention-deficit/hyperactivity disorder (ADHD) and treated with methylphenidate; once SDB is diagnosed and treated, behavior improves and stimulants sometimes can be discontinued. 2,4,5 The frequency with which ADHD and SDB occur together, in the community, remains unknown, but ADHD affects more than 5% of children, 6 undiagnosed SDB affects at least 1% to 3%, 7,8 and surveys about symptoms of the two conditions suggest that the comorbidity may be relevant to a substantial minority of children with hyperactive behavior. 9 The mechanism by which SDB may contribute to hyperactive behavior is unknown. Children deprived of sufficient amounts of sleep or adequate quality of sleep may fidget, spring out of chairs, shift attention, and create excitement in an effort to remain awake and learn. Sleep disruption is implicated as a cause of hyperactivity in SDB by observations that other sleep disorders, such as periodic limb movement disorder (PLMD) Accepted for publication December 2000 Address correspondence to: Ronald D. Chervin MD, MS, Michael S. Aldrich Sleep Disorders Laboratory, University Hospital 8D8702, Box 0117, 1500 E. Medical Center Dr., Ann Arbor, MI ; Tel: (734) ; Fax: (734) ; chervin@umich.edu SLEEP, Vol. 24, No. 3, higher than those of children without SDB (59.0±15.1, [54.8, 63.1]). Hyperactivity showed no significant associations with the rate of apneas and hypopneas, minimum oxygen saturation, or most negative esophageal pressure (p>0.10), but was associated with the presence of 5 or more PLMS per hour (p=0.02). The rate of PLMS showed a linear association with hyperactivity among those subjects with SDB (p = 0.002), but no association among those subjects without SDB (p = 0.64). Conclusions: These findings suggest that hyperactive behavior is common among children referred for suspected SDB, regardless of the presence or severity of SDB. Current observations cannot prove causality, but they are consistent with the hypothesis that PLMS may contribute to hyperactivity and SDB may act as an effect modifier. Key words: Sleep-disordered breathing; obstructive sleep apnea; upper airway resistance syndrome; polysomnography; hyperactivity; attentiondeficit/hyperactivity disorder; ADHD; periodic leg movements; periodic limb movement disorder; children and narcolepsy, 13,14 also are associated with hyperactive behavior. As in SDB, treatment of PLMD or narcolepsy often leads to behavioral improvement. 15,16 In contrast to children, adults with SDB most often complain of excessive daytime sleepiness. 17 Once again, however, the mechanism by which SDB causes daytime symptoms is unclear: both hypoxemia and the frequency of apneic events presumably a reflection of the severity of sleep disruption are independently associated with objective measures of sleepiness, but the magnitudes of these associations are not large. 18 Analogous studies in children, to compare behavioral outcomes with polysomnographic measures of SDB, have not been performed. Finally, the argument that SDB actually causes hyperactive behavior in children, by any mechanism, remains far from proven. Few studies have used validated behavioral measures to document largely anecdotal observations. Clinical series have not included behavioral data on children referred to sleep centers but not found to have SDB. Hyperactive children may be referred for sleep studies more frequently than others because they frequently have disturbed sleep No double-blind, single-blind, placebo controlled, or randomized treatment trials have documented improvement in hyperactivity as a result of treatment for SDB. If SDB does cause hyperactivity in children, an association between levels of hyperactivity and polysomnographic measures of SDB would be expected. We used the well-validated hyperactivity index of the Connors Parental Rating Scale 23,24 to assess hyperactivity in a series of children evaluated for suspected SDB, and then examined what polysomnographic measures of SDB were most closely and independently associated with hyperactiv-

2 ity. As PLMD is common in SDB, at least among adults, we also tested for associations between the frequency of periodic leg movements during sleep and levels of hyperactivity. METHODS Subjects Children were included in this prospective, observational, Institutional Review Board-approved study if the parents or guardians signed an informed consent, subjects 10 years of age or older signed an assent, and subjects met the following criteria: 1) referral to our sleep laboratory for a full night of nocturnal polysomnography between August 23, 1996 and March 29, 2000 to assess for suspected SDB; 2) age 2.0 to 18.0 years; 3) hyperactivity scale completed by a parent or guardian at the time of polysomnography; 4) no tracheostomy normally open during sleep; and 5) no concurrent cognitive, physical, or medical disability, as reported by the parent on a laboratory questionnaire, that would be likely to preclude interpretation of the hyperactivity score. The questionnaire included a medical history; list of current medications; specific inquiry about any physician-suspected or diagnosed psychological, psychiatric, emotional, and behavioral problems; and an immediate family history for behavioral or psychiatric problems. To avoid creation of a sample with little resemblance to that seen in clinical practice, subjects were still included if they were diagnosed with sleep disorders other than SDB, or with behavioral problems judged unlikely to prevent behaviors addressed by the hyperactivity measure described below. A total of 113 patients met criteria and were included in this study. Measures Hyperactive behavior was assessed with the hyperactivity index of the commonly-used Connors Parental Rating Scale. 23,24 Raw scores were converted to age-adjusted t-scores, which range from 0 to 100 and average 50 in population samples. Conversion was based on standard age-appropriate data from children aged 3 through 17 years; data for the three to five-year-old group were used, as an approximation, for study subjects who were 2.0 to 2.9 years old. The Connors Scale is not used to establish a diagnosis of attention-deficit/hyperactivity disorder, but the hyperactivity index t-scores provide a widely recognized measure of behavior in which either one s.d. above the population mean (score=60) or two SD s (score=70) is used to define an abnormally hyperactive group. Nocturnal polysomnography included four electroencephalographic leads (C3-A2, C4-A1, O1-A2, O2-A1 of the international electrode placement system), 2 electro-oculographic leads (right and left outer canthi), chin and bilateral anterior tibialis surface electromyograms, 2 electrocardiographic leads, nasal and oral airflow (thermocouples), thoracic and abdominal excursion (piezoelectric strain gauges), and finger oximetry. In a minority of cases (n=19), and at the referring physician s request, esophageal pressure was monitored with a water-filled catheter 25 that does not have significant adverse effects on children s sleep. 26 Sleep stages were scored in 30-second epochs according to standard criteria 27 by technologists who, after an extensive training program, had correctly scored at least 90% of epochs in SLEEP, Vol. 24, No. 3, a set of reliability records. An apnea was defined as 10 or more seconds of complete airflow cessation during sleep, regardless of any change in oxygen saturation. An hypopnea was defined as a 10-second reduction in airflow, chest excursion, or abdominal excursion that led to either a 4% or greater oxyhemoglobin desaturation, an arousal, or an awakening. Some pediatric sleep laboratories now score apneas or hypopneas that are less than 10 seconds long, in part because normal respiratory rates can be considerably higher in young children than in adults. However, this justification would not apply well to the older children and teenagers included in our sample. We suspect that in the current study, esophageal pressure monitoring allowed identification of subtle SRBDs in some of the younger children for whom shorter apneic event criteria might otherwise have been important. For this study, we classified subjects as having SDB if they had a rate of apneas and hypopneas (apnea/hypopnea index, or AHI) that was at least five per hour of sleep, or if their peak negative end-inspiratory esophageal pressure was -20 cm of water or more negative. These criteria were chosen to provide internally uniform limits that would be reasonably applicable to our wide range of subject ages, and with the realization that better cutoffs are difficult to define at present in the absence of the necessary outcome-based, age-stratified studies. 28 Although one frequently cited study of 50 normal children found an apnea index of 1 or more to be unusual, 29 this study did not define what rate of apneas causes adverse outcomes and did not address the role of hypopneas, which at least in adults appear to have outcome-related significance comparable to that of apneas. 18,30 We chose to include hypopneas in our criterion for OSA, and we used an AHI threshold of 5 instead of an apnea index of 1. Although esophageal pressures in excess of -20 cm of water do not define upper airway resistance syndrome, no widely accepted alternative definition is available. One frequently referenced publication described the syndrome as a combination of excessive daytime sleepiness and gradually worsening pressures that lead to frequent arousals, 31 but published data have not 1) defined a minimum required rate of arousals above that seen in normal adult sleep, 32 2) established a criterion pressure that distinguishes abnormal vs. normal pressures, 3) indicated what proportion of an abnormal record must exhibit abnormal pressures, or 4) established what length of recording, prior to an arousal, should exhibit continuously worsening pressures in order to attribute the arousal to increased breathing effort. Furthermore, some publications 33,34 and our own experience have suggested that some patients have excessively negative pressures without frequent arousals but still experience improvement in symptoms after upper airway obstruction is relieved. As a criterion for inclusion in the current SDB group when OSA was absent, we therefore adopted a simple, objective, but somewhat conservative criterion: a minimum esophageal pressure of -20 cm of water. 35 This cut-off level is twice the limit considered to be pathological by some researchers. 31 Arousal frequency was not scored in these studies, and Multiple Sleep Latency Tests (MSLTs) were not generally administered, but surrogate measures for sleep disruption and sleepiness were available, namely sleep efficiency (total sleep time divided by time in bed) and nocturnal sleep latency. At least among adults, the nocturnal sleep latency is the best nocturnal correlate of MSLT results. 36

3 Table 1 Subject characteristics * Variable All Subjects Subjects with SDB Subjects without SDB (n = 113) (n = 59) (n = 54) Age (years) 9.9 ± ± ±4.4 No. Male (%) 73 (65) 39 (66) 34 (63) AHI 8.1 ± ± ±1.3 No. With OSA (%) 49 (43) 49 (83) 0 (0) Minimum O2 (%) 88.9 ± ± ± 3.0 Most Negative Pes (cm of water) ** ± ± ± 1.8 No. With High Upper Airway Resistance (%) *** 14 (74) 14 (93) 0 (0) PLMI 4.9 ± ± ±13.6 No. With PLMD (%) 29 (26) 16 (27) 13 (24) Sleep Efficiency 85.7 ± ± ± 12.4 Sleep Latency (min) 18.2 ± ± ± 38.3 Hyperactivity Index T-score 59.2 ± ± ± 15.1 No. Hyperactive (%) **** 27 (24) 15 (25) 12 (22) No. With History of ADHD (%) 20 (18) 13 (22) 7 (13) No. With Stimulant Use (%) 20 (18) 13 (22) 7 (13) * AHI = apnea/hypopnea index, minimum O 2 = minimum oxygen saturation, Pes = esophageal pressure, PLMI = periodic leg movement index, OSA = obstructive sleep apnea, PLMD = periodic limb movement disorder, ADHD = attention-deficit/hyperactivity disorder ** N = 19 subjects, 15 with SDB and 4 without SDB *** Among the subjects whose esophageal pressure was monitored **** Hyperactivity index t-score >70. p<0.05, for difference between subjects with and without SDB, by t-test for continuous data or chi-square test for dichotomous data Periodic leg movements during sleep (PLMS) were scored when they met criteria for duration (0.5 to 5 seconds), periodicity (5 to 120 seconds between each movement), and number (at least 3 in a row). For this study, PLMD was considered present when the rate of leg movements per hour of sleep (periodic limb movement index, PLMI) was at least 5; 10,37 previous work has shown that this level of periodic leg movements is rare in normal children. 12 Analysis Results were summarized as means ± SD or by frequencies. Simple and multiple linear regression models were used to test for associations between hyperactivity scores as outcomes, explanatory variables derived from polysomnography, and age and sex. Associations were tested first in the entire group of subjects, and then among those with and without SDB, separately, because mechanisms that underlie hyperactivity in children with SDB may differ from those in children without SDB. The high frequency of hyperactivity among children virtually assures that even if SDB contributes to this behavior in some study subjects, others without SDB may have the behavior for other reasons, and their data could then obscure the association of interest among children with SDB. All analyses were performed with SAS, version 6.12 (SAS Institute Inc., Cary, NC). In tests of statistical significance, the level was set at RESULTS Subjects Subjects ages ranged from 2.8 to 18.0 years; only one child was less than three years old, and the mean age was 9.9±4.0 years (Table 1). Seventy-three (65%) of the subjects were male. Fortynine subjects (43%) had OSA and an additional 10 subjects (9% of all subjects) had high upper airway resistance in the absence of OSA. As expected, in comparison to the 54 subjects without SDB, the 59 with SDB showed higher rates of apneas and hypopneas, lower minimum oxygen saturation, and more negative esophageal pressures. The two groups showed no statistically significant differences in age, sex, PLMI, sleep efficiency, or sleep latency (Table 1). At the time of polysomnography, most (n=40) subjects without SDB as defined in this study were thought to have primary snoring, potentially subtle SDB despite negative findings, or no sleep disorder. Other subjects without SDB received clinical diagnostic impressions of symptomatic PLMS or restless legs syndrome (n=7), parasomnias (n=2), delayed sleep phase syndrome (n=2), mild central sleep apnea (n=1), unspecified insomnia (n=1), and subtle SDB vs. narcolepsy (n=1). All diagnostic impressions were those of board-certified sleep specialists whose practices included children, but identification of symptomatic PLMS and restless legs syndrome can be particularly challenging in children and these two diagnoses were not based strictly on a specific set of research criteria. A total of 20 subjects had received a diagnosis of ADHD or were thought to have hyperactive behavior by their physicians, and 24 subjects had behavioral or psychiatric problems of other kinds, including anxiety disorders (n=3), bipolar disorder (n=3), SLEEP, Vol. 24, No. 3,

4 Table 2 All subjects (n = 113): Simple linear regressions of hyperactivity scores on each listed variable * Variable β s.e. F p-value R 2 AHI Minimum O 2 (%) Most Negative Pes (cm of water) OSA ** PLMI PLMD ** Sleep Efficiency Sleep Latency (min) Age (years) Male Sex History of ADHD < Stimulant Use * AHI = apnea/hypopnea index, minimum O 2 = minimum oxygen saturation, Pes = esophageal pressure, PLMI = periodic leg movement index, OSA = obstructive sleep apnea, PLMD = periodic limb movement disorder ** Present = 1, vs. absent = 0 and conditions that were described rather than labeled by parents (n=18). Neurological problems were reported in 29 children, most of whom had seizures (n=15) or headaches (n=5). Family histories were notable for ADHD (n=6 relatives of five children) and other learning or psychiatric problems (n=10 problems in nine relatives). Sixty-one children regularly took medications, the most common of which included stimulants (n=20 children), antiepileptics (n=15), antidepressants (n=10), neuroleptics (n=3), inhalers (n=4), antihistamines (n=6), diuretics (n=4), calcium channel blockers (n=3), and beta blockers (n=5). Among the 27 children with hyperactivity scores more than 2 SDs above normal, 9 were treated with stimulants and 18 were not. The proportion of hyperactive children who were treated with stimulants was identical among those with and without SDB (5 of 15 children and 4 of 12 children, respectively). Within the hyperactive group, nine children with SDB and five children without SDB took no stimulant or other psychoactive medication. Correlates of Hyperactivity The mean hyperactivity index t-score showed a statistically significant elevation from normal (50), at 59.2 ± 16.8 (95% C.I. [56.1, 62.3]). The mean scores were no different between subjects with and without SDB. Hyperactivity scores > 2 SDs above population means were somewhat more frequent among SDB subjects compared to non-sdb subjects (Table 1), but the difference was not statistically significant (χ 2 =0.16, p=0.69). Similar results were obtained for a history of attention-deficit/hyperactivity disorder and for current stimulant use. Regression of hyperactivity scores on AHI, minimum oxygen saturation, most negative esophageal pressure, PLMI, sleep efficiency, sleep latency, age, or sex showed no significant associations except that hyperactivity was associated with male sex (R 2 = 0.049, p = 0.02, Table 2). Regression of hyperactivity scores on SDB (presence vs. absence) also showed no significant association. Among the 29 children with PLMI 5, PLMI still showed no association with hyperactivity (R 2 = 0.000, p = 0.690, not shown in Table 2). Among these children, 11 (38%) had hyperactivity scores more than 2 SDs above normal, whereas among the 84 children with PLMI < 5, 16 (19%) had hyperactivity scores in this range. Regression of hyperactivity scores on PLMD (PLMI 5 vs. PLMI < 5) within the entire sample suggested that although a linear ( dose-dependent ) relationship with PLMI had not been found, the presence of PLMD was associated, on average, with an 8-point higher hyperactivity score (R 2 =0.045, p=0.02, Table 2). When male sex was added to the model, PLMD was still associated with higher hyperactivity scores (Part R 2 =0.034, p=0.047). For comparison, Table 2 also shows the strength of the associations between hyperactivity scores and two expected correlates: history of attentiondeficit/hyperactivity disorder and stimulant use. Controlling for stimulant use caused essentially no change in the strengths of the associations between hyperactivity and AHI, minimum oxygen saturation, most negative esophageal pressure, SDB, PLMI, or PLMD. To help exclude the possibility that the above analyses failed to detect a non-linear association, subjects were categorized based on AHI into four groups of similar size, but regression of hyperactivity on these categories still failed to reveal a significant association. A quadratic model of AHI also showed no significant association. Similar negative results were obtained with categorical and quadratic models of minimum oxygen saturation and esophageal pressure. Children With SDB Among the 59 children who had SDB, hyperactivity again showed no association with AHI, minimum oxygen saturation, most negative Pes (n = 15), sleep efficiency, or sleep latency (Table 3). However, hyperactivity was significantly associated with PLMI, PLMD, older age, and male sex. On average, an increase in the PLMI of 1 movement per hour was associated with a 1-point increase on the hyperactivity scale T-score (R 2 =0.159, p=0.002), and the presence of PLMD was associated with a 13-point increase in the hyperactivity score (R 2 =0.099, p=0.015; Figure 1A). In a multiple regression that adjusted for age and male sex, hyperactivity retained a significant association with PLMI (Part R 2 = 0.098, p = 0.010). SLEEP, Vol. 24, No. 3,

5 Table 3 Subjects with SDB (n=59): Simple linear regressions of hyperactivity scores on each listed variable * Variable β s.e. F p-value R 2 AHI Minimum O2 (%) Most Negative Pes (cm of water) OSA ** PLMI PLMD ** Sleep Efficiency Sleep Latency (min) Age (years) Male Sex * AHI = apnea/hypopnea index, minimum O 2 = minimum oxygen saturation, Pes = esophageal pressure, PLMI = periodic leg movement index, OSA = obstructive sleep apnea, PLMD = periodic limb movement disorder ** Present = 1, vs. absent = 0 Children Without SDB Among the 54 children who did not have SDB, hyperactivity showed no significant association with AHI, minimum oxygen saturation, most negative Pes (but n=4), sleep efficiency, sleep latency, age, male sex, PLMI, or PLMD (all p>0.10, Figure 1B). The association between hyperactivity and PLMI remained insignificant when two possible outliers with PLMI> 60 were removed from the analysis (F=1.037, p=0.3134, R 2 =0.020). An interaction model for all subjects (n=113) verified that SDB (as measured by AHI) was a statistically significant effect modifier in the relationship between hyperactivity and PLMI (Part R 2 =0.086, p=0.002 for the interaction term): increased A HIT PLMI was associated with increased hyperactivity among subjects with higher AHI but not among those with lower AHI (Figure 1). Adjustment for age and sex did not change this result appreciably (Part R 2 =0.075, p=0.003). In an analogous interaction model for minimum oxygen saturation and PLMI, the interaction term showed a trend toward significance (Part R 2 =0.034, p=0.052), even after adjustment for age and sex (Part R 2 =0.032, p=0.056). DISCUSSION This study of children with suspected SDB suggests that ageadjusted hyperactivity scores are high among children with B HIT PLMI PLMI Figure 1 Effect modification by sleep-disordered breathing (SDB). When SDB was present (A), higher periodic leg movement indices (PLMI) were associated with higher hyperactivity index t-scores (HIT), but when SDB was absent (B), no such association existed. SLEEP, Vol. 24, No. 3,

6 polysomnographically-confirmed SDB, but no higher than among children without SDB. Levels of hyperactivity were not associated with measures of SDB severity, but were associated with the presence of PLMS. Upon more detailed analysis, the association between hyperactivity and PLMS derived entirely from the results of the children who had SDB; no such association could be detected among children without SDB. More severe SDB, as measured by AHI, linearly increased the strength of the association between PLMS and hyperactivity. These results which to the authors knowledge represent the first quantitative behavioral and polysomnographic comparisons in children are surprising in several respects. Conclusions that can be drawn from correlative data in a clinical sample have several notable limitations, but the findings may have important implications for the hypothesis that inadequate sleep, common to several types of sleep disorders, contributes to hyperactivity in children. The high level of hyperactivity in subjects not found to have SDB was unexpected. Previous clinical series that described hyperactive behavior in children with polysomnographicallyconfirmed SDB did not discuss referred children who did not have SDB. 1,2 In one study, 11 snoring children without SDB showed similar hyperactivity scores to those of 12 children with SDB, but the assessment for SDB relied primarily on oximetry and video data. 4 Another unexpected observation was the extent to which PLMD, in this sample of children studied for suspected SDB, was associated with hyperactivity. However, this result is consistent with previous studies, among different samples of children, that suggest a strong association between PLMD and ADHD, 10,12,42 especially when PLMD is severe. 11 In a consecutive series of children with attention-deficit/hyperactivity disorder, 26% were identified by parental report as having periodic leg movements, studied with polysomnography, and confirmed to have PLMIs > Although the current data showed no direct association between SDB and hyperactivity, SDB did appear to be an important effect modifier 43 of the association between PLMS and hyperactivity: among subjects with SDB, but not those without SDB, PLMI showed a dose-dependent association with hyperactivity. This previously unreported finding was not hypothesized prior to the study, and therefore requires future confirmation. If verifiable, the finding may have implications for mechanisms that underlie reports of associations between sleep disorders and hyperactivity. Our study also has several limitations that stem from techniques used and subjects enrolled. First, some of our subjects who did not have SDB had other sleep-related diagnoses, and all were referred for sleep-related symptoms. Many of the subjects had psychiatric problems, neurological problems (especially seizures), and medication regimens for which we could not control. The complexity of this patient population may reflect, in part, the location of our sleep laboratory in a neurology department of a large tertiary referral center. In addition, despite use of both age-appropriate T-scores and statistical adjustment for age effects, the broad age range included in this study suggests that hyperactivity, even as measured on a single instrument, is likely to represent somewhat different constructs at the two ends of the age spectrum. Second, SDB may have remained undetected in some cases by SLEEP, Vol. 24, No. 3, the methods and criteria used in this study. In adults, the AHI is not strongly correlated with objective measures of sleepiness, 36,38 and in children, upper airway resistance syndrome (with AHIs close to zero) and OSA reportedly are associated with similar frequencies of hyperactivity. 1,2 These observations suggest that the measure of SDB severity most relevant to behavioral outcomes may not be the AHI we calculated. In young children, a shorter criterion for apnea and hypopnea length, 39 commensurate with increased respiratory rates, may produce a more relevant AHI. Furthermore, airflow monitoring with nasal pressure cannulae are now advocated in adults to increase sensitivity for hypopneas. 40 Failure to detect subtle hypopneas in some patients may have allowed some hypopneas to be scored as PLMS, as the two phenomena sometimes can be difficult to distinguish. Future studies that incorporate more sensitive measures of SDB might allow currently undetected relationships to emerge between SDB, PLMS, and behavior. On the other hand, no outcome-based studies have yet demonstrated the advantage of shorter criteria for apneas and hypopneas, 41 and no studies of nasal pressure in children, who are often mouth-breathers, have been published. Even if sub-optimal in younger children, the AHI derived from 10-second events and thermocouple recordings would still be expected to show high (if imperfect) correlations with the AHI derived from other criteria or equipment. Finally, in a subsample of 19 children in the current study, even a gold-standard assessment of respiratory effort esophageal pressure monitoring failed to show any association with hyperactivity. These observations support the validity of our central finding that levels of hyperactivity were not associated with SDB severity. Figure 2 shows several potential mechanisms that could theoretically explain associations between sleep disorders and hyperactive behavior in children. In model A, SDB, PLMS, and other sleep disorders contribute to hyperactive behavior, perhaps through intervening variables such as sleep disruption or deprivation. Our current data, which show no association between SDB and hyperactivity, make the hypothesis that SDB contributes directly to the abnormal behavior less tenable. In addition, the lack of association between hyperactivity and either sleep efficiency or nocturnal sleep latency though not optimal measures of sleep disruption and sleepiness fail to support the idea that sleep disruption or sleepiness acts as an intervening variable. In model B, a third, as yet unidentified variable causes both PLMS and hyperactive behavior. Neurological disorders, antiepileptic agents, and other medications all of which were common in our clinical sample may affect sleep and behavior, and could potentially explain comorbidity of PLMS and hyperactive behavior. Correlative data such as ours cannot distinguish models A and B. However, model B is difficult to reconcile with reports that other sleep disorders besides PLMS are associated with hyperactive behavior; that experimental sleep deprivation, inadequate sleep, or sleepiness can lead to such behavior; and that treatment for a wide range of sleep disorders leads to behavioral improvement. 2,16,44-46 In models C and D, PLMS are hypothesized to contribute to hyperactive behavior or hyperactive behavior is hypothesized to produce PLMS, but in each case SDB or other sleep disorders are shown as effect modifiers, as suggested by our data for SDB. Only model C, however, is consistent with past reports which suggest that treatment for sleep disorders ameliorates hyperactiv-

7 A. SDB Hyperactive Behavior PLMS Hyperactive Behavior Sleep Disorders Hyperactive Behavior B. Hyperactive Behavior Third Variable PLMS C. SDB, Other Sleep Disorders PLMS D. SDB, Other Sleep Disorders PLMS ity. The designs of these studies have prevented firm conclusions, but we speculate that future data on these important issues will support model C. Our findings have several important implications for clinical practice and research. The results of this study suggest that SDB is not a direct cause of hyperactivity in children, though its presence may modify a relationship between PLMS and daytime behavior. Clinicians who find PLMS and SDB on a polysomnogram often dismiss the PLMS as an incidental finding, and treat only the SDB. Our findings raise the possibility that treatment of either problem could improve a child s daytime behavior. However, our study was observational in design; and associations cannot establish causality. Future research should endeavor to clarify the central neurophysiology of PLMS and hyperactive behavior, to further examine the possibility that PLMS contribute to hyperactive behavior, and to explain the behavioral effects of nocturnal sleep disruption. ACKNOWLEDGMENTS Submitted by NINDS (K02-NS02009). The authors wish to thank Judy Wiebelhaus, RPSGT, for technical assistance with these studies. REFERENCES Hyperactive Behavior Hyperactive Behavior Figure 2 Several hypothetical models that could explain associations between sleep disorders and hyperactivity in children are shown. In Model A, various sleep disorders can contribute to hyperactive behavior, perhaps through an intervening variable such as sleep disruption. In Model B, a third variable (i.e., some as yet unidentified factor) produces both hyperactive behavior and PLMS. In Model C, PLMS contribute to hyperactive behavior; SDB and perhaps other disorders that disrupt sleep act as effect modifiers and permit the effect of PLMS on behavior to be expressed. In Model D, hyperactive behavior leads to PLMS, and other sleep disorders again act as effect modifiers. Data from the current study arestudy, in combination with those from previous reports, are more consistent with Models C and D, but notd than with A or B. SDB = sleep-disordered breathing; PLMS = periodic leg movements during sleep. 1. Guilleminault C, Korobkin R, Winkle R. A review of 50 children with obstructive sleep apnea syndrome. Lung 1981;159: Guilleminault C, Winkle R, Korobkin R, Simmons B. Children and nocturnal snoring evaluation of the effects of sleep related respiratory resistive load and daytime functioning. Eur J Pediatr 1982;139: SLEEP, Vol. 24, No. 3, Hansen DE, Vandenberg B. Neuropsychological features and differential diagnosis of sleep apnea syndrome in children. Journal of Clinical Child Psychology 1997;26: Ali NJ, Pitson D, Stradling JR. Sleep disordered breathing: effects of adenotonsillectomy on behaviour and psychological functioning. Eur J Pediatr 1996;155: Goldstein NA, Post JC, Rosenfeld RM, Campbell TF. Impact of tonsillectomy and adenoidectomy on child behavior. Arch Otolaryngol Head Neck Surg 2000;126: Szatmari P. The epidemiology of attention-deficit hyperactivity disorder. Child Adolesc Psychiatr Clin North Am 1992;1: Ali NJ, Pitson DJ, Stradling JR. Snoring, sleep disturbance, and behaviour in 4-5 year olds. Arch Dis Child 1993;68: Gislason T, Benediktsdottir B. Snoring, apneic episodes, and nocturnal hypoxemia among children 6 months to 6 years old. An epidemiologic study of lower limit of prevalence. Chest 1995;107: Chervin RD, Dillon JE, Bassetti C, Ganoczy DA, Pituch KJ. Symptoms of sleep disorders, inattention, and hyperactivity in children. Sleep 1997;20: Picchietti DL, Walters AS. Restless legs syndrome and periodic limb movement disorder in children and adolescents: comorbidity with attention-deficit hyperactivity disorder. Child Adolesc Psychiatr Clin North Am 1996;5: Picchietti DL, Walters AS. Moderate to severe periodic limb movement disorder in childhood and adolescence. Sleep 1999;22: Picchietti DL, England SJ, Walters AS, Willis K, Verrico T. Periodic limb movement disorder and restless legs syndrome in children with attention-deficit-hyperactivity disorder. Journal of Child Neurology 1998;13: Guilleminault C, Pelayo R. Narcolepsy in prepubertal children. Annals of Neurology 1998;43: Dahl RE, Holttum J, Trubnick L. A clinical picture of child and adolescent narcolepsy. Journal of the American Academy of Child & Adolescent Psychiatry 1994;33: Walters AS, Mandelbaum DE, Lewin DS, Kugler S, England SJ, Miller M. Dopaminergic therapy in children with restless legs/periodic limb movements in sleep and ADHD. Pediatr Neurol 2000;22: Dahl RE. Narcolepsy in children and adolescents. Child Adolesc Psychiatr Clin North Am 1996;5: Chervin RD, Guilleminault C. Obstructive sleep apnea and related disorders. Neurologic Clinics 1996;14: Chervin RD, Aldrich MS. Characteristics of apneas and hypopneas during sleep and relation to excessive daytime sleepiness. Sleep 1998;21: Ring A, Stein D, Barak Y, Teicher A, Hadjez J, Elizur A, Weizman A. Sleep disturbances in children with attention-deficit/hyperactivity disorder: a comparative study with healthy siblings. Journal of Learning Disabilities 1998;31: Owens JA, Maxim R, Nobile C, McGuinn M, Msall M. Parental and self-report of sleep in children with attention-deficit/hyperactivity disorder. Arch Pediatr Adolesc Med 2000;154: Kaplan BJ, McNicol J, Conte RA, Moghadam HK. Sleep disturbance in preschool-aged hyperactive and nonhyperactive children. Pediatrics 1987;80: Corkum P, Tannock R, Moldofsky H. Sleep disturbances in children with attention-deficit/hyperactivity disorder. Journal of the American Academy of Child & Adolescent Psychiatry 1998;37: Connors CK. A teacher rating scale for use in drug studies with children. Am J Psychol 1969;126: Conners CK, Barkley RA. Rating scales and checklists for child psychopharmacology. Psychopharmacology Bulletin 1985;21: Asher MI, Coates AL, Collinge JM, Milic-Emili J. Measurement of pleural pressure in neonates. J Appl Physiol 1982;52: Chervin RD, Aldrich MS. Effects of esophageal pressure monitoring on sleep architecture. Am J Resp Crit Care Med 1997;156: Rechtschaffen A, Kales A. A manual of standardized terminology,

8 techniques and scoring system for sleep stages of human subjects. Los Angeles, CA: Brain Information Service/Brain Research Institute, UCLA, Greene MG, Carroll JL. Consequences of sleep-disordered breathing in childhood. Current Opinion In Pulmonary Medicine 1997;3: Marcus CL, Omlin KJ, Basinski DJ, Bailey SL, Rachal AB, Von Pechmann WS, Keens TG, Davidson-Ward SL. Normal polysomnographic values for children and adolescents. Am Rev Respir Dis 1992;146: Gould GA, Whyte KF, Rhind GB, Airlie MA, Catterall JR, Shapiro CM, Douglas NJ. The sleep hypopnea syndrome. Am Rev Respir Dis 1988;137: Guilleminault C, Stoohs R, Clerk A, Cetel M, Maistros P. A cause of excessive daytime sleepiness: the upper airway resistance syndrome. Chest 1993;104: Mathur R, Douglas NJ. Frequency of EEG arousals from nocturnal sleep in normal subjects. Sleep 1995;18: Guilleminault C, Stoohs R, Clerk A, Simmons J, Labanowski M. Excessive daytime somnolence in women with abnormal respiratory efforts during sleep. Sleep 1993;16:S137-S Guilleminault C, Stoohs R, Kim Y, Chervin R, Black J, Clerk A. Upper airway sleep-disordered breathing in women. Ann Intern Med 1995;122: Chervin RD, Hedger KM, Dillon JE, Pituch KJ. Pediatric Sleep Questionnaire (PSQ): validity and reliability of scales for sleep-disordered breathing, snoring, sleepiness, and behavioral problems. Sleep Med 2000;1: Chervin RD, Kraemer HC, Guilleminault C. Correlates of sleep latency on the multiple sleep latency test in a clinical population. Electroencephalogr Clin Neurophysiol 1995;95: American Sleep Disorders Association. International classification of sleep disorders, revised: diagnostic and coding manual. Rochester, MN: American Sleep Disorders Association, Chervin RD, Aldrich MS. The relation between MSLT findings and the frequency of apneic events in REM and NREM sleep. Chest 1998;113: American Thoracic Society: Medical Section of the American Lung Association. Standards and indications for cardiopulmonary sleep studies in children. Am J Resp Crit Care Med 1996;153: American Academy of Sleep Medicine Task Force. Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. Sleep 1999;22: American Thoracic Society. Cardiorespiratory sleep studies in children: establishment of normative data and polysomnographic predictors of morbidity. Am J Resp Crit Care Med 1999;160: Picchietti DL, Underwood DJ, Farris WA, Walters AS, Shah MM, Dahl RE, Trubnick LJ, Bertocci MA, Wagner M, Hening WA. Further studies on periodic limb movement disorder and restless legs syndrome in children with attention-deficit hyperactivity disorder. Movement Disorders 1999;14: Hennekens CH, Buring JE. Epidemiology in medicine. Boston, MA: Little, Brown and Company, Randazzo AC, Muehlbach MJ, Schweitzer PK, Walsh JK. Cognitive function following acute sleep restriction in children ages Sleep 1998;21: Dahl RE. The impact of inadequate sleep on children s daytime cognitive function. Seminars in Pediatric Neurology 1996;3: Dahl RE, Pelham WE, Wierson M. The role of sleep disturbances in attention deficit disorder symptoms: a case study. J Pediatr Psychol 1991;16: SLEEP, Vol. 24, No. 3,