Does phenobarbital used for febrile seizures cause sleep disturbances?

OriginalArticles

Does Phenobarbital Used for Febrile Seizures • 9 Cause Sleep Disturbances. D e b o r a h G. Hirtz, M D * , Ta C h u a n C h e n , PhD+, K a r i n B. N e l s o n , MD~, S t e p h e n S u l z b a c h e r , PhD~, J a c q u e l i n e R. F a r w e l l , MD% a n d J o n a s H. E l l e n b e r g , PhD+

The effect of phenobarbital on total sleep time, night awakenings, and lengthy awakenings was examined as part of a randomized trial of children with febrile seizures; information about sleep patterns was gathered by parental observation. Children were between ages 8-36 months at enrollment and were examined subsequently for 21/2 years. Night awakenings were not more common in children assigned to phenobarbital except for those who were poor sleepers at the beginning of the study. Total sleep time was no different in children assigned to phenobarbital than in those assigned to placebo. It is concluded that sleep problems reported in most young children with febrile seizures treated with phenobarbital did not exceed those reported in children treated with placebo, but a subset of predisposed children did experience an increase in night awakenings. Hirtz DG, Chen TC, Nelson KB, Sulzbacher S, Farwell JR, Ellenberg JH. Does phenobarbital used for febrile seizures cause sleep disturbances? Pediatr Neurol 1993;9: 94- 100.

Introduction Parents of young children treated with daily phenobarbital for prevention of febrile seizure recurrences commonly complain about irritability and sleep disturbances [1-4]. Some reports have also noted lengthy night awakenings in treated children [3,5[. Febrile seizures usually occur in children 6 months to 4 years old, an age during which sleep and behavior problems are c o m m o n and often unpredictable. Age-related sleep disruptions could easily be confused with medication side effects. Most infants sleep through the night by 6 months of age, but in the second year of life almost one-half of those who had previously slept through the night revert to nighttime awakenings [6]. Night awakenings may occur through the age of 3 years in as many as one-fourth of

From the *Developmental Neurology Branch; tBiometry and Field Studies Branch; and ~NeuroepidemiologyBranch; National Institutes of Neurological Disorders and Stroke, NIH; Bethesda, Maryland; §Child Psychiatry and Behavioral Medicine; and 1Department of Child Neurology; University of Washington; Seattle, Washington.

94 PEDIATRIC NEUROLOGY Vol. 9 No. 2

children [7,8]. Disruption of previously established sleep patterns at this age may often occur as p a n of normal developmental patterns [9] or in response to a change in the child's physical or emotional environment. A number of factors have been related to night awakenings, including events in pregnancy, labor, and delivery [ 10-t2], temperament [12-15], family stress [12-16], breast feeding [17-20], night feeding, and cosleeping [ 17-211. Phenobarbital causes adults to fall asleep faster, decreases nighttime wakefulness, and suppresses REM sleep [22-24]. In adult epileptic patients taking phenobarbital, sedation and drowsiness appear acutely but disappear with chronic use [25]. Instead of sedation, children tend to experience paradoxical excitation with phenobarbital. Typical complaints include daytime irritability, fussiness, and hyperactivity [1,26], as well as disruption of nighttime sleep. During the course of a randomized, double-blind, placebo-controlled trial examining behavioral and cognitive effects of phenobarbital used for prevention of febrile seizure recurrence [27], parents were instructed to keep sleep logs for the 3 days before each visit. These records were analyzed to determine whether the phenobarbital group disclosed differences in sleep patterns from those observed in the placebo group. The sleep records of the children with febrile seizures were also compared to those of age-matched, seizure-free control children, obtained over corresponding time intervals.

Methods In our study, 217 children with febrile seizures from the greater Seattle area, between ages 8-36 months, were enrolled in the study over a 3-year period. Febrile seizures were defined as seizures occurring in the presence of fever but without concurrent infection of the central nervous system or a previous history of nonfebrile seizures. Children who had a single, brief febrile seizure and no identified risk factors tbr subsequent seizures were excluded; eligibility for the trial was based on characteristics of the child or the seizure that would have allowed that child to be considered for treatment according to local practice. These characteristics included age younger than 1 year, at least 2 febrile seizures, a history of epilepsy in an immediate family member, pre-existing

Communications should be addressed to: Dr. Hirtz; Developmental Neurology Branch, NINDS-NIH; 7550 Wisconsin Avenue; Rm. 8C-02; Bethesda, MD 20892. Received September 15, 1992; accepted January 4, 1993.

Table 1. Age of children at each visit

Controls Ave. Age No. (mos)

Febrile Seizure Patients Placebo Phenobarbital Ave. Age Ave. Age No. (mos) No. (mos)

Study entry

141

19.3

96

20.5

88

19.3

6 weeks

127

20.6

79

21.6

74

20.8

6 months

121

25.2

80

27.2

70

25.7

12 months

125

31.4

82

32.7

72

31.7

18 months

111

37.5

74

38.8

63

37.8

24 months

117

43.8

81

44.8

68

43.2

30 months

116

49.9

78

51.8

67

50.0

Visit*

* Follow-up visits were made, on the average, 2 weeks later than the expected schedule. neurologic abnormality, or a seizure that was focal, long (> 15 min), or recurred multiple times within 24 hours [27]. After informed consent was obtained, children were randomly assigned to either phenobarbital (4-5 mg/kg/day) or placebo. Medication was to be given dally for 2 years and then gradually withdrawn over 1-2 months and a final visit was made at 21/2 years. Blood levels were drawn for all children taking either study medication at each visit and spurious values were reported to study physicians for those on placebo. Children were also studied with cognitive and behavioral measures at study entry, 6 weeks after starting medication, and then every 6 months. A concurrent group of 150 seizure-free, age-matched control children were enrolled and tested with the same measures at the same intervals as the children with febrile seizures. Parents of the patients and controls were requested to keep a 3-day log of their child's sleep/wake activity before the initial evaluation and each of the next 6 visits. The first follow-up visit took place at 6 weeks, the next at 6 months, and every 6 months thereafter. Parents recorded, in half-hour time intervals, 3 consecutive sleep/wake cycles totaling 72 hours, indicating periods of sleep, wakefulness, or wakefulness with crying or fussing. The primary variables analyzed included sleep disturbances and total sleep time, defined as follows: (1) Sleep disturbances included episodes of waking, with or without crying or fussing, during the nighttime sleep period (i.e., the time from initial sleep at night to the morning awakening). This variable was dichotomized as frequent (2 or more episodes per 3-day cycle) or infrequent (1 or no episodes per 3-day cycle); and, (2) Total sleep time per 24 hours (nighttime sleep plus nap time) was the mean number of hours of sleep per 24-hour period for each 3-day cycle. Treatment comparisons considered each child to be in the treatment group to which he or she was originally assigned. Initial assessment of differences in rates of sleep disturbance among the two treatment groups and control group were accomplished using chi square tests with 2 degrees of freedom and individual group comparisons were performed using chi square analysis with 1 degree of freedom only when the overall chi square was significant at P < .05. A log linear model method, which included demographic and other study entry characteristics, was also used to investigate the effect of phenobarbital on sleep disturbances [28,29]. A general linear model approach [30,31] was employed to investigate the regression of the continuous variable total sleep time on treatment group, total sleep time at baseline, age, and other demographic variables. Given the many statistical tests made within this large dataset, some were likely to achieve nominal levels of statistical significance by chance alone. We did not adjust significance levels using multiple comparison techniques which would lessen the chance of observing some

clinically important risk factors [32]; however, when appropriate, statistical significance was required of simultaneous comparisons of multiple groups prior to making intergroup comparisons.

Results T h e r e w e r e 2 1 7 c h i l d r e n w i t h febrile s e i z u r e s e n r o l l e d in the study, 108 o f w h o m w e r e a s s i g n e d to p h e n o b a r b i t a l a n d 109 to p l a c e b o . A c c o r d i n g to c a l c u l a t i o n s b a s e d o n life-table m e t h o d s , b y t h e 1-year visit the p r o b a b i l i t y t h a t a s u b j e c t w o u l d c o n t i n u e to r e c e i v e the a s s i g n e d m e d i c a t i o n was 0.82 in the p h e n o b a r b i t a l g r o u p a n d 0.65 in t h e p l a c e b o g r o u p ; at 2 years, the c o m p a r a b l e figures w e r e 0 . 6 6 a n d 0.46, respectively. S o m e o f t h e s e c h i l d r e n w h o d i s c o n t i n u e d a s s i g n e d m e d i c a t i o n w e r e c r o s s o v e r s to t h e o t h e r s t u d y m e d i c a t i o n , a n d others s t o p p e d t a k i n g e i t h e r study m e d i c a t i o n [27]. B e c a u s e a c o m p l e t e set o f sleep r e c o r d s for all visits was n o t a v a i l a b l e for m a n y patients, statistical a n a l y s e s w e r e c r o s s - s e c t i o n a l b y visit, r a t h e r than longitudinal. E i g h t y - e i g h t c h i l d r e n a s s i g n e d to p h e n o b a r b i t a l a n d 96 a s s i g n e d to p l a c e b o h a d sleep r e c o r d s at s t u d y entry. A p p r o x i m a t e l y 8 0 % o f t h e s e c h i l d r e n p r o v i d e d sleep r e c o r d s at o n e or m o r e o f the visits a f t e r s t u d y e n t r y (Table 1) a n d t h e s e p a t i e n t s w e r e i n c l u d e d in the a n a l y s e s for total sleep t i m e a n d sleep d i s t u r b a n c e s . T h e p r o t o c o l r e q u i r e d t h a t the first d o s e o f m e d i c a t i o n b e g i v e n after the initial sleep r e c o r d s w e r e c o m p l e t e d . In a p p r o x i m a t e l y 2 5 % o f patients in the p h e n o b a r b i t a l g r o u p t h e initial 3 - d a y sleep records may have included some days after initiation of the drug. T h e results o f the a n a l y s e s w e r e n o t d i f f e r e n t w h e n the s u b g r o u p o f c h i l d r e n w h o s e b a s e l i n e sleep records m a y h a v e b e e n c o m p l e t e d after the date o f the initial visit w e r e e x c l u d e d . A p p r o x i m a t e l y 8 2 % o f t h e p h e n o b a r bital a n d p l a c e b o g r o u p s a n d 8 8 % o f the c o n t r o l g r o u p h a d s l e e p r e c o r d s for at least 2 o f the 3 f o l l o w - u p visits in the first s t u d y year. T h e m e a n age at e n t r y for t h e p l a c e b o g r o u p w a s 20.5 m o n t h s (range: 8-35 m o n t h s ) a n d for the p h e n o b a r b i t a l

Hirtz et al: Phenobarbital and Febrile Seizures

95

group 19.3 months (range: 8-34 months; Table l). There were more males than females in both groups, but the differences were not significantly different for either treatment group (placebo: 56% vs 44%; phenobarbital: 58% vs 42%). The socioeconomic levels were also comparable for both groups. The control group was comparable in age and slightly higher in socioeconomic level. The frequency of missing data for sleep disturbances in the first year, whether because of missed visits, dropping out, or failure to complete sleep records, did not differ between the placebo and phenobarbital groups. The number of sleep disturbances at study entry, sex, age, and socioeconomic index (SEI) [33] were comparable for patients analyzed and those excluded because of incomplete records. Total S l e e p Time. For normal controls and both seizure groups, total sleep time decreased over the study period as the ages of children increased, chiefly due to decreasing nap time (Fig 1). After adjustment for SEI, age, sex, and initial differences in total sleep time, the placebo and phenobarbital groups did not differ at any visit from each other or from normal controls with regard to total sleep time, night sleep time, or nap time. S l e e p D i s t u r b a n c e s . Children were grouped according to the frequency of nighttime sleep disturbances, dichotomized as frequent (i.e., 2 or more episodes) or less than 2

13

0 Normal Control A Placebo

o L~

• Phenobarbital

O

12

O Total Sleep

0

¢n

0 A

-A

-r

.c_

o

O

o Night Sleep

•

k: 03

It L~

m Nap Sleep

0

I

0

I

I

'

'

I

6

. . . . .

I , , , , ,

12

I

,

J

18

~ , ,

I , ,

24

,

,

,

I

30

Study Visit (Months)

Figure 1. Average number of hours o f sleep per 24 hours: phenobarbital (0), placebo (A), and normal control (0). Differences existed at the initial visit among the 3 groups for total sleep time [F(2,319) = 5.40; P < .01] and night sleep time [F~2.319; = 8.26; P < .01], although the individual comparisons between placebo and phenobarbital were not statistically different•

96 PEDIATRICNEUROLOGY Vol.9 No. 2

episodes of waking during the total night hours for the 72-hour cycle reported before each visit. Ahout 80% t~t the sleep disturbances occurred between midnight and f~ a.m.; the great majority of the remainder occurred befi)re midnight. At study entry, frequent sleep disturbances were observed in 43.2% of the phenobarbital group, 30.2% of the placebo group, and 30.5% of control group. These differences were not significant (Z2(2) = 4.6; P = 0.1). Analysis using a log linear model, which included treatment group, age at time of visit, sex of child, SEI, and presence of frequent sleep disturbances at study entry, indicated that the number of sleep disturbances at study entry was an important factor associated with sleep disturbances in the first study year (6 weeks: P < .001:6 months: P < .001; 12 months: P = .086). For controls and both seizure groups, children who did not have frequent sleep disturbances at study entry remained without frequent sleep disturbances at later visits (Fig 2A), whereas in children who were poor sleepers at study entry (Fig 2B), the percentage of those who continued to have frequent sleep disturbances declined during the study as the children became older. In the latter group, children assigned to phenobarbital had more observed sleep disturbances for all visits in the first year of the study than those assigned to placebo or normal controls. A log linear model incorporating the covariates above was utilized to assess the adjusted differences in sleep disturbances at the different visits. At the 6-week visit, which best identified short-term effects, 63% of the phenobarbital group and 56% of the placebo group had frequent sleep disturbances (P = 0.58). The largest observed difference occurred at the 6-month visit, with 56% of the phenobarbital group and 32% of the placebo group having frequent sleep disturbances (P = .08), but this difference was not statistically significant. Because only 70-80% of children bad sleep records available for all 3 visits in the first study year, cross-sectional comparisons may have lacked sufficient power to detect differences during any individual visit; therefore, we considered differences in the percentage of children in each group with frequent sleep disturbances for at least 2 of the 3 follow-up visits in the first study year. No overall difference existed in the proportion of children with frequent sleep disturbances among the 3 groups, except in those with frequent initial sleep disturbances, among whom the treatment effect was highly significant (P = .002). After controlling fbr age at study entry, sex, and SEI, a large difference was observed between the phenobarbital and placebo groups (59% vs 27%; P = .01) in reference to frequent sleep disturbances during the first year of study. At the 18- and 24-month visits, the number of children with sleep disturbances was low in all groups with no differences observed. In children reported to be taking their medication, we assessed whether the amount of exposure to phenobarbital was related to the degree of sleep disruption. In the first weeks on the drug, adjustments in dosage were often

70O Normal Control

8o 60

A Placebo

g

• Phenobarbital

=

50

~

40

'~

30

.--- 20

A

A u

,

a~

0

10 0

I

I

I

I

I

l

I

l

I

I

l

0

0

l

6

0

L~

O

0

I

I

I

I

i

[

12

•

0

& I

I

I

I

I

l

18

i

I

I

I

i

l

24

30

Study Visit (Months) Sample Size Control

88

83

89

81

84

83

Placebo

54

58

57

53

56

55

Pheno.

39

36

40

37

36

37

A 70

8 =~ 60

~

60

_~

40

~

3O

z

20

N

10

ffl

O Normal Control •

A

/ ~ Placebo

•

•

Phenobarbital

O

A m

6 O I

I

I

I

I

l

6

I

I

l

I

I

|

12

zx I

I

I

I

I

i

18

I

I

I

I

I

[

I

I

I

I

I

]

24

30

Study Visit (Months) Sample Size Contml

39

38

36

30

33

33

Placebo

25

22

25

21

25

23

Phen~

35

34

32

26

32

30

B

Figure 2. Percentage of children with 2 or more episodes of night sleep disturbances 3 nights prior to each visit, illustrated for baseline sleep disturbance groups. Figure 2A documents children with one or no sleep disturbances and 2B those with two or more at study entry. For each visit, symbols represent mean values for all children in each group: phenobarbital (0), placebo (A), and normal control (0).

made; therefore, during the first scheduled return visit at 6 weeks, the actual blood level on the day of the visit was used as the measure of drug exposure. For the subsequent visits, we derived an estimated average blood level over each 6-month time period preceding the visit using all reported blood levels, parents' reports of compliance, and records of medication changes. The groups were divided

by estimated average blood levels of phenobarbital which were either _< 15 ~tg/mg or > 15 ~tg/mg. The level of phenobarbital exposure was not associated with sleep disturbances during the first year in children with few sleep disturbances at study entry. For the children with frequent initial sleep disturbances, we observed an inverse relationship between phenobarbital exposure

Hirtz et al: Phenobarbital and Febrile Seizures

97

lOO~ g

90

so ~

70

~

6o

'~

•

50

0

40

[]

Normal Control

[]

High Phenobarbital (>15#g/ml)

•

Low Phenobarbital (~ 15 ~tg/ml (~), average phenobarbital < 15 ~g/ml (0), and normal control (0). The computed Z2(/) values for comparison of the differences between lowand high-phenobarbital exposure at 6 weeks, 6 months, and 24 months were higher than the tabled values at P = .05.

and sleep disturbance (Fig 3) which was significant for some but not all visits. These results are based on subgroup analyses with attendant biases and caution should be used in generalizing from the results. Lengthy Awakenings. For this measure, the number of lengthy night awakenings from all available visits was used. Night awakenings lasting longer than 1 hour were uncommon and tended to decrease in frequency through the course of the study (Table 2). Lengthy awakenings occurred more frequently in febrile seizure children than in control children at study entry and in the first 12 months (P < .001). Lengthy awakenings were not statistically different at any visit between the phenobarbital and placebo groups. Disturbances lasting longer than 2 hours occurred even less often and were not more common in children assigned to phenobarbital versus placebo.

Fussiness Before Sleep. Parents reported whether children were awake and fussy before falling asleep for the night. At study entry, there were significant differences among the 3 treatment groups (phenobarbital 20.1%, placebo 18.6%, controls 12.2%; Z2(2) = 109; P < .01), with febrile seizure children tending to have more fussiness in the 30 min prior to sleep than the controls. These complaints decreased to about 15% by 1 year in all 3 groups, and by 2 years to 10% or less. No differences in presleep fussiness occurred between the phenobarbital and placebo groups. Discussion

The peak age of febrile seizure onset is t8 to 22 months. That is also about the age at which sleep disturbances commonly appear in children who have pre-

Table 2. Number of patient days with at least 1 episode of night sleep disturbance longer than 1 hour

Controls (N* = 148) Days % All visits

Febrile Seizure Patients PhpAloPlacebo (N* = 108) barbital (N* = 103) Days % Days %

2,790

2.4

1,907

4.2

1,731

5. I

Study entry

510

2.9

334

6.6

304

8.6

6 weeks

474

2.3

304

5.9

294

4.8

1,392

2.5

934

5.5

880

5.2

888

1.8

639

1.1

547

3.1

First 12 mos Second 12 mos

* Night awakenings from all available visits were used. One patient assigned to placebo, 5 patients assigned to phenobarbital, and 2 controls had no sleep data.

98

PEDIATRIC NEUROLOGY

Vol. 9 No. 2

viously slept through the night. Only a blinded, controlled study can distinguish the effects of medication from spontaneous perturbations common in children of this age. In this study, we found less disturbance of sleep than we had anticipated on the basis of commonly reported parental complaints. Perhaps it is analogous that, although hyperactivity is commonly attributed to phenobarbital, it was not associated with phenobarbital treatment in the doubleblind, placebo-controlled, randomized study by Camfield et al. [3]. Behavioral disturbances troubling to parents often appear during the "terrible twos;" when this behavior occurs near the time phenobarbital is prescribed, parents may attribute the change to the medication. Definitions of sleep disturbances vary, but usually night awakenings 2-4 times a week are considered a problem [34]. About 30% of control children entering this study had frequent nighttime sleep disturbances which is consistent with most surveys [7]. The children in the febrile seizure group were more likely to have reported sleep disturbances at study entry than were the controls, perhaps due to their recent illness. An excess of sleep disturbances observed in febrile seizure children assigned to phenobarbital was observed only in those who had exhibited sleep disturbances at the beginning of the study. Another example of the interaction of medication with intrinsic vulnerability was reported in a randomized study of children with febrile seizures treated with phenobarbital, in which the only characteristic of the child or the seizure which was related to the occurrence of a behavior disorder was hyperactivity before the first febrile seizure [1]. Similarly, in a study of older children treated with phenobarbital, depression was observed primarily in children with a predisposition consisting of a personal or family history of an affective disorder [35]. The relationship of sleep disturbances to the amount of drug exposure appeared to be paradoxical, with lower exposure to medication associated with more frequent sleep disturbances, an effect observed only in the group with more frequent sleep disturbances at study entry. We cannot distinguish a pharmacologic effect from factors leading to low compliance. These results must be interpreted with great caution and considered exploratory because they are derived from a subgroup of children whose variations in blood levels may have been related to various factors that could affect sleep disturbances, such as willingness to take medication in the face of perceived side effects; however, they do raise the interesting possibility that lower levels of exposure are associated with a higher rate of sleep disturbance compared to higher levels of exposure in susceptible children. Unlike Camfield et al. [3], we found that lengthy night awakenings were rare and neither lengthy awakenings nor fussiness before sleep was more frequent at any visit in those assigned to phenobarbital as compared to placebo. The slight, absolute increase in lengthy awakenings observed in children with febrile seizures as compared to

controls in the first study year may have been due to family disruption from the concern over febrile seizures. Crossover between phenobarbital and placebo was not infrequent in this study and could potentially mask side effects of medication by including some children in the treated group who were not treated and vice-versa; however, the "intent to treat" analysis is appropriate here, and the occurrence of crossover is consistent with actual events in clinical practice. When analysis is not performed according to "intent to treat," the benefits of randomization are lost and unknown factors affecting the outcome in question may be unbalanced in the two treatment groups. Although most infants continue to awaken at night, by the time they are a few months old, one-half of them are able to return to sleep on their own [36]. Parents clearly do not observe all awakenings and parental observation of sleep disturbances is not as accurate as polysomnographic recordings; however, in a clinical setting involving possible side effects of medication, the parents' perceptions of alterations in sleep patterns are critical because any perceived side-effects influence compliance or may contribute to requests that medication be changed. For the group of children assigned to phenobarbital treatment, we did not observe an effect on frequency of lengthy awakenings, presleep fussiness, or total sleep time. We did observe more frequent nighttime sleep disturbances in the first study year in those children assigned to phenobarbital who had more frequent sleep disturbances at study entry. In this double-blind, placebo-controlled study, phenobarbital was associated with more frequent disruptions of nighttime sleep only in predisposed children. Physicians who consider prescribing this medication for young children may want to ask questions regarding sleep problems before determining treatment.

We gratefully acknowledge the assistance of Martha Erickson, Mary Voeller, RN, Lee Adelman, RN, and Kay Frey, RN of the University of Washington, and Mary Livingston, Gail Carter, and Jack Panossian of the National Institutes of Neurological Disorders and Stroke.

References

[1] Wolf SM, Forsythe A. Behavior disturbance, phenobarbital, and febrile seizures. Pediatrics 1978;61:728-31. [2] Lee K, Melchior JC. Sodium valproate versus phenobarbital in the prophylactic treatment of febrile convulsions in childhood. Eur J Pediatr 1981;137:151-3. [3] Camfield CS, Chaplin S, Doyle AB, Shapiro SH, Cummings C, Camfield PR. Side effects of phenobarbital in toddlers: Behavioral and cognitive aspects. J Pediatr 1979;95:361-5. [4] Thorn I. A controlled study of prophylactic long-term treatment of febrile convulsions with phenobarbital. Acta Neurol Scand 1975; 60(Suppl):67-73. [5] Heckmatt JZ, Houston AB. Failure of phenobarbitone to prevent febrile convulsions. Br Med J 1976;1:559-61. [6] Moore T, Ucko LE. Night waking in early infancy: Part 1. Arch Dis Child 1957;32:333-42.

Hirtz et al: Phenobarbital and Febrile Seizures 99

[7] Ferher R. Sleep, sleeplessness, and sleep disruptions m infants and young children. Ann Clin Res 1985;17:227-34. [8] Johnson MC. Infant and toddler sleep: A telephone survey of parents in one community. J Dev Behav Pediatr 1991;12:108-14. [9] Crowell J, Keener M, Ginsburg N, Anders T. Sleep habits in toddlers 18 to 36 months old. J Am Acad Child Adolesc Psychiatry 1987;26:510-5. [10] Jones NB, Rossetti Ferreira MC, Farquar M, MacDonald BL. The association between perinatal factors and later night waking. Dev Med Child Neurol 1978;20:427-34. [11] Bernal JE Night waking in infants during the first 14 months. Dev Med Child Neurol 1973;15:760-9. [12] Riehman N. A community survey of characteristics of one- to two-year-olds with sleep disruptions. J Am Acad Child Psychiatry 1981 ; 20:281-91. [13] Carey WB. Night waking and temperament in infancy. J Pediatr 1974;84: 756-8. [14] Richman N. Sleep problems in young children. Arch Dis Child 1981 ;56:491-3. [15] Keener MA, Zeanah CH, Anders TE Infant temperament, sleep organization, and nighttime parental interventions. Pediatrics 1988;81: 762-71. [16] Kataria S, Swanson MS, Trevathan GE. Persistence of sleep disturbances in preschool children. J Pediatr 1987;110:642-6. [17] Elias M E Nicolson NA, Bora C, Johnston J. Sleep/wake patterns of breast-fed infants in the first 2 years of life. Pediatrics 1986;77: 322-9. [18] Eaton-Evans J, Dugdale AE. Sleep patterns of infants in the first year of life. Arch Dis Child 1988;63:647-9. [19] Carey WB. Letter: Breast feeding and night waking. J Pediatr 1975;87: 327. [20] W r i g h t E Macleod HA, Cooper MJ. Waking at night: The effect of early feeding experience. Child Care Health Dev 1983;9:309-19. [21] Van Tassel EB. The relative influence of child and environmental characteristics on sleep disturbances in the first and second years of life. J Dev Behav Pediatr 1985;6:81-6. [22] K a r a c a n I, Orr W, Roth T, et al. Dose-related effects of phenobarbitone on human sleep-waking patterns. Br J Clin Pharmacol 1981; 12:303-13.

100

PEDIATRIC NEUROLOGY

Vol. 9 No. 2

[23] Prinz PN, Vitiello MV, Roehrs TA, Linnoil~a M Weitzm:m tq) Effect of phenobarbital on sleep and nighttime plasma ~owth h~wmonc and cortisul levels. Can J Physiol Pharmacol 1981;59:II ~9 15. [24] Wolf P, Roder-Wanner UU. Brede M. lnfluencu t~t"therapeutic' phenobarbital and phenytoin medication on the polygraphic sleep ~f patients with epilepsy. Epilepsia 1984;25:467-75 [25] Barzaghi N. Gatti G. Manni R, et al. Time-dependent pharu~acodynamic effects of phenobarbital in humans. Ther Drug Moni~ 1~)89: I 1:661-6. [26] Hellstrom B, Barlach-Christoffersen M. hffluence ol phenobarbital on the psychomotor development and behaviour in preschool children with convulsions. Neuropediatrics 1980;11:151-60. [27] Farwell JR, Lee YJ, Hirtz DG, Sulzbacher SI, Ellenberg JH, Nelson KB. Phenobarbital for febrile seizures- Effects on intelligence and on seizure recurrence. N Engl J Med 19¢X):322:364-9+ [28] Grizzle JE, Starmer CE Koch GG. Analysis of categorical data by linear models. Biometrics 1969;25:489-504. [29] SAS Institute. Statistics. In: SAS Users Guide. Cary: SAS institute, 1982;257-85. [30] lbid p.139-99. [31] Searle SR. Linear models. New York: John Wiley, 1971:340-61. [32] Mantel N. Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer lnst 1959;22: 719-48. [33] Hoilingshead AB, Redlich FC. Social class and mental illness: A community study. New York: John Wiley, 1958;398-407. [34] Richman N. Surveys of sleep disorders in children in a general population. In: Guilleminault C, ed. Sleep and its disorders in children. New York: Raven Press, 1987;115-27. [35] Brent DA, Cmmrine PK, Varma R, Brown RV. Allan MJ. Phenobarbital treatment and major depressive disorder in children with epilepsy: A naturalistic follow-up. Pediatrics 1990;85:1086-91. [36] Anders TE Halpern LF, Hua J. Sleeping through the night: A developmental perspective. Pediatrics 1992;90:554-60.