The Prevalence and Characteristic of Sleep Problems among Children with Fetal Alcohol Spectrum Disorders (FASD).

Fetal alcohol spectrum disorders (FASD) is a group of conditions resulting from prenatal alcohol exposure (PAE). Patients with FASD experience a variety of neuropsychological symptoms resulting from central nervous system impairment. Little is known about sleep disorders associated with PAE. The objective of this study was to investigate the prevalence and characterize sleep problems among individuals with FASD.


Background
Fetal alcohol spectrum disorders (FASD) is an umbrella term covering different diagnoses resulting from prenatal alcohol exposure (PAE). With the estimated worldwide prevalence of 22.77 (0 to 176.77)(1) FASD is one of the most common neurodevelopmental conditions. The principal clinical characteristic of FASD is a brain impairment associated with alcohol neurotoxicity. (2) The variety of brain functions or domains can be affected including executive functions, cognition, social skills, abstract reasoning, attention de cit/hyperactivity; adaptive behavior, social skills, psychomotor coordination (3) etc. yet each individual with FASD presents a different clinical picture. Many patients complain not only of neurobehavioral issues but also of disorders of urination(4), defecation (5), nutrition (6) and sleep, also known as "neglected problems". It is now well established from a variety of studies, that sleep disorders are much more common among children with neurodevelopmental conditions.(7) Sleep disorders affect daily performance of children and can affect the diagnostic process creating a complex clinical picture. (8) Clinical experience, anecdotal reports and qualitative studies (9,10) suggest that sleep disorders are a major complaint among individuals with FASD. Although several quantitative studies have been carried out on this matter, there is still very little scienti c understanding of the nature of sleep disorders in FASD.
An objective of this study was to determine the prevalence of sleep disorders and characterize what sleep behaviors and sleep aspects are altered the most among persons with FASD.

Material And Methods
The study was conducted as a prospective study in the years 2018-2020.

Subjects
The study group consisted of 40 patients of FASD Diagnostic Center in St. Louis Children Hospital. The inclusion criteria were: FASD diagnosis established by a multidisciplinary team according to Hoyme 2016 criteria (11), age 3-18 at the moment of the enrollment. The exclusion criteria were: comorbid pulmonary disease causing a distorted sleep (severe asthma, cystic brosis), another birth defect, a known seizure disorder. The control group was selected from patients of the Gastroenterology Department of St. Louis Children Hospital diagnosed with a small intestine bacterial overgrowth (SIBO) and successfully treated.
The participation was offered during a follow-up visit, after the patients were asymptomatic regarding abdominal pain for at least 3 months. The control group consisted of 18 males and 22 females, the median age was 10 (8; 13). The exclusion criteria for the control group were: con rmed or unknown prenatal alcohol exposure, microcephaly, height/weight below 10th percentile, presence of facial dysmorphic features characteristic for FASD (smooth philtrum, narrow upper lip, short palpebral ssure), comorbid pulmonary disease causing distorted sleep (severe asthma, cystic brosis), comorbid neurologic or neurodevelopmental condition causing a distorted sleep (f.e. ADHD, spinal muscular atrophy). From each group, 40 consecutive patients eligible for the study were offered the participation. No patient met exclusion criteria. 22 patients from the study group and 7 from the control group were eligible for the second phase of the study, however, 3 patients from the study group and 2 from the control group discontinued the participation due to a caregiver's refusal (Fig. 1). Therefore, to compare the results of PSG, a control group from the retrospective data from the laboratory was formed (n = 105).
The group was composed of the patients who had the PSG performed among typically developing children, mainly due to the request of their caregivers or primary care physicians. The inclusion criteria consist of: lack of comorbidities, genetic syndromes and negative result of polysomnography in the term of sleep disordered breathing (a number of breathing events per hour of sleep less than one). The group was treated as a population sample representative for the laboratory and reference intervals for sleep related parameters were calculated after removal of 10% extreme values.

Methods
The study consisted of two phases. A written consent was obtained by the researchers from the parents and children above 12 years of age. Children younger than 12 were informed about the study and all their questions were answered by a researcher. Parents/caregivers of the patients eligible for participation in the study were asked by an independent researcher to ll in the Child's Sleep Habits Questionnaire (CSHQ) -an abbreviated form (Polish version). (12) The questionnaire is an established tool for screening pediatric sleep problems. It consists of 33 questions covering the following aspects of sleep: bedtime resistance, sleep onset delay, sleep duration, sleep anxiety, night wakings, parasomnias, sleep distorted breathing and daytime sleepiness. Based on the answers of the caregivers a total sleep disturbance score was calculated. In consistency with the previous publications(12) the cutoff was established at 41 points, and patients whose score was above the cutoff were asked to participate in the second phase of the study.
In the second phase of the study in-lab attended polysomnography (PSG) was performed (Alice 6 Philips Respironics) according to AASM recommendations (13) The measurements consisted of electroencephalographic channels (O1/A2, O2/A1, C3/A2, C4/A1, F3/A2, F4/A1), left and right electrooculograms, chin electromyogram, left and right tibialis electromyogram, electrocardiogram, ventilatory monitoring using thermal and pressure sensors, breathing effort monitoring using thoracic and abdominal belts, pulse oximetry, snoring and body position monitoring. Hooking-up of the subjects was performed directly before the start of the measurement, usually between 7-9 p.m. The subjects were woken up around 6 a.m. All recordings were manually scored, sleep staging and scoring breathing events were performed using pediatric criteria.(13) Sleep-stage percentages were calculated on the basis of the total sleep time (TST) and for wake periods as percentage of time in bed (TIB) -time measured from lights out to lights on. Sleep e ciency was calculated as the sleep stages percentage of TIB. Sleep latency was de ned as a time from lights-off to the beginning of the rst epoch considered as sleep (N1, N2, N3 REM). Stage shift means a change from one sleep stage to another. Arousals and breathing events were presented as common indices -the number of events per hour of sleep. Respiratory disturbance index (RDI) calculated as the number of respiratory events (apneas all types, hypopneas, RERA's) per hour of sleep. In the case of our study RDI effectively equals to apnea hypopnea index due to the lack of RERA's observations in all three examined groups.

Statistical analysis
The number of patients and percentage of the respective group were reported for the categories. Contingency tables were analyzed with Pearson chi-squared test. Age-adjusted logistic regression was used to con rm the difference in CSHQ results between FASD patients and controls. Continuous variables were exposed as median and lower and upper quartiles because not all the data were normally distributed. For comparison with previous sleep studies, selected parameters were exposed as a mean and standard deviation. Wilcoxon (Mann-Whitney) rank sum test was used to examine differences between two groups, for a greater number of groups, Kruskal-Wallis test was applied. The relation between continuous variables were examined using a linear regression model. Calculations were performed using (14)(Microsoft R Open 3.5.3) environment for statistical computing and Statistica 12.0 (StatSoft, Tulsa, OK, USA).

Sample characteristics
The median age of the FASD group was 8 years (6; 11), the group consisted of 19 males and 21 females (Table 1). Twelve percent of the patients were with at least one of the biological parents, 35% were in a foster care, whilst 52% were in adoptive families. Within FASD, the most common diagnosis was ARND (42%), followed by pFAS (32%) and FAS (25%). When compared with the control group, patients from the FASD group had an equal sex distribution and body mass index (BMI) but were slightly younger (Table 1).

CSHQ
The number of participants with sleep disturbances, i.e. total sleep disturbance score above the cutoff, was markedly higher in the study group as compared to controls (55% vs. 20%, Table 2). The ageadjusted odds ratio for a positive result in CSHQ (i.e. total sleep disturbance score > 41) was 4.31 (95% con dence interval: 1.54-12.11; p = 0.005) for FASD patients as compared to the control group. BMI was not associated with the total sleep disturbance score > 41 among children with FASD (p = 0.09 after age adjustment).There were no signi cant differences in CSHQ results (subscales and total sleep disorder score) between children diagnosed with FAS, pFAS and ARND, although those with pFAS tended to show the highest scores ( Fig. 2A). Moreover, the diagnosis was not signi cantly associated with the percentage of positive CSHQ results [4 (40%) of patients with FAS, 9 (69%) of patients with pFAS, and 9 (53%) of patients with ARND; p = 0.4].There were no signi cant differences in CSHQ results between FASD children depending on the type of a family, however, those in biological families tended to show the lowest scores (Fig. 2B). The type of a family was not signi cantly associated with the positive result in CSHQ among patients with FASD [2 (40%) of children in biological families, 9 (64%) in foster families, and 11 (52%) in adoptive families; p = 0.6]. Signi cant differences between the study and control groups were observed in the following subscales: sleep onset delay, night wakings, parasomnias, sleep disordered breathing, and daytime sleepiness ( Table 2). Children from the study group had a distorted sleep architecture in comparison with the patients from the laboratory control group representing a population sample ( Table 3). The number of stage shifts was greater, the proportion of time spent in the N1 stage of sleep and proportion of summary time of N3 and REM was signi cantly higher among the patients with FASD than among the patients from the laboratory control group (Table 3)). Children from the FASD group and children from the control group experienced more arousals during the sleep than the children from the laboratory control group (Table 3). There were no signi cant differences in comparison of respiratory indexes between FASD and control groups although indices in both groups were higher than previously published data for a normal population (Table 4). There was no relation between the arousals index and RDI in the FASD group (R 2 < 0.01, p = 0.73). The similar lack of relationship was observed between the obstructive apnea index and central apnea index (R 2 < 0.18, p = 0.07) and between the hypopnea index and central apnea index (R 2 = 0.14, p = 0.62) in FASD group.    (12,17), quali ed only the children who scored above 41. However, 7 children from the FASD group (17.5%) and 7 children from the control group (17.5%) had a Total score equal with 41. The subscales revealed that sleep onset delay, night wakings, parasomnias, sleep disordered breathing and daytime sleepiness occur more frequently among individuals with FASD. The trend, although without statistical signi cance, was also observed regarding sleep anxiety and bedtime resistance. Scher et al (19)(20)(21) demonstrated that infants with prenatal alcohol exposure tend to present fragmented sleep and experience more arousals, however, these studies were performed with a different methodology.
A similar nding was documented by Volgin et al. on an animal model of PAE. (22) In our study sleep in FASD subjects was less stable than in the control group which is mainly expressed in a greater number of stage shifts, an increased amount of stage N1 and a higher number of arousals. The lack of relation between arousals and breathing events suggests that the subset of arousals has other than a respiratory cause. Interestingly, in comparison to general population samples (23,24) (Table 4) the results are not consistent. Montgomery Downs reported the amount of N1 sleep higher than in our subjects but Goodwin reported it approximately at the same level. The amount of N3 is comparable to the Montgomery-Downs sample but much higher than in the Goodwin sample. The percentage of REM sleep is the lowest in our FASD group in comparison to both studies. The amount of this stage is also lowered in our control groups, so the proportion of laboratory related and FASD related factors is di cult to be examined.
However, insu ciency of REM sleep is one of the major causes of neurocognitive problems. The number of arousals is lower in comparison to both studies.
Both examined groups (FASD and control) had much more detected breathing events than reported in general population samples. It is clearly visible in all compared indices (Table 4), however for some of them (hypopnea index) it could be a combined result of our groups characteristics and differences in the scoring criteria. The changes in the hypopnea de nition which were introduced in version 2 of AASM manual lead to an increased number of detected events. According to the authors' knowledge there is no information about how those changes affect hypopnea index results in children population but the hypopnea index in our study is larger by an order of magnitude which is more than a potential impact of de nition changes. The central apnea index in the FASD group is greater by factor of two, even though current children central apnea de nition is more restrictive.  (27,28) This is the rst study to comprehensively assess the sleep problems among children with FASD in a relatively big sample with a control group. Due to organizational reasons the enrollment to the both groups was performed simultaneously which resulted in the negligible age difference between the groups. However, patients from both groups were within the age frame for CSHQ (3-18 years) and all differences observed between the groups were independent of age. CSHQ is a subjective sleep problems assessment which might be considered a limitation of this study. Only the patients screened by CSHQ were offered the objective sleep evaluation. The majority of parents in the study group were either foster or adoptive parents yet, the patients from the control group were in biological families. It is well established that foster and adoptive parents of FASD children experience a lot of distress and they present a tendency to an overprotective attitude(29) which can be a potential source of bias in the selfreport. Moreover, the 3 patients from the study group and 3 from the control group withdrew their consent and discontinued the study after the invitation to the second phase of the study. However, even with this limitation many patients completed the PSG and the results can be considered representative.
Notwithstanding the limitations, the study offers the overview of the frequency and nature of sleep problems among individuals with FASD. Further experimental investigations are needed to determine the mechanisms that contribute to this phenomenon as well as the treatment options. From the clinical perspective it is important for the physicians and psychologist taking care for patients with FASD to include the question about the sleep quality in the history taking and try to address this issue.

Conclusion
In conclusion, individuals with FASD have an increased risk of sleep disorders among which distorted sleep architecture and apneic/hypopneic events are the most common. The study was approved by the regional ethical committee by Regional Board of Physicians (approval number: 123/kbl/oil/2018). The study complies with the Declaration of Helsinki.

Consent for publication
Not applicable

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Total sleep disturbance score in the studied children with FASD according to diagnosis in comparison with healthy controls (A) and according to the type of a family (B). There were no signi cant differences between FAS, pFAS and ARND patients (p=0.4) and between FASD patients in biological, foster or adoptive families (p=0.6). The data are shown as median (central line), interquartile range (box) and raw data (points).