Public Health Significance
Craniosynostosis refers to premature closure of the cranial sutures during the perinatal period. Isolated single-suture craniosynostosis (SSC) is a non-syndromic form impacting only one suture. There is evidence that children with craniofacial anomalies have a higher risk for SDB as compared with unaffected children; however, it is not known whether this also applies to isolated SSC.Sleep-disordered breathing (SDB) occurs in 0.7% to 13.0% of childrenand is associated with behavioral, neurocognitive and cardiovascular consequences.We compared parent-reported SDB-related symptoms in children with and without isolated SSC, hypothesizing that those with SSC would have increased SDB-related symptoms as compared with controls. This preliminary study may have implications for clinical care by identifying a treatable condition (i.e. SDB) in an already vulnerable pediatric population.
Method
Among children with and without SSC, 6 SDB symptoms were assessed by parent report: restless sleep, arousal, snoring, snorting/gasping, apneas, and daytime sleepiness. Current sleep problems (yes/no) and ever sleep problems (yes/no) were identified based on the constellation of SDB-related outcomes. Snoring, the most common symptom of SDB, was categorized as never, 0-1 nights/week, and 2+nights/week. Logistic and multinomial logistic regression models were applied to assess the association between sleep outcomes and SSC, adjusted for sex, age, race/ethnicity and family socioeconomic status (SES).
Results
Current sleep problems were reported slightly more often in children with SSC (19%) than in controls (14%; adjusted odds ratio (aOR)=1.6, 95% CI 0.9-2.8). There was no difference in report of having ever had sleep problems between cases and controls. In the multinomial regression model, the overall association of SSC and SDB was statistically significant (p=0.02). Snoring was more often reported as 2+nights/weekby parents of SSC cases (13%) than controls (4%) (versus never, aOR=3.4, 95% CI=1.4-7.9); whilesnoring 0-1 nights/week was reported similarly inboth groups.
Conclusion
Children with SSC had significantly increased presence of snoring during sleep compared to controls. This suggests that children with isolated SSC may be at increased risk for SDB and further suggests that awareness and clinical evaluation of SDB are important for children with SSC. Further study using standardized assessments of SDB are needed.
TABLE OF CONTENTS
1.0Introduction
1.1CRANIOSYNOSTOSIS
1.2Sleep-disordered breathing
1.3Sleep Disordered Breathing and Craniosynostosis
1.4PRESENT STUDY
2.0METHODS
2.1participants
2.2Data Collection
2.3Statistical analyses
3.0results
4.0discussion
bibliography
List of tables
Table 1. Sociodemographic Characteristics of SSC Cases and Controls at Follow-Up
Table 2. Results of Logistic Regression Comparing the Odds of Current and Ever Sleep Problems in SSC Cases versus Controls
Table 3. Results of Multinomial Logistic Regression to Assess the Association between Snoring and SSC
Table 4. Comparison of Fit Statistics for Model Selection in Latent Class Analysis
Table 5. Latent Classes Identified for Participants with SSC Cases and Controls
List of figures
Figure 1. Current / Ever Sleep Problem across Suture Involvement
Figure 2. Snoring across Suture Involvement
1
1.0 Introduction
1.1CRANIOSYNOSTOSIS
Craniosynostosisis a craniofacial anomaly defined as premature fusion of cranial sutures (e.g. sagittal, coronal,metopic, and/or lambdoidal) during the perinatal period, affectingthe growth of the face and head. It is usually diagnosed by computed tomography (CT) scans and three-dimensional reconstruction, along with medical history evaluation and careful physical examination.
Craniosynostosis is classified as syndromicor non-syndromic/isolated. Syndromiccraniosynostosis manifests as part of a congenital syndrome such as Crouzon, Apert, Pfeiffer, and Saethre-Chotzen syndromes, which may include additional abnormal development and appearance ofthe jaw, lip/palate, eye and/or trunk.1,2Some known causal genetic mutations are associated with syndromic craniosynostosis, including FGFR-1, FGFR-2, FGFR-3, TWIST-1 and EFNB-1, though not always identified.1,3In comparison, non-syndromic/isolatedcraniosynostosis typically occurs without other congenital anomalies.4Meanwhile, craniosynostosis is classified as single-suture or multiple-suture, depending on the number of cranial sutures involved.Isolated craniosynostosis may impact only one suture5, in which case it is called isolated single-suture craniosynostosis (SSC). Among sutures, sagittal is the most commonly impacted in SSC6. Although estimates vary in different populations, prevalence of all types of craniosynostosis ranges from 3 to 14 per 10,000 live births, and single-suture craniosynostosis is present in approximately 5 in 10,000 live births7,8.
Craniosynostosis does not only result in abnormal appearance of the head, but also causes functional impairment.Several studies indicated that SSC was associated with neurocognitive, psychological and behavioral problems9,10. Surgical treatment, usually a single cranioplasty, is commonly performed for children with isolated SSC, before 1 year of age to release the fused suture11.Ideally, both children using surgical or non-surgical treatment should be managed by a multidisciplinary team4.
1.2Sleep-disordered breathing
Sleep-disordered breathing (SDB)refers to several chronic conditions categorized by partial or complete breathing obstruction or cessation during sleep, which affects functional ability and quality of life12,13,14. Studies indicate that childhood SDB is not only associated with behavioral and neurocognitive problems, such as decreased attention, learning difficulties, and anxious/depressive symptoms, but also long-term cardiovascular consequences15,16,17,18.
There is a wide range of the estimated prevalence of SDB in children, rangingfrom 0.7% to 13.0%19depending on diagnostic method.Polysomnography is the gold standard to measure and diagnose sleep disorders objectively. However, polysomnography is not feasible to be used to identify sleep problems in large pediatric epidemiological studies, because of its high price andtime burden forparticipants. Thus, validated parent-report questionnaires, such as the Pediatric Sleep Questionnaire20 andSleep Disturbance Scale for Children21,are commonly employed to identify children with suspected SDB22.
Diverse manifestations are present for children with SDB, such as nighttime snoring, daytime sleepiness, restless sleep, frequent arousals,snorting/gasping, witness apneas, dry mouth/mouth breathing23. The constellation of symptoms change with age; yet nighttime snoring is regarded as the most common presentation of SDB at any age among children.
The causes of SDB are multifactorial. One main reason is the narrowairway structure due toenlarged adenoids or/and tonsils,overweight, or insufficient maxillary growth, which may obstruct air flow during sleep24,25. Meanwhile, neuromuscular activation and inflammation can also lead to the development of SDB23. Compared to the general population, individuals withcraniofacial and neuromuscular anomalies have an increased risk of SDB23.
For children, the most common treatment of SDB is adenoidectomy and tonsillectomy. After adenoidectomy and tonsillectomy, children should be followed up using polysomnography. If SDB is not sufficiently treated and residual SDB occurs, other orthodontic treatment, distractor orexpansion can be applied according to individual’s situation26,27.
1.3Sleep Disordered Breathing and Craniosynostosis
There is evidence that children with craniofacial anomalies have a higher risk for SDB as compared with unaffected children; however, the literature is sparse, and focuses on cleft lip/palate, craniofacial microsomia, and syndromic craniosynostosis28,29,30,31,32. Few published studies have examined sleep outcomes in children with either non-syndromic craniosynostosis or isolated SSC. One previous study indicated that both children with syndromic and non-syndromic craniosynostosis had increased risk of developing of SDB33. However, this study is limited by the small sample size (4 with syndromic, and 10 with non-syndromic craniosynostosis), no comparison group, and referral bias. To our knowledge, there are no studies comparing sleep outcomes in children with SSC as compared with unaffected children and thus the association between SSC and SDB is yet to be clarified.
1.4PRESENT STUDY
The current study aimed to assess whether children with isolated SSC are at higher risk for SDB as compared with controls. We hypothesized that those with SSC would have increased SDB-related symptoms as compared with controls. We compared parent-reported SDB-related symptoms in children with and without isolated SSC ina multi-center cohort.
2.0 METHODS
2.1participants
Thiscross-sectional study includes 184 children with isolatedSSC and 184 unaffected controls. These children areparticipants in an ongoing, multisite,follow-up study of neurodevelopment in children with and without isolated SSC34,35. When initially enrolled in the original study, cases were infants diagnosed (by computed tomographic scans)with one of the following types of isolated SSC: sagittal, metopic, right unicoronal, left unicoronal and lambdoid.Controls were healthy infants without congenital anomalies. Cases and controls were frequency matched by age, sex, race/ethnicity,and family socioeconomic status (SES). Exclusion criteria for the originalenrollment included age 30 months, gestational age <34 weeks, major medical comorbidities(e.g. cardiac defects, seizure disorders), co-occurrence of major malformations, or >3 minor congenital anomalies36.Informed consent was obtained before enrollment with approval of institutional review boards (IRB) of each four site (Seattle, Chicago, St. Louis and Atlanta). Participation in the current cross-sectional analysiswas limited to those who enrolled in the original study and completed the follow-up visit.
2.2Data Collection
SDB-related outcomes, sleep-related medical history and demographic information were obtained by staff-administered interview. First, parents of all participants were asked whether their child had sleep problems and whether these problems were new, continuing or reported at the previous study visit. For those who reported new or continuing sleep problems, 6 supplemental questions related to SDB symptoms were further assessed, including restless sleep, arousal, snoring, snorting/gasping, apneas, and daytime sleepiness.Choices for each response were “usually: 5-7 times per week”, “sometimes: 2-4 times per week”, “rarely: 0-1 time per week”, and “never”.Demographic information, such as sex, date of birth, race/ethnicity, site and family SES, wascollected at baseline and updated by parentreport at follow-up. Additionally, sleep-related medical history regarding airway intervention (tonsillectomy and adenoidectomy) and respiratory diseases, including asthma, allergies,and respiratory infections (e.g., bronchitis, bronchiolitis, and pneumonia) was obtained duringthe interview.
2.3Statisticalanalyses
Descriptive analyses were used to illustrate the sociodemographic characteristics of children with and without SSC, including age, sex, race/ethnicity, family SESand site of data collection. Among the children with SSC, type of suture involvement (e.g., sagittal, metopic, right and left unicoronal, lambdoid), and sociodemographic characteristics bysuture involvementwere assessed as well. Counts and percentages were calculated for discrete variables. Means, standard deviations, medians and ranges were calculated for continuous variables.
Logistic regression and multinomial logistic regressionwere performed to explore whether children with SSC have increased risk to have sleep problems than unaffected children. For the purpose of analyses, current sleep problems (yes/no) were identified based on endorsement of ‘having new or continuing sleep problems’, or endorsement of at least one SDB symptom. Meanwhile, we classified participants as having ever sleep problems (yes/no) if the child had current sleep problems or previously reported sleep problems. Besides the above two outcome variables, we also separately assessed whether snoring was associated with SSC. Participants with “sometimes” and “usually” responseswere combined into one group to avoid the issue of small cell sizes; therefore,snoring was recoded as never, 0-1 night/week, and 2+nights/week. To compare the odds of sleep problems between case and control groups, logistic regression was conducted for binary variables (current sleep problem and ever sleep problem).Multinomial logistic regression model was used for multi-level outcome variable (snoring) with “Never” as the reference group.For all adjusted regression models, potential confounding variables (sex, race/ethnicity, age, and SES)were included as covariates. Odds ratios and their 95% confidence intervals (CIs) were computed. P-values less than 0.05 were regarded as statistically significant.
To identify whether reported snoring frequency was distributed similarly among cases with different cranial suture involvement (sagittal, metopic, right and left unicoronal, lambdoid), proportions of snoring frequency responseswerecompared across suture type using Fisher’s exact test, taking into account that expected counts in some cells were less than 5.
Latent class analysis37wasemployed to explore the underlying structure of the sleep patterns among children with and without isolated SSC based on the 6reported SDB-related items.To determine the number of latent classes, we fitted a model with two classes and repeated with one additional class iteratively to comparethe fit statistics, including the likelihood-ratio G2 statistic, Akaike’s Information Criterion (AIC) and Bayesian Information Criterion (BIC). The optimallatent class model and the composition of each class were described.
Sensitivity Analyses
To assess potential bias introduced by specific conditions, sensitivity analyses were conducted by excluding children who had probable and known causal mutations, suffered from respiratory diseases (e.g., asthma, allergy,respiratory infections), and had ever undergone airway interventions (tonsillectomy and adenoidectomy), respectively.We repeated the main regression analysesin the subsamplessuccessively, and then compared the results obtained in the subsamples and in the full sample.
All analyses above were performed using SAS 9.4.PROC LCA developed by the Pennsylvania State University was used for latent class analysis.
3.0 results
Participants included 184 children with isolated SSC and 184 unaffected controls, with the mean age of 7.5 and 7.4 years, respectively. (Table 1) As the SSC case and control groups were frequency matched bysex, race/ethnicity and family SES, these characteristics were distributed similarly in the two groups. In both SSC cases and controls, of the majority were male (64%, 63%)and white (80%, 74%). Also, most SSC cases and controls had high family SES, classified as I or II based on Hollingshead Classification (71%, 85%).
Within case group, sagittal synostosis (42%)wasthe mostcommon form,whereas lambdoid (7%) wasthe least common. (Table1)Sociodemographic characteristicsdiffered by suture type:there were more femalesamongchildren with left and right unicoronal SSC,butmore malesfor other types of SSC. Race/ethnicity and age were similaracross different suture types. The composition of family SES were overall similar, thoughthe proportion of children with high family SES (I or II) in left unicoronal group (55%) was lower than other suture types. No children with left unicoronal were enrolled in St. Louis and Atlanta.
Table 1. Sociodemographic Characteristics of SSC Cases and Controls at Follow-Up
TotalControls
N (%)++ / Total
Cases
N (%) / Sagittal
N (%) / Metopic
N (%) / Right
Unicoronal
N (%) / Left
Unicoronal
N (%) / Lambdoid
N (%)
Number / 184 / 184 / 77 (42) / 49 (27) / 24 (13) / 22 (12) / 12 (7)
Age
Mean(SD) / 7.4 (0.5) / 7.5 (0.6) / 7.5 (0.7) / 7.4 (0.3) / 7.4 (0.3) / 7.4 (0.3) / 8.1 (0.8)
Range / 7.0: 11.1 / 6.9: 11.5 / 7.0: 11.5 / 7.0: 8.4 / 7.0: 8.5 / 6.9: 8.1 / 7.2: 9.3
Sex
Female / 69 (38) / 67 (36) / 18 (23) / 14 (29) / 13 (54) / 17 (77) / 5 (42)
Male / 115 (63) / 117 (64) / 59 (77) / 35 (71) / 11 (46) / 5 (23) / 7 (58)
Race/Ethnicity
White / 137 (74) / 148 (80) / 64 (83) / 38 (78) / 20 (83) / 16 (73) / 10 (83)
Non-White / 47 (26) / 36 (20) / 13 (17) / 11 (22) / 4 (17) / 6 (27) / 2 (17)
SES+
I:55-60 (high) / 54 (29) / 44 (24) / 21 (27) / 11 (22) / 3 (13) / 5 (23) / 4 (33)
II:40-54 / 103 (56) / 86 (47) / 33 (43) / 24 (49) / 17 (71) / 7 (32) / 5 (42)
III:30-39 / 14 (8) / 32 (17) / 13 (17) / 9 (18) / 3 (13) / 4 (18) / 3 (25)
IV:20-29 / 10 (5) / 20 (11) / 10 (13) / 4 (8) / 1 (4) / 5 (23) / 0 (0)
V:8-19 (low) / 3 (2) / 2 (1) / 0 (0) / 1 (2) / 0 (0) / 1 (5) / 0 (0)
Site
Seattle / 72 (39) / 76 (41) / 34 (44) / 20 (41) / 6 (25) / 12 (55) / 4 (33)
Chicago / 76 (41) / 66 (36) / 16 (21) / 21 (43) / 14 (58) / 10 (45) / 5 (42)
St. Louis / 9 (5) / 17 (9) / 12 (16) / 1 (2) / 3 (13) / 0 (0) / 1 (8)
Atlanta / 27 (15) / 25 (14) / 15 (19) / 7 (14) / 1 (4) / 0 (0) / 2 (17)
+SES, socioeconomic status, Hollingshead Four Factor Index of social status (Hollingshead, 1975)
++Percentages may not add up to 100% because of rounding.
Current / Ever Sleep Problems
Current sleep problems were reported slightly more often in children with SSC (19%) than in controls (14%; adjusted odds ratio (aOR)=1.6, 95% CI 0.9-2.8, p=0.13) (Table 2). There was no difference in report of having ever had sleep problems between cases (25%) and controls (23%; aOR=1.1, p=0.59).
Table 2. Results of Logistic Regression Comparing the Odds of Current and Ever Sleep Problems in SSC Cases versus Controls
ControlN (%) / Case
N (%) / Crude
Odds Ratio
(95% CI) / Adjusted
Odds Ratio+
(95% CI) / p-value++
Current Sleep Problem / No / 159 (86) / 149 (81) / 1.5 (0.8, 2.5) / 1.6 (0.9, 2.8) / 0.13
Yes / 25 (14) / 34 (19)
Ever Sleep Problem / No / 142 (77) / 138 (75) / 1.1 (0.7, 1.8) / 1.1 (0.7, 1.9) / 0.59
Yes / 42 (23) / 45 (25)
+Adjusted for sex, age, race/ethnicity and family SES
++P-value based on the adjusted regression models
Snoring
The overall association of SSC and snoring was statistically significant (p=0.02) in the multinomial regression model. Snoring was more often reported as 2+ nights/week by parents of children with SSC (13%) than controls (4%) (aOR vs. never=3.4, 95% CI=1.4-7.9, p=0.01); Snoring was reported as 0-1 night/week in 5% and 8% of cases and controls, respectively(aOR vs. never=0.8, 95% CI=0.3-2.0, p=0.62). (Table 3)
Table 3. Results of Multinomial Logistic Regression to Assess the Association between Snoring and SSC
ControlN (%) / Case
N (%) / Crude
Odds Ratio
(95% CI) / Adjusted
Odds Ratio+
(95% CI) / p-value++
Snoring / Never / 161 (88) / 149(82) / Referent / Referent / -
0-1 nights/week / 14 (8) / 9(5) / 0.7 (0.3, 1.7) / 0.8 (0.3, 2) / 0.62
2+ nights/week / 8 (4) / 23(13) / 3.1(1.3, 7.2) / 3.4(1.4, 7.9) / 0.01
Note: P-value for the overall model based on Walt test was 0.02.
+Adjusted for sex, age, race/ethnicity and family SES
++P-value based on the adjusted regression models
Comparisonby Suture Involvement
Among children with isolated SSC, there were no statistically significant overall differences in the proportions of children with current sleep problems, ever sleep problems, and snoring frequency across suture involvement, and the p-values of the overall test (Fisher’s exact test) were 0.07, 0.39 and 0.52, respectively.However, it is notable that children with left unicoronal SSC had a higher proportion of sleep-related problems than other suture types. 41% of children with left unicoronal SSC were reported having current sleep problem and ever sleep problem, and 37% of children with left unicoronal SSC were reported having snoring symptoms. (Figures 1, 2)