Behavioral Adjustment of Preschool Children with and without Craniofacial Microsomia

Abstract

Objective:

The study aim was to assess behavioral adjustment in preschool children with and without craniofacial microsomia (CFM).

Design:

Multisite cohort study of preschoolers with CFM (“cases”) or without CFM (“controls”).

Participants:

Mothers (89%), fathers (9%), and other caregivers (2%) of 161 preschoolers.

Outcome Measure:

Child Behavior Check List (CBCL 1.5–5); linear regressions with standardized effect sizes (ES) adjusted for sociodemographic confounds.

Results:

CBCLs for 89 cases and 72 controls (average age 38.3±1.9 months). Children were male (54%), white (69%), and of Latino ethnicity (47%). Cases had microtia with mandibular hypoplasia (52%), microtia only (30%), or other CFM-associated features (18%). Nearly 20% of cases had extracranial anomalies. Composite CBCL scores were in the average range compared to test norms and similar for cases and controls. On the subscales, cases’ parents reported higher Anxious/Depressed scores (ES = 0.35, p = 0.04), Stress Problems (ES = 0.40, p = 0.04), Anxiety Problems (ES = 0.34, p = 0.04), and Autism Spectrum Problems (ES = 0.41, p = 0.02); however, the autism subscale primarily reflected speech concerns. Among cases, more problems were reported for children with extracranial anomalies and certain phenotypic categories with small effect sizes.

Conclusions:

Behavioral adjustment of preschoolers with CFM was comparable to peers. However, parental reports reflected greater concern for internalizing behaviors, thus anxiety screening and interventions may benefit children with CFM. Among cases, more problems were reported for those with more complex presentations of CFM. CFM-related speech problems should be distinguished from associated psychosocial symptoms during developmental evaluations.

Introduction

Craniofacial microsomia (CFM), also known as hemifacial microsomia (HFM), refers to a spectrum of congenital differences resulting from variation in the embryologic development of the first and second branchial arches (Jones, 2006). Characteristics include microtia (hypoplasia of the external ear) and aural atresia (absence of the external ear canal); facial asymmetry related to maxillary and/or mandibular hypoplasia; preauricular or facial tags; ocular and orbital anomalies; facial nerve paresis; and systemic findings, primarily vertebral, genitourinary, cardiovascular, and in the central nervous system (Poswillo, 1973; Lauritzen, 1985; Gorlin, 1990; Cohen et al., 2017). CFM occurs in approximately 1 in 3,500 to 5,600 live births (Barisic et al., 2014), with a higher prevalence of microtia among individuals of Hispanic/Latino and Native American descent (Canfield et al., 2009; Luquetti et al., 2011).

Although the CFM phenotype has been well characterized in terms of craniofacial morphology, its neurobehavioral correlates are less certain (Heike et al., 2014), particularly those related to behavioral and psychosocial adjustment. Some children with CFM would seem particularly susceptible to internalizing problems due to multiple risk factors, including speech and hearing differences (Hollody and Kampos, 2005; Collett et al., 2019; le Clercq et al., 2019) and atypical facial appearance that could elicit negative peer attention (Feragen and Stock, 2017). Cranial nerve anomalies that can restrict facial expressiveness and affective communication are also present in a subgroup of individuals with CFM (Cline et al., 2014; Manara et al., 2015). Some investigators have speculated that CFM, as well as other diagnoses sharing similar embryogenesis with cranial nerve dysfunction (e.g., Moebius syndrome), may be associated with increased risk for an autistic spectrum disorder (Johansson et al., 2007).

Empirical data relevant to these hypotheses are limited (Feragen and Stock, 2017), primarily because CFM is a rare condition and most studies of social-behavioral functioning have involved diagnostically heterogeneous samples of children with more prevalent craniofacial diagnoses (e.g., oral clefts). Diagnostic subgroups in this research typically have included some children with CFM, but usually not in sufficient numbers or at a level of analytic detail to allow for reliable conclusions (e.g., Snyder and Pope, 2010). For example, greater difficulties with social competence were reported for 11 children with CFM (Padwa et al., 1991). Alternatively, in a small sample of twins, investigators observed no meaningful differences in the social or academic competencies of six children with CFM compared to their unaffected twin (Maris et al, 1999).

Two studies have reported behavioral assessment data from a relatively large cohort of children with CFM and a demographically-comparable control group (Werler et al., 2004; Dufton et al., 2011; Wallace et al., 2018). When psychosocial status was first assessed around age 7, parents of children with and without CFM reported similar levels of behavior problems and social competence (Dufton et al., 2011). However, teachers reported more frequent internalizing behavior problems on average for children with CFM, along with lower social competence and less peer acceptance. These group differences were small for most variables. Among children with CFM, slightly worse behavioral outcomes were observed for girls, children born to mothers age 25 or younger, and those who had eye anomalies (e.g., orbital hypoplasia or displacement) and/or any of the less common features associated with the CFM spectrum (e.g., cleft lip/palate, renal, or cardiac anomalies).

Wallace et al. (2018) followed the same cohort into adolescence at an average age of 13 and repeated the earlier behavioral assessments, with the addition of adolescents’ self-reports of their own behavioral and social functioning. Little difference between youth with and without CFM was observed on measures of behavioral pathology (including items related to autistic behaviors), but small group differences were noted in various aspects of social functioning. For example, item analyses revealed that compared to controls, both parents and adolescents with CFM reported less frequent participation in social activities, fewer friends, less time with friends, more frequent teasing, and rejection by peers. Most of these group differences, though of relatively small magnitude, were consistently reported across parents, teachers, and adolescents. In contrast to findings at age 7 (Dufton et al., 2011), no associations between cases’ physical features and social-behavioral outcomes were found.

Other relevant psychosocial research includes studies of individuals with non-syndromic microtia. Studies completed in the U.S. and international studies show that children with microtia exhibit more behavior problems and social problems than same-aged peers (Jiamei et al., 2007; Li et al., 2010; Steffen et al., 2010; Kristiansen et al., 2013; Johns et al., 2017; Mandelbaum et al., 2017). Findings suggest that, while youth with CFM are not more likely than their peers to have specific psychological diagnoses, they may have elevated behavior problems, difficulties with socialization, and are likely to experience more frequent negative peer interactions (e.g., bullying, teasing). Based on this research, psychosocial support has been included in standards of care for patients with CFM and related conditions (e.g., the UK Care Standards for the Management of Patients with Microtia and Atresia; Henderson and Moffat, 2019).

It is important to note that nearly all previous studies focused on school-age or older children with CFM. The behavioral and social development of preschool or younger children with CFM has not been well described, which has precluded a prospective understanding of the possible precursors of later social functioning in this population. For example, it is unclear if there are indicators of socialization issues among preschoolers with CFM prior to their exposure to classroom peer groups at school entry. Some studies suggest that this is a distinct possibility. In a study of young children with microtia (Johns et al., 2017), teasing was reported to begin between the ages of 2 to 5 years and to increase with age, with all the children in this sample reported having been teased by age 6. Similar early teasing patterns were reported by Luquetti et al. (2018) in a larger sample of children with CFM.

The present study was designed to address existing research gaps by collecting parent reports of behavioral and social functioning at approximately age 3 within a cohort of children with CFM (“cases”) and children without a craniofacial condition (“controls”). The primary aims of this observational study were to determine whether preschoolers with CFM exhibit more frequent behavioral-social problems than demographically-similar peers as well as assess the association between psychosocial adjustment and phenotype and hearing status among cases. In secondary analyses, child sex and maternal age at birth were examined, which have previously been reported to moderate outcomes for school-age children with CFM (Collett et al., 2011; Dufton et al., 2011).

Methods

Participants

Infants between the ages of 12 and 24 months were recruited to participate in an observational, longitudinal, multi-center project called Craniofacial microsomia: Longitudinal Outcomes in Children pre-Kindergarten (CLOCK), which tracked the neurodevelopmental, speech and hearing outcomes, and phenotypic features of infants with and without CFM. Participants were enrolled between 2013 and 2017 from the following craniofacial centers: Children’s Hospital Los Angeles, Children’s Hospital of Philadelphia, Seattle Children’s Hospital, University of Illinois-Chicago (including Shriners Hospital for Children, Chicago), and University of North Carolina at Chapel Hill. Infants were enrolled and received a baseline developmental assessment at a mean age of 14 months (see Speltz et al., 2018 for details). The current study describes the caregiver reports of child behavior between the ages of 36 and 45.6 months, with an average age of 38.3 months. This research was approved by the Institutional Review Boards at all participating centers. All parents gave informed consent for their child to participate in the study.

Cases were recruited from each site’s craniofacial centers, hospital-based clinics seeing infants or young children with CFM (e.g. hearing screening programs, otolaryngology programs), and research study websites (e.g. clinicaltrials.gov). Details of the study methods have been previously published (Luquetti et al., 2019). To be eligible, cases had to meet at least one of the following inclusion criteria: (1) microtia or anotia; (2) facial asymmetry and preauricular tag; (3) facial asymmetry and facial tag; (4) facial asymmetry and epibulbar dermoid; (5) facial asymmetry and lateral oral cleft; (6) preauricular tag and epibulbar dermoid; (7) preauricular tag and lateral oral cleft; (8) facial tag and epibulbar dermoid; and/or (9) lateral oral cleft and epibulbar dermoid. The congenital anomalies could be either unilateral or bilateral. Infants were excluded if they had (1) an abnormal karyotype; (2) a diagnosis of a syndrome that involves microtia and/or underdevelopment of the jaw, such as Townes-Brocks, Treacher-Collins, branchiootorenal, Nager, or Miller syndrome; (3) a major medical or neurological condition (e.g., cancer, cerebral palsy) at time of recruitment; (4) increased risk for less than full compliance with study procedures and follow-ups (e.g., child in foster care or a ward of the state); (5) delivery at < 34 weeks estimated gestational age; (6) a sibling already participating in the CLOCK study; and/or (7) consenting parent spoke neither English nor Spanish. Of the 219 potentially eligible cases that were approached, 108 (49%) were enrolled at Time 1. The two most common reasons for non-participation were (1) declinations to participate by a parent or caregiver and (2) the study team was unable to contact the family (Luquetti et al., 2019). Eighty-nine children with CFM (82% of those enrolled) attended their Time 3 visit.

Eligible control participants were identified and frequency-matched to the case group based on infant age and sex, language spoken in the home (English or Spanish), and family socioeconomic status (SES). SES was calculated using the Hollingshead (1975) Four-Factor Index using a combination of parental marital/cohabitation status, highest education level, employment status, and occupational category, with continuous scores ranging from 8 to 66. Controls were recruited from local pediatric practices close to the hospitals that served case families and from existing infant registries when available (see Luquetti et al., 2019). Exclusion criteria for controls included: (1) meeting one or more of the exclusionary criteria for cases; and (2) diagnosis or history of any disorder, condition, or injury that would affect facial features (e.g., craniofacial malformation or deformation; facial surgery or trauma). Of the 148 potentially eligible controls that were approached, 84 (57%) were enrolled at Time 1 and 72 (85% of those enrolled at Time 1) participated at Time 3.

Measures

Child Behavior Checklist: Ages 1½ to 5 (CBCL).

Following a neurodevelopmental testing session (for details see Collett et al., 2019), parents completed the CBCL in their preferred language (English or Spanish). The CBCL has been used widely in pediatric samples and has strong psychometric properties (Achenbach and Rescorla, 2001; Ivanova et al., 2010; Pontoppidan, 2017). The CBCL consists of 99 items of child behaviors observed in the past two months with responses of 0 (not true), 1 (somewhat or sometimes true), or 2 (very true or often true). Composite scores on the CBCL are generated for Internalizing, Externalizing, and Total Problems. Subscale scores are generated within these composites. Internalizing subscales include: Emotionally Reactive, Anxious/Depressed, Somatic Complaints, and Withdrawn. Externalizing subscales include: Attention Problems and Aggressive Behavior. Additionally, subscale scores for Sleep Problems and Stress Problems are included in the Total score. The remaining subscales are based on DSM-related constructs, including: Depressive Problems, Anxiety Problems, Autism Spectrum Problems, Attention Deficit/Hyperactivity (ADHD) Problems, and Oppositional Defiant Problems. T-scores (mean = 50, SD = 10) were derived for scales based on separate norms for males and females from a nationally representative sample (Achenbach and Rescorla, 2001). Higher scores indicate higher frequencies of observed problem behaviors. For the composite scales, T-scores of 60 or higher are “borderline to clinically significant” and T-scores of 65 or higher on the subscales are in the “borderline to clinically significant” range. Raw scores were used in regression analyses for the subscales, following recommendations by Achenbach and Rescorla (2001).

Hearing status.

Cases’ hearing status was based on audiological data obtained as part of routine clinical care and phenotypic data, e.g., presence or absence of aural atresia, which is a reliable marker of conductive hearing loss (Mitchell et al., 2017). One of the investigators (SN), an audiologist with craniofacial expertise, reviewed the best available audiologic and phenotypic data to determine hearing status as: (1) “No hearing loss”, defined as the absence of atresia and a normal audiometric finding; (2) “Unilateral hearing loss”, which referred to single-sided hearing loss based on audiometry or the presence of unilateral atresia; or (3) “Bilateral hearing loss” meeting at least one of the following criteria: (a) presence of bilateral atresia, (b) audiological evidence of hearing loss in both ears, and/or (c) the presence of unilateral atresia plus audiological evidence of hearing loss in the contralateral ear.

Phenotype.

Phenotypic classifications were based on the integration of data collected from a health history interview with parent, medical record abstraction, and analysis of multiple standardized craniofacial photographs taken of each participant reviewed by two craniofacial experts (CH and DL; for details see Luquetti et al., 2019). In order to evaluate differences among children with CFM, three phenotypic subgroups were created based on shared clinical characteristics: 1) microtia only (in the absence of mandibular hypoplasia, epibulbar dermoids, lateral clefts, preauricular or facial tags, small and/or displaced orbit, nerve palsies); 2) microtia with mandibular hypoplasia; and 3) other combinations of CFM-associated malformations (two or more were required). Please see Heike et al. (2016) for more detailed information on the phenotypic assessment protocol and rationale for developing these subgroups.

Statistical Methods

The distributions of demographic characteristics and psychosocial adjustment scores were calculated for cases and controls separately, including the percent of children that were in the borderline to clinically significant range on the CBCL (T-score ≥ 60 for composite scales and T-score ≥ 65 for subscales). Logistic regression was used to estimate the odds (OR) of meeting borderline to clinically significant thresholds, with controls as the referent. Linear regression was used to compare children with and without CFM on continuous measures of behavioral adjustment. Standardized mean difference effect sizes were calculated using the sample distribution for each score. All analyses were adjusted for the child’s age at assessment (continuous), sex (male vs. female), SES (continuous), study site (categorical), race (white vs. non-white), and family language use (any Spanish vs. only English). The composite scales (Internalizing, Externalizing, and Total) were the primary outcomes, with subscale analyses used to explore any specific areas of behavioral concern. Additionally, for subscales significantly associated with case status, case-control differences were examined for specific behavioral items using chi-squared analyzes to compare the proportions of children whose parents reported that a behavior problem was “somewhat or sometimes true” or “very true or often true” versus “not true.” Linear regressions with controls as the referent category also examined differences in CBCL composite scores across the three phenotypic categories, hearing status, and by the presence/absence of extracranial anomalies. In secondary analyses, linear regression was used to evaluate whether case-control differences in CBCL scores varied in relation to the child’s sex and maternal age at birth.

Due to the exploratory nature of this study and the lack of previous data related to the specific questions addressed, analyses were not adjusted for multiple comparisons. Rather than viewing p-values as dichotomous tests of significance, the magnitude and precision of observed group differences were examined with effect sizes.

Results

As seen in Table 1, participants included 89 cases and 72 controls at an average of 38.3 months of age (SD = 1.9). The CBCL was completed primarily in English (72%) by participants’ mothers (89%). Relative to controls, cases were more likely to be male (63% vs. 43%), white (78% vs. 58%), and speak some Spanish in the home (33% vs. 22%). Cases had a history of unilateral hearing loss (76%), bilateral hearing loss (6%), or no hearing loss (18%). Cases were classified as having microtia with mandibular hypoplasia (52%), microtia only (30%), or with additional CFM-associated features (18%). Extracranial findings were present in 19% of cases, the most common being congenital heart defects (n = 11), vertebral or rib defects (n = 4), kidney anomalies (n = 3), or central nervous system defects (n = 2). Within the three phenotype groups, extracranial findings were present in 15% of children with microtia; 15% of those with microtia and mandibular hypoplasia; and 38% of children with additional CFM-associated features.

Table 1.

Participant characteristics

Characteristic	Controls (n = 72)		Cases (n = 89)
	M	SD	M	SD
Age at assessment (months)	38.6	2.0	38.1	1.8
Hollingshead Socioeconomic Status	38.3	14.5	35.8	13.7
	N	%	N	%
Sex
Male	31	43.1	56	62.9
Female	41	56.9	33	37.1
Race
White	42	58.3	69	77.5
Black/African American	4	5.6	1	1.1
Asian	0	0.0	6	6.7
American Indian/Alaskan Native	3	4.2	2	2.2
Other race	4	5.6	2	2.2
Multiracial	19	26.4	7	7.9
Ethnicity
Hispanic/Latino	27	39.1	49	56.3
Not Hispanic/Latino	42	60.9	38	43.7
Language
English	56	77.8	60	67.4
Spanish	6	8.3	7	7.9
Combination English/Spanish	10	13.9	22	24.7
CBCL Respondent
Mother	65	90.3	78	87.6
Father	6	8.3	8	9.0
Other	1	1.4	3	3.4
Phenotype (cases only)
Microtia only	-	-	27	30.3
Microtia with mandibular hypoplasia	-	-	46	51.7
Other CFM-associated features	-	-	16	18.0
Hearing Loss
None	-	-	16	18.0
Unilateral	-	-	68	76.4
Bilateral	-	-	5	5.6
Study Site
Children’s Hospital Los Angeles	7	9.7	37	41.6
Children’s Hospital of Philadelphia	5	6.9	1	1.1
Seattle Children’s Hospital	48	66.7	29	32.6
University of North Carolina, Chapel Hill	6	8.3	12	13.5
University of Illinois, Chicago	6	8.3	10	11.2
Developmental Interventions
Physical Therapy	1	1.4	9	10.1
Occupational Therapy	4	5.6	8	9.0
Developmental Services	2	2.8	10	11.2
Speech, Language, or Hearing Services	7	9.7	61	68.5

Abbreviations: CBCL, Child Behavior Check List; CFM, craniofacial microsomia; SD, standard deviation

Case-Control Differences

For both cases and controls, mean CBCL composite scores were all in the average range compared to test norms (Table 2). However, 19% of cases had scores in the borderline to clinically significant range for Internalizing Problems, compared to 13% of controls. Total Problems were above clinical cutoff at a similar rate (14% of cases and 13% of controls) and Externalizing Problems were above the clinical range cutoff at slightly lower rates (8% of cases and 11% of controls). Controls generally had a lower proportion of children above the clinical cutoff for the subscales; however, odds ratios with controls as the referent group adjusted for sociodemographic variables and study site were generally low (Table 2).

Table 2.

CBCL T-scores distribution and odds ratios of meeting borderline clinical cutoff

	T-Scores						Odds Ratios
	Controls (n = 72)			Cases (n = 89)			Controls	Cases
	M	SD	% > Cutoffa	M	SD	% > Cutoffa		Odds Ratiob	95% CI		p
Composite Scales
Internalizing Problems	47.7	10.4	12.7	49.7	11.0	19.1	Ref	1.62	0.59	4.50	0.35
Externalizing Problems	47.5	10.5	11.3	48.2	8.8	7.9	Ref	0.77	0.23	2.58	0.67
Total Problems	47.4	10.5	12.7	49.3	9.9	13.5	Ref	1.13	0.39	3.23	0.83
Subscales
Emotionally Reactive	53.4	5.3	8.5	54.0	5.7	10.1	Ref	1.83	0.56	5.95	0.31
Anxious/Depressed	52.4	4.6	5.6	53.4	5.2	5.6	Ref	1.27	0.28	5.70	0.76
Somatic Complaints	54.3	5.6	9.9	54.5	6.3	16.9	Ref	2.41	0.85	6.84	0.10
Withdrawn	54.0	5.1	2.8	55.6	7.3	9.0	Ref	4.43	0.55	35.50	0.16
Sleep Problems	53.9	5.6	5.6	53.5	4.6	2.3	Ref	0.91	0.13	6.22	0.93
Attention Problems	54.0	5.9	8.5	54.2	5.4	5.6	Ref	0.50	0.12	2.18	0.36
Aggressive Behaviors	53.2	5.5	7.0	52.6	4.7	4.5	Ref	0.79	0.18	3.53	0.75
Stress Problems	53.4	5.5	4.2	54.9	6.4	9.0	Ref	2.57	0.54	12.21	0.23
Depressive Problems	53.5	5.3	4.2	54.2	5.8	4.5	Ref	0.91	0.15	5.52	0.92
Anxiety Problems	53.3	6.2	7.0	53.9	6.1	6.7	Ref	1.57	0.39	6.26	0.52
Autism Spectrum Problems	54.4	6.4	8.5	56.5	7.7	15.7	Ref	2.25	0.72	6.98	0.16
ADHD Problems	53.1	6.0	5.6	52.7	4.2	2.3	Ref	0.48	0.06	3.61	0.47
Oppositional Defiant Problems	53.4	5.2	5.6	53.5	5.1	5.6	Ref	1.25	0.29	5.36	0.77

^aCutoff for borderline clinically significant range is T-score ≥ 60 for Internalizing Problems, Externalizing Problems, and Total Problems and is a T-score ≥ 65 for remaining scales

^bAdjusted for age at (continuous months), sex, SES (continuous), study site, race (white vs. non-white), language (any Spanish vs. only English

Table 3 displays the linear regression results for case-control differences in CBCL scales, after adjustment for demographic confounds. Cases scored higher (i.e., exhibited more behavior problems) than controls on all of the CBCL composite scales, though effect sizes were modest, and the confidence intervals included the null. Differences were most apparent for Internalizing (ES = 0.30, p = 0.08) and Total Problems (ES = 0.29, p = 0.06), while the difference for Externalizing was negligible (ES = 0.14, p = 0.34). On the CBCL subscales, cases exhibited more behavior problems than controls, though again differences were mostly modest and imprecise. As seen in Table 3, case-control differences were present for Anxious/Depressed (ES = 0.35, p = 0.04), Stress Problems (ES = 0.40, p = 0.04), Anxiety Problems (ES = 0.34, p = 0.04), and Autism Spectrum Problems (ES = 0.41, p = 0.02). For an item included on both the Anxious/Depressed and Anxiety Problems scales, children with CFM were more likely to be described as “clinging to adults” or being “too dependent” (70% vs. 42%, χ² = 12.14, p < .001). On the Autism Spectrum Problems scale, parents of children with CFM reported higher rates of their child “not getting along with” other children (27% vs. 11%, χ² = 6.46, p = 0.01) and speech problems (45% vs. 17%, χ² = 14.56, p < .001). None of the individual items for Stress Problems differed significantly between cases and controls.

Table 3.

CBCL differences between cases and controls by linear regression

	Controls	Cases (n = 89)
	(n = 72)	Unadjusted				Adjusteda
		ESb	95% CI		p	ESb	95% CI		p
Composite Scales
Internalizing Problems	Ref	0.19	−0.13	0.51	0.24	0.30	−0.04	0.64	0.08
Externalizing Problems	Ref	0.07	−0.22	0.36	0.64	0.14	−0.15	0.44	0.34
Total Problems	Ref	0.18	−0.13	0.48	0.26	0.29	−0.01	0.59	0.06
Subscales
Emotionally Reactive	Ref	0.11	−0.22	0.43	0.52	0.27	−0.08	0.62	0.13
Anxious/Depressed	Ref	0.27	−0.05	0.59	0.10	0.35	0.02	0.69	0.04
Somatic Complaints	Ref	0.02	−0.31	0.36	0.88	0.11	−0.27	0.49	0.56
Withdrawn	Ref	0.32	−0.06	0.70	0.11	0.24	−0.14	0.61	0.22
Sleep Problems	Ref	−0.02	−0.32	0.27	0.89	0.20	−0.11	0.51	0.20
Attention Problems	Ref	0.06	−0.25	0.36	0.71	0.13	−0.18	0.44	0.42
Aggressive Behaviors	Ref	0.04	−0.24	0.33	0.76	0.11	−0.19	0.42	0.47
Stress Problems	Ref	0.26	−0.07	0.59	0.13	0.40	0.03	0.77	0.04
Depressive Problems	Ref	0.20	−0.13	0.52	0.24	0.18	−0.15	0.50	0.29
Anxiety Problems	Ref	0.18	−0.12	0.49	0.23	0.34	0.02	0.65	0.04
Autism Spectrum Problems	Ref	0.38	0.06	0.71	0.02	0.41	0.07	0.76	0.02
ADHD Problems	Ref	0.07	−0.22	0.36	0.62	0.14	−0.15	0.43	0.34
Oppositional Defiant Problems	Ref	0.14	−0.16	0.44	0.36	0.27	−0.05	0.59	0.10

^aAdjusted for age at (continuous months), sex, SES (continuous), study site, race (white vs. non-white), language (any Spanish vs. only English)

^bstandardized effect size

Reflecting the large proportion (82%) of cases with hearing loss, variance based on the hearing status of cases relative to controls essentially replicated the case-control differences, with findings for elevated Stress Problems (ES = 0.44, p = 0.02) and Autism Spectrum Problems (ES = 0.41, p = 0.03). That is, there were too few cases without hearing loss for a meaningful analysis of case-control differences as a function of their hearing status.

Analyses by Phenotype

Case-control differences were most prominent among children with other CFM-associated features, with effect sizes ranging from 0.32 to 0.52 (p-values = 0.14 to 0.33) for the Internalizing, Externalizing, and Total Problems composites (Table 4). More concerns were reported by parents of children with other CFM-associated features on the Emotionally Reactive scale (ES = 0.80, p = 0.03) and Stress Problems scale (ES = 0.89, p = 0.02) compared to controls. Cases with microtia only scored higher (greater concern) than controls on the Anxious/Depressed subscale (ES = 0.58, p = 0.01) and for Autism Spectrum Problems (ES = 0.52, p = 0.04).

Table 4.

CBCL differences based on phenotype of cases compared to controls by linear regression

	Controls	Cases
	(n = 72)	Microtia Only (n = 27)				Microtia with Mandibular Hypoplasia (n = 46)				Other CFM-Associated Features (n = 16)
		ESa	95% CI		p	ESa	95% CI		p	ESa	95% CI		p
Composite Scales
Internalizing Problems	Ref	0.38	−0.05	0.82	0.08	0.19	−0.20	0.57	0.35	0.52	−0.16	1.21	0.14
Externalizing Problems	Ref	0.21	−0.15	0.57	0.25	0.05	−0.28	0.38	0.77	0.32	−0.33	0.97	0.33
Total Problems	Ref	0.36	0	0.73	0.05	0.18	−0.17	0.53	0.31	0.50	−0.19	1.18	0.16
Subscales
Emotionally Reactive	Ref	0.23	−0.17	0.63	0.26	0.11	−0.29	0.52	0.58	0.80	0.07	1.53	0.03
Anxious/Depressed	Ref	0.58	0.16	0.99	0.01	0.23	−0.11	0.58	0.19	0.39	−0.42	1.19	0.35
Somatic Complaints	Ref	0	−0.49	0.48	0.99	0.11	−0.35	0.56	0.65	0.29	−0.44	1.02	0.44
Withdrawn	Ref	0.11	−0.39	0.60	0.68	0.06	−0.36	0.49	0.77	0.91	−0.22	2.04	0.12
Sleep Problems	Ref	0.22	−0.16	0.59	0.26	0.18	−0.18	0.55	0.32	0.24	−0.31	0.78	0.40
Attention Problems	Ref	0.29	−0.11	0.69	0.16	−0.02	−0.36	0.32	0.91	0.34	−0.40	1.08	0.37
Aggressive Behaviors	Ref	0.08	−0.28	0.44	0.66	0.04	−0.30	0.38	0.82	0.37	−0.25	0.99	0.25
Stress Problems	Ref	0.14	−0.27	0.55	0.51	0.35	−0.07	0.77	0.11	0.89	0.14	1.64	0.02
Depressive Problems	Ref	0.18	−0.25	0.60	0.42	0.03	−0.32	0.39	0.85	0.60	−0.28	1.48	0.19
Anxiety Problems	Ref	0.40	0	0.79	0.05	0.22	−0.11	0.54	0.19	0.59	−0.16	1.34	0.12
Autism Spectrum Problems	Ref	0.52	0.04	1.01	0.04	0.31	−0.09	0.71	0.13	0.56	−0.15	1.27	0.12
ADHD Problems	Ref	0.30	−0.09	0.70	0.13	0.06	−0.25	0.38	0.71	0.16	−0.47	0.80	0.61
Oppositional Defiant Problems	Ref	0.27	−0.12	0.67	0.18	0.20	−0.19	0.59	0.32	0.48	−0.13	1.08	0.13

^astandardized effect size adjusted for age (continuous months), sex, SES (continuous), study site, race (white vs. non-white), language (any Spanish vs. only English)

Cases with extracranial findings scored higher across all outcomes relative to controls (Table 5), with significant differences observed for Anxiety Problems (ES = 0.92, p = 0.01) and Autism Spectrum Problems (ES = 0.71, p = 0.04). In cases without extracranial anomalies, the Anxious/Depressed scale (ES = 0.32, p = .04) and Stress Problems (ES = 0.44, p = .02) were elevated relative to controls.

Table 5.

CBCL differences based on extracranial anomalies present in cases compared to controls by linear regression

	Controls	Cases
	(n = 71)a	No Extracranial Anomalies ( n = 72)				Extracranial Anomalies (n = 17)
		ESb	95% CI		p	ESb	95% CI		p
Composite Scales
Internalizing Problems	Ref	0.24	−0.10	0.58	0.17	0.60	−0.04	1.24	0.07
Externalizing Problems	Ref	0.15	−0.15	0.46	0.31	0.21	−0.35	0.76	0.46
Total Problems	Ref	0.26	−0.04	0.55	0.09	0.52	−0.10	1.13	0.10
Subscales
Emotionally Reactive	Ref	0.24	−0.11	0.60	0.18	0.54	−0.06	1.13	0.08
Anxious/Depressed	Ref	0.32	0.01	0.63	0.04	0.62	−0.09	1.34	0.09
Somatic Complaints	Ref	0.04	−0.35	0.42	0.86	0.43	−0.35	1.20	0.28
Withdrawn	Ref	0.09	−0.23	0.42	0.58	0.68	−0.24	1.59	0.15
Sleep Problems	Ref	0.15	−0.16	0.46	0.34	0.51	−0.02	1.04	0.06
Attention Problems	Ref	0.10	−0.22	0.42	0.55	0.25	−0.31	0.80	0.39
Aggressive Behaviors	Ref	0.12	−0.19	0.42	0.45	0.22	−0.32	0.76	0.42
Stress Problems	Ref	0.44	0.07	0.81	0.02	0.48	−0.14	1.10	0.13
Depressive Problems	Ref	0.13	−0.18	0.44	0.42	0.54	−0.16	1.24	0.13
Anxiety Problems	Ref	0.22	−0.07	0.51	0.14	0.92	0.25	1.58	0.01
Autism Spectrum Problems	Ref	0.33	−0.01	0.67	0.06	0.71	0.04	1.38	0.04
ADHD Problems	Ref	0.14	−0.15	0.44	0.35	0.16	−0.37	0.69	0.55
Oppositional Defiant Problems	Ref	0.30	−0.04	0.64	0.09	0.24	−0.29	0.76	0.37

^aone case with an extracranial anomaly excluded from analysis

^bstandardized effect size adjusted for age (continuous months), sex, SES (continuous), study site, race (white vs. non-white), language (any Spanish vs. only English)

Secondary Analyses

As seen in Table 6, case-control differences tended to be larger for males than females. On the composite scales, case-control differences for males ranged from ES = 0.26 to 0.40 (p-values = 0.09 to 0.27). For females, effect sizes were small in magnitude for the Internalizing, Externalizing, and Total Problems composites (ES = −0.02 to 0.23, p-values = 0.36 to 0.91). On the CBCL subscales, scores were higher for males with CFM relative to control males for the Emotionally Reactive (ES = 0.52, p = 0.04), Stress Problems (ES = 0.45, p = 0.03), and Anxiety Problems subscales (ES = 0.47, p = .03). Males also received higher scores on the Autism Spectrum Problems scale (ES = 0.41, p = 0.08), though the confidence interval included the null. There were no consistent subscale differences among females with CFM relative to female controls. When comparing scores for cases born to mothers who were 25 or younger to those older than 25, there were no significant differences found relative to controls (results not shown).

Table 6.

CBCL differences based on sex of cases compared to controls by linear regression

	Females						Males
	Controls (n = 41)	Cases (n = 33)				Controls (n = 31)	Cases (n = 56)
		ESa	95% CI		p		ESa	95% CI		p
Composite Scales
Internalizing Problems	Ref	0.23	−0.26	0.71	0.36	Ref	0.36	−0.13	0.84	0.15
Externalizing Problems	Ref	−0.02	−0.40	0.36	0.91	Ref	0.26	−0.19	0.71	0.27
Total Problems	Ref	0.14	−0.26	0.54	0.49	Ref	0.40	−0.05	0.85	0.09
Subscales
Emotionally Reactive	Ref	−0.04	−0.57	0.49	0.89	Ref	0.52	0.04	1.00	0.04
Anxious/Depressed	Ref	0.37	−0.11	0.86	0.14	Ref	0.37	−0.06	0.81	0.09
Somatic Complaints	Ref	0.24	−0.30	0.78	0.39	Ref	0.02	−0.54	0.57	0.96
Withdrawn	Ref	0.17	−0.37	0.72	0.53	Ref	0.25	−0.27	0.78	0.35
Sleep Problems	Ref	0.07	−0.38	0.52	0.76	Ref	0.32	−0.12	0.75	0.16
Attention Problems	Ref	0	−0.45	0.45	0.99	Ref	0.22	−0.17	0.62	0.27
Aggressive Behaviors	Ref	−0.01	−0.42	0.40	0.97	Ref	0.23	−0.23	0.68	0.33
Stress Problems	Ref	0.27	−0.31	0.85	0.37	Ref	0.45	0.04	0.85	0.03
Depressive Problems	Ref	0.23	−0.26	0.71	0.37	Ref	0.16	−0.26	0.58	0.46
Anxiety Problems	Ref	0.19	−0.28	0.65	0.43	Ref	0.47	0.06	0.88	0.03
Autism Spectrum Problems	Ref	0.37	−0.21	0.94	0.22	Ref	0.41	−0.04	0.86	0.08
ADHD Problems	Ref	0.03	−0.39	0.45	0.89	Ref	0.23	−0.18	0.65	0.27
Oppositional Defiant Problems	Ref	0.15	−0.28	0.57	0.50	Ref	0.33	−0.18	0.83	0.21

^astandardized effect size adjusted for age (continuous months), sex, SES (continuous), study site, race (white vs. non-white), language (any Spanish vs. only English)

Discussion

Few studies have focused on the psychosocial adjustment of preschoolers with CFM and this study contributes to a better understanding of early childhood behaviors from their parents’ perspective. The current findings generally replicate those reported previously from a large, multisite study of an older and less socioeconomically diverse sample of youth with CFM (Dufton et al., 2011; Wallace et al., 2018). On the CBCL continuous composite scores for Internalizing, Externalizing, and Total Problems, minimal average differences were found between preschoolers with CFM and demographically similar children without CFM. However, 19% percent of cases were above the CBCL clinical cutoff score for internalizing problems, which is consistent with teacher reports of school age children with CFM in the Dufton et al. (2011) study. In addition, subscale and item analyses revealed differences of moderate effect size between cases and controls related to the internalizing areas of anxiety and stress, such as being perceived as too dependent by their parents. These findings require replication, ideally using a multi-method approach to evaluate features of anxiety (e.g., parent report, observational measures of behavioral inhibition). This finding, paired with comparable findings in older children with CFM, suggests a need for increased clinical vigilance and surveillance of internalizing behavior problems in the early screening of preschoolers in this population. As reported in the general population for children ages 2 to 8 (Cree et al., 2018), young males with CFM may be at increased risk for behavioral concerns relative to female peers.

Although the CBCL Autism Spectrum Problems subscale was higher for CFM cases than controls, the two elevated items within this subscale indicated that this result was driven by higher parental reports of difficulties getting along with other children and speech problems, a finding similar to one noted by Dufton et al. (2011) in their study of 7-year-old children with CFM. It may be that the speech difficulties reported contribute to parental observations of difficulties with peers. Like other behavior rating scales, the CBCL relies on empirically (or factor) derived subscales and scale labels can sometimes be misleading (Leary et al., 2010). In this case, without further review of items, the findings may be misleading for children with CFM and other conditions known to affect speech intelligibility. Given the rate of autism in the US (about 1 in 40 children; Kogan et al., 2018), a small percentage of children with CFM will likely meet diagnostic criteria for the autism spectrum. When symptoms of autism are suspected, providers should supplement administrations of a broad screening measure, such as the CBCL, with more detailed, autism-specific measures (Zwaigenbaum, 2015). For example, clinicians can administer interviews or assessments, such as the Autism Diagnostic Inventory – Revised (Rutter et al., 2003) or the Autism Diagnostic Observation Schedule – Second Edition (Lord et al., 2012), or parental reports can be gathered with measures like the Social Responsiveness Scale, Second Edition (Constantino, 2012).

Minimal variation in psychosocial adjustment was observed based on phenotype, a finding also reported by Wallace et al. (2018) among adolescents with CFM. When cases in each of the phenotypic categories were compared with controls, differences were found on several subscale scores, together indicating slightly increased vulnerability for cases with isolated microtia, as well as those with CFM features other than the combination of microtia with mandibular hypoplasia. However, these analyses were complicated by relatively small sample sizes and differing proportions of children with extracranial malformations. For example, 38% of the children with other features of CFM also had extracranial malformations, compared to 15% in the microtia only and microtia with mandibular hypoplasia groups. These findings, as well as those from previous studies (Dufton et al., 2011), suggest that CFM phenotype is only weakly related to behavioral outcomes. The impact of CFM phenotype might play a greater role in accounting for variation in academic skills and outcomes, as suggested by findings of associations between CFM phenotype and speech and language skills in children and adolescents with CFM (Collett et al., 2011; Collett et al., 2019) and reading and writing scores among adolescents with CFM (Speltz et al., 2017).

One limitation of this study was the enrollment rate. Although this rate was relatively low, it was consistent with similar pediatric studies (e.g., Karlson and Rapoff, 2009) and there were few significant demographic differences between participants and nonparticipants (Luquetti et al., 2019). In addition, due to the high proportion of cases with hearing loss (82%), the subsample of those without hearing loss was too small for meaningful analysis. Given the exploratory nature of this initial study of behavioral functioning in preschoolers with CFM, analyses were not adjusted for multiple comparisons to minimize Type II errors (i.e., ‘missing’ a potentially important difference that would warrant future study). Other study limitations include the reliance on parent-reported behavior problems alone, which can vary from children’s self-report in general (e.g., Levi and Drotar, 1998) and with prior CFM samples (e.g., Khetani et al., 2013). Children’s perspectives were excluded because of the challenges in reliable self-report for children this age, as well as avoiding overtiring children during an extensive neurodevelopmental assessment battery. Innovative methods have been used to collect information about preschooler’s self-perceptions, such as structured play interviews (Marsh et al., 2002), as well as other approaches designed for young children, which should be considered for use in future studies of young children with CFM (e.g., Varni et al., 2001; Ren et al., 2012; von Baeyer et al., 2017). Given the young age of the participants, there were too few children with daycare providers or preschool teachers to meaningfully analyze teacher report data. Additionally, observational measures used to assess potential areas of vulnerability (e.g., anxiety/inhibition) may help to better understand how these manifest in young children with CFM. There are multiple measures of toddler social-emotional development with psychometric properties similar to those of the CBCL, with different potential strengths and weaknesses depending on the intended use of the measure (e.g., Pontoppidan et al., 2017). The CBCL was selected for this study in order to compare results with those from previous studies of children with CFM (Collett et al., 2011; Dufton et al., 2011; Wallace et al., 2018) and with other pediatric populations. In future studies, the inclusion of a craniofacial-specific validated measure may provide further delineation of diagnosis-specific functioning (Stock and Feragen, 2019). Similarly, the integration of qualitative (e.g., parent focus groups and/or interviews) and quantitative assessments in future research could provide a better understanding of the impact of particular child behaviors on family functioning, including parenting practices and sibling relationships.

In conclusion, the overall behavioral adjustment of preschoolers with CFM was in the average range in relation to test norms and comparable to demographically-similar peers without CFM. However, categorical analyses suggest that approximately 1 in 5 children with CFM show borderline to clinically significant scores on the Internalizing Problems composite scale and several subscales showed greater vulnerability for cases’ internalizing behaviors. While these findings had low to moderate effect sizes, they were consistent with the results of previous studies with older children and adolescents with CFM (Dufton et al., 2011; Wallace et al., 2018). This suggests that screening and interventions for children with CFM and their families may need to be tailored for young children with social concerns or anxious behavior, which may be informed by interventions developed with behaviorally inhibited preschool children (e.g., Muris et al., 2010). Parent-focused intervention models (e.g., Rapee et al., 2005; Yap et al., 2016) may also be useful, especially as parents of children with CFM have also been found to experience anxiety (Ongkosuwito et al., 2018). As some of these features, along with speech and language delays, may be confused with symptoms of autism, providers should exercise caution when interpreting the results of broad screening measures in children with CFM and specific autism assessment measures should be used when indicated.