Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Apr 1.
Published in final edited form as: Am J Med Genet B Neuropsychiatr Genet. 2019 Feb 1;180(3):204–212. doi: 10.1002/ajmg.b.32714

The Genetic and Environmental Structure of Fear and Anxiety in Juvenile Twins

Chelsea Sawyers 1,*, Thomas Ollendick 2, Melissa A Brotman 3, Daniel S Pine 3, Ellen Leibenluft 3, Dever M Carney 1, Roxann Roberson-Nay 1, John M Hettema 1
PMCID: PMC6414251  NIHMSID: NIHMS1013285  PMID: 30708402

Abstract

Fear and anxiety are conceptualized as responses to acute or potential threat, respectively. Adult twin studies found substantial interplay between genetic and environmental factors influencing fear disorders (phobias) and anxiety disorders. Research in children, however, has largely examined these factors independently. Thus, there exists a substantial knowledge gap regarding the underlying etiologic structure of these closely-related constructs during development. Symptom counts for five fear (criticism, the unknown, death, animal, medical) and four anxiety (generalized, panic, separation, social) dimensions were obtained for 373 twin pairs ages 9–14. Multivariate twin modeling was performed to elucidate the genetic and environmental influences distributed amongst these dimensions. The best fitting model contained one genetic, two familial environmental, and two unique environmental factors shared between fear and anxiety symptoms plus dimension-specific genetic and unique environmental factors. Although several environmental factors were shared between fear and anxiety dimensions, one latent factor accounted for genetic influences across both domains. While adult studies find somewhat distinct etiological differences between anxiety and phobic disorders, the current results suggest that their relative genetic and environmental influences are not as clearly demarcated in children. These etiological distinctions are more nuanced, likely contributing to the highly diffuse symptom patterns seen during development.

Keywords: Anxiety, Fear, Twin Study, Behavioral Genetics, Heritability

Fear and anxiety are adaptive responses to acute or potential threat, respectively (Steimer, 2002). When their symptoms become dysregulated, excessive, or interfere with functioning and quality of life, they are considered clinical phobias or other anxiety disorders, respectively (American Psychiatric Association, 2013). As disorders of threat response with some shared features, psychiatric nosology traditionally includes phobias within the anxiety disorder domain together with generalized anxiety and panic. Both domains have roots in childhood, but commonly expand and persist into adolescence and adulthood, accounting for a substantial proportion of lifetime psychiatric illness (Kessler et al., 2005). However, due to their individually broad but partially distinguishable features and complex unfolding across development, researchers have often separately investigated various aspects of their symptomatology at different ages.

Fear represents the emotional-behavioral response to the perception of immediate danger, leading one to avoid the threat for discernible survival value (Panksepp, 2010). Phobias are common throughout adulthood (Kessler, 2005), and girls consistently report more fears than boys during childhood and adolescence with reliable patterns of developmental changes (Ollendick & Muris, 2015). Twin studies conducted by our group have found phobias in adults to be moderately heritable (30–40%) with phobia subtypes having overlapping genetic and environmental influences as well as subtype-specific factors (Kendler, Myers, Prescott, & Neale, 2001; Kendler Neale, Kessler, Heath, & Eaves, 1992). These overlapping influences help explain the high comorbidity amongst fears and phobias (Loken, Hettema, Aggen, & Kendler, 2014). Individual fear symptoms in children are also moderately heritable with modest familial environmental influences shared between twins and a greater role for unique environment (Eaves & Silberg, 2008; Eley, Rijsdijk, Perrin, O’Connor, & Bolton, 2008; Lichtenstein & Annas, 2000). A few studies investigating the comorbidity structure of fear symptoms in children reported a common genetic factor that influenced all symptom types in addition to fear-specific factors (Eaves & Silberg, 2008; Lichtenstein & Annas, 2000).

Anxiety disorders (ADs) often have a basis in normal anxious concerns; however, the degree of anxiety and associated symptoms becomes excessive, uncontrollable, and impairing to an individual’s life. The ADs are the most prevalent class of psychiatric disorders in US adults (Kessler, 2005). Similar to fears, females have higher rates of ADs throughout both childhood and adulthood (Grills-Taquechel & Ollendick, 2012). Twin studies demonstrate ADs are also moderately heritable (Hettema, Neale, & Kendler, 2001). ADs are highly comorbid with each other, and adult twin studies suggest this comorbidity may be due, in part, to genetic risk factors shared between disorders (Hettema, Prescott, Myers, Neale, & Kendler, 2005; Scherrer et al., 2000). This comorbidity pattern is also seen in children, where 40% to 60% of children with one AD are estimated to meet criteria for additional ADs (Kashani and Orvaschel, 1990; Leyfer, Gallo, Cooper-Vince, & Pincus, 2013). While the heritability of ADs has been studied extensively in adults, the genetic liability to childhood ADs has primarily focused on more anxiety-related symptoms or behaviors, often in the context of other internalizing syndromes like depression.

In a non-twin study, Muris and colleagues (1998) report substantial correlations between subscales of the Fear Survey Schedule for Children-Revised (FSSC-R; Ollendick, 1983) and the Screen for Child Anxiety-Related Emotional Disorders (SCARED; Birmaher et al., 1997). However, only a few studies of children have examined the liability structure of DSM-based anxiety dimensions (Ogliari et al., 2006) or phobic fear symptoms (Stevenson et al., 1992; Eaves & Silberg, 2008). A recent longitudinal twin study examined internalizing symptoms more broadly, reporting increasingly shared risk factors between depression and SCARED anxiety scales with age, (Waszczuk, Zavos, Gregory, & Eley, 2014). However, no child twin studies have explored the potential sources of shared etiology of each separate anxiety dimension together with various specific fear dimensions as assessed in the FSSC-R. Given the partial distinction between risk factors for phobias and other ADs in adults, it is important to examine their childhood precursors. Furthermore, early fear and anxiety disorders are the strongest predictors of later psychiatric comorbidity (Kessler et al., 2012). Therefore, this study aimed to explicate the shared and specific risk factors (genetic, familial environment and unique environment) and their roles amongst fear and anxiety domains in youth. Based on the prior literature, significant genetic, familial environment and unique environment are predicted for fears and anxiety examined separately. Due to the moderate level of phenotypic correlation between the FSSC-R and SCARED subscales described above, differential overlap between genetic and environmental factors across these scales is predicted. By leveraging symptom sum scores rather than diagnostic criteria, the power to detect meaningful patterns of shared and specific variance is maximized.

Methods

Editorial Policies and Ethical Considerations

The Institutional Review Boards at VCU and NIMH approved this study, and parents of all participants provided informed consent before participating.

Participants

The twins included in these analyses comprised the VCU Juvenile Anxiety Study (VCU-JAS; Carney et al., 2016). Using twin families recruited by the Mid-Atlantic Twin Registry (MATR; Lilley and Silberg, 2013), VCU-JAS enrolled twins aged 9–14 across two sites (VCU and the National Institute of Mental Health; NIMH) to participate in a study of internalizing phenotypes. This age range focuses on middle-school aged children in the pre-adolescent period of increasing risk for internalizing disorders. Only Caucasian twins were recruited to minimize heterogeneity within the sample for the genetic aims of the overall study. The Institutional Review Boards at VCU and NIMH approved this study, and parents of all participants provided informed consent before participating. Self-report data available for this study came from 746 youths (N=130 monozygotic (MZ) twin pairs and N=243 dizygotic (DZ) twin pairs) consisting of 388 female and 358 male twins. Zygosity was determined using parental responses to standard questions about physical appearance of the twins and DNA testing as described in detail elsewhere (Carney et al., 2016).

Measures

Fear

The FSSC-R (Ollendick, 1983) is a widely used questionnaire for assessing common fears in children (Gullone, 1999; Myers & Winters, 2002; Ollendick & Muris, 2015). It uses a 3-point Likert scale for each of 80 feared stimuli or situations. The shortened 25-item form (FSSC-RSF) has a 5-factor structure similar to the full scale with the following means (standard deviations; Muris, Ollendick, Roelofs, & Austin, 2014): fear of failure and criticism 8.51 (2.43), fear of the unknown 7.68 (2.38), fear of animals 7.28 (2.02), fear of danger and death 12.96 (3.41), and medical fears 6.85 (2.06). A sum score was calculated for all subscales with minimal missingness (<2%).

Anxiety

The SCARED was developed to screen for ADs within clinical samples (Birmaher et al., 1999, 1997) but has also been widely used in community and research studies (Verhulst and van der Ende, 2006; Ramsawh et al., 2012). It assesses five types of childhood anxiety symptoms: panic/somatic 5.17 (3.86) , generalized anxiety 5.94 (3.69), social anxiety 5.20 (3.42), and separation anxiety 6.00 (3.29), as well as school avoidance which was not used in these analyses. The 41-item version (Birmaher et al., 1999) assesses symptoms on a 3-point Likert scale. A sum score was calculated for each of the four subscales with minimal missingness (<2%).

Statistical analyses

Analyzing the similarity of MZ and DZ twins elucidates the roles of genetic (A), familial environmental (C), and unique environmental (E) effects. Variance is partitioned into underlying genetic and environmental influences by leveraging the difference in genetic relatedness between twin types. A reflects the latent cumulative effects of individual genetic loci influencing a trait. Familial environment (C) captures non-genetic influences that make twins more similar to each other compared to the general population. Unique environment (E) describes influences that contribute to the differences seen between co-twins including measurement error. Analyses were conducted with a maximum-likelihood approach using the OpenMx package (Neale et al., 2016).

In multivariate structural equation modeling, ACE components can be specific to each subscale (e.g., As1 in Figure 1) as well as common to multiple subscales (e.g., Ac1). Given the substantial effects of age and sex on fear and anxiety measures, these two predictors were included as covariates in the primary analyses. Significance of individual model parameters were tested by comparing constrained submodels to a saturated model (i.e., all paths freely estimated) via the Akaike information criterion (AIC), and a likelihood ratio χ2 test used to determine if the constrained model fits the data significantly worse than the saturated. AIC is based on twice the difference in log likelihood between higher order and submodels with a penalty for degrees of freedom, with lower AIC denoting a better balance of model fit and parsimony (Williams & Holahan, 1994).

Figure 1.

Figure 1.

Open in a new tab

Best fitting model for FSSC-RSF subscales. The model contains 1 common additive genetic factor (Ac1), 1 common familial environmental factor (Cc1), and 1 common unique environmental factor (Ec1). Only subscale specific additive genetic and unique environmental factors were found to be significant and retained in the final model. Path coefficients representing standardized estimates are listed above 95% confidence intervals for each path for fear of failure and criticism (CRIT), fear of the unknown (UNKN), fear of animals (ANML), fear of danger and death (DEATH), and medical fears (MED). Triangles in the middle figure denote age and sex moderators on the means for all subscales with 95% CIs listed below the standardized path estimate.

An increasingly complex sequence of hypothetical models was fit to the data. Multivariate independent pathway models were estimated separately for the Fear and Anxiety dimensions to determine their overall etiologic structure. The final model, an independent pathway model covering both Fear and Anxiety tested the need for multiple sets of common ACE sources of covariance to explain the observed data.

Results

Fear

The Fear section of Table 1 displays the fit statistics for the modeling of the FSSC-RSF. To test for genetic and environmental factors common to all subscales, we began with a model including single common plus specific A, C, and E factors with age and sex covariates for each subscale mean. Significance of common and then specific factors were sequentially tested by iteratively constraining parameters to zero. As indicated in Table 1, Model 1b was determined to be the best fitting and most parsimonious model. It included a single set of common ACE factors and subscale specific A and E factors. Females had the expected pattern of higher mean scores compared to males across all subscales, depicted in Figure 1 as paths from the sex moderator (triangle) loading on each subscale. A slight decrease in means as age increases for all subscales except for criticism is consistent with observed decreases in most childhood fears over development. The influence of common and specific genetic factors accounted for 10–34% of the total of variance of each subscale with the remaining variance accounted for primarily by subscale specific, unique environmental factors.

Table 1.

Model-Fitting Results for Multivariate Independent Pathway Models of Fear and Anxiety

Fear
Model Common
Factors		 Specific
Factors		 EP
df/Δ		 Model Fit
−2LL/Δ AIC/Δ P
1 A1C1E1 All ACE 45 3612 15553.6 15643.6 -
1a A1C1E1 All AE 40 5 0.0 −10.0 .999
1b A1C1E1 All CE 40 5 5.0 −5.0 .413
1c A1C1E1 All E 35 10 15.6 −4.4 .110
2 C1E1 All ACE 40 5 16.1 6.1 .000
3 A1E1 All ACE 40 5 7.2 −2.8 .204
4 A1C1 All ACE 40 5 161.2 151.2 .000
5 - All ACE 30 15 902.1 872.1 .000
Anxiety
1 A1C1E1 All ACE 36 2900 14824.0 14896.0 -
1a A1C1E1 All AE 32 4 0.0 −8.0 .999
1b A1C1E1 All CE 32 4 13.6 5.7 .008
1c A1C1E1 All E 28 8 28.6 12.7 .000
2 C1E1 All ACE 32 4 9.4 1.4 .005
3 A1E1 All ACE 32 4 12.1 4.1 .001
4 A1C1 All ACE 32 4 127.4 119.4 .000
5 - All ACE 24 12 720.1 696.1 .000
Open in a new tab

Abbreviations: EP=estimated parameters, df/Δ= degrees of freedom for full model and change in degrees of freedom for submodels, −2LL/Δ = twice the negative log likelihood of full model and change in −2LL of submodels compared to full, AIC/Δ= Akaike information criterion of full model and change in AIC of submodels compared to full model. For AIC and −2LL, smaller or more negative Δs indicate a better fit compared to the full model (Model 1). p is related to the statistical difference of −2LL values between full and sub models. Bold designates the overall best fitting model.

Anxiety

The Anxiety section of Table 1 displays the fit statistics for the modeling of the SCARED. Similar to Fear, we found model 1b fit best with single set of common ACE factors and subscale specific A and E. Age and sex influenced the means of each subscale in a similar pattern to Fear. Generalized anxiety was the exception for which the opposite age trend was found, consistent with clinically observed increased risk of generalized anxiety with age. The total influence of all genetic factors accounted for 18–35% of each subscale’s variance with remaining variance accounted for primarily by subscale specific unique environmental factors. The genetic factor common to all subscales accounted for 5–19% of the variance for panic, generalized anxiety and separation anxiety. Only social phobia did not share genetic influences with the other symptoms. Figure 2 depicts the path estimates from the best fitting model for Anxiety.

Figure 2.

Figure 2.

Open in a new tab

Best fitting model for SCARED subscales. The model contains 1 common additive genetic factor (Ac1), 1 common familial environmental factor (Cc1), and 1 common unique environmental factor (Ec1). Only subscale specific additive genetic, and unique environmental factors were found to be significant and retained in the final model. Path coefficients representing standardized estimates are listed above the 95% confidence intervals (CIs) for each path for panic disorder (PAN), generalized anxiety disorder (GAD), social phobia (SOC), and separation anxiety (SEP).

Modeling Across Fear and Anxiety

To examine whether the etiological overlap in children is similar to that reported for adults, we started with a model containing two sets of latent common ACE factors. To ensure model identification as well as a unique solution for each set of ACE common factors, we constrained one loading on each set of latent factors in order to designate one set as the ‘Anxiety’ set and the other as the ‘Fear’ set. I.e., all SCARED subscales load on the ‘Anxiety’ ACE factors and all subscales but criticism from FSSC-RSF do so as well, and the reverse with all FSSC-RSF and all SCARED except generalized anxiety loading onto the ‘Fear’ ACE set (Illustrated in the C and E diagrams in Figure 3). This allowed for all scales to remain in the analyses while also providing a unique analytic solution. Parameter estimates and model fit did not significantly differ when we examined constraining alternative pathways. When testing a single common risk source (i.e., Ac), we allowed all subscales to load onto that single source. While one of the two latent Ac factors could be eliminated without a significant deterioration in fit, we were unable to remove any of the four common environmental factors (i.e., keeping two each for Cc and Ec). The best fitting was Model 3b in Table 2 with similar effects of age and sex as before. The proportion of variance in liability accounted for by each source of variance is listed in Table 3. Figure 3 illustrates the larger role of shared genetic influences on the Anxiety subscales, with limited influence of familial environment common to all subscales and the largest proportion of unique environmental influences originating from subscale specific factors. This is reflected in Table 3 where the total genetic influences across shared and specific components account for 15–40% of the variance, whereas the total variance accounted for by common and specific familial environment is markedly lower (0–17%); the remainder was accounted for by some common but predominantly specific unique environment (47–74%).

Figure 3.

Figure 3.

Open in a new tab

Best fitting model for Fear and Anxiety dimensions. The model contains one common additive genetic factor (Ac1), two common familial environmental factors (Cc1 and Cc2), and two common unique environmental factors (Ec1 and Ec2). Only subscale specific additive genetic and unique environmental factors were found to be significant and retained in the final model. The darker lines indicate a stronger influence of the latent factor on the observed variable. Lighter lines indicate a standardized path estimate less than 0.10. Table 3 provides the proportion of variance accounted for by each of these pathways.

Table 2.

Model Fitting Results for Independent Pathway Model Including All Symptom Clusters

Model Common
Factors		 Specific
Factors		 EP df/ Δ Model Fit
−2LL/ Δ AIC/ Δ P
1 A1A2C1C2E1E2 All ACE 102 6481 30146.6 30350.6 -
2 A1A2C1E1E2 All ACE 95 7 −85.2 −99.2 .999
3 A1C1C2E1E2 All ACE 95 7 −121.8 −135.8 .999
3a A1C1C2E1E2 All CE 86 16 −106.5 −138.5 .032
3b A1C1C2E1E2 All AE 86 16 −121.8 −153.8 .999
3c A1C1C2E1E2 All E 77 25 −77.6 −127.6 < .001
4* A1A2C1C2E1 All ACE 95 7 - - -
5 A1C1E1 All ACE 81 21 −41.2 −83.2 .999
6 C1E1 All ACE 72 30 50.3 18.3 .014
7 A1E1 All ACE 72 30 24.2 −35.8 .762
8 A1C1 All ACE 72 30 234.9 174.9 < .001
9 - All ACE 54 48 1853.2 1757.2 < .001
Open in a new tab

Table 3 shows the fit statistics for all models tested. Common Factors are the ACE latent factors shared between all observed variables. Specific Factors are the nine sets of ACE latent factors that each only load onto one observed variable, respectively. Specific factors were tested by dropping a class at a time (e.g., all specific A latent factors dropped at same time). Bold designates the overall best fitting model.

*

Model 4 was unable to converge to a final solution

Table 3.

Proportion of Variance in Liability to Anxiety and Fear Symptom Clusters from Common and Specific Genetic and Environmental Risk Factors*

Symptom
Cluster		 Genetic Factors Familial Environmental Factors Unique Environmental Factors
Ac As Total Cc1 Cc2 Total Ec1 Ec2 Es Total
PAN .28 .12 .40 .03 .02 .05 .16 .00 .39 .55
GAD .50 .03 .53 .00 - .00 .12 - .35 .47
SEP .21 .12 .33 .02 .10 .12 .21 .04 .30 .55
SOC .14 .26 .40 .02 .00 .02 .18 .02 .38 .58
CRIT .37 .00 .37 - .00 .00 - .19 .44 .63
UNKN .14 .00 .14 .02 .15 .17 .04 .27 .38 .69
ANML .03 .18 .21 .06 .01 .07 .00 .18 .54 .72
DEATH .14 .02 .16 .10 .00 .10 .00 .15 .59 .74
MED .15 .04 .19 .15 .00 .15 .00 .26 .30 .66
Open in a new tab

Panic disorder (PAN), generalized anxiety disorder (GAD), social phobia (SOC) and separation anxiety (SEP), fear of failure and criticism (CRIT), fear of the unknown (UNKN), fear of animals (ANML), fear of danger and death (DEATH), and medical fears (MED), Ac (Common A factor), As (Specific A factor), Cc1 (First Common C factor) Cc2 (Second Common C factor), Ec1 (First Common E factor), Ec2 (Second Common E factor) Es (Specific E factor). Bolded columns designate proportion of total variance accounted for by the combined common and specific etiological sources of variance for each subscale.

*

Best-fitting IPM model 3b from Table 3

Discussion

We used multivariate modeling to examine the structure of genetic and environmental risk factors that underlie the associations between Fear and Anxiety domains in a juvenile twin sample. We first separately examined each domain’s genetic and environmental factors. Fear and Anxiety each displayed an overall similar etiologic covariance structure across their respective dimensions that included moderate influences of genetic plus familial and unique environmental factors common to all symptoms (Figure 2). The remaining influences were due to subscale specific genetic and unique environmental effects. Prior childhood twin studies of Fear have primarily focused on the etiology of particular fears (Stevenson et al., 1992) and their longitudinal changes over development (Eaves & Silberg, 2008) or the etiology shared between fears or phobias (Lichtenstein & Annas, 2000). Whether measuring diagnoses or symptom counts, extant studies reported moderate genetic influences that were partially shared with other fears or phobias plus a predominance of unique environmental influences. Our finding of significant fear-specific genetic and unique environmental with substantially less familial environmental influences reflects prior studies examining these domains independently (Bolton et al., 2006; Eley et al., 2008; Lichtenstein & Annas, 2000; Waszczuk et al., 2014).

Within Anxiety symptom dimensions, a latent genetic factor Ac accounted for a modest proportion of variance shared by all SCARED subscales except social phobia. This is generally consistent with findings by Ogliari and colleagues (2010). Differences between these two studies are largely accounted for by our finding of an additional common familial factor Cc with modest influence on the covariance of subscales that the other study did not include in their final models. Overall, our findings of significant moderate genetic and unique environmental influences are similar to previous independent studies of social phobia, separation anxiety, and panic in children (Bolton et al., 2006; Eley et al., 2008; Scaini et al., 2012).

Considering previous twin research that examined Fear and Anxiety separately in children, our study provides novel insights into the potential etiological underpinnings responsible for the high comorbidity observed between phobic fear and anxiety domains in youth. Within the combined model (Figure 3), risk across both domains was variably influenced by a single genetic factor (Ac) in addition to more domain-specific familial environment (Cc1 and Cc2), plus environment unique to each individual (Ec1 and Ec2). The proportion of variance accounted for by Ac is lower for fear symptoms (3–37%) than for anxiety symptoms (14–50%). Only Ac accounts for greater than 30% of the variance for any of the subscales affecting both fear of criticism and generalized anxiety, the most genetically influenced of each scale (37% and 50% heritability, respectively). The pattern of ‘same genes but, different environments’ found in this study is also found in the relationship between two highly comorbid disorders in adults, major depression and generalized anxiety.(Kendler et al., 1992; Kendler, Heath, Martin, & Eaves, 1987) Phobic fears, anxiety, and depression are all closely related internalizing domains, so this shared genetic influence Ac may be indexing a broader risk factor than just Fear and Anxiety such as a global predisposition for negative affective syndromes. Previous research has demonstrated the overlapping etiology of these internalizing syndromes in children and adolescents (Waszczuk et al., 2014), whereas the adult literature shows only a partial overlap between genetic risk for internalizing disorders (Hettema et al., 2005).

While our best fitting model included multiple C and E factors, they only partially distinguished between Fear and Anxiety domains. Although significant in the separate analyses of Fear and Anxiety, when examined in the combined model, familial environment was not as strongly influential, and the pattern that emerged was not simply overall Fear versus Anxiety. As Figure 3 shows, Cc1 specifically affects fear of death and medical fears with only modest influences on Anxiety. At the level of content, both death and medical fears may relate to threats to bodily integrity. On the other hand, only separation anxiety and fear of the unknown load onto Cc2 with negligible remaining effects. One may speculate these might represent responses related to separation from a caregiver who provides security from unknown threatening situations. These are distinguishable from potential identifiable threats such as animals and social interactions. Notably, familial environment plays little role in comorbidity of internalizing disorders in adults (Kendler, 1996). As seen previously, middle childhood seems to be a developmental stage in which family influences peak on many behavioral outcomes.

The only common factors to show an arguable overall distinction between the domains were the unique environmental influences. Figure 3 shows all the Anxiety dimensions clustering together on Ec1 and all the Fear dimensions on Ec2 with minimal cross loadings from the other domain (represented by dashed lines in the figure). Fear is a more primitive, instinctive defensive reaction primarily involving the amygdala and its recruitment of other subcortical regions that develop early, while Anxiety requires more complex responses dependent upon cortical involvement which develops later (LeDoux & Pine, 2016). Furthermore, given that normative fears are variably expressed within certain developmental windows (fear of strangers, separation, the dark, animals, etc.), it is hypothetically more likely for their environmental influences to cluster according to exposures by age that make them more highly correlated with each other than with environmental influences on anxiety symptoms. That is, while their predisposing genetic influences largely overlap, their environmental influences may be differentially moderated by age and neurodevelopmental stage.

Our results further suggest that the etiological structure of Fear and Anxiety in children is not as clearly differentiated as in adults. Prior adult twin studies suggest a continuity of risk factors from fears to the corresponding phobias (Kendler et al., 1992). Furthermore, both adult phenotypic (Krueger, 1999) and twin (Hettema et al., 2005) studies find correlated but partially distinct structural relationships between phobias and other anxiety disorders. Thus, while our finding of moderate levels of genetic influences common to all subscales is not unexpected, the degree of sharing seen here is notable. This reflects, and likely helps explain, clinical observations in which children are substantially more likely to have a complex, changing pattern of syndromes compared to adults (Grills-taquechel et al., 2012; Kashani et al., 1990; Kessler, 2005; Leyfer et al., 2013).

The results of this analysis should be interpreted within the context of several limitations. First, fear and anxiety symptoms are transient in nature, especially within children who commonly outgrow them. While the use of symptom measures limits generalizability beyond the age range they are designed to investigate, most studies in children use measures similar to these, allowing a relative comparison across the current literature. Although it might limit generalizability to clinical samples, such a dimensional approach increases the statistical power to detect the influences of etiologic significance over use of less prevalent psychiatric diagnoses. Second, we opted to control for age and sex effects in the analyses rather than split our sample and conduct age- or sex-specific analyses. Furthermore, this sample does not possess sufficient power to examine qualitative sex effects on sources of variance; we estimate a peak power of 0.45 to detect a correlation of 0.2 between male and female genetic factors. However, previous studies have indicated conflicting results regarding age and sex having a moderating effect on the variance of fear and anxiety measures (Eaves & Silberg, 2008; A Ogliari et al., 2006; Scaini et al., 2012); thus, we covaried for them at the means level to minimize these biases. Finally, generalizability is also limited due to the exclusive use of Caucasian twin pairs driven by the aim to minimize genetic heterogeneity introduced when sampling from multiple ethnicities. Most prior twin studies were also conducted in Caucasian twins, maximizing our comparability with them.

Conclusions

The findings of this analysis have implications for investigating the risk mechanisms underlying fear and anxiety symptoms in childhood and beyond. From a trans-diagnostic perspective, these findings help explain and potentially validate the high rates of internalizing comorbidity found in children. Longitudinal research in developmental psychopathology could also benefit from incorporating both threat response domains given their close links in childhood. Studies in adults show a clearer distinction between the two domains and their sources of covariation, while their expression in children is more diffuse and malleable. From an etiological perspective this could be due to the greater degree of shared genetic influences expressed during child development coupled with developmentally specific environmental influences that help disentangle fear and anxiety. A longitudinal study extending into late adolescence would further inform the temporal unfolding of fear and anxiety risk factors as they merge into those seen in adulthood. Finally, more effective gene-finding investigations of fear and anxiety disorders such as genome-wide association studies (GWAS) could be designed to incorporate these shared genetic influences. Two prior GWAS, one for childhood anxiety related behaviors (Trzaskowski et al., 2013) and another for adult anxiety disorders (Otowa et al., 2016), have utilized this strategy. Similar efforts are being undertaken on a much larger scale by the Psychiatric Genomics Consortium Anxiety Disorders Working Group (PGC-ANX).

Acknowledgements

The authors wish to acknowledge Andrea Molzhon, PhD, Elizabeth Moroney, and Laura Machlin for overseeing the collection and processing of these data. This work was supported by the National Institute of Mental Health (C.S.: R01MH098055, T32MH020030; JMH and RRN: R01MH098055; DSP: NIMH-IRP-ziamh002781). The authors have no conflicts of interest to report.

This work was supported by the National Institute of Mental Health (C.S.: R01MH098055, T32MH020030; JMH and RRN: R01MH098055; DSP: NIMH-IRP-ziamh002781).

References

  1. American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th ed.). Arlington, VA.
  2. Birmaher B, Brent DA, Chiappetta L, Bridge J, Monga S, & Baugher M (1999). Psychometric properties of the Screen for Child Anxiety Related Emotional Disorders (SCARED): a replication study. Journal of the American Academy of Child and Adolescent Psychiatry, 38(10), 1230–1236. doi: S0890–8567(09)63237–8. [DOI] [PubMed] [Google Scholar]
  3. Birmaher B, Khetarpal S, Brent D, Cully M, Balach L, Kaufman J, & Neer SM (1997). The Screen for Child Anxiety Related Emotional Disorders (SCARED): scale construction and psychometric characteristics. Journal of the American Academy of Child and Adolescent Psychiatry, 36(4), 545–553. doi: 10.1097/00004583-199704000-00018. [DOI] [PubMed] [Google Scholar]
  4. Bolton D, Eley TC, O’Connor TG, Perrin S, Rabe-Hesketh S, Rijsdijk F, & Smith P (2006). Prevalence and genetic and environmental influences on anxiety disorders in 6-year-old twins. Psychological medicine, 36(3), 335–344. doi:S0033291705006537. [DOI] [PubMed] [Google Scholar]
  5. Carney DM, Moroney E, Machlin L, Hahn S, Savage JE, Lee M, Towbin KA, Brotman MA, Pine DS, Leibenluft E, Roberson-Nay R, & Hettema JM (2016). The Twin Study of Negative Valence Emotional Constructs. Twin research and human genetics : the official journal of the International Society for Twin Studies, 19(5), 456–464. doi: 10.1017/thg.2016.59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Eaves LJ, & Silberg JL (2008). Developmental-genetic effects on level and change in childhood fears of twins during adolescence. Journal of child psychology and psychiatry, and allied disciplines, 49(11), 1201–1210. doi: 10.1111/j.1469-7610.2008.01956.x. [DOI] [PubMed] [Google Scholar]
  7. Eley TC, Rijsdijk FV, Perrin S, O’Connor TG, & Bolton D (2008). A multivariate genetic analysis of specific phobia, separation anxiety and social phobia in early childhood. Journal of abnormal child psychology, 36(6), 839–848. doi: 10.1007/s10802-008-9216-x. [DOI] [PubMed] [Google Scholar]
  8. Grills-Taquechel AE, & Ollendick TH (2012). Phobic and anxiety disorders in children and adolescents Advances in psychotherapy. Cambridge, MA: Hogrefe & Huber Publishers. [Google Scholar]
  9. Gullone E (1999). The assessment of normal fear in children and adolescents. Clinical Child and Family Psychology Review, 2, 91–106. [DOI] [PubMed] [Google Scholar]
  10. Hettema JM, Neale MC, & Kendler KS (2001). A review and meta-analysis of the genetic epidemiology of anxiety disorders. The American Journal of Psychiatry, 158(10), 1568–1578. [DOI] [PubMed] [Google Scholar]
  11. Hettema JM, Prescott CA, Myers JM, Neale MC, & Kendler KS (2005). The structure of genetic and environmental risk factors for anxiety disorders in men and women. Archives of General Psychiatry, 62(2), 182–189. doi: 62/2/182. [DOI] [PubMed] [Google Scholar]
  12. Kashani JH, & Orvaschel H (1990). A community study of anxiety in children and adolescents. The American Journal of Psychiatry, 147(3), 313–318. doi: 10.1176/ajp.147.3.313. [DOI] [PubMed] [Google Scholar]
  13. Kendler K (1996). Major depression and generalized anxiety disorder same genes , (partly) different environments- Revisited. The British journal of psychiatry, 168(30), 68–75. [PubMed] [Google Scholar]
  14. Kendler KS, Myers J, Prescott CA, & Neale MC (2001). The genetic epidemiology of irrational fears and phobias in men. Archives of General Psychiatry, 58(3), 257–265. doi:yoa20158. [DOI] [PubMed] [Google Scholar]
  15. Kendler KS, Neale MC, Kessler RC, Heath AC, & Eaves LJ (1992). The genetic epidemiology of phobias in women. The interrelationship of agoraphobia, social phobia, situational phobia, and simple phobia. Archives of General Psychiatry, 49(4), 273–281. [DOI] [PubMed] [Google Scholar]
  16. Kessler RC (2005). Prevalence, severity, and comorbidity of twelve-month DSM-IV disorders in the National Comorbidity Survey Replication (NCS-R). Arch. Gen. Pscyhiatry, 62, 617–627. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kessler RC, Avenevoli S, McLaughlin KA, Green JG, Lakoma MD, Petukhova M, Pine DS, Sampson NA, Zaslavsky AM, & Merikangas KR (2012). Lifetime co-morbidity of DSM-IV disorders in the US National Comorbidity Survey Replication Adolescent Supplement (NCS-A). Psychological Medicine, 42(9), 1997–2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, & Walters EE (2005). Lifetime Prevalence and Age-of-Onset Distributions of Distributions of DSM-IV Disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry, 62(June), 593–602. doi: 10.1017/S0033291712000025. [DOI] [PubMed] [Google Scholar]
  19. Krueger RF (1999). The structure of common mental disorders. Archives of general psychiatry, 56(10), 921–6. [DOI] [PubMed] [Google Scholar]
  20. LeDoux JE, & Pine DS (2016). Using Neuroscience to Help Understand Fear and Anxiety: A Two-System Framework. American Journal of Psychiatry, 173(11), 1083–1093. doi: 10.1176/appi.ajp.2016.16030353. [DOI] [PubMed] [Google Scholar]
  21. Leyfer O, Gallo KP, Cooper-Vince C, & Pincus DB (2013). Patterns and predictors of comorbidity of DSM-IV anxiety disorders in a clinical sample of children and adolescents. Journal of anxiety disorders, 27(3), 306–311. doi: 10.1016/j.janxdis.2013.01.010. [DOI] [PubMed] [Google Scholar]
  22. Lichtenstein P, & Annas P (2000). Heritability and prevalence of specific fears and phobias in childhood. Journal of child psychology and psychiatry, and allied disciplines, 41(7), 927–937. [PubMed] [Google Scholar]
  23. Lilley EC, & Silberg JL (2013). The Mid-Atlantic Twin Registry, revisited. Twin research and human genetics : the official journal of the International Society for Twin Studies, 16(1), 424–428. doi: 10.1017/thg.2012.125. [DOI] [PubMed] [Google Scholar]
  24. Loken EK, Hettema JM, Aggen SH, & Kendler KS (2014). The structure of genetic and environmental risk factors for fears and phobias. Psychological medicine, 44(11), 2375–2384. doi: 10.1017/S0033291713003012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Muris P, Gadet B, Moulaert V, & Merckelbach H (1998). Correlations between Two Multidimensional Anxiety Scales for Children. Perceptual and Motor Skills, 87(1), 269–270. doi:S0005–7916(98)00023–8. [DOI] [PubMed] [Google Scholar]
  26. Muris P, Ollendick TH, Roelofs J, & Austin K (2014). The short form of the fear survey schedule for children-revised (FSSC-R-SF): an efficient, reliable, and valid scale for measuring fear in children and adolescents. Journal of anxiety disorders, 28(8), 957–965. doi: 10.1016/j.janxdis.2014.09.020. [DOI] [PubMed] [Google Scholar]
  27. Myers K, & Winters NC (2002). Ten-year review of rating scales. II: Scales for internalizing disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 41(6), 634–659. doi:S0890–8567(09)61019–4. [DOI] [PubMed] [Google Scholar]
  28. Neale MC, Hunter MD, Pritikin JN, Zahery M, Brick TR, Kirkpatrick RM, Estabrook R, Bates TC, Maes HH, & Boker SM (2016). OpenMx 2.0: Extended Structural Equation and Statistical Modeling. Psychometrika, 81(2), 535–549. doi: 10.1007/s11336-014-9435-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Ogliari A, Citterio A, Zanoni A, Fagnani C, Patriarca V, Cirrincione R, Stazi MA, & Battaglia M (2006). Genetic and environmental influences on anxiety dimensions in Italian twins evaluated with the SCARED questionnaire. Journal of anxiety disorders, 20(6), 760–777. doi:S0887–6185(05)00104–0. [DOI] [PubMed] [Google Scholar]
  30. Ogliari A, Spatola CA, Pesenti-Gritti P, Medda E, Penna L, Stazi MA, Battaglia M, Fagnani C (2010). The role of genes and environment in shaping co-occurrence of DSM-IV defined anxiety dimensions among Italian twins aged 8–17. Journal of Anxiety Disorders, 24(4), 433–439. doi: 10.1016/j.janxdis.2010.02.008. [DOI] [PubMed] [Google Scholar]
  31. Ollendick TH (1983). Reliability and validity of the Revised Fear Survey Schedule for Children (FSSC-R). Behavior Research and Therapy, 21, 685–692. [DOI] [PubMed] [Google Scholar]
  32. Ollendick TH, & Muris P (2015). The Scientific Legacy of Little Hans and Little Albert: Future Directions for Research on Specific Phobias in Youth. Journal of Clinical Child and Adolescent Psychology, 44(4), 689–706. doi: 10.1080/15374416.2015.1020543. [DOI] [PubMed] [Google Scholar]
  33. Otowa T, Hek K, Lee M, Byrne EM, Mirza SS, Nivard MG, … Hettema JM (2016). Meta-analysis of genome-wide association studies of anxiety disorders. Molecular Psychiatry, 21(10), 1391–1399. doi: 10.1038/mp.2015.197 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Panksepp J (2006).Emotional Endophenotypes in Evolutionary Psychiatry. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 30(5), 774–784. doi: 10.1016/j.pnpbp.2006.01.004. [DOI] [PubMed] [Google Scholar]
  35. Ramsawh HJ, Chavira DA, Kanegaye JT, Ancoli-Israel S, Madati PJ, & Stein MB (2012). Screening for adolescent anxiety disorders in a pediatric emergency department. Pediatric emergency care, 28(10), 1041–1047. doi: 10.1097/PEC.0b013e31826cad6a. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Scaini S, Ogliari A, Eley TC, Zavos HM, & Battaglia M (2012). Genetic and environmental contributions to separation anxiety: a meta-analytic approach to twin data. Depression and anxiety, 29(9), 754–761. doi: 10.1002/da.21941. [DOI] [PubMed] [Google Scholar]
  37. Scherrer JF, True WR, Xian H, Lyons MJ, Eisen SA, Goldberg J, Lin N, & Tsuang MT (2000). Evidence for genetic influences common and specific to symptoms of generalized anxiety and panic. Journal of affective disorders, 57(1–3), 25–35. doi:S0165032799000312. [DOI] [PubMed] [Google Scholar]
  38. Steimer T (2002). The biology of fear- and anxiety-related behaviors. Dialogues in Clinical Neuroscience, 4(3), 231–249. doi: 10.1097/ALN.0b013e318212ba87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Stevenson J, Batten N, & Cherner M (1992). Fears and fearfulness in children and adolescents: a genetic analysis of twin data. Journal of child psychology and psychiatry, and allied disciplines, 33(6), 977–985. [DOI] [PubMed] [Google Scholar]
  40. Trzaskowski M, Eley TC, Davis OS, Doherty SJ, Hanscombe KB, Meaburn EL, … Plomin R (2013). First genome-wide association study on anxiety-related behaviours in childhood. PloS One, 8(4), e58676 10.1371/journal.pone.0058676 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Verhulst FC, & van der Ende J (2006). Assessment scales in child and adolescent psychiatry. CRC Press. [Google Scholar]
  42. Waszczuk MA, Zavos HM, Gregory AM, & Eley TC (2014). The phenotypic and genetic structure of depression and anxiety disorder symptoms in childhood, adolescence, and young adulthood. JAMA psychiatry, 71(8), 905–916. doi: 10.1001/jamapsychiatry.2014.655. [DOI] [PubMed] [Google Scholar]
  43. Williams L, & Holahan PJ (1994). Parsimony-based fit indices for multiple-indicator models: Do they work? Structural Equation Modeling, (1), 161–189. [Google Scholar]

RESOURCES