Most multivariate behavior genetic contributions to the classification of mood and anxiety disorders have been based on three general approaches: (1) split an existing diagnostic category into subtypes and examine differences in the causes of the resulting subtypes; (2) look for common causes that cut across the symptoms of several diagnostic categories, using multivariate behavior genetic methods; or (3) combine the first two approaches (i.e., break down a given diagnostic category into clusters of symptoms and then examine their common causes) to determine whether the etiologies of various symptom clusters may be partially overlapping and partially independent.
One example of the splitting approach is the study by Kendler et al. (1996; see earlier), who reported that genetic and environmental influences on various subtypes of depressive illness appeared to have different etiologies. Research with adults suggests that genes may act in a nonspecific way to confer risk for a variety of mood and anxiety syndromes, and the particular syndrome manifested depends either on the degree of genetic risk (e.g., as indexed by number of ill relatives) or on individual-specific, nonshared environmental influences. On the other hand, a few studies (reviewed later) suggest that the mood and anxiety syndromes are etiologically heterogeneous, and different sets of genes and/or environmental risk factors influence different patterns of symptoms within a diagnostic grouping. In this way, behavior genetic research can contribute to the development of an etiologically based classification of psychopathological syndromes. Following, we review multivariate behavior genetic studies that have contributed to our understanding of the boundaries between different mood and anxiety disorders by examining their common and unique causes.
Adults. Kendler et al. (1992c) reported that additive genetic influences largely explained the observed covariation of symptoms of GAD and major depression in women and were entirely shared between the two disorders; nonshared environmental influences accounted for the remaining unique liability variance for these disorders. The finding of a common genetic factor for GAD and major depression (with no unique genetic influences) was replicated by Roy, Neale, Pedersen, Mathe, and Kendler (1995) in an extension to a mixed male and female adult sample. Kendler et al. (1993c) also found that the observed covariation of major depression and phobia symptoms in women was largely attributable to additive genetic influences and that nonshared environmental influences contributed to the unique variance in liability to each of the two disorders. When the same authors investigated sources of overlap among phobias (agoraphobia, social phobia, animal phobia, and situational phobia; Kendler et al., 1992d), they found that the familial aggregation of these phobias in women could be explained by common genetic and nonshared environmental influences and that additional unique genetic and nonshared environmental influences contributed to the variance in liability to each of the specific subtypes of phobia. For most of the phobias (except animal phobia), however, the common nonshared environmental factor contributed a greater share of the liability variance than the common genetic factor. Using the same Virginia-based twin sample, Kendler et al. (1995) examined the causes of comorbid-ity among panic disorder, phobias, GAD, major depression, alcoholism, and bulimia nervosa in women. They reported that the genetic influences responsible for the overlap between major depression and GAD in female twins appeared to be distinct from the genetic influences responsible for the overlap between panic disorder and phobias. Based on the finding of more than one shared genetic factor, the authors speculated that the anxiety disorders may be etiologically heterogeneous. Although provocative, this hypothesis seems hard to reconcile with the earlier findings by this group on common genetic influences on depression and phobias (Kendler et al., 1993c).
Additional evidence for a genetic basis to the overlap of anxiety and depressive symptoms comes from studies using the Australian National Health and Medical Research Council twin Registry, a nationwide, volunteer registry. Jardine, Martin, and Henderson (1984) found that genetic influences accounted for about 50-60% of the observed co-variance between Neuroticism (a personality dimension associated with risk for mood and anxiety disorders), symptoms of anxiety, and symptoms of depression, and that the remaining covariation was attributable to nonshared environmental influences and/or measurement error. Kendler, Heath, Martin, and Eaves (1987) reanalyzed the same data (focusing on anxiety and depressive symptoms but not Neuroticism), and concluded that common genetic influences largely-explained the covari-ance between anxiety symptoms and depressive symptoms, and that nonshared environmental influences primarily contributed to the unique variance in anxiety and depressive symptoms.
The relationship between unipolar affective disorder and bipolar affective disorder has also been examined using twin data. Kendler, Pedersen, Neale, and Mathe (1995) reported that their data were consistent with a multiple threshold model in which the two forms of affective disorder are alternative manifestations of a single continuum of liability, rather than separate phenomena with distinct causes, and that bipolar disorder represents a more severe point on the continuum. Kar-kowski and Kendler (1997) analyzed data from the Virginia-based female twin sample described ear lier and reported that the data supported the conceptualization of unipolar and bipolar affective disorder as points of differing severity along the same continuum of illness (i.e., the two disorders have common causes). The application of the multiple threshold model to the bipolar-unipolar distinction is also consistent with two other pieces of data: (1) epidemiological studies showing that bipolar disorder occurs much less frequently in the population than unipolar depressive disorder, and (2) family studies showing a higher rate of affective disorders in the relatives of bipolar versus unipolar probands (Tsuang & Faraone, 1996). Both the greater familial loading of affective disorders in bipolar patients compared to unipolar patients and the lower population prevalence of bipolar compared to unipolar disorder are consistent with a model in which bipolar disorder is only manifest when a higher threshold of liability (resulting from a high degree of genetic and/or environmental risk) is surpassed.
Evidence from multivariate behavior genetic analyses in adult twin samples suggests that familial causes contribute substantially to the diagnostic overlap (or symptom covariation) among the anxiety and depressive disorders. In addition, the results of these studies suggest that the familial causes underlying the covariation among anxiety and depressive symptoms in adults are largely or entirely genetic. Now we review the results of the few multivariate behavior genetic studies of mood and anxiety disorders in children and adolescents to examine the extent to which the causes of co-morbidity are similar to those found in adults. Children and Adolescents. In contrast to the growing body of data on genetic and environmental influences on the covariation between anxiety and depressive symptoms in adults, very few mul-tivariate behavior genetic studies of similar disorders or symptoms have been conducted with children and adolescents. Recently, two multivari-ate behavior genetic studies examined the causes of covariation among symptoms of various internalizing disorders in children. Eley and Stevenson (1997; see Eley, 1997) found that genetic influences accounted for 80% of the covariation between children's reports of anxiety and depressive symptoms and that shared environmental influences explained the remaining 20% of the covariation. Feigon et al. (1997b) examined the causes of covariation among three different types of anxiety disorder symptoms (separation anxiety symp toms, avoidant disorder symptoms, and somatic complaints) and found that the covariation among the three types of symptoms was largely attributable to common environmental influences (primarily common shared environmental influences). There was no evidence for common genetic influences on symptom covariation, that is, genetic influences contributed only to the variance that was unique to each of the three types of symptoms.
Thus, studies obtained evidence for common shared environmental influences on symptom overlap, although the magnitude of these influences differed greatly between the two studies. An obvious difference, of course, is the finding of substantial common genetic influences on symptom overlap in Eley and Stevenson's but not Feigon et al.'s study. Several differences in design may have contributed to the differences in the findings of these two studies. One difference is that Eley and Stevenson (1997) were interested in the causes of covariation between anxiety and depressive symptoms, whereas Feigon et al.'s (1997b) study was concerned with the causes of covariation among various types of anxiety symptoms. A second difference is that Feigon et al. used parent ratings of symptoms and Eley and Stevenson used child self-report measures. Both studies, however, support the notion that familial causes (whether genetic or shared environmental) are likely to be most important in accounting for the overlap among anxiety and depression disorders in children, as well as in adults. Additional data from new twin samples are needed to clarify the causal mechanisms that underlie comorbidity between anxiety and depressive disorders in children and adolescents.
Was this article helpful?