The central importance of anxiety in common mental disorders

Despite the fact that the diagnosis of ‘anxious depression’ does not exist in either major classification, many investigators have recognised that it exists (Angst and Dobler-Mikola, 1985; Clayton et al., 1991; Fava et al., 2004; Rao and Zisook, 2009). Ormel et al. (1993) have shown that anxiety and depressive symptoms show synchrony of change, and Malhi et al. (2002) have stressed the importance of recognising the anxious face of depression. This has been at least partly remedied in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), where clinicians are required to use an ‘anxiety specifier’ whenever depression is diagnosed, relating to whether the depressive symptoms are accompanied by current anxious symptoms.

Anxious symptoms cut across all common mental disorders (CMDs) except non-anxious depression and, in the case of both unipolar and bipolar depression, tend to make the outlook worse and suicide more likely (Goldberg and Fawcett, 2012). Leckman et al. (1983a) studied over 1300 first-degree relatives of 215 unipolar depressives, and found that if an anxiety disorder was present in addition to the depression, the relatives were at increased risk of depression, whether or not the anxious symptoms were present at the same time as the depressive symptoms. In contrast, the relatives of individuals with non-anxious depression were at no greater risk of depression, with or without anxiety, than normal controls. A related study showed that when depression was associated with symptoms of panic, relatives were more than twice as likely to have major depression, panic disorder, phobia and/or alcoholism than the relatives of probands with major depression without any anxiety disorder (Leckman et al., 1983b). A later family study showed that anxious depressives were more likely to have alcohol abuse, anxious personality and dramatic personality than pure depressives or controls, and were also much more likely to be diagnosed with alcohol abuse (Reich, 1993).

Goldberg et al. (2013) have shown that the severity of anxious forms of depression is not entirely due to the presence of the additional anxious symptoms, since, even if these are discounted, the symptoms of depression are more likely to be particularly severe (e.g. early morning waking, weight loss, psychomotor retardation, diurnal variation of mood and guilt). Where the psychopathology among the parents of people with depression is concerned, those with non-anxious depression come from parents with depression on its own or accompanied by other anxiety disorders, excluding generalised anxiety disorder (GAD), while those with anxious depression come from parents with co-morbid GAD and major depression, with GAD alone or with other anxiety disorders on their own, and with mania. Finally, anxious and non-anxious depressions have different personality characteristics: anxious depressives are more likely to be harm avoidant and less likely to be novelty-seeking than controls, while non-anxious depression are no more harm avoidant than controls, but are more likely to have a high value for novelty seeking (Goldberg et al., 2013).

Both the International Classification of Diseases (ICD) and the DSM classifications conflate these two forms of depression, since the concept of ‘major depression’ defines a heterogeneous collection of depressive states, creating a bogus order out of anarchy. Our diagnostic rules come at a price: not only are anxious symptoms ignored as though they do not exist, but even worse we include dissimilar depressive states in treatment studies, rather than acknowledging that inert, apathetic depressions respond differently than anxious, agitated depressions (Fava et al., 2008). Since non-anxious depressions respond better to treatment than anxious depressions, their presence in a study that does not discriminate between the two forms of depression will provide a more optimistic outcome compared to a study that considered only anxious depression.

Our classifications allow us to consider anxiety with depression only if it has lasted at least 6 months when we arbitrarily declare that the patient has two different illnesses. Those described as depression without GAD may have been experiencing anxious symptoms for anything up to 6 months, so the classifications do not really recognise a concept of non-anxious depression. In general medical settings, if the duration of anxiety is reduced to 1 month, the prevalence of both anxiety and anxious depression greatly increase; however, depression without anxiety becomes the least common mental disorder (Goldberg et al., 2011). Two cohort studies (Moffitt et al., 2007; Richards and Goldberg, 2008) have shown that those who develop co-morbid major depressive disorder (MDD) and GAD have had more adverse early stressors than either disorder on its own. The family studies quoted above indicate that they also are likely to have other family members with anxiety disorders, and are likely to have higher scores on neuroticism, harm avoidance or negative affect inventories.

While neuroticism has an important heritable component, it also has a variable component affected by adverse environmental conditions and current episodes of common mental disorders, and Ormel et al. (2013b) conclude that ‘the utility of neuroticism is mainly that it is a highly efficient marker of risk, both genetic and non-genetic, for developing psychopathology’. The heritable component has not been eliminated from the gene pool because it relates to preference in threat-management strategies, with individuals high in neuroticism favouring strategies that avoid or pre-empt threat (Matthews, 2004).

Ormel et al. (2013a) conclude that neuroticism:

reflects individual differences in brain circuits involved in perception of and cognitive control over negative stimuli … the neural evidence matches the psychological findings, which associate neuroticism with a negative bias in attention, interpretation and recall of information, increased reactivity, and ineffective coping, and is consistent with findings of decreased cardiovascular flexibility.

Since major depression and GAD have the same set of genes, it remains to ask what these common genes control. They might control a general tendency to experience internalising disorders, or alternatively they might be largely concerned with anxiety. Flint (2004) favours the first possibility, arguing that the genes act on ‘measures of emotionality’, and shows that the same genes influence variation in both rodent and human phenotypes. There is also some evidence in favour of the latter possibility. This would see the common genes basically affecting anxiety, with the important caveat that anxiety is itself a major influence on vulnerability to depression – so the two alternatives are really not that different. Some support for the anxiety hypothesis comes from Silberg et al.’s (2001) demonstration that the genetic vulnerability to depression (in girls) manifests as anxiety early in adolescence, playing a role in depression only during middle-late adolescence.

In addition to the genes that are shared between GAD and major depression, Hettema et al. (2005) have shown that another set of genes partially determine phobic illnesses, but major importance must be assigned to the common genes that contribute the heritable component to neuroticism. The effects of current anxiety appear to be as important in bipolar disorders as in the common mental disorders, and they are associated with a wide range of other mental disorders. When non-anxious people become depressed, adverse environmental circumstances can still release depression, but the possible additional effects of a set of genes relating specifically to depression remain to be determined.