Prepared by Michael B. First, M.D., DSM Consultant to the American Psychiatric Institute for Research and Education (APIRE), a subsidiary of the American Psychiatric Association
The third diagnosis-related research planning conference in the "Future of Psychiatric Diagnosis: Refining the Research Agenda" series, focusing on Stress-induced and Fear Circuitry Disorders, was held on June 23rd and 24th , at the APA Headquarters in Arlington, VA. Conference co-chairs were Dennis Charney, MD from Mount Sinai School of Medicine in New York City, and Gavin Andrews, MD from the University of New South Wales in Sydney, Australia, and invited participants included 26 scientists from the United States and abroad.
The conference began with presentations designed to provide overviews of the four DSM/ICD disorders covered under the rubric Stress-induced and Fear Circuitry Disorders: (i.e., Posttraumatic Stress Disorder, Panic Disorder, Social Phobia [Social Anxiety Disorder] and Specific Phobia), specifically focusing on some of the diagnostic issues and controversies that have arisen since DSM-IV and ICD-10 were published in the early 1990's.
Matthew Friedman MD, PhD (Hanover, NH) and Elie Karam MD (Beirut, Lebanon) began their presentations on Posttraumatic Stress Disorder (PTSD) with a focus on some of the problems and shortcomings of the current diagnostic criteria. For example, the initial criterion describing exposure to the traumatic event has been criticized for being too restrictive (i.e., should suffering with severe medical problems or psychosis qualify as an event) and for omitting other symptoms occurring at the time of exposure (e.g., dissociation, disgust, physiological symptoms). Furthermore, parameters such as the frequency of occurrence of required PTSD symptoms have a large impact on rates of PTSD and need to be looked at in a variety of settings. In addition, requirements for dysfunction and for symptoms being present from the three clusters (i.e. re-experiencing the trauma, avoidance of stimuli associated with trauma, and post-trauma arousal) as opposed to requiring symptoms from only two clusters might not be essential (at least in some settings) in predicting the comorbidity of PTSD with other clinically significant mental disorders disorders such as Major Depressive Disorder. Other areas in need of research include subsyndromal PTSD (is PTSD better characterized dimensionally?), complex PTSD (the validity of a syndrome occurring in response to chronic stress such as childhood sexual abuse or torture of political detainees) and possible cross-cultural variability. PTSD can also be characterized as a disorder of memory in which traumatic memories either become indelible or else cannot be retrieved. Although structural brain abnormalities (e.g., smaller hippocampal volume) are well documented, it remains unclear whether these findings are a risk factor for developing PTSD or occur as a result (perhaps related to chronic disturbances in the hypothalamic/pituirary/adrenal [HPA] axis). Finally, since only a small minority of individuals exposed to severe trauma develop PTSD, research must focus on resilience as well as risk and vulnerability factors.
Carlo Faravelli MD (Florence, Italy) and Toshi Furukawa, MD, PhD (Nagoya, Japan) next presented on Panic Disorder. There are several important issues regarding the relationship between panic attacks, panic disorder, and agoraphobia. Panic attacks, independent of whether the person has full panic disorder, have been shown to be a marker for greater severity of disorders co-occurring with the panic attacks, suggesting that panic attacks should be given greater diagnostic status, perhaps as a subtype that would be included other disorders (e.g., specific phobia with panic attacks). Although the definition of panic disorder requires that panic attacks come on "out of the blue", evidence suggests that for the majority of patients, the panic attacks are preceded by mild phobic or hypochondriacal symptoms. It may therefore be that panic attacks are better characterized as "unpredictable" rather than "unexpected." Evidence that limited symptom attacks may be as disabling as full-blown panic attacks supports a panic spectrum concept in which panic symptoms are dimensional. More research is needed to clarify the causal relationship between panic attacks and agoraphobia; DSM-IV adopts the view that phobias occur as a consequence of panic attacks whereas the European view is that phobic cognition precedes panic attacks, which then exacerbates the phobic symptoms. Although agoraphobia without panic is rare in clinical samples, it is much more common in community surveys suggesting that it may on a continuum with panic disorder with agoraphobia. Evidence of high comorbidity between panic and bipolar disorder suggests that there may be two kinds of panic disorder: "neurotic panic," which occurs in association with anxiety and depressive disorders, and "bipolar panic." Dr. Furukawa also stressed the importance of interoceptive hypersensitivity (fear of bodily sensations) and in-situation safety behaviors (subtle avoidances) in the assessment of panic symptomatology.
Susan Bogels PhD (Maastricht, The Netherlands) and Murray Stein MD (La Jolla, CA) gave presentations covering various aspects of Social Phobia. Evidence suggests moderate levels of heritability for Social Phobia; however, there is even higher heritability of fear of negative evaluation and behavioral inhibition, traits that predispose to Social Phobia and other anxiety disorders. Although rearing experiences influence the development of Social Phobia, the type of rearing that protects against social phobia in sensitive children is unclear. Peer neglect also appears to play a role in the development of Social Phobia in some individuals. Dr Bogels suggested research into possible new Social Phobia subtypes that might differ based on origins and treatment response. The subtypes would be based on the content of the fear of rejection, including fear of being negatively evaluated because of bodily symptoms (blushing, trembling, sweating), appearance (hair, body shape, clothes), performance (sport, music, speech), and negative feelings about self (being boring, weak, stupid). Dr. Stein noted that the current "generalized" subtype has better reliability than social phobia itself and has prognostic and treatment utility. Additional research is also needed to help set the boundary with normality (i.e., shyness) based on harmful outcomes.
Hans-Ulrich Wittchen, PhD (Dresden, Germany) and Paul Emmelkamp, PhD (Wassenaar, the Netherlands) presented on Specific Phobias. Although the Specific Phobia subtypes have some clinical utility with respect to types of behavioral interventions, there is overall little evidence from etiological and experimental research that these subtypes are sufficiently different to support such a subtle distinction. However, the number of different specific phobias an individual has or the presence of panic attacks may be a vulnerability marker. Beyond diagnostically non-specific vulnerability-stress models, there is little research available about what causes specific phobias to develop. Since onset is frequently pre-pubertal, retrospective data regarding onset of the phobias are problematic. Suggested conditioning pathways (for example the development of phobias secondary to exposure to traumatic event or secondary to learning by observation of others with phobias) seem not to be sufficient explanations. Family genetic and twin studies do suggest a modest genetic influence. Specific phobias are the most common internalizing disorder in childhood and early adolescence. There is a considerable degree of continuity of this disorder into adulthood, although not necessarily as a phobia. Spontaneous remissions of specific phobias are rare; generally most will go on to have some type of anxiety disorder. Furthermore, 50% of individuals with specific phobia in early adolescence will develop a depressive disorder later in life suggesting that specific phobias are a significant risk factor for adult depression. As is the case with the other childhood internalizing disorders, very few individuals receive mental health treatment for their phobias, although CBT for specific phobia is one of the most effective treatments documented. Treatment occurs typically after the development of secondary complications, usually the onset of a depressive episode.
The remaining presentations focused on topics that cut across the various anxiety disorders. The first presentations, by Richie Poulton, PhD (Dunedin, New Zealand) and Danny Pine, MD (Bethesda, MD), addressed the lifetime stability of anxiety disorders. Dr. Poulton presented longitudinal data from the Dunedin cohort, which was formed in the early 1970's and evaluated every few years from age 3 up to age 32: 60-80% of those with a mental disorder at age 26 had a mental disorder at age 18. With the regard to the stress-induced and fear circuitry disorders, the study indicated continuity of the behavior over time, suggesting that it makes sense to reframe most "adult" anxiety disorders as extensions of childhood anxiety disorders. The Dunedin study also allowed for an examination of the risk factors for the development of these disorders (e.g., the risk of developing PTSD is related to early poor cognitive functioning, having a "difficult" temperament, and having been exposed to physical abuse before age 11). Of note, the stress-induced and fear circuitry disorders have both shared and specific risk factors and the specific disorders are distinguishable in terms of their early childhood risk antecedents. In his presentation, Dr. Pine noted that most adult disorders start in childhood but that most childhood disorders, especially specific anxiety disorders, are transient. He suggested that research that aims to understand brain function and the genetic underpinning of childhood anxiety disorders (including gene and environment interactions) requires mapping brain intermediate phenotypes (like behavioral inhibition and threat bias) rather than the anxiety disorders themselves.
Abby Fyer, MD (New York, NY) and Tim Brown, PhD reviewed the genetic and epidemiological data addressing the issue of whether the anxiety disorders form a cohesive group that is distinct from other disorders. Traditional descriptive validity measures (e.g. age at onset, gender distribution, course, treatment response and comorbidity) suggest both overlap and distinctness both among the four anxiety disorders and between anxiety and affective disorders. Data on heritability follow a similar pattern. All anxiety disorders demonstrate moderate heritability, but twin and family studies provide, evidence for both shared (i.e., a common "proneness" to develop anxiety disorder that is inherited) as well as specific genetic and environmental contributions. These data are consistent with the hypothesis suggested by Steve Hyman in the Research Agenda for DSM-V that the different anxiety disorders will each represent clusters of illnesses with overlapping or even distinct genetic and non-genetic risk factors that converge to produce patterns of pathogenesis, symptoms, and course that have a close family resemblance. William Lawson, MD, PhD (Washington, DC) followed up with a presentation on issues of race and ethnicity, noting that there are differences in prevalence among certain anxiety disorders in certain minorities (for example, the NCS-R showed less prevalence among blacks) but raised questions about sampling issues (i.e., there is little minority representation in clinical trials or in biological studies).
Jack Gorman, MD (New York, NY) and Thalia Eley, PhD (London, UK) addressed gene and environment interactions in their presentations. Animal data provide a number of examples in which environmental interventions have a profound impact on genetically-determined temperament to produce significant behavioral and neurobiological changes. For example, consider two species of macaque monkey: pigtail and bonnet with different behavioral characteristics, neurobiology, and rearing patterns. The pigtail species is characterized by hostile behavior and higher corticotrophin releasing factor (CRF) levels and the mothers are very protective of their young, not letting them have any interactions with peers. In contract, the bonnet monkeys are calm, have lower CRF levels, and are reared to encourage peer interaction. Stressing the mothers of newborn bonnets (by exposing them to a variable foraging demand) results in the offspring acting like pigtail monkeys, demonstrating that exposure to a significant environmental stress can alter the natural course of a behavioral phenotype. Of note, in contrast to animals where the negative effects of stress appear permanent, there is much more plasticity in humans; parents can correct for it. Studying the interaction between genes and environment is important because of the known environmental stressors for anxiety and the fact that although genes for depression and anxiety are largely shared, environmental effects are more specific. Gene-environment interactions can be studied indirectly by using twin/adoption data and more directly using molecular genetics approaches. Of particular interest in humans is the serotonin transporter (SERT) gene. Having a version of the gene with two short alleles is linked to alcohol preference, ACTH response to separation, and effects on amygdala when scary pictures are shown. Furthermore if individuals with this double short allele are exposed to stressful life events, they are at increased risk for developing major depression, demonstrating how specific genes can interact with environmental events to produce psychopathology.
Rene Hen, PhD (New York, NY) and Steve Hyman, MD (Cambridge, MA) presented on what is currently known about the neural mechanisms underlying fear and anxiety. The relationship between genes and neural circuits is complex. For example, mutations in the gene encoding the serotonin transporter which is expressed in many brain circuits will have consequences in multiple structures, producing very different phenotypes than mutations in genes expressed only in a particular brain region like the amygdala or hippocampus. Furthermore, besides gene-circuit interactions, there are also development-gene-circuit interactions; wherein genes are expressed in a particular developmental trajectory. For example, the consequences of a commonly studied mutation in the promoter of the serotonin transporter gene appear to result from abnormal expression of this protein during development rather than in adulthood. Another example from the preclinical literature is the 5-HT1A receptor which when blocked during development results in increased anxiety in tests that are hippocampal-dependent (contextual fear conditioning) but not in tests that are amygdala dependent (cued fear conditioning) . It would be interesting to apply such behavioral paradigms to human to assess whether specific anxiety disorders are similarly hippocampal or amygdala dependent.
There are several possible approaches for classifying anxiety disorders: by clinical similarities (e.g., symptoms, signs, and course as is done in current DSM/ICD), by treatment response (which is treacherous, given non-specificity of treatment), by etiology (risk genes and gene/environment interaction), and by pathophysiology (neural and cognitive/affective processes that give rise to the symptoms and signs). Although the latter two have clear advantages over the current descriptive approach, it should be noted that multiple etiologies may produce similar pathophysiologies which are not perfectly coupled to symptoms, signs, and impairment. Furthermore, evidence suggests that DSM/ICD mood and anxiety disorders have overlapping pathophysiologies, including the high rates of comorbidity of between anxiety and mood disorders, sharing the same factor in factor analyses of symptom data sets, sharing genetic risk factors in twin studies and in association studies, sharing developmental antecedents (e.g., behavioral inhibition), and sharing response to antidepressants and benzodiazepines; all suggesting that DSM is not correctly "carving nature at the joints" as originally hoped. A better approach is to consider a functional view of the emotion-processing circuits, perhaps defining anxiety disorders in terms of deficits in circuits that appraise the survival relevance value of stimuli in the environment, activate appropriate physiologic, cognitive, and behavioral responses, facilitate memory formation in multiple circuits, or produce subjective feelings and conscious cognition.
Edna Foa, PhD (Philadelphia, PA) and Richard McNally, PhD (Cambridge, MA) presented on the role of cognition in the development of anxiety disorders. Cognitive processes are mechanisms related to the detection, encoding, storage, retrieval, and utilization of information. Cognitions can be divided into two kinds: evaluations, thoughts, appraisals, or beliefs measured through self-report and cognitive biases measured mostly by behavior elicited through experimental paradigms. Cognitions are thought to play a causal role in the etiology and maintenance of anxiety disorders via negative evaluations and cognitive biases. Anxiety disordered individuals overestimate the probability and cost of negative outcomes in a way that is specific to their disorder, negatively evaluate neutral stimuli or responses; focus on threat information; and remember more negative information. Studies conducting longitudinal observations on PTSD patients demonstrate that negative post-trauma evaluations correlated with PTSD severity, with cognitive predictors accounting for 75% of the variance in PTSD severity compared to 40% with other established variables. However, without studies evaluating cognitions before the trauma had occurred, it is not possible to know whether negative cognitions cause chronic PTSD or are the consequence of it. With panic disorder, high scores on the anxiety sensitivity index (fear of bodily sensations) predicted later panic attacks. Treatment studies examining differences between treated and untreated patients show a reduction of negative evaluations and cognitive biases. These data speak more to mechanisms of maintenance rather than causality. Finally, experimental paradigms that manipulate disorder-related cognitions influence symptom severity (e.g., perceived control during induction of panic-related symptoms reduces panic attacks in panic disordered individuals). The data that most directly speak to the causal role of cognitions in anxiety come from experiments which induce attentional and interpretation biases and then measure reactions to stress. Individuals who were trained to have negative biases show heightened anxiety during stressful events than do untrained individuals.
Ron Rapee, PhD (Sydney, Australia) and Richard Bryant, PhD (Sydney, Australia) presented on the relationship of stress and psychosocial factors in the development of anxiety disorders. It is important to distinguish between distal causes (in which there is a long temporal period between stressor and effect) and proximal causes (in which there is a short temporal relationship); a lot of psychosocial causes that are distal are better considered to be vulnerability factors. Overall, the extent to which stress induces anxiety will depend on vulnerability factors (genetic, cognitive, etc). Prospective studies indicate importance of vulnerability factors influencing response to stressors. Twin and family studies demonstrate that both shared factors (e.g., family factors in children) and unique environmental factors (e.g., exposure to trauma) are important but that shared factors are less important in adult anxiety disorders suggesting that different types of etiological factors may be involved at different developmental stages. Furthermore, environment may be responsible for the specificity of anxiety disorders since genetic factors appear to be relatively non-specific. Consistent gender differences have been seen in various countries for GAD but not Panic Disorder, suggesting that these differences are not likely to be a cultural effect. There are, however, cultural effects on frequency. Social phobia is diagnosed less in Taiwan and Korea than in the West but symptom levels of social anxiety appear to be higher in the East, which may reflect difference in perceived impact (i.e., social anxiety is seen as positive in the East by parents). Regarding stressors, childhood sexual abuse has been linked with a wide range of disorders, including anxiety. Peer victimization may be more likely associated with social anxiety disorder. There is little indication that parent-child relationship factors are causal to any particular disorder; their effects are mostly non-specific.
Scott Rauch, MD (Boston, MA) and Wayne C. Drevets, MD (Bethesda, MD) presented on neural circuits and anxiety disorders, noting that given the early stages of research in this area, it would be premature to reorganize the DSM classification of anxiety disorders based on brain imaging results alone. Human brain imaging will likely provide a critical translational bridge from animal research to clinical populations. In animals, classical fear conditioning involves habituation, conditioning, and extinction, with the amygdala playing a critical role in threat and fear response. In humans, the role of amygdala and hippocampus is analogous but although the prefrontal cortex is involved, the precise location of homologous subregions is less clear. Indelible learning about threat-related stimuli involves amygdala and hippocampus; exaggerated fear responses primarily involve amygdala, impaired extinction involves prefrontal cortex and amygdala; deficient ability to suppress attention/response to disorder-relevant stimuli involves rostral anterior cingulate cortex, and contextual conditional and inability to appreciate safe contexts involves hippocampus. To date, initial brain imaging findings in anxiety disorders suggest group differences in amygdalo-cortical structure and function. Many of these findings, however, are not highly reliable or specific and further research is needed to disentangle neural substrates or risk.
Rachel Yehuda, PhD (New York, NY) presentation focused on the neurochemistry and neuroendocrinology of anxiety disorders. There are currently no biomarkers for any of the anxiety disorders, although there is some evidence for alterations involved in stress when challenge strategies are used. There are a number of possible reasons why we are not getting a clear neurochemical/endocrine signal that will support the fear-stress diagnostic grouping. These include: 1) methodological artifacts (e.g., peripheral markers may not relate brain processes, study design issues like power considerations); 2) problems with the current classification system (e.g., over-inclusive or imperfect diagnostic criteria); or 3) possible flaws in the conceptual foundation (e.g., rather than defining PTSD as an abnormal response to extreme stress, it may represent a specific phenotype that results from an inability to recover from the normal effects of trauma, perhaps due to inadequate glucocorticoid signaling that has pre-traumatic origins).
After these presentations, the participants formed two working groups (one covering Panic and PTSD, the other Social and Specific Phobias) which were given the tasks of identifying strengths and weaknesses in the current diagnostic criteria, identifying promising new hypotheses to make the criteria more valid, and devising a research agenda for the accomplishment of these goals. Research agenda items that emerged from these discussions included: 1) examine anxiety disorder patients and perform morphometric analysis of amygdala, hippocampus, and prefrontal cortex; 2) select cognitive paradigms that map stages of fear conditioning onto neural circuits; 3) correlate the imaging findings with patient symptoms and course; 4) collect existing family and twin data to examine genetic hypotheses about shared risk factors and symptoms; 5) based on imaging studies, design genetic studies of the most robust intermediate phenotypes; 6) consider disentangling symptoms and signs from severity and impairment when doing studies of patients with anxiety symptoms; 7) conduct longitudinal epidemiological and clinical studies using multiple measurement domains including imaging, genetics, environmental and personality/other risk factors and identify predictors for continuation of childhood disorders into adulthood; and 8) identify new physiological measures such as blushing for social phobia.
The conference concluded with a presentation by Edward Nunes, MD (New York, NY) examining the relationship between anxiety and substance use. Evidence suggests that there may be shared risk factors (e.g., children with anxious/depressed symptoms may have a higher risk for subsequent substance use). Furthermore, animal models of addiction demonstrate that exposure to stress increases drug self-administration and prompt reinstatement of previously extinguished drug-seeking behavior. Neuroimaging studies of cue-elicited drug craving suggest activation of anterior cingulate and amygdala (brain areas involved in fear response). The acute effects of many substances are anxiolytic or anxiogenic (e.g., intoxication with alcohol, sedatives, opiates, cocaine, and other stimulants is anxiolytic, withdrawal from alcohol, sedatives, opiates is anxiogenic) perhaps providing a model to better understand neural mechanisms of anxiety.
The full presentations were published in a monograph by American Psychiatric Publishing, Inc., which can be purchased here.