When . . . you see someone pale with worry . . . this man is disordered in his desires and aversions.
—Epictetus, Discourses
Let’s say a patient comes to you with a recent history of a single florid panic attack, in the context of giving a speech before an audience of 2000 colleagues. I don’t mean a case of the “jitters”—I mean a 10-minute episode of palpitations, shaking, sweating, choking, dizziness, derealization, and the belief that he is dying. Since you are remarkably empathic, and have had some public speaking anxiety yourself, you think, “I can understand how someone could have an attack like that, under those circumstances.
Let’s hold off on suggesting any formal diagnoses (a panic attack is not a formal DSM diagnosis—only a “building block” for one). Was this episode normal and non-disordered anxiety, because it is understandable to you? What about a patient with the identical set of symptoms, in the context of, say, hanging by his fingers over the edge of a cliff? If you can understand the occurrence of a panic attack in this context, was it therefore normal? These may sound like very theoretical questions, but they go to the heart of what we think of as normal or disordered, in both psychiatry and general medicine. How we answer these questions also has important implications for what we mean by the term “false positive” in psychiatry, and what categories we create for DSM-5
At the last annual meeting of the APA—where I had just spoken on the experiential differences between grief and major depression—a very well-respected senior researcher in the audience rose to comment, evidently quite perturbed. He expressed great surprise at my claim that an explanatory context shouldn’t determine our clinical assessment of disorder or abnormality. My critic gave the example of someone who has a full-blown panic attack while hanging by his fingers, over a steep cliff. Surely, he insisted, context is critically important in such a case. After all, the context explains the person’s panic attack, and thus renders the attack non-pathological.
This is a perfectly plausible position, and probably represents the prevailing opinion among the general public. Indeed, many clinicians may be inclined to say, “Hey, I’d have a panic attack, too, if I were hanging by my fingers, over a cliff!” So calling such a panic attack normal is just common sense. Maybe so—but as Einstein once reminded us, “Common sense is the collection of prejudices acquired by age eighteen.” Science is the systematic testing of “common sense” assumptions against the range of alternative theories.
In my view, the hypothetical panic attack on the precipice is inherently pathological and disordered. And this “disorderness”—that state in which healthy and adaptive organismic function is disrupted—is not mitigated by any explanatory context. Furthermore, I want to suggest that “explanatory context” is usually a misleading guidepost, in so far as the determination of disorderness is concerned. It leads us to erroneous conclusions in other areas of psychiatry, besides panic attacks, such as whether to regard bereavement-related major depressive syndromes as instances of normal sadness or of bona fide MDD.1
And so, I want to suggest that the general concept of disorderness in psychiatry ought to be—with very few exceptions—non-contextual. But before my psychodynamically oriented colleagues recoil in horror, I hasten to add that context is critically important in working psychotherapeutically with patients. After all, psychotherapeutically speaking, there is a world of difference between a severely depressed patient who has just lost a loved one and an equally depressed patient who is being investigated for bank fraud—although, in my view, both are in a disordered state and deserve professional treatment.
The background story: panic on the precipice
Recently, I came across an article that may have been the genesis of my distinguished colleague’s “hanging off the edge of a cliff” scenario. In his 2007 review of the book The Loss of Sadness,2 Dr Kenneth Kendler3 wrote:
If an individual experience[s] a full-blown panic attack when . . . he looses his grip and falls 40 feet before his rope catches him . . . no psychiatrist I know would consider this to be a psychopathological phenomenon. A panic attack is not—in and of itself—psychopathological. It only becomes pathology when it occurs in certain contexts—at times and in places when it should not. Thus the diagnostic status of panic disorder is inherently contextual. It is not a disorder in and of itself but only in certain contexts. . . .
Later in his review, Kendler alludes to what he takes to be a unanimous consensus among psychiatrists, “. . . our all agreeing that the climber dangling from the rope has a clearly ‘understandable’ and hence non-pathological panic attack.” He then contrasts panic attacks with, for example, a bizarre delusion, such as, “A hard drive has been installed in my head by aliens. . . .” He regards the latter as inherently pathological. But, regarding the panic attack, is Kendler correct?
On a purely pragmatic view of psychopathology, I believe he is correct. Undoubtedly, no psychiatrist would say to our mountain climber, after his cliff-hanger panic attack, “You need psychiatric treatment. Please set up an appointment with me right away!” Nor would many competent psychiatrists say, “You are likely to need psychotherapy and perhaps medication, given that you experienced this panic attack.”
So, in terms of clinical praxis, Kendler is right to claim that the panic-on-the-precipice scenario is not an instantiation of psychopathology—at least in the sense that the term “psychopathology” is typically used in the psychoanalytic literature; ie, as a disturbance of internalized objects, unresolved unconscious conflicts, use of primitive ego defenses, etc. Kendler is also technically correct in noting that panic disorder (like its building block, the panic attack) is “inherently contextual,” in the limited sense that DSM-IV and DSM-5 criteria require that panic disorder be characterized by recurrent and unexpected, ie, spontaneous, panic attacks. Unexpected attacks are, in essence, contextless attacks—ones that come on “out of the blue.” The clear implication is that there is such a thing as expected panic attacks. In DSM-IV, “expected” implies that the attack is associated with a situational trigger, such as a cue or reminder of a previous trauma—ie, the attack occurs in an understandable context.
It’s not clear how the framers of DSM-IV or DSM-5 would classify the panic attack in our cliff-hanger scenario. But there are no compelling clinical reasons for viewing a context-based panic attack as non-pathological. Thus, I believe Kendler erred in suggesting that the mountain climber’s panic attack was “. . . ‘understandable’ and hence [a] non-pathological panic attack.” The problem is with the use of the word “hence.” That an event is understandable does not, by itself, render the event non-pathological. (Kendler, of course, is well aware of this with respect to major depressive symptoms in the context of recent bereavement, and has so argued on the DSM-5 Web site.4) Similarly, I believe Jerome Wakefield errs when he comments on the Kendler scenario, arguing that the mountain climber’s panic attack “was normal because that is precisely the context in which such intense anxiety experiences were biologically designed to occur [italics added].”5
I know of no empirical evidence that human beings are biologically “designed” to experience panic attacks in any circumstance or context—precipice or no precipice. Nor am I aware of any evidence that such intense anxiety in objectively dangerous situations is somehow advantageous to the human organism. In my view, panic attacks do not demonstrate biological design, but biology gone awry. We should not confuse anxiety with fear, which is a realistic and adaptive emotion in the face of some objective, external threat—such as a Mack truck heading straight for your car. Unlike ordinary fear, panic attacks do not prepare the endangered person for appropriate defensive action—rather, they usually incapacitate him.
Richard Maddock has pointed out that it is theoretically possible for a patient to meet DSM-IV panic attack criteria with only 4 of 13 possible symptoms (eg, tachycardia, sweating, sensations associated with increased respirations, and a fear of dying). In theory, under some threatening circumstances, these particular fear-related symptoms might be adaptive—but patients with such limited panic symptoms are almost never seen in clinical practice. Maddock notes: “Although the simple DSM-IV definition of a panic attack can capture some adaptive fear or stress responses, I believe this has no bearing on clinical practice” (personal communication, November 25, 2012).
It is erroneous to claim that labeling a panic attack as “pathological” or “disordered” represents a “false positive,” if the attack occurs in an understandable context.5 Indeed, the entire notion of a false positive in psychiatry rests on an unproved ontological assumption—ie, that there exist natural types of disease entities (taxons) defined by necessary and sufficient criteria, against which diagnostic claims may be deemed false.6 Lilienfeld notes, “Such terms as ‘false positives’ and ‘overdiagnosis’ carry no ontological meaning in the absence of a taxon [a genuine category that exists in nature], as they presume the existence of at least some true breaking point in nature.”7
One possible reason for confusion among contextualists is the overlap in DSM criteria for panic attack with what traditionally has been called the fight or flight response, or the general adaptation syndrome (GAS), first characterized by Selye.8 Most physiologists would indeed regard the GAS as an evolution-based adaptation to acute stress. But despite some overlapping features with panic attacks (eg, adrenergic activation, tachycardia, increased respiratory rate, sweating), the GAS is a fundamentally different process. For example, the GAS usually lacks such panic-specific features as a feeling of choking, chest pain, nausea, dizziness, fear of going crazy, derealization, or paresthesias—none of which appears adaptive.
Although research is still incomplete, there is reason to believe that the physiology of a panic attack differs from that of the prototypical fight or flight response. Maddock, one of the foremost researchers in the area of panic disorder, notes:
. . . panic attacks are dysfunctional, while ordinary fight or flight responses are generally adaptive. From the perspective of physiological data supporting this distinction, one difference immediately comes to mind. In the GAS response, elevated [serum] cortisol is the norm. However, elevated cortisol is distinctly the exception during panic attacks (personal communication, November 16, 2012).
Indeed, panic attacks appear to share more features with acute coronary syndrome—basically, myocardial ischemia—than with the GAS.9 Moreover—unlike the GAS—panic attacks predict onset and severity of psychopathology beyond anxiety disorders.10 Whereas the GAS is adaptive—at least, in its earliest stages—it is far from clear that any panic attack, under any circumstances, is ever normal or adaptive. Insofar as it is experienced as terrifying, crippling, death-dealing, or debilitating, a panic attack is always pathological (from pathos, meaning “suffering”) and disordered. That said, a single panic attack does not qualify as a discrete disorder, nor does it merit diagnosis of a specific disease entity. Thus, to diagnose panic disorder, DSM-5 requires additional features, such as recurrent attacks and maladaptive changes in behavior.
The second part of this article further explores the role of context in psychiatric diagnosis.
Acknowledgments—My sincere appreciation to Dr Richard Maddock and Dr Sidney Zisook for their helpful comments and suggestions on early drafts or portions of this essay.
