Three decades of research on post-traumatic stress disorder (PTSD) has informed our understanding of post-traumatic psychiatric morbidity. At the same time, the very research spurred by PTSD's introduction in DSM–III has come to challenge almost every aspect of the construct's originating assumptions. ^{Reference Rosen and Lilienfeld1,Reference Rosen, Spitzer and McHugh2}

PTSD and DSM–V

The current state of affairs surrounding PTSD is reflected in conflicting proposals for how the syndrome should be operationalised in the forthcoming and fifth edition of the DSM (DSM–V), due for publication in 2013. Proposals for how to distinguish traumatic events (Criterion A) from more ordinary stressors have included encouragements to better adhere to current definitions, ^{Reference Weathers and Keane3} modifications to current wording, ^{Reference Kilpatrick, Resnick and Acierno4–8} and the radical suggestion that Criterion A should be eliminated entirely. ^{Reference Brewin, Lanius, Novac, Schnyder and Galea9} In a recent posting on the internet, members of the DSM–V workgroup on PTSD proposed the approach of modifying earlier definitions. In their proposed draft criteria, now available for public comment, the subjective component of a traumatic event (Criterion A2), first introduced in DSM–IV, ¹⁰ is eliminated altogether, whereas objective aspects of life-threatening trauma are reinforced. We applaud these suggestions because they may reduce the problem of ‘criterion creep’ that has been associated with recent DSM operationalisations of PTSD. ^{Reference Rosen11}

Still, the committee's proposal for Criterion A does not resolve or even address serious problems with PTSD's underlying assumption of a distinctive and specific etiology. ^{Reference Rosen and Lilienfeld1} It also is telling that committee members perceived the need to specify that watching television or films should not qualify as a traumatic event. This rather remarkable qualification represents a milestone in the history of psychiatry: heretofore it was not thought necessary to specify that post-traumatic psychiatric disorder did not result from exposure to media (e.g. television shows) that individuals freely choose to watch. That such a statement was believed necessary is further testament to the ‘Criterion A problem’. ^{Reference Weathers and Keane3}

In the absence of a coherent position on the question of specific etiology – a position that the DSM–V proposal does not address directly or indirectly – the validity of PTSD largely rests on the distinctiveness of its clinical syndrome. Yet PTSD's symptom criteria (Criteria B–D) remain as controversial as Criterion A, largely because of substantial overlap with other disorders (e.g. specific phobia, depression, dissociative disorders). ^{Reference Rosen, Spitzer and McHugh2} To address this concern, some have proposed the elimination of non-distinctive symptoms. ^{Reference Spitzer, First and Wakefield7} At the same time, the question regarding which of PTSD's symptoms best identifies the syndrome remains unclear. There has been the suggestion that re-experiencing symptoms (Cluster B: e.g. nightmares) are central because they involve content related to the traumatic event. ^{Reference McNally5} Others have found that re-experiencing symptoms are non-specific stress responses associated with multiple disorders. ^{Reference Rosen and Lilienfeld1} Further, North et al proposed that the hallmark symptoms of post-traumatic morbidity involve avoidance and emotional numbing (Cluster C). ^{Reference North, Suris, Davis and Smith6} A more sweeping alternative suggests that PTSD often results from emotions such as anger, guilt and shame, and therefore is not primarily a fear- or anxiety-based condition. ^{Reference Resick and Miller12} This viewpoint argues for an entirely new classification category that encompasses a spectrum of traumatic stress disorders, one of which would be PTSD. Most recently, committee members for the DSM–V provided a listing of 21 possible symptoms and signs, grouped into four (rather than the current three) clusters (intrusion symptoms, avoidance, negative affect, hyperarousal). ⁸

It is instructive to recall that the PTSD clinical syndrome was first operationalised in DSM–III ¹³ by only 12 symptoms, grouped into three clusters. This arrangement yielded 135 combinations by which an individual could meet the minimum requisite symptom criteria. In DSM–IV, ¹⁰ 17 symptoms were grouped in the same three clusters, with minimum criteria yielding 1750 combinations. The current proposal for DSM–V, in which 21 symptoms are grouped into four clusters, allows for 10 500 ways to meet minimum requisite criteria! This expansion is beyond anything experienced for other diagnoses. Minimum criteria for diagnosing major depressive episodes, for example, allowed for 70 combinations in DSM–III, 112 combinations in DSM–IV, and essentially no new combinations in DSM–V. Minimum criteria for diagnosing generalised anxiety disorder allowed for 4 combinations in DSM–III, 20 combinations in DSM–IV, and a proposed reduction to 8 combinations in DSM–V. Once again, PTSD is sui generis in the DSM with regard to the expansion of its diagnostic criteria and continued blurry boundaries. ^{Reference Spitzer, First and Wakefield7} By vastly increasing permitted heterogeneity at the phenotypic level, DSM–V risks increasing etiological heterogeneity, while providing no resolution to the symptom overlap conundrum.

A sound scientific alternative

Continuing controversy over how to operationalise PTSD in DSM–V has led to the suggestion that the diagnosis might best be relegated to the manual's appendix for experimental criteria sets. ^{Reference Rosen and Lilienfeld1} A concern that such a move would lead to the construct's demise is not warranted, as illustrated by strong interest in Spitzer's proposal for binge eating disorder despite its placement in the appendix of DSM–IV. Yet another approach that makes use of the DSM's appendix for experimental criteria sets is illustrated by the diagnosis of dysthymic disorder. With that diagnosis, an alternative criterion set was listed in the appendix for experimental sets, while extant criteria for dysthymic disorder remained in the main text of DSM–IV.

We believe that use of the DSM's appendix for experimental criteria sets can operationalise PTSD in a manner that encourages research and allows for treatment of a wide range of post-traumatic reactions, while delaying scientifically premature acceptance of any specific proposal. This approach can also serve to remind clinicians that PTSD in its present form should not be reified to the status of a distinct disorder in nature, at least until such time that we better understand the full range of normal and disordered reactions that occur after traumatic and other high-magnitude stressors. ^{Reference Spitzer, First and Wakefield7,Reference Bryant, Rosen and Frueh14,Reference McHugh15}