In an enlightening meta-analysis, Pappa & Dazzan (Reference Pappa and Dazzan2009 a) presented data from a series of medication-naive schizophrenia patients, and reported cumulative results to suggest that spontaneous movement abnormalities (i.e. dyskinesia and Parkinsonisms) are part of pathogenic disease progression of psychosis. In support, our group observed that among a group of high-risk individuals, after controlling for baseline symptoms and medications, the presence of spontaneous dyskinesia predicted an exacerbation in psychotic symptomatology 1 year later (Mittal et al. Reference Mittal, Tessner, Trottman, Dhruv, Esterberg, Simeonova, McMillan, Murphy, Saczawa and Walker2007). Given that spontaneous dyskinesia and Parkinsonisms may be associated with constitutional vulnerability to schizophrenia, it is clearly worthwhile to discuss this ‘marker’ within the framework of designing a new DSM.

In their recent letter, van Harten & Tenback (Reference van Harten and Tenback2009) have suggested several excellent points when weighing the benefits of including spontaneous movement abnormalities in DSM-V (in terms of prevalence, predictive value, biological basis, and the extent to which specificity of the symptom adds to the value of the criterion). The authors note that the prevalence of spontaneous movements (estimated to range from 13% to 20% for dyskinesia; 18–28% for hypokinetic/Parkinsonian signs based on instrumental measurements) (rates obtained using clinical scales are lower: 9% for dyskinesia; 17% for Parkinsonian signs; see Pappa & Dazzan, Reference Pappa and Dazzan2009 b for details) approaches the ideal base rate for a criteria ‘A’ classification, and exceeds several less common existing criteria (e.g. thought disorder, catatonia, affect abnormalities). They also suggest that the hypothesis of a shared DA dysfunction provides a compelling rationale in terms of a sufficient biological basis.

However, there are also several problems with the suggestions posed by the authors. Specifically, with regard to predictive value, the authors suggest that only in schizophrenia and possibly schizotypal personality disorder (SPD) do medication-naive patients exhibit these movements. van Harten & Tenback (Reference van Harten and Tenback2009) then go on to note, ‘Another factor that supports the inclusion of movement disorders in antipsychotic-naive patients with schizophrenia as an A criterion for schizophrenia is that is it highly specific. All other DSM criteria of schizophrenia are non-specific and non-pathognomic, i.e. many symptoms are also prevalent in affective disorders' suggesting that because movement abnormalities are unique to schizophrenia, as opposed to other disorders (that may include psychotic features), the movements are diagnosis-specific. Further, they argue that because spontaneous movement abnormalities are relatively easy to assess in a variety of clinical settings, they may serve as a useful marker and effective DSM-V criterion.

We offer the following cautions about this proposal. First, while there is evidence to suggest that hyperkinetic and hypokinetic movements are unique to the psychosis spectrum among psychiatric disorders, there has not been sufficient research comparing rates of movement abnormalities in patients with different subtypes of psychosis (e.g. schizophrenia, bipolar with psychotic features, depression with psychotic features), and consequently no strong empirical basis for the claim of diagnostic specificity. Indeed, the scant available evidence suggests that movement abnormalities may be common among several disorders that share psychotic features. For example, Mittal & Walker (Reference Mittal and Walker2007) found that the presence of dyskinetic movements in high-risk populations predicted conversion to a range of psychotic disorders, including schizophrenia, schizoaffective disorder, depression with psychotic features, and bipolar disorder with psychosis. In another study examining rates of spontaneous Parkinsonisms among antipsychotic-naive patients with different psychotic disorders, researchers observed these signs across psychotic spectrum disorders, and in fact, the Parkinsonisms were significantly more prevalent in individuals with affective psychosis (i.e. bipolar disorder and depression with psychotic features) and schizoaffective disorder, when compared to patients with non-affective psychosis such as schizophrenia (Chong et al. Reference Chong, Subramaniam and Verma2005). Taken together, evidence suggests that spontaneous movement abnormalities have poorer specificity than proposed by van Harten & Tenback (Reference van Harten and Tenback2009) and this limits the value of these signs as a potential A criterion, although it does not necessarily preclude the inclusion of these spontaneous movements under a different diagnostic classification strategy.

With regard to van Harten & Tenback's (Reference van Harten and Tenback2009) suggestion to place movement abnormalities in the criteria ‘A’ category, we would also raise some questions about classification and stability. For example, should spontaneous dyskinesia and spontaneous Parkinson's signs be considered in the same or in separate categories? Although these two classes can co-occur due to the nature of direct and indirect striatal-pallido pathways (Duval et al. Reference Duval, Fenney, Jog, Groenewgen, Voorn, Berendse, Mulder and Cools2009), each is presumed to reflect different aspects of striatal dopamine (DA) activity, where hyperkinetic movements (tremor, athetoid movements, chorea, ballism) are associated with high levels of DA, and hypokinetic (Parkinsonian type movements: akinesia, bradykinesia, rigidity) are associated with low striatal DA (Delong & Wichman, Reference Delong and Wichman2007). It is important to note that both categories are indicative of broad basal ganglia DA dysfunction, and this may be the pathogenic factor at play. However at present there is insufficient empirical evidence informing our understanding of common and distinct neurological underpinnings between the two movement subtypes and the etiology of psychosis, and until these relationships are clearer, we should not rule out or rule in any definitive category split.

Finally, several more general points should be considered when discussing an A-criterion movement abnormality for schizophrenia. First, it appears that there are non-specific genetic risk factors for psychosis, broadly defined, rather than for schizophrenia specifically (Cardno et al. Reference Cardno, Rijsdijk, Sham, Murray and McGuffin2002). Second, there is significant heterogeneity among psychotic patients (with and without schizophrenia) in the inherited and acquired genetic factors that confer vulnerability to their illness (Riley & Kendler, Reference Riley and Kendler2006; McClellan et al. Reference McClellan, Susser and King2007). These and other findings suggest that current DSM categories are not valid with respect to etiological distinctions. More specially, the evidence suggests that schizophrenia, bipolar disorder with psychotic features, and major depression with psychotic features share a variety of etiological factors, both genetic and environmental (Cardno et al. Reference Cardno, Rijsdijk, Sham, Murray and McGuffin2002; Pini et al. Reference Pini, de Queiroz, Dell'Osso, Abelli, Mastrocinque, Saettoni, Catena and Cassano2004). Finally, while it is true that movement abnormities have been observed at a rate that is comparable to some of the current diagnostic criteria for schizophrenia, patients with clinical and/or instrumentally measured movement abnormities, nonetheless, likely constitute only a subgroup (Neumann & Walker, Reference Neumann, Walker, Ollendick and Prinz1996). It is plausible that this subgroup is also characterized by some distinct etiologic factors and pathophysiological processes. From a dimensional perspective of psychotic disorders (see van Os & Kapur, Reference van Os and Kapur2009), these points highlight the potential for spontaneous movement abnormalities to be considered separate dimension as opposed to a new A criterion for schizophrenia.

Although the inclusion of movement abnormalities as a DSM-V criterion for schizophrenia may not be justified, the use of this phenomenon as a biomarker for the new ‘psychosis risk syndrome’ is promising. Because researchers have developed standard measures for diagnosing high-risk syndromes (e.g. the Structured Interview for Prodromal Syndromes; Miller et al. Reference Miller, McGlashan, Rosen, Somjee, Markovich, Stein and Woods2002), it is now possible to identify groups of high-risk adolescents/young adults of whom approximately 35% will convert to a psychotic disorder in a 2-year period (Cannon et al. Reference Cannon, Cadenhead, Cornblatt, Woods, Addington, Walker, Seidman, Perkins, Tsuang, McGlashan and Heinssen2008). As a majority of these individuals are not yet medicated, the movement abnormalities may serve an excellent marker of emerging pathology (see Pappa & Dazzan, Reference Pappa and Dazzan2009 b for a detailed discussion of issues of differentiating spontaneous versus medication-induced Parkinsonisms and dyskinesia). Indeed, our research has shown that the relationship between movements and symptoms increases in magnitude as high-risk individuals progress through the prodromal period towards the mean age of onset (Mittal et al. Reference Mittal, Neumann, Saczawa and Walker2008) and further, the presence of baseline movement abnormalities differentiates those high risk individuals who eventually convert to psychosis (Mittal & Walker, Reference Mittal and Walker2007).

We encourage further discussion and serious consideration of research evidence, as well as practicality; an empirically informed DSM-V stands to benefit clinicians, researchers, and patients.