Biomarkers in schizophrenia: we need to rebuild the Titanic

Lawrie et al have made a remarkable effort to take stock of the number of current clinical and neurobiological measures which may serve as potential objective diagnostic and prognostic markers, and may move future clinical, therapeutic and pathophysiological re- search in schizophrenia in a promising direction. They argue that replacing current diagnostic criteria for this illness by another set of subjective criteria would be comparable to rearranging deck chairs on the Titanic. We cannot agree more, and believe that we should look at salvaging the Titanic itself.

The major challenge in developing biomarkers of diagnostic value lies in the limitations of the current diagnostic and classificatory approaches. While the current diagnostic approaches have clearly improved reliability of diagnoses with the recurrent revisions of the DSM, the validity of disorders such as schizophrenia remains in question.

First, the content validity of the schizophrenia construct is seriously limited by the substantive heterogeneity of the disorder in cross-sectional presentations, neurobiological characteristics as well as the etiological factors implicated 1. It is commonplace in the schizophrenia literature for authors to invoke heterogeneity as an explanation for inconsistent findings. Heterogeneity must be viewed as a problem to be addressed rather than as an explanation or a solution, and is the strongest reason to revisit the long-entrenched and inadequate conceptualization of this disease entity 2.

The predictive validity of schizophrenia as a construct is hampered by the fact that the longitudinal course of this illness is highly variable 3,4,5, as is the response to different treatments 6. Furthermore, the discriminate validity of schizophrenia is limited by its blurred boundary with other major disorders such as bipolar disorder. Overlap between these disorders is seen in neurobiology, genetics, symptomatology as well as treatment response, posing a central challenge to the century-old Kraepelinian view that these are distinct illnesses 7. At the heart of this debate is the core concept of schizoaffective disorder as an entity that combines the features of both illnesses. Psychiatric disorders generally do not meet the time-honored dictates that symptom constellations (syndromes) would have specific pathology which would lead to specific etiology. In this context, Robins and Guze 8 proposed four tenets of a valid psychiatric diagnosis. These include the need for a distinct signature in phenomenology, course, family history, and biology. Schizoaffective disorder fails to meet these criteria, being characterized by having overlaps with schizophrenia and bipolar disorder in each of these domains 6,9. The interface between schizophrenia and the continuum of “health” is also fuzzy, leading to the intermediate syndromes of schizotypal and brief psychotic disorders.

An increasingly held view is that the pathophysiological heterogeneity of schizophrenia may be resolved by elu- cidating independent families of intermediate phenotypes that traverse across structural, functional, neurochemical and molecular domains, and map on to psychopathological dimensions, but are agnostic to diagnostic categorization 10. As progress is made toward these goals, it is possible that the current entity of schizophrenia will be deconstructed and rebuilt as phenotypically overlapping, but etiopathologically distinct component entities. Biomarkers of the kind Lawrie et al review may be of better value to identify and “diagnose” such entities, perhaps in the not too distant future.

Lawrie et al suggest other key scena- rios beyond diagnosis where current clinical practice operates in the dark: early detection and predicting response to treatment. It is in these venues that the application of our understanding of the pathophysiology of schizophrenia may make an earlier impact in the clinical world. The ability to identify the cohort that is likely to develop these disorders may enable effective preventive interventions with non-pharmacologic means such as cognitive behavior therapy and cognitive remediation, and pharmacological interventions such as omega3 fatty acids and low-dose atypical antipsychotics. One can also envision in the relatively near future biomarker screens that may predict treatment response and side effects irrespective of diagnosis.

In conclusion, the paper by Lawrie et al provides a useful, quantitative appraisal of the state of our understanding of schizophrenia as we now know it and how that understanding impacts clinical care. Some of the more prognostic issues outlined by the authors regarding early identification and prediction of outcome have the potential to be dramatically affected by our ability to understand this disease in biological terms. We are in agreement with their conclusion that diagnosis by biomarkers is not currently feasible and would add that, for various reasons mentioned above, this particular issue may not be where our biological understanding of schizophrenia makes the most immediate impact in the clinical world. That may change, however, as our current Titanic-like construct of schizophrenia gives way to component entities defined across phenomic, genomic, enviromic and endophenomic dimensions 11.