Abstract
It is a pleasure to have this opportunity to comment on the findings regarding the metabolic risk in psychosis from the second Australian survey (Galletly et al., 2012). The survey has the following distinctive features:
A rigorous and standardised epidemiological methodology in a large population-based sample of adults with psychotic disorders, which in contrast to most previous studies is an unbiased population survey. The sample size (n = 1286 with fasting bloods) is almost double the number of subjects of the very influential CATIE study (n = 687) (McEvoy et al., 2005). The specific diagnosis distribution (ICD-10) was: schizophrenia (n = 857, 47%), schizoaffective disorder (n = 293, 16.1%), bipolar mania (n = 319, 17.5%), depressive psychosis (n = 81, 4.4%), delusional disorders and other non-organic psychoses (n = 92, 5%) and subjects that did not meet full criteria for psychotic disorder (n = 183, 10.1%).
The inclusion criterion was psychosis, rather than a specific diagnosis (e.g. schizophrenia), so the findings can inform service planning and delivery for all people with psychosis.
The frequency of metabolic syndrome in the participants doubled the figure observed in the Australian general population. A significant proportion of subjects were either untreated or unresponsive.
The authors of this study identify that antipsychotic drug (AP) use in 81.6% of the participants is a critical contributing factor to this phenomenon. These agents are, and will be, key components in the treatment of psychosis. Hence, I will focus on this topic and will remark upon some fundamental conclusions from a set of conferences promoted by the University of Alabama, in Newark, NJ, USA, in July 2011(www.norc.uab.edu/contactus).
I will briefly describe some cornerstone studies that have impacted upon our knowledge of this topic so as to allow a better appreciation of the second Australian survey.
Selected cornerstone studies in the field of APs and metabolic dysfunction
The early recognition of the impact of the APs on body weight (BW) (Plananski, 1958).
Isolated reports on this topic in a silent period (Silverstone et al., 1988; Stanton 1995).
The meta-analysis of Allison et al. (1999), showing a differential effect of the APs on BW.
The rigorous study of Dixon et al. (2000), showing a high prevalence of type 2 diabetes in AP-treated subjects.
A model of continuous AP administration in rodents that better resembles the human treatment setting (Remington et al., 2011) and the remarkable effects of the APs in dogs and worms (Ader et al., 2005; Donohoe et al., 2009).
The CATIE study, confirming Allison’s findings (McEvoy et al., 2005).
Studies aimed to dissect the genetic influence on AP effects on BW and metabolism (Lee and Bishop, 2011; Windemuth et al., 2012) and in the response to metformin (Fernandez et al., 2010, 2012).
The study in humans of the neural pathways involved in feeding regulation during olanzapine administration (Mathews et al., 2012).
The development of psycho- educational (Alvarez-Jimenez et al., 2008) and pharmacological strategies (mainly metformin and topiramate) (Maayan et al., 2010) to attenuate the AP-induced metabolic dysregulation.
The study of Tiihonen et al. (2009), suggesting a protective effect of APs on mortality in schizophrenia and the rigorous criticisms by De Hert et al. (2010).
The study of Francey et al. (2010), exploring the proposal of delaying AP treatment in first-episode psychosis and improving non-pharmacological approaches.
The synthesis of de Léon (2012), claiming that evidence-based and personalised medicine need to work together.
None of these studies are devoid of criticism, and most are incomplete, but they demonstrate wonderfully the formidable effort of creative scientists working in a field that was marginal in medicine until recent years.
The Australian study now confirms that a significant proportion of subjects with psychosis have a metabolic dysfunction. This agrees with our own study on the other side of the world (Baptista et al., 2011). Even though we did not find significant differences among our patients with psychosis and the general population regarding the whole metabolic syndrome, we found significant differences in some of its constituting variables such as glucose levels and waist circumference. The two studies are not strictly comparable because: (a) we discriminated by specific mental disorders; (b) concurrently assessed a probabilistic sample of the Venezuelan general population; and (c) there are significant differences in the ethnic distribution.
Hence, more than half a century after the initial reports, we now know that some clinically effective APs induce more metabolic dysfunction than others and that this dysfunction has both genetic and environmental contributions, the latter being related to components such as the quality of diet and physical activity. In the end, the prevalence of some sort of metabolic derangement could be two- or three-fold higher in patients with any psychosis than in the general population.
Jose de Léon (2012) recently brought our attention to the growing tension between evidence-based medicine (EBM) and personalised medicine (PM). While this is mostly evident in the pharmacological field, it also has implications for research in pathophysiology and, importantly, for mental health service provision. Research protocols (mainly randomised clinical trials) and treatment prevention programmes on an epidemiological scale aspiring to achieve the label of EBM will face the challenge of accounting for the (very important) ‘outliers’ and the recently recognised issue of drug response heterogeneity. PM, which is basically what every skilled clinician practises every day, meets the challenge of creating reliable research protocols that produce some practical generalisations and heuristic hypotheses.
Outliers are a critical issue in the field of AP-induced metabolic dysfunction. For example, some subjects undergoing prolonged treatment with clozapine or olanzapine may have stable BW and metabolic profile; however another patient may have marked metabolic adverse effects, e.g. a 24-year-old lady who gained 48 kg after 1 year of olanzapine administration and had a significant triglyceride increase after 3 weeks of treatment. In some individuals metformin can significantly reverse AP-induced BW gain and hyperglycaemia, whereas in others it may have a paradoxical effect. Some patients report a remarkable appetite increase after olanzapine administration, whereas others are unable to report any appetite change, etc.
A database such as that obtained in the Australian survey appears ideal for promoting a dialogue between EBM and PM. On the one hand it will allow for EBM-based counselling to a fragile clinical population needing medical attention. On the other hand, with the assistance of behavioural and molecular genetics, it will identify ‘outliers’ to aid us in the search for specific and/or unusual pathways and treatments for metabolic dysfunction in particular mental disorders (e.g. schizophrenia vs. bipolar disorders), and for specific agents (e.g. clozapine, olanzapine, valproate, mirtazapine, lithium, etc.). This is what I consider a foundational study: bringing us the best of basic epidemiology with wide applications, and setting the stage for clinicians and neuroscientists to test specific hypotheses on pathogenesis and therapeutics.
See Research by Galletly et al., 2012, 46(8): 753–761
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.