Understanding disproportionately high cardiometabolic risk in Australians treated for psychosis

Prior to 2004 there was relatively little awareness of the importance of cardiometabolic risk factors in patients with severe mental illness (SMI). Studies relying on convenience monitoring under emphasised the magnitude of the problem and any suspicion of an association between antipsychotics and diabetes was dismissed because of an over reliance upon retrospective methodology in the largest studies (Jin et al., 2004). The year 2004 became a watershed in no small part due to the publication of the American Diabetes Association/American Psychiatric Association consensus document. Since 2004, authors of over 200 studies have systemically tested patients with SMI for cardiometabolic risk factors, dyslipidaemia, diabetes and, less commonly, smoking and physical activity. Across all studies the proportion suffering metabolic syndrome using International Diabetes Federation (IDF) criteria has recently been summarized as one in three patients with schizophrenia (Mitchell et al., 2011). Beyond the metabolic syndrome itself patients endure a high risk of all major metabolic components, each contributing uniquely to adverse outcomes. It is easy to dismiss these high rates as wholly due to an unhealthy lifestyle at one extreme or wholly due to innate illness characteristics of schizophrenia itself. Neither view is likely to be true. A minority with unhealthy lifestyles do not develop these complications. Moreover, relatively low rates of metabolic complications are seen in those experiencing a first episode, particularly unmedicated patients, suggesting blaming schizophrenia as a disease entity is not an adequate explanation (Mitchell et al., 2012a). That said, significant metabolic risk factors such as obesity and smoking accrue very quickly during the early phases of treatment, particularly for those who start antipsychotics with a low weight, low fitness, poor diet and little healthy living advice (Saddichha et al., 2008). These effects are by no means confined to patients with schizophrenia, patients with bipolar disorder treated with antipsychotics appear to be equally afflicted; but this group has rarely been compared head-to-head with other types of SMI (Vancampfort et al., 2012).

Despite this accumulation of evidence over the last 10 years, we have been lacking large scale data involving patients typically seen in clinical practice. Indeed, it is unusual for systematic medical surveillance to take place outside of a clinical trial. Monitoring is typically haphazard in most mental health settings, although some innovative centres have developed metabolic clinics (Mitchell et al., 2012b). In this journal, Galletly and colleagues (2012) report on cardiometabolic monitoring in the second Australian Survey of Psychosis. This is one of the largest and most complete datasets of metabolic monitoring of patients with SMI ever assembled. Some will criticise this study because of the 40% participation rate; but this is actually impressive given the combined difficulty of gathering comprehensive medical results and completing semi-structured diagnostic interviews. Results from this study fill in several important gaps in our knowledge by studying clinically representative patients, comprehensively measuring metabolic risk factors (including physical activity) and by measuring fasting blood samples, a useful criterion reference when considering diabetes and impaired glucose tolerance. Results confirm and extend those mentioned above. Metabolic syndrome, using the same IDF criteria (but with an adapted cut-off for abdominal obesity), was found in 55% of patients which is in the worst 10^th percentile of all studies conducted worldwide. This markedly high rate of metabolic syndrome is exceeded only by studies of very vulnerable groups who are at special risk by virtue of ethnicity or by their receipt of long-term antipsychotic medication (Lin et al., 2009; Kato et al., 2004). In this second Australian Survey of Psychosis, rates of individual cardiometabolic risk factors were also strikingly high. Sixty-seven percent were smokers, 82% had abdominal obesity and 97% had low or very low levels of physical activity. Twenty-nine percent had impaired fasting glucose, which can be interpreted as a large group of patients who are in transition to diabetes at approximately 5-10% per year (Tabák et al., 2012). In terms of frank cardiometabolic disease, the second Australian Survey of Psychosis found 20% with diabetes/hyperglyaemia and 18% with cardiovascular disease, although it should be noted that these figures were collected by self-report. Nevertheless, these results put Australian patients at the top of the list of medically vulnerable patients with SMI, a position hitherto held by the United States. Australian clinicians must urgently ask why their patients have such profound rates of comorbidity and also what is being done to address the situation? At the time of testing 22%-43% of those with metabolic abnormalities were not taking appropriate treatment. This is a recurring theme. Correll et al. (2007) and Bernardo et al. (2009) found similar rates of under-treatment in the US.

Despite the wealth of valuable data in this study, some critical points are not addressed and deserve further attention. The first question is, do results hold true for all diagnostic groups? My suspicion is that they will once duration of illness, age, gender and antipsychotic treatment have been taken into account. Second, do results vary by region, or by ethnicity? Indigenous Australian people are known to have high cardiometabolic risk, and this risk appears to have been rising in recent years (Wang et al., 2010; Li et al., 2012). In all previous studies the metabolic risk profile of patients with SMI reflected the background population rate that those patients were living in. Could this be part of the explanation underlying these important findings? Third, and in my opinion most important, what is the influence of antipsychotic medication? Eighty-two percent of this sample were taking antipsychotics and therefore 18% were not. Analysis of currently unmedicated patients is useful but a crucial observation can be made in the rarefied cohort of patients who are drug naïve at the time of study. Only by studying the metabolic and lifestyle profiles of antipsychotic naïve patients with SMI can we establish how much psychiatrists have unwittingly contributed to the unhealthy cardiometabolic comorbidity present in many, perhaps most, of our patients.