Australasian Science: Australia's authority on science since 1938

How Early Can We Predict and Prevent Psychosis?

There is growing evidence that subtle changes in brain function can be identified long before the onset of psychotic symptoms. Credit: vchalup/Adobe

There is growing evidence that subtle changes in brain function can be identified long before the onset of psychotic symptoms. Credit: vchalup/Adobe

By Scott R. Clark, K. Oliver Schubert & Bernhard T. Baune

The addition of a simple blood test could improve predictions of a first psychotic episode.

Psychosis is a loss of contact with reality that manifests in abnormal perception, thinking and behaviour. These abnormalities can include hallucinations (e.g. hearing non-existent derogatory voices), delusions (e.g. of government surveillance and persecution), disorganised movement, poor motivation, slowed thinking, and loss of expression of emotions.

In severe psychosis, a person’s speech and therefore thinking may become so disorganised that topics change from moment to moment without relationship. Behaviour can also be disorganised, from low levels of self-initiated activity to furious and bizarre movement without a specific goal.

Psychotic illness may be episodic or chronic, and can occur in both severe mood disorders such as bipolar disorder or primary psychotic disorders such as schizophrenia. Up to one-quarter of patients receive little or no benefit from regular antipsychotic medication, but many patients are able to live relatively symptom-free if treated appropriately.

The 12-month prevalence of psychotic illness managed within Australian public mental health services has recently been estimated at 4.5 per 1000 people. Up to half of these patients have reported a suicide attempt in their lifetime; more than 60% report only partial recovery or continuous chronic illness; 32% have a severe dysfunction in the quality of self-care; and 85% rely on a government pension as their main source of income.

Psychotic illness is associated with high rates of comorbid chronic medical conditions, particularly those associated with metabolic syndrome, which occurs in more than 50% of psychotic patients. Seventy-five per cent of psychotic patients are overweight or obese, nearly half have raised cholesterol or triglycerides, and one-third have raised fasting glucose. Life expectancy in those with schizophrenia is decreased by up to 20 years, largely due to cardiovascular disease.

The symptoms of psychosis are highly distressing and impact on a person’s mental and physical health, day-to-day function and normal developmental milestones, such as completing school, getting a job and living independently. There is emerging evidence that treatments such as psychotherapy and some medications can prevent a first psychotic episode in young people who display risk factors. Thus there is a clear need for early intervention in psychotic illness.

Current models of psychosis prediction rely on assessments of subthreshold psychotic symptoms, day-to-day function and genetic risk based on family history obtained during the patient interview. These methods are conservative, and less than one-third of patients identified as high risk actually develop psychosis in long-term studies. Consequently, medical management follows a wait-and-watch approach to avoid unnecessary medication side-effects and diagnostic labelling, thus preventing effective early treatment.

However, there is growing evidence that subtle changes in brain function can be identified long before the onset of psychotic symptoms by measuring brain waves, imaging brain structure and activity, and identifying the presence of specific proteins and genes. Individually, though, none of these biomarkers accurately predicts the risk of psychosis or the response to treatment.

Our recent study, published in Translational Psychiatry, reports a new model that combines clinical information from the patient interview with blood levels of fatty acids. This model accurately identified almost three-quarters of high-risk patients who experienced a first psychotic episode within a year of their first assessment.

Our preliminary study used data from 40 European patients who were interviewed and then followed for signs of psychosis for 1 year as part of a clinical trial. We analysed a large number of factors that may have contributed to risk of psychosis. These could be grouped into historical factors, symptoms and function at the start of the study, and biological markers including brain wave patterns, protein markers of oxidation and fatty acid levels. Evidence suggests that high levels of oxidation may be responsible for damage to nerve cells, while fatty acids are required for normal neuronal function, are at low levels in active cases of psychotic illness, and are protective against psychotic illness.

We found that the most accurate model included a history of drug use, high levels of psychotic symptoms and poor function, and low blood levels of total fatty acid and nervonic acid levels at baseline. Interestingly, nervonic acid is a component of the fatty myelin sheath that surrounds nerve cells and is required for normal nerve function. Fatty acid supplementation (e.g. in fish oil) is thought to be protective against psychotic symptoms.

Our model identified 73% of high-risk patients that went on to suffer a psychotic episode. In comparison, standard assessments only identified 28% of patients who later experienced psychosis. Our results suggest that such a model could help to identify those at highest risk of psychosis, allowing more assertive early treatment.

One of the major goals of our research is to improve the prediction of mental health outcomes by combining new biomarkers with standard measures obtained during the patient interview. This approach represents a significant advance on current medical practice, where assessment of psychosis is based on patient interview alone, and brain scans and blood tests are only used to exclude significant physical illness that can cause psychotic symptoms.

We have used simple Bayesian mathematical techniques to calculate the probability of a first psychotic episode in a step-wise fashion as new information becomes available. Our stepwise approach is similar to the decision-making process in psychiatry, and therefore these models should be easier to implement into clinical practice.

While our results come from a small sample of patients, our findings suggest that standardised bedside assessment of history, symptoms and function, combined with a simple blood test for fatty acid levels, could greatly improve predictions of a first psychotic episode. This could allow earlier use of effective treatments that may delay or even prevent the onset of psychosis and significantly reduce the individual and social burden of psychotic illness.

Further testing of this model is required to confirm that our results can be replicated in larger samples. Such studies could take 5–10 years to produce definitive results.

Dr Scott R. Clark is a consultant psychiatrist in a community clinic, and a clinical academic for The University of Adelaide, where he is co-convener of psychiatry courses in the Medicine program. Dr K. Oliver Schubert is a consultant psychiatrist in an acute inpatient service, and a clinical senior lecturer for The University of Adelaide. Prof Bernhardt T. Baune is Chair of Psychiatry and leads the Adelaide Integrated Mental Health Biobank and Biomarker Centre at The University of Adelaide.