Antipsychotic drugs: challenges and future directions

Some sixty years on from the first use of chlorpromazine to treat schizophrenia, it is worth reflecting on where we have come from. Back in the 1950s, it was not known that dopamine was a neurotransmitter, how antipsychotics worked, what symptoms they worked on, or indeed if they worked at all1. Now we know that dopamine is a neurotransmitter, antipsychotics are all dopamine receptor blockers and, as Correll et al2 nicely review, large randomized, double‐blind placebo‐controlled trials have unequivocally demonstrated that they work both to treat acute psychotic episodes and to reduce relapse rates over the short to medium term.

Recent meta‐analytic data generated from over sixty years of placebo‐controlled trials estimate the standardized mean difference (SMD) between antipsychotics and placebo to be 0.38, with a greater effect seen on positive symptoms (SMD=0.45) than negative symptoms (SMD=0.35), quality of life (SMD=0.35) or depression (SMD=0.27)3. Such effect sizes are comparable to or larger than those found for treatments used for many common physical health conditions, including angiotensin‐converting enzyme (ACE) inhibitors for reducing cardiac events and mortality due to hypertension (SMD=0.16) and statins for reducing the risk of cardiac disease and stroke (SMD=0.15)4. Clearly, we have come a long way from the 1950s, but, despite these robust data on antipsychotics, many fundamental gaps in knowledge remain.

One glaring gap highlighted in this Forum is that as of yet we are unable to say conclusively what the optimum length of treatment with antipsychotic medication should be, once a patient has recovered from an acute episode. In current practice, many patients are treated with antipsychotic medication long‐term if not lifelong, in an attempt to prevent the frequency and severity of relapses that can be so disruptive to a person's life.

Where patients are symptom free but experiencing side effects, such as weight gain, that may shorten life as well as affect its quality, the risk‐benefit balance for relapse prevention is finely poised. Yet, as Correll et al highlight, there is little evidence from randomized, double blind controlled studies to support prophylactic treatment beyond two‐three years. Whilst some naturalistic studies do provide support for treatment beyond this term, the inherent limitations of these designs mean that the question remains unresolved, and guidelines cannot be conclusive.

This is a challenge to the field which needs to be met. We will need longer and, crucially, larger randomized controlled studies. This will not be easy, but other fields have risen to the challenge. For instance, in the case of the examples discussed above, statins and ACE inhibitors, there are now a number of randomized placebo‐controlled trials with several thousand patients. These studies are roughly two orders of magnitude larger and five to ten times longer than the typical long‐term randomized controlled study in schizophrenia. These large sample sizes give the power to have extended follow‐up and account for treatment changes and drop‐out. It is likely that we will need new ways of working, including international academic consortia as well as partnership with the pharmaceutical industry and governments, to achieve such large‐scale studies.

Correll et al also highlight heterogeneity in schizophrenia, something that is increasingly becoming apparent in the neurobiology underlying the disorder as well as its clinical manifestations, course and treatment response5.

Treatment resistance is probably the most clinically important manifestation of heterogeneity in patients with schizophrenia, and remains a major issue that continues to provoke debate over its pathophysiology, diagnosis and clinical management6. About a third of patients are thought to have treatment resistant illnesses, and around 15% show treatment resistance from illness onset7. Moreover, we have no way to identify the individuals whose illness will benefit from antipsychotic treatment.

Thus, large numbers of patients currently receive antipsychotic treatment although their illness is unlikely to respond to dopamine antagonists. The solutions to this will likely be found in part through identifying biomarkers that allow disease stratification, for example by the likelihood of response to dopamine receptor antagonists and, in the future, novel non‐dopamine receptor blocking medication.

As both trial data and clinical experience show, current antipsychotic treatment works most effectively in reducing the positive symptoms of schizophrenia, whereas the negative and cognitive symptoms often remain problematic. Cognitive symptoms in particular are associated with poor functional outcomes in schizophrenia8, yet our current treatments do nothing for them. In fact, there is evidence to suggest that dopamine antagonists may cause secondary negative and cognitive symptoms in people with schizophrenia9. Put simply, taking an antipsychotic may be unpleasant for some patients, and lead to secondary symptoms. This highlights the third challenge to the field: the need to develop treatments that are more than just antipsychotic and that patients are happy to take in the long term if necessary.

The final challenge is that our current antipsychotic medications are not disease modifying. Pre‐synaptic striatal dopamine dysfunction is thought to drive the symptoms of schizophrenia10, yet all of our current antipsychotic drugs act post‐synaptically. Thus, they block the consequences of pre‐synaptic dopamine dysfunction but do not address the underlying dopamine dysfunction, which remains present even after long‐term treatment. This provides a neurobiological explanation for why patients may relapse on stopping antipsychotic treatment.

Targeting the upstream abnormality and/or the factors that lead to it is an alternative approach that could both be better tolerated and more effective in the long term. Broadly speaking, the glutamatergic and GABAergic systems have excitatory and inhibitory effects, respectively, on the dopamine system. Genetic studies measuring copy number variants in patients with schizophrenia11 suggest that abnormalities in both neurotransmitter systems may be critical to the upstream regulation of dopamine. Findings like these suggest that targeting GABA and glutamate control of subcortical dopamine function could modify the pathophysiology, and potentially even be disease modifying. The interaction between psychosocial factors and biological changes12 also highlights the potential for psychological treatments to be disease modifying.

It is clear that we have come a long way from the 1950s in terms of both understanding of the pathophysiology of schizophrenia and its treatment, and this has thrown up new questions and issues. Antipsychotic drugs are likely to remain a crucial part of our therapeutic arsenal for years to come, so it behoves us to address the questions that remain.

Oliver H. Howes1, 2, 3, Stephen J. Kaar1
1Institute of Psychiatry, Psychology & Neuroscience, King's College London, London, UK; 2Medical Research Council London Institute of Medical Sciences, London, UK; 3Institute of Clinical Sciences, Imperial College, Hammersmith Hospital Campus, London, UK