In my book The Bitterest Pills, I wrote of how many psychiatrists simply do not want to face up to the harms their treatments can produce. This is illustrated by the way the psychiatric establishment tried to avoid the implications of tardive dyskinesia, by suggesting it was a symptom of ‘schizophrenia’ and ignoring the evidence that tardive dyskinesia involves cognitive impairment. Similarly, when it became obvious that some of the second generation antipsychotics caused massive weight gain and metabolic disturbance, mainstream psychiatric journals published articles suggesting that diabetes was linked with schizophrenia as well.
The suggestion that antipsychotics reduce brain volume is not new. Psychiatrist Peter Breggin made this claim over thirty years ago1, but was dismissed as a crank. Over the last 10 years, however, the evidence has become irrefutable. Some leading members of the psychiatric establishment, like the UK’s Robin Murray, have even publicised their concern about it2.
Coupled with and possibly linked to this evidence of harm, doubts have started to mount that the benefits of long-term treatment with antipsychotics have not been as firmly established as is generally believed. The inadequacies of randomised trials of maintenance treatment have been highlighted, and the fact that there is little data on the overall impact of drug treatment when it is taken over the long periods of time that many patients are prescribed it for. Some evidence points towards the possibility that some people may do better if they stop or reduce their antipsychotic treatment, rather than continue on it long-term. These points have been raised by several mainstream psychiatrists3, as well as the usual suspects, including me4!
The paper by Donald Goff and seven other leading psychiatrists published in the American Journal of Psychiatry on May 5th is an attempt to rebut these concerns and re-establish the good reputation of antipsychotics5. I am shocked by how the article dismisses concerns about long-term treatment and evidence of brain impacts. It is riddled with distortion, ignores the most pressing criticisms, and is shot through with the unexamined presumption that the multitude of problems currently labelled as schizophrenia or psychosis will one day be revealed to be due to a specific brain abnormality that is targeted by antipsychotics.
While I would not dispute the usefulness of antipsychotics for the treatment of acute psychosis (in many, though not all situations), decades of research into early intervention has not demonstrated that early antipsychotic treatment improves long-term outcomes. Goff et al’s suggestion that the Norwegian Early Detection study (known as the TIPS study) showed improved long-term outcomes with earlier antipsychotic treatment is not borne out by the paper they cite6. Although the paper shows lower levels of negative symptoms, cognitive and depressive symptoms at the two year follow-up in people from the area with the Early Detection programme compared to those in areas without the programme, baseline data demonstrate that people in the Early Detection area had milder conditions, with fewer negative symptoms to begin with. In fact negative symptoms declined more in people from the area without the Early Detection programme! Moreover, the Early Detection group showed no benefit in terms of remission, relapse or positive symptoms.
Goff et al state that “the effectiveness of maintenance treatment for prevention of relapse has been well established,” but they do not acknowledge how there are no prospective randomised trials of maintenance treatment, only studies of maintenance treatment versus usually sudden withdrawal of maintenance treatment. Thus they completely fail to address concerns that effects of withdrawal of long-term treatment inevitably confound such studies. They also fail to mention the dearth of long-term data from randomised trials. Only 6 of the 65 trials in Leucht et al’s 2012 meta-analysis lasted longer than one year7.
They correctly point out that results of naturalistic studies, like Martin Harrow’s long-term follow-up of people from Chicago8, and the Finnish cohort study 9, are affected by the fact that patients who stop antipsychotics successfully are likely to have less severe conditions. However, they also claim that two other naturalistic studies found “improved outcomes in people with schizophrenia in people who continued antipsychotic treatment with those who did not.” Yet the papers cited refer to quite different sorts of studies, with neither presenting any data on global outcomes or social functioning, and only one involving the sort of follow-up periods of the Finnish and Chicago studies. One of the papers referred to looked at rehospitalisation rates within 3 years of antipsychotic discontinuation in a national database10, and the other looked at mortality rates11 (also see the extensive critique of this paper by De Hert et al, 2010). 12
The most worrying thing about the Goff et al paper, however, is the minimisation of the evidence that antipsychotics produce brain shrinkage. First the authors claim that shrinkage of brain grey matter has been shown to be part of schizophrenia. They trot out the old adage that brain differences were detected long before the introduction of antipsychotics. The paper they cite here is a post mortem study published in 1985, long after antipsychotics had been introduced13. The air encephalography studies that properly pre-date the introduction of antipsychotics involved long-term institutionalised patients who had been heavily treated with various sedative drugs along with physical treatments such as ECT and insulin coma therapy. Although these are commonly referred to as evidence that people with schizophrenia have smaller brains and larger brain ventricles, in fact the only two which had proper control groups showed no difference between brains of people with schizophrenia and brains of people without 14 15.
In any case, the presence of differences between the brains of people with schizophrenia and controls does not establish that there is progression of brain volume loss, which is what has been clearly demonstrated in people and animals taking antipsychotics. There are no studies that show progressive brain changes in people diagnosed with schizophrenia or psychosis in the absence of antipsychotic treatment. The authors cite one report from a group in Edinburgh suggesting progressive brain loss in people at ‘high risk of psychosis’ prior to receiving antipsychotic treatment16. However, this study finds subtle changes in some regions in a very small number of patients, not the cortex-wide loss of grey matter observed in people and animals taking antipsychotics.
Goff et al also cite a paper which showed decreases in grey matter volume in eight people following antipsychotic discontinuation compared to eight people who continued taking antipsychotics. However, the changes were localised to the putamen and nucleus accumbens, components of the basal ganglia, which other studies have shown to be enlarged during treatment with antipsychotics. Far from concluding that the study is evidence “that volume changes reflect progression of illness” as Goff et al suggest (P 5), the authors of the discontinuation paper concluded that “discontinuation reverses effects of atypical medication”17.
When describing the animal studies that show antipsychotic induced brain volume reductions18 19, Goff et al suggest that “the relevance of findings in rodents and monkeys to the treatment of psychosis in humans is unclear, both because of species related differences, and because animals lack the pathophysiology of schizophrenia. It is possible that antipsychotics have deleterious effects on normal brain, but protective effects in the presence of schizophrenia-related neuropathology” (P 6). Anything is possible, but this is just a leap of faith, and one that is completely at odds with the Hippocratic oath to ‘first do no harm!’ Monkeys and rats were chosen for these studies because of their similarities to human biology. There is no good reason at present to assume that the effects demonstrated in these animals would not occur in a similar species, i.e. us! And there is no evidence that antipsychotics have different effects on the brains of people with and without schizophrenia — although of course such evidence would be very difficult to obtain. I agree with Goff et al that these changes have not been definitively linked with effects on actual mental functioning, and that we need further data on this, but as they correctly suggest, ‘most but not all’ current studies show that brain volume reductions are correlated with decreased intellectual performance.
I still think antipsychotics can be useful, and that the benefits of treatment can sometimes outweigh the disadvantages, even in the long-term for some people. However, it does no one any service to pretend that they are innocuous substances that somehow magically transform (hypothetically) abnormal schizophrenic brains back to normal. Psychiatrists need to be fully aware of the detrimental effects of antipsychotics on the brain and body. They also need to acknowledge the way these drugs make life so miserable for many people, even for some who might have been even more distressed were they to be without them, something that is well described in Miriam Larssen-Barr’s recent blog here on Mad in America. Psychiatrists need to support people to evaluate the pros and cons of antipsychotic treatment for themselves and to keep doing this as they progress through different stages of their problems. To do this they need to be able to acknowledge the real nature of these drugs, and not sweep inconvenient truths under the carpet!