Researchers from Germany, Greece and the U.S. reviewed the literature on relapse at 7 to 12 months following initiation of antipsychotic treatment. They conclude that maintenance treatment with antipsychotics benefits patients with schizophrenia. An accompanying article (by Jim van Os) in The Lancet notes that the study “breaks a trend” in recent reviews that casts doubt on both pharmacological and non-pharmacological treatments relative to cheaper and less complex approaches.
Cost-benefit analysis over the long term is not demonstrated by this study, van Os points out. He also notes the “reassuring” finding that relapse rates were not higher after acute antipsychotic withdrawal than after tapered withdrawal. Further, he notes, the study defines relapse in terms of positive symptoms, while issues of structural brain changes, affective dysregulation, cognitive alterations, social participation, and motivational impairment are equally if not more significant in terms of long-term recovery. Van Os concludes “The road to new developments will be long, tortuous, and not commercially rewarding for the foreseeable future, but will probably be more productive than behavioural studies of antipsychotic activity in rats.”
Antipsychotic Drugs for Prevention of Relapse (Reply by Jim van Os and Oliver D. Howes to “Antipsychotic drugs versus placebo for relapse prevention in schizophrenia: a systematic review and meta-analysis”, in print in The Lancet. Available online May 2, 2012.
“In The Lancet, Stefan Leucht and colleagues1 summarise 50 years of evidence and show that antipsychotic drugs reduce the rate of relapse in patients with schizophrenia at 1 year (risk ratio 0·40, 95% CI 0·33–0·49). This finding is consistent with a previous report,2 in which the authors stated that psychiatric drugs have similar efficacy to most general medical drugs. The present review1 breaks a trend, because systematic re-evaluations in the past 2 years  and  have cast doubt on the perceived effectiveness of both pharmacological (eg, antidepressant drugs) and non-pharmacological treatments (eg, cognitive behavioural therapy in depression) in psychiatry, in the sense that cheaper and less complex treatments might be as effective, or that effects apply to subgroups of patients only.  and 
“Although the evidence for antipsychotic drugs seems robust, several important issues remain. We do not know whether patients prefer drugs to placebo, and to what degree treatment affects social participation. The clinical relevance of the findings is also limited by what is thought of as relapse in psychotic illness, which is mainly classed as recrudescence of psychosis (delusions and hallucinations) rather than presence of negative symptoms and cognitive alteration. Furthermore, although of longer duration than typical 6-week studies, the median follow-up time during which relapse rates were established (26 weeks, range 1·75–156)1 is only a fraction of the time patients need support from services, obscuring the extent of antipsychotic-related morbidity and mortality in the long term.
“Antipsychotic drugs are associated with important side-effects other than those investigated in trials, such as possible structural brain changes,  and  and persistent dysphoria and motivational impairment.7 Combined with Leucht and co-workers’ finding1 that differences between drugs and placebo become smaller as lengths of studies increase, these potential side-effects suggest that long, comprehensive investigations are needed for a complete and balanced cost-benefit analysis of long-term use of antipsychotic drugs. However, high dropout rates mean that long-term studies will be challenging. All the studies included by Leucht and colleagues used inert placebos, and open-label studies showed greater differences between antipsychotic drug and placebo than did blinded studies. Therefore, effect sizes might be overestimated; meta-analytic evidence has established that drug side-effects can allow participants or investigators to identify whether drug or placebo is being used and thus might bias results of conventional trials with inert placebos.8
“Schizophrenia is associated with presynaptic dopamine dysfunction—both increased dopamine release and capacity for dopamine synthesis—which is linked to onset and relapse.9 Because antipsychotic drugs share the ability to block the D2 dopamine receptor, the findings reported by Leucht and colleagues1 make sense from a neurobiological perspective, but also draw attention to the fact that drugs are acting downstream of the pathological changes and address only psychosis as a symptom. Disease-modifying interventions are necessary across all domains of psychopathology, targeting the pathological changes (eg, modification of dopamine synthesis or firing) or upstream mechanisms. Upstream interventions might target higher-order neurobiological systems (eg, glutamate), as well as higher-order neurobiological, psychological, and societal-level risk. Interventions aiming to reduce risk would target mechanisms of interaction between genetic and non-genetic causal exposures10 (eg, progressive dopamine sensitisation with repeated exposure to dopamine-releasing drugs,11 behavioural sensitisation as a result of childhood trauma,  and  or dysfunctional match between individual-level and neighbourhood-level social characteristics, resulting in exclusion or demoralisation.14)
“Leucht and co-workers’ findings raise other mechanistic questions: what does long-term antipsychotic treatment do to the dopamine system and why do some patients relapse despite such treatment? Although D2 receptor density is unaltered in patients before they start antipsychotic treatment, studies in rats have shown that treatment upregulates the D2 receptor and some evidence suggests that this increase also occurs in patients.9 Such upregulation has long been thought to result in supersensitivity psychosis—ie, increased dopamine receptor sensitivity after antipsychotic treatment resulting in rapid relapse—so the report that relapse rates were not higher after acute antipsychotic withdrawal than after tapered withdrawal1 is reassuring. However, we still need to understand why many patients relapse despite adhering to antipsychotic treatment. Is relapse due to changes in the dopamine system leading to breakthrough psychosis, or some other mechanism?
“Although antipsychotic drugs might reduce relapse of psychosis, the cost–benefit analysis for long-term use is imprecise at best, and no evidence that other, more disabling, domains of psychopathology such as cognitive alterations or motivational impairment are similarly alleviated is available. A concerted effort is needed to establish how effects, non-effects, and side-effects of antipsychotic drugs are mediated in relation to the comorbid expressions of psychosis, affective dysregulation, cognitive alterations, and motivational impairment that constitute multidimensional psychotic syndromes.15 Results of studies combining treatment experiments with aetiological experiments might allow derivation of new insights for novel pharmacological and non-pharmacological treatments and testing for consumer preference. The road to new developments will be long, tortuous, and not commercially rewarding for the foreseeable future, but will probably be more productive than behavioural studies of antipsychotic activity in rats.”