There was standing room only in the hall. The speakers for and against the motion were eloquent. And, if the main purpose of a debate is to change at least some minds, it was a success.

Before hearing the session’s arguments, around a quarter of the packed audience backed the idea of discontinuing antipsychotics in the remitted patient. At the end of the debate, that proportion had increased appreciably – but only to around a third. The clear majority against discontinuing treatment remained.

Stefan Priebe (Queen Mary, University of London, UK) began by pointing out that the wording of the question is “can” we discontinue treatment, not “should” we discontinue. And the answer is self-evidently that we can, since many patients do in fact discontinue, sometimes with the support of their psychiatrist.

Benefits vs side-effects

We sometimes have to look through the meta-analyses to see the individual patient

That said, the bulk of the audience – along with Stefan Leucht, of the Technical University, Munich, Germany -- who opposed the motion, probably had more than a technicality of wording in mind. The central argument was about whether the benefits of continuing treatment – notably the reduced rate of relapse, at least in the short term -- were more than offset by the side-effects. If so, a trial of discontinuation would be a rational approach, should the patient wish it. But the data suggests benefits outweigh costs.

Stefan Priebe was prepared to accept that there is reasonable evidence – at a group level – that continuation of anti-psychotic treatment leads to fewer relapses. But, he argued, we sometimes have to look through the meta-analyses to see the individual patient. Even at the group level, the reduced relapse risk is only partial. And, unfortunately, we have proved conspicuously unsuccessful at identifying clinical factors which predict continued remission rather than relapse despite maintained treatment.

We sometimes have to look through the meta-analyses to see the individual patient

The Dutch study by Lex Wunderink and colleagues, which randomised remitted first episode patients to either a guided discontinuation strategy or maintenance therapy, illustrates the complexity of the studies on which we base our judgments.

Complex studies, difficult judgments

The relationship between anti-psychotics and loss of brain volume is confounded by disease severity

Among patients randomised to discontinue, the relapse rate at 18 months was double that in patients assigned to maintenance (43% vs 21%). But the comparison is not a clean one: only 20% of the discontinuation group achieved this goal long-term; half had to restart antipsychotics within two years because of recurrent symptoms; and discontinuation was not feasible at all in 50%.

To further muddy the waters, while the relapse rate was higher initially in the discontinuation group, the curves tended to come together with time. And, at seven years, patients assigned to the discontinuation strategy showed better long-term outcome: the proportion with symptomatic and functional remission was higher (40% vs 18%). Most of the difference was accounted for by functional status.

But Professor Leucht was not prepared to concede that this difference could be causally attributed to randomisation to the discontinuation strategy. For the five years between first follow up and long-term outcome assessment, there are no data on treatment. “Anything might have happened” to the patients in either group, he argued. “Even ECT (Electroconvulsive therapy).”

He was more inclined to rely on the recent Leucht et al meta-analysis which covered more than six thousand patients included in 65 trials and found that the one-year relapse rate with anti-psychotic medication was 27% while that on placebo was 64%. Significantly fewer patients on continued medication were readmitted to hospital. The conclusion: maintenance anti-psychotics benefit patients.

The relationship between anti-psychotics and loss of brain volume is confounded by disease severity.

Side effects, and shrinking the brain

Adverse effects, though, are the other side of the coin. These vary in severity but are practically always present, diminishing quality of life, Professor Priebe argued. Weight gain compromises physical health. Use of anti-psychotics is associated with significantly greater risk of sudden cardiac death. And long-term use causes loss of brain volume.

Professor Leucht accepted some but not all of these points. Suicide, he argued, is a disease-related event that is at least as catastrophic as sudden death. Loss of brain volume is an intrinsic part of the schizophrenia process, and the relationship with antipsychotics and drug dose is confounded by disease severity. Besides which, the loss of brain volume associated with taking antipsychotics is of questionable clinical significance.

A fine balance

Philip Gorwood (Descartes University, Paris, France), who commented on the debate, suggested that these kinds of issue reflect both the art and the science in psychiatry. In principle, the question of whether or not to continue anti-psychotics should be decided on the data. But not all the data point in the same direction. And – in any case – we have to adapt our advice to the circumstances of the individual patient, which is always going to be a matter of balanced judgement.