On 2015 Oct 28, Peter Gøtzsche commented:

The authors mention in the abstract of this huge review (55 trials and 5,506 patients) without any reservation that akathisia didn’t occur more often in the chlorpromazine group than in the placebo group. The largest trial that contributed data to this outcome even found significantly less akathisia in the active group than in the placebo group, relative risk 0.57, 95% CI 0.37 to 0.88. Since we know that antipsychotics cause akathisia and that placebo cannot cause akathisia, this result speaks volumes about how flawed trials in schizophrenia generally are. What was seen in the placebo group were cold turkey symptoms caused by withdrawal of the antipsychotics the patients had received before randomisation.

I believe this fundamental problem renders the review unreliable and I refer readers to another review. The most reliable placebo controlled trials are those of first episode schizophrenia where none of the patients have ever received drugs before. There is a Cochrane review that approaches this ideal, but even this review is biased, as the trials are not limited to first episode patients; the review includes studies “with a majority of first and second episode schizophrenia spectrum disorders” (1).

The authors of that Cochrane review pointed out that the available evidence doesn’t support a conclusion that antipsychotic treatment in an acute early episode of schizophrenia is effective. They felt this was worrying given the widespread use of antipsychotics in the acute treatment of early episode schizophrenia-type psychoses, and also because the use of antipsychotics for millions of people with an early episode appears based on the trials for those with multiple previous episodes (which we know are highly flawed).

I suppose this means that we don’t have the evidence to support using antipsychotics at all. In fact, despite the trials being flawed by the cold turkey design, what was seen in recent placebo controlled trials in submissions to the FDA was only a 6 point improvement on the Positive and Negative Syndrome Scale (PANSS) (2,3), far below the minimally relevant clinical effect on this scale, which is about 15 points (4). Benzodiazepines should be preferred, when patients need to be calmed down, as they are far cheaper and less toxic than antipsychotics and also seem to sedate patients better (5).

1 Bola J, Kao D, Soydan H, et al. Antipsychotic medication for early episode schizophrenia. Cochrane Database Syst Rev 2011;6:CD006374.

2 Moncrieff J. The bitterest pills. Basingstoke: Palgrave Macmillan; 2013.

3 Khin NA, Chen YF, Yang Y, et al. Exploratory analyses of efficacy data from schizophrenia trials in support of new drug applications submitted to the US Food and Drug Administration. J Clin Psychiatry 2012;73:856–64.

4 Leucht S, Kane JM, Etschel E, et al. Linking the PANSS, BPRS, and CGI: clinical implications. Neuropsychopharmacology 2006;31:2318-25.

5 Dold M, Li C, Tardy M, et al. Benzodiazepines for schizophrenia. Cochrane Database Syst Rev 2012;11:CD006391.

On 2015 Oct 28, Peter Gøtzsche commented:

The authors mention in the abstract of this huge review (55 trials and 5,506 patients) without any reservation that akathisia didn’t occur more often in the chlorpromazine group than in the placebo group. The largest trial that contributed data to this outcome even found significantly less akathisia in the active group than in the placebo group, relative risk 0.57, 95% CI 0.37 to 0.88. Since we know that antipsychotics cause akathisia and that placebo cannot cause akathisia, this result speaks volumes about how flawed trials in schizophrenia generally are. What was seen in the placebo group were cold turkey symptoms caused by withdrawal of the antipsychotics the patients had received before randomisation.

I believe this fundamental problem renders the review unreliable and I refer readers to another review. The most reliable placebo controlled trials are those of first episode schizophrenia where none of the patients have ever received drugs before. There is a Cochrane review that approaches this ideal, but even this review is biased, as the trials are not limited to first episode patients; the review includes studies “with a majority of first and second episode schizophrenia spectrum disorders” (1).

The authors of that Cochrane review pointed out that the available evidence doesn’t support a conclusion that antipsychotic treatment in an acute early episode of schizophrenia is effective. They felt this was worrying given the widespread use of antipsychotics in the acute treatment of early episode schizophrenia-type psychoses, and also because the use of antipsychotics for millions of people with an early episode appears based on the trials for those with multiple previous episodes (which we know are highly flawed).

I suppose this means that we don’t have the evidence to support using antipsychotics at all. In fact, despite the trials being flawed by the cold turkey design, what was seen in recent placebo controlled trials in submissions to the FDA was only a 6 point improvement on the Positive and Negative Syndrome Scale (PANSS) (2,3), far below the minimally relevant clinical effect on this scale, which is about 15 points (4). Benzodiazepines should be preferred, when patients need to be calmed down, as they are far cheaper and less toxic than antipsychotics and also seem to sedate patients better (5).

1 Bola J, Kao D, Soydan H, et al. Antipsychotic medication for early episode schizophrenia. Cochrane Database Syst Rev 2011;6:CD006374.

2 Moncrieff J. The bitterest pills. Basingstoke: Palgrave Macmillan; 2013.

3 Khin NA, Chen YF, Yang Y, et al. Exploratory analyses of efficacy data from schizophrenia trials in support of new drug applications submitted to the US Food and Drug Administration. J Clin Psychiatry 2012;73:856–64.

4 Leucht S, Kane JM, Etschel E, et al. Linking the PANSS, BPRS, and CGI: clinical implications. Neuropsychopharmacology 2006;31:2318-25.

5 Dold M, Li C, Tardy M, et al. Benzodiazepines for schizophrenia. Cochrane Database Syst Rev 2012;11:CD006391.