On 2015 Jan 30, Peter Gøtzsche commented:

This meta-analysis was fatally flawed

Based on FDA trial data, Arif Khan et al. report that antipsychotic drugs lower total mortality in schizophrenia by more than 50% and that the drugs also lower suicides (1). These results agree very poorly with the fact that people with schizophrenia live about 20 years less than other people while almost all of them receive antipsychotics; that many experience excessive weight gains and diabetes because of the antipsychotics, which shorten their life substantially; and that a meta-analysis of old people found that double as many died when they got antipsychotics than when they got placebo (2).

It is difficult to understand how Khan, who has been principal investigator of more than 340 clinical trials sponsored by more than 65 pharmaceutical companies and 30 contract research organizations (1), can publish a meta-analysis that is so intensely misleading as this one (1). The fatal flaw in their analysis is that the authors used person-years instead of using persons. As pointed out in a letter to the editor, a person-year is not just a person-year (3). The average duration of placebo exposure was very short, only 33 days, as compared to 132 days on drug (1). Khan et al. seem to have included not only the double-blind phases of the trials but also safety extension studies (in which patients only receive active drug of course). Such analyses are notoriously unreliable, as those who continue on active drug are those who tolerate it. We need to look at deaths in relation to number of randomised patients, and then we can see that antipsychotics kill people, which we knew beforehand. The relative risk is 1.65, i.e. a 65% increase in total mortality on drug compared with placebo, and for suicides, the relative risk is 2.83.

Conflicts of interest: none.

1 Khan A, Faucett J, Morrison S, et al. Comparative mortality risk in adult patients with schizophrenia, depression, bipolar disorder, anxiety disorders, and attention-deficit/hyperactivity disorder participating in psychopharmacology clinical trials. JAMA Psychiatry 2013;70:1091-9.

2 Schneider LS, Dagerman KS, Insel P. Risk of death with atypical antipsychotic drug treatment for dementia: meta-analysis of randomized placebo-controlled trials. JAMA 2005;294:934–43.

3 Nordentoft M, Laursen TM. Was Risk of Suicide Underestimated? JAMA Psychiatry 2014;71:716.

On 2015 Jan 30, Peter Gøtzsche commented:

This meta-analysis was fatally flawed

Based on FDA trial data, Arif Khan et al. report that antipsychotic drugs lower total mortality in schizophrenia by more than 50% and that the drugs also lower suicides (1). These results agree very poorly with the fact that people with schizophrenia live about 20 years less than other people while almost all of them receive antipsychotics; that many experience excessive weight gains and diabetes because of the antipsychotics, which shorten their life substantially; and that a meta-analysis of old people found that double as many died when they got antipsychotics than when they got placebo (2).

It is difficult to understand how Khan, who has been principal investigator of more than 340 clinical trials sponsored by more than 65 pharmaceutical companies and 30 contract research organizations (1), can publish a meta-analysis that is so intensely misleading as this one (1). The fatal flaw in their analysis is that the authors used person-years instead of using persons. As pointed out in a letter to the editor, a person-year is not just a person-year (3). The average duration of placebo exposure was very short, only 33 days, as compared to 132 days on drug (1). Khan et al. seem to have included not only the double-blind phases of the trials but also safety extension studies (in which patients only receive active drug of course). Such analyses are notoriously unreliable, as those who continue on active drug are those who tolerate it. We need to look at deaths in relation to number of randomised patients, and then we can see that antipsychotics kill people, which we knew beforehand. The relative risk is 1.65, i.e. a 65% increase in total mortality on drug compared with placebo, and for suicides, the relative risk is 2.83.

Conflicts of interest: none.

1 Khan A, Faucett J, Morrison S, et al. Comparative mortality risk in adult patients with schizophrenia, depression, bipolar disorder, anxiety disorders, and attention-deficit/hyperactivity disorder participating in psychopharmacology clinical trials. JAMA Psychiatry 2013;70:1091-9.

2 Schneider LS, Dagerman KS, Insel P. Risk of death with atypical antipsychotic drug treatment for dementia: meta-analysis of randomized placebo-controlled trials. JAMA 2005;294:934–43.

3 Nordentoft M, Laursen TM. Was Risk of Suicide Underestimated? JAMA Psychiatry 2014;71:716.