On 2016 Jan 12, Peter Gøtzsche commented:

The odds ratio for death was 1.95 (95% confidence interval 1.21 to 3.14), when psychotic patients received 3 or more antipsychotics simultaneously instead of one, thus doubling the mortality. However, the authors adjusted for somatic comedication and the adjusted odds ratio was 1.16 (0.68 to 2.00). This is a fatal error. The death rate soared, the more drugs for somatic illnesses, the patients received (27 times if they received at least 10), and the use of, for example, cardiovascular drugs were 37% and 16% among those who died and those who survived (controls), respectively. The use of diabetes drugs was 12% and 6%, respectively. Since increased doses and the use of several antipsychotics simultaneously increase the incidence of somatic illnesses, it is blatantly wrong to adjust for the use of somatic comedication, as this is part of the causal chain from psychosis to death. In this manner, the adjustment removes a relation between polypharmacy and death that actually exists.

Better studies have shown that polypharmacy with antipsychotics increases deaths, as expected. Some of these studies are mentioned in the current study’s Discussion section.

Peter C Gøtzsche, Professor and Director, Nordic Cochrane Centre

On 2016 Jan 12, Peter Gøtzsche commented:

The odds ratio for death was 1.95 (95% confidence interval 1.21 to 3.14), when psychotic patients received 3 or more antipsychotics simultaneously instead of one, thus doubling the mortality. However, the authors adjusted for somatic comedication and the adjusted odds ratio was 1.16 (0.68 to 2.00). This is a fatal error. The death rate soared, the more drugs for somatic illnesses, the patients received (27 times if they received at least 10), and the use of, for example, cardiovascular drugs were 37% and 16% among those who died and those who survived (controls), respectively. The use of diabetes drugs was 12% and 6%, respectively. Since increased doses and the use of several antipsychotics simultaneously increase the incidence of somatic illnesses, it is blatantly wrong to adjust for the use of somatic comedication, as this is part of the causal chain from psychosis to death. In this manner, the adjustment removes a relation between polypharmacy and death that actually exists.

Better studies have shown that polypharmacy with antipsychotics increases deaths, as expected. Some of these studies are mentioned in the current study’s Discussion section.

Peter C Gøtzsche, Professor and Director, Nordic Cochrane Centre