Authors' reply

Banerjee & Basu are correct to point out that our study does not have a case-control design in the purest epidemiological sense, that is a study in which patients who have developed a disease are identified and their past exposure to aetiological factors is compared with that of controls. Our study is conceptually similar to the case-control design, although we accept that both our study and case-control studies are inherently vulnerable to methodological weaknesses as discussed recently (Lee et al, 2007). We selected individuals who had had a diagnosis of severe mental illness and antipsychotic treatment to ascertain whether this population was at increased risk for cardio-vascular and metabolic disease compared with a control group.

We also accept, and acknowledge in our paper, that the sample size is relatively small. However, we found highly statistically significant differences between patients and controls across a number of outcomes. The analysis of effect of diagnosis, type of antipsychotic medication and smoking status was not a primary aim but was a secondary analysis. Increasing the sample size further might have added power to the study to detect differences in these variables. Notwithstanding, emerging evidence from studies in people with bipolar disorders points to an excess of cardiovascular and metabolic disease comparable to that in schizophrenia, suggesting that the similar rates across our diagnostic groups is a true finding.

Banerjee & Basu question the appropriateness of the term ‘severe mental illness’. The vast majority (75.6%) of patients had a diagnosis of schizophrenia, bipolar disorder or schizoaffective disorder. Many in the ‘other’ category (comprising 24.4%) had experienced psychotic depression and other severe depressive disorders requiring antipsychotic treatment. Although the use of the word ‘severe’ may be questioned, we are confident that this is appropriate.

We await larger, prospective studies designed specifically to tease out the aetiological factors that contribute to an excess of cardiovascular and metabolic risk and ultimately an excess mortality rate in this population. We hope our study has gone some way to highlighting the need for vigilant monitoring and appropriate intervention in this high-risk group.