To the Editor: The drug company-funded article Diabetes, psychotic disorders and antipsychotic therapy: a consensus statement by Lambert and Chapman1 tests some ethical boundaries.

Firstly, although categorised as a position statement, it does not express the position of any professional body.

The cumbersome title was probably inspired by a recent landmark review of the same topic.2 But there, the American Diabetes Association (ADA), with others, recommended that prescribers bear in mind the strong association of olanzapine and clozapine with diabetes when choosing an antipsychotic drug.

The MJA article does not dispute the ADA data, which rate a brief mention. But — and here is the second, more serious, ethical issue — the key ADA recommendation, that the varying diabetogenic potential of antipsychotic drugs should be a factor to consider when choosing an antipsychotic drug, is not mentioned at all. Instead, the article appears to be trying to use the weight of an impressive consultation process to influence prescribers to ignore the present state of knowledge.

One cannot take issue with the Australian recommendations, which echo those of an important North American 2002 conference.3 But the funding source means that the omission of the ADA recommendations is an ethical problem: one must question why they have been omitted.

To end on a positive note, the Australian recommendation for a prospective comparison trial of antipsychotic drugs for weight gain and diabetes is certainly a good one. It is high time that a university took this up —and published the results without drug company assistance.

In reply: Our article did not ignore the relative diabetic potential of various antipsychotic drugs. We wrote “clozapine and olanzapine are associated with greater weight gain and a higher occurrence of diabetes and dyslipidaemia than risperidone and quetiapine”, and so on.1

However, the article was not aimed to influence prescribing habits, but rather to alert health professionals to the relative metabolic risks both inherent in people with psychosis and arising from treatment with antipsychotic drugs in general, and the consequent need for constant assessment. The article tried to emphasise the need to balance risk versus benefit. We also tried to underline the need to effectively treat the illness (ie, the psychosis) with the most appropriate agent. Ignoring this would be akin to not using corticosteroids in severe asthma for fear of metabolic consequences.

Finally, the article is a pointer to the full consensus statement, which is available on the Internet at www.psychiatry.unimelb.edu.au/open/diabetes_consensus/ and may be downloaded without charge.