A wide variety of antipsychotic drugs has been used in the treatment of schizophrenia and mood disorders. Antipsychotic drugs are split into first and second generation, or typical and atypical antipsychotics. Second generation drugs have only been around for some 20 years and have been favoured due to a better side effect profile. 

Drug induced hyperglycaemia
A variety of drugs has been implicated in precipitating diabetes mellitus (see Figure 1).

Figure 1 | Pharmacological agents implicated in the development of diabetes mellitus

Case reports linking atypical antipsychotics with hyperglycaemia have appeared for some years going back to 1994 for clozapine   1 , 1998-99 for olanzapine   2  and 1999 for quetiapine   3 . Reports linking older antipsychotic drugs such as chlorpromazine to increased blood glucose levels   4  and aggravation of existing diabetes   5  go back to the 1950s and 1960s   6 .

A study set in 1999 reviewed the records of 38,632 outpatients with schizophrenia receiving either atypical (58.6%) or typical antipsychotics (41.4%)   7 . The overall prevalence of diabetes was high at 19%, and in those under the age of 60, there was an increased risk of developing diabetes for those who were on clozapine, olanzapine and quetiapine (but not risperidone), compared to those on other antipsychotic agents.

A study, using the UK-based General Practice Research Database (GPRD)   8 , showed that the incidence of diabetes was higher among users of atypical than conventional antipsychotics (10/1,000 person years for clozapine; 5.4/1,000 for risperidone; 5.1/1,000 for conventional agents). Compared to those not on any antipsychotic treatment (and adjusted for age, sex and use of drugs known to affect glucose metabolism), olanzapine was an independent risk factor for diabetes, but risperidone and conventional agents were not (olanzapine odds ratio (OR) 5.8 (2.0-16.7); risperidone OR 2.2 (0.9-5.2); and conventional agents OR 1.4 (1.1-1.7)). When compared to those treated with conventional agents, olanzapine, but not risperidone, was again an independent risk factor.

A retrospective study examined claims data from a health plan of nearly two million members from 1997 to 2000   9 . Over 16,000 people with psychosis were identified, and those treated with antipsychotics (n=6,582) were compared to those not treated with antipsychotics (n=10,296). Major depression was the most common diagnosis in both groups (76% versus 38%). Treatment with olanzapine was a significant predictor of the development of diabetes, but other antipsychotics showed no association with diabetes.

However, this finding was not supported by another similar study   10 . In another study of a huge claims database, which handles 300 million prescription claims per year for over 50 million members in the USA   11 , those on antipsychotic medications were 3-4 times more likely to develop diabetes (after adjusting for age and sex), compared to the general population. Of note, there was no difference in the risk of developing diabetes between those treated with typical and atypical drugs.

A meta-analysis of 14 studies (10 retrospective cohort and four case control studies) included 256,000 people with schizophrenia or related disorders   12 . Comparisons between atypical and conventional antipsychotics revealed a strong association between atypical agents and the development of diabetes (clozapine OR 1.37 (1.25-1.52), olanzapine OR 1.26 (1.10-1.46) and risperidone OR 1.07 (1.00-1.13)). Although there was a trend towards diabetes development for quetiapine, results did not reach statistical significance.

When compared to those not receiving antipsychotics, there was a significantly increased risk of developing diabetes for clozapine (OR 7.44 (1.59-34.75)) and a borderline significant increase in risk for olanzapine (OR 2.31 (0.98-5.46)). Head to head comparisons showed a possible increase in diabetes risk associated with olanzapine, when compared to risperidone (OR 1.13 (0.99-1.28)), to clozapine (OR 1.41 (1.08-1.83)), and to quetiapine (OR 1.17 (1.00-1.37)).

Although the vast majority of cases of diabetes associated with antipsychotic drugs are type 2 diabetes, there are a number of case reports linking these drugs with diabetic ketoacidosis (DKA)   13 . DKA is a life-threatening, acute metabolic disturbance, which is usually associated with type 1, not type 2, diabetes.

The many studies examining the link between antipsychotic use and development of diabetes tend to have similar limitations. That is, they are often retrospective, not randomized control studies and do not always take into account risk factors such as race, body mass index (BMI) and family history. Other factors include not allowing for lack of compliance   14 , which is a problem in this population, changing medications during the study period and detection bias   15 .

Potential causes of drug-induced hyperglycaemia
The specific biochemical links between psychosis, antipsychotic drugs and diabetes remain poorly understood. However, it is abundantly clear that people with psychoses such as schizophrenia have markedly suboptimal lifestyles (see Figure below). They also have lower standards of living, being less educated and more likely to suffer from poverty and unstable living conditions   16 . They do not exercise as much; their diet is poor and they have high rates of smoking   17   18 . Studies examining diet found a lower consumption of dietary fibre, fruit and vegetables compared to a matched control group. They also appear to consume more sugar and fat, which seem to be independently associated with the severity of symptoms particularly in schizophrenia   19 . Poor dietary choices may however, also be related to the use of antipsychotic medication which increases appetite   20 .

Figure 2 | Factors associated with higher morbidity and mortality among people with psychotic conditions 

Screening for glucose and other metabolic disorders
The American Diabetes Association (ADA) has issued a consensus statement outlining screening, which should be done on commencement of antipsychotic therapy, and periodically thereafter   21  (see Figure).

Figure 3 | ADA Consensus monitoring protocol for patients initiating antipsychotic medication   21 

Other recommendations have been for testing to be done at the initiation of therapy, and then repeated at four months and annually thereafter   22 . An Australian Consensus working group has addressed the practical problems of screening this population, including taking into account the ethnicity of the person and recommends finger-prick testing and the use of a modified OGTT where necessary, where standard investigations are not possible   23 .

Monitoring should also be individually tailored together with an emphasis on modification of lifestyle factors such as diet and exercise.