OBJECTIVE—Type 2 diabetes is more prevalent in the indigenous Polynesian population of New Zealand (Maori) than in Europeans. The aim of this study was to determine whether insulin resistance in Maori psychiatric patients was associated with antipsychotic treatment and to investigate the mechanism of an association.
RESEARCH DESIGN AND METHODS—Thirty adult Maori psychiatric patients receiving antipsychotic medication for >6 months and 30 healthy, age-, sex-, and BMI-matched control subjects were enrolled. Early morning fasting blood samples were analyzed for plasma levels of glucose, insulin, A1C, triglycerides, total cholesterol, IGF-1, cortisol, cortisol-binding globulin (CBG), and adiponectin.
RESULTS—The patient group had significantly higher median fasting insulin plasma levels than the control group (P = 0.002), which were independent of BMI, age, and sex. In addition, the patient group had significantly higher total cortisol (P = 0.03) and lower CBG levels (P = 0.004) than the control group, resulting in significantly higher levels of free cortisol (P = 0.004). The patient group was also significantly more hypoglycemic (P = 0.026) and hypertriglyceridemic (P = 0.028) than the control group. There was no significant difference in BMI, waist circumference, A1C, total cholesterol, IGF-1, or adiponectin levels between the two groups.
CONCLUSIONS—An increase in insulin resistance is seen in Maori psychiatric patients treated with antipsychotic medication. Therefore, Polynesian ethnicity should be considered in prescribing practice and general care of this group. In addition, the hypothalamic-pituitary-adrenal axis may have an important role in the mechanism by which this insulin resistance develops.
Antipsychotic medications are used to treat the symptoms of psychosis, such as delusions, perceptual disturbance, thought disorder, and disorganized behavior. Psychosis is common in schizophrenia and bipolar disorder and is also seen in major depressive disorder, dementia, and some personality disorders. In New Zealand, in 2005, 8,750 prescriptions for antipsychotic medication were written for every 100,000 population (1). Unfortunately, many studies have associated the treatment of patients with antipsychotic medication with the development of impaired glucose tolerance, insulin resistance, and type 2 diabetes (2–4), but the mechanism underlying this association is not known. However, before the advent of antipsychotic medication in the 1960s, studies indicated that patients with schizophrenia had higher rates of diabetes than the normal population (5). This suggests that environmental or genetic factors may also be important in the development of insulin resistance in psychiatric patients.
Type 2 diabetes affects 21.1% of the indigenous, Polynesian population of New Zealand (Maori) compared with only 7.5% of New Zealand Europeans (6,7). Therefore, type 2 diabetes is a major cause of morbidity and mortality in Maori (6), and Maori may be vulnerable to the effects of antipsychotic medication on glucose metabolism. In our study, both patients and control subjects were drawn from the Maori population to determine whether treatment with antipsychotic medication was associated with an increase in insulin resistance in this population. We also investigated possible mechanisms by measuring plasma factors that have previously been implicated in the development of insulin resistance, i.e., IGF-1, adiponectin, and cortisol.
RESEARCH DESIGN AND METHODS
Before carrying out this study, written informed consent was obtained from each individual as approved by the Canterbury Ethics Committee. Ethnicity was assessed by self-report according to the 1996 New Zealand census question (8), and each subject had at least one Maori parent.
The patient group consisted of 30 Maori psychiatric patients recruited from both acute inpatient psychiatric wards and community outpatient psychiatric clinics. All patients had a psychotic illness (according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, text revision [DSM-IV-TR]: diagnoses of schizophrenia, DSM-IV-TR 295.0 (n = 24), schizoaffective disorder, DSM-IV-TR 295.7 (n = 3), bipolar I disorder, DSM-IV-TR 296.7 (n = 2), and psychotic disorder NOS [not otherwise specified], DSM-IV-TR 298.9 (n = 1). Patients were included if they were aged between 18 and 60 years, did not have a prior diagnosis of type 2 diabetes, had received antipsychotic medication for >6 months, and were not treated with other medications known to affect weight, such as sodium valproate. Patients had been treated with the antipsychotic medications clozapine (n = 10), olanzapine (n = 12), risperidone (n = 6), quetiapine (n = 1), and flupentixol (n = 1). The median time span for treatment was 6.5 months (range 6–11 months).
The control subjects (n = 30) were drawn from a group of Maori (n = 40) who had responded to an offer of free testing for diabetes and volunteered for the study. Maori volunteers were excluded if they were familially related to other subjects, had a prior diagnosis of type 2 diabetes, or reported a significant current or chronic illness. Members of the patient group were pair wise matched with the volunteer who was closest in age, sex, and BMI.
Clinical measurements and procedures
Participants were asked to fast for 12 h before giving a blood sample between 0730 and 0900 h. Glucose tolerance status was determined using World Health Organization criteria (9). Insulin resistance was determined using the homeostasis model of assessment (HOMA) computer model with results expressed as percentage sensitivity (HOMA%S) (10). Free cortisol was determined according to Coolens et al. (11). The height and weight of subjects were measured to the nearest 0.5 cm and 0.1 kg, respectively. BMI was calculated as weight in kilograms divided by the square of height in meters (2) and was used as an estimate of overall adiposity.
Laboratory procedures
Plasma levels of glucose, total cholesterol, and triglycerides were determined using an Aeroset analyzer (Abbott, North Chicago, IL). A1C was measured using the Bio-Rad Variant II high-performance liquid chromatography system. Insulin levels were measured using an immunoassay analyzer (IMX; Abbott). Total cortisol and cortisol-binding globulin (CBG) levels were measured in-house by radioimmunoassay and enzyme-linked immunosorbent assay, respectively (12). Plasma IGF-1 concentrations were measured, in duplicate, using a radioimmunoassay (Diagnostic System Laboratories). Plasma adiponectin concentrations were measured, in duplicate, by radioimmunoassay (LINCO Research, St. Charles, MO).
Statistical analysis
The Wilcoxon signed-rank test for nonparametric data was used to compare the patient and control groups, and results are presented as medians [interquartile range]. A value of P ≤ 0.05 was considered to be statistically significant. To examine the relationships between measured variables, Spearman's correlation coefficients were determined.
RESULTS
Median fasting plasma insulin levels were significantly higher in the patient group (84 pmol/l [interquartile range 44.8–190]) than in the control group (38.5 pmol/l [28.9–77.0]; P = 0.002) (Table 1). HOMA%S values were significantly lower in the patient group (52.9 [25.5–108.9]) than in the control group (117.2 [58.4–156.7]; P = 0.006) (Table 1). Plasma triglyceride levels were significantly higher in the patient group (1.7 mmol/l [1.3–2.6]) than in the control group (1.25 mmol/l [0.9–1.8]; P = 0.028). Median fasting plasma glucose levels were lower in the patient group (4.6 mmol/l [4.1–5.0]) than in the control group (5.1 mmol/l [4.7–5.4]; P = 0.026); however, levels for both groups were in the clinically healthy range.
Plasma total cortisol was significantly higher in the patient group (436.5 nmol/l [interquartile range 370.3–510.3]) than in the control group (374.5 nmol/l [323.3–449.3]; P = 0.03). Plasma CBG levels were significantly lower in the patient group (447.5 nmol/l [349.3–580.8]) than in the control group (580.5 nmol/l [528.3–707.5]; P = 0.004). Therefore, the median plasma free cortisol level was significantly higher in the patient group (54.4 nmol/l [43.7–74.4]) than in the control group (33.6 nmol/l [22.0–46.7]; P = 0.004).
There was no significant difference in BMI or waist circumference between the groups; however, both the patient and control groups were overweight (13). There was no significant difference in sex, age, A1C, cholesterol, IGF-1, and adiponectin between the two groups (Table 1). In addition, there was no significant difference between patients subgrouped according to whether they were taking olanzapine (n = 12) clozapine (n = 10), or risperidone (n = 6) with respect to free cortisol (P = 0.49), CBG (P = 0.29), and insulin (P = 0.35).
Free cortisol positively correlated with plasma insulin levels (rs = 0.262, P = 0.043) and negatively with fasting plasma glucose levels (rs = −0.366, P = 0.004). The correlation between free cortisol and insulin in patients did not differ significantly from the correlation in control subjects (P = 0.058).
CONCLUSIONS
A number of ethnic groups, such as the Pima Indians of Arizona and the Polynesians of the Pacific, have a much higher risk for type 2 diabetes than do Caucasians (6,14). The total Polynesian population worldwide is about 1.5 million (15,16) and includes native Hawaiians, Tongans, Samoans, and New Zealand Maori. Maori comprise ∼15% (600,000) of the total New Zealand population (8) and arrived in New Zealand, from Southeast Asia via the South Pacific, ∼900 years ago. Type 2 diabetes affects 21.1% of Maori compared with only 7.5% of New Zealand Europeans (6,7), and insulin-resistant diabetic subjects are two to five times more likely to die from cardiovascular disease (CVD) than those without diabetes (17–19). In addition, increased triglyceride concentrations found in populations in the Asia-Pacific region increased the relative risk for coronary heart disease by 1.33-fold (20). Therefore, the overall health of Polynesians is vulnerable to any additional factors that could increase the development of type 2 diabetes.
Excess body mass, insulin resistance, and elevated triglycerides are determining factors for the metabolic syndrome (9,21), which in turn is associated with an increased risk for diabetes (22). Both the patient and control groups in our study were classed as overweight, which is consistent with previous reports that increased BMI is more prevalent in Maori (63%) than in Europeans (21%) (6). However, the patient group was more insulin resistant and hypertriglyceridemic than the control group, implying additional risks of future diabetes and CVD for Maori treated with antipsychotic medication. It has been suggested that the association of antipsychotic treatment with both weight gain and diabetes in patients is confounded by the lack of exercise and alterations in diet common to psychiatrically ill patients. However, in our study, patients and control subjects were matched for BMI, which indicated that the increased insulin resistance is additional to that conferred by weight gain. The increased insulin resistance was also independent of age or sex.
In our study, Maori patients treated with antipsychotic medication had HOMA%S values similar to those reported in Northern Europeans treated with antipsychotic medication (23,24). This finding suggested that individual Maori are not more prone to antipsychotic medication–associated insulin resistance than individual Europeans.
The majority of the patient group were treated with a dibenzodiazapine antipsychotic medication (clozapine or olanzapine, 70%) and had schizophrenia (80%), reflecting their recruitment from inpatient and outpatient psychiatric units, which in New Zealand are tasked with the treatment of severe mental illness. Both clozapine and olanzapine have been suggested by others to be more diabetogenic and obesogenic than other antipsychotic medications (3,23,25). Our study was insufficiently powerful to determine significant differences in effects among individual antipsychotic medications.
Glucocorticoid hormones (mainly cortisol in humans) are so named because it was recognized long ago that one of their actions is on carbohydrate metabolism (26). These hormones are produced in the adrenal cortex under the control of the hypothalamic-pituitary-adrenal axis and at times of stress provide a longer-term signal to damp many of the acute responses to illness and “reset” metabolism in favor of providing substrates for oxidative metabolism. In a recent study of antipsychotic-naive patients with schizophrenia, it was found that these patients were more insulin resistant and also more hypercortisolemic than healthy age- and sex-matched control subjects (27). Because hypercortisolemia (Cushing's syndrome) leads to insulin resistance, glucose intolerance, high blood pressure, and triglyceridemia, the authors suggested that hypercortisolemia could be the primary defect that leads to the development of the insulin resistance in antipsychotic-naive patients with schizophrenia (27). However, in studies examining the effects of antipsychotic medication on patients with schizophrenia, changes in total plasma cortisol levels have generally not been measured or have been an inconsistent finding (27,28).
CBG binds 75% of circulating cortisol in plasma and is produced by the liver. Total plasma cortisol is made up of both bound and free fractions, and free cortisol is the form that is active, because it crosses cell membranes and interacts with receptors. Therefore, measuring both plasma CBG levels and total plasma cortisol levels allows free plasma cortisol to be estimated, which is a better measure of cortisol activity than the measurement of total plasma cortisol alone (29). Our results indicated that both total cortisol and free cortisol levels were higher in the plasma of the patient group than in the plasma of the control group. In addition, free cortisol positively correlated with plasma insulin levels, which is consistent with a cause-and-effect relationship. This finding suggested that factors such as antipsychotic medication or stress could directly reduce CBG plasma levels, leading to hypercortisolemia, and subsequent increased insulin resistance. However, total daily exposure to cortisol may not be accurately reflected by a single morning measurement, and obtaining samples at other times of the day could add to the validity of this relationship.
The insulin receptor/insulin receptor substrate-1/phosphoinositol 3-kinase signaling system has been implicated in the mechanism of corticosteroid-induced insulin resistance. In liver and muscle, dexamethasone treatment resulted in a reduction in insulin-stimulated insulin receptor substrate-1–associated phosphoinositol 3-kinase, suggesting that it may play a role in the pathogenesis of insulin resistance at the cellular level in an animal model (30,31). However, it remains possible that the environmental stress resulting from psychotic episodes or unknown genetic variations peculiar to individuals susceptible to psychotic episodes is the primary cause of low CBG and hypercortisolemia rather than direct effects of antipsychotic medication. The glycemic effects of having psychotic episodes versus the effects of antipsychotic medication could be examined by determining whether the prevalence of insulin resistance is as common in patients with psychotic illness treated with antipsychotic medication as in patients with nonpsychotic psychiatric illness treated with antipsychotic medication. In an animal study, dogs challenged with antipsychotic medication developed insulin resistance, which indicated that the association of insulin resistance and antipsychotic medication is unlikely to be entirely related to a psychotic illness (32).
The lower median value for glucose in the patient group was surprising but was within the healthy range (<7 mmol/l). In addition, there was no significant difference in A1C levels between patients and control subjects, and the trend was in the expected direction (slightly higher A1C levels in patients).
There was no difference in IGF-1 levels between Maori receiving antipsychotic medication and the control subjects. Low plasma levels of IGF-1 are associated with insulin sensitivity and are considered to be a marker for insulin resistance (33). Although IGF-1 is primarily a growth factor, it has been proposed to also play a role in insulin resistance via muscle IGF-1 receptor enhancement of glucose uptake (34). Our study is consistent with previous reports indicating that antipsychotic medication–associated insulin resistance in psychiatric patients does not involve IGF-1 (35).
Similarly, there was no difference in plasma adiponectin between Maori treated with antipsychotic medication and the control group. Adiponectin is a hormone produced by adipocytes that acts both in peripheral tissues and in the central nervous system to regulate peripheral glucose levels and body weight (36,37). Adiponectin levels are reduced in patients with insulin resistance and type 2 diabetes (38). The results of our study are consistent with previous reports (39) but conflict with other reports of increased levels of plasma adiponectin in patients treated with olanzapine and risperidone (40).
In summary, treatment with antipsychotic medication (predominantly olanzapine and clozapine) is associated in psychotically ill Maori with significant increased insulin resistance and hypertriglyceridemia, both of which are risk factors for type 2 diabetes and CVD. In addition, these findings were not solely related to BMI or central obesity. Because Maori have a much higher risk than the general New Zealand population for development of type 2 diabetes, any additional increase in insulin resistance and elevated triglycerides is particularly detrimental to the health of this group. In addition, our results indicated that these patients had both higher morning total and free cortisol plasma levels, because of lower CBG levels and that the free plasma cortisol level was positively correlated with the plasma insulin levels. Therefore, activation of the hypothalamic-pituitary-adrenal axis may be important in the development of insulin resistance in patients with a psychotic illness who are given antipsychotic medication.
Clinical characteristics of subjects
Parameters . | Patient group . | Control group . | P value . |
---|---|---|---|
n | 30 | 30 | |
Sex (male/female) | 26/4 | 26/4 | |
Age (years) | 32.5 (25.0–41.5) | 37.7 (25.8–45.2) | 0.095 |
BMI (kg/m2) | 29.5 (25.5–33.1) | 30.6 (26.7–32.9) | 0.199 |
Waist (cm) | 96.3 (90.0–105.5) | 100 (92.3–106.3) | 0.393 |
Fasting insulin (pmol/l) | 84.0 (44.8–190.0) | 38.5 (28.9–78.0) | 0.002* |
HOMA%S | 52.9 (25.5–108.9) | 117.2 (58.4–156.7) | 0.006* |
Triglycerides (mmol/l) | 1.7 (1.3–2.6) | 1.25 (0.9–1.8) | 0.028* |
Fasting glucose (mmol/l) | 4.6 (4.1–5.0) | 5.1 (4.7–5.4) | 0.026* |
Total cortisol (nmol/l) | 436.5 (370.3–510.3) | 374.5 (323.3–449.3) | 0.03* |
CBG (nmol/l) | 447.5 (349.3–580.8) | 580.5 (528.3–707.5) | 0.004* |
Free cortisol (nmol/l) | 54.4 (43.7–74.4) | 33.6 (22.0–46.7) | 0.004* |
A1C (%) | 5.7 (5.4–6.0) | 5.6 (5.3–5.8) | 0.124 |
Total cholesterol (mmol/l) | 4.9 (3.9–5.5) | 5.1 (4.4–5.8) | 0.107 |
IGF-1 (μg/l) | 197.0 (138.3–251.0) | 201 (146.3–228.0) | 0.861 |
Adiponectin (μg/l)) | 3.6 (2.5–6.2) | 5.2 (2.6–8.6) | 0.061 |
Parameters . | Patient group . | Control group . | P value . |
---|---|---|---|
n | 30 | 30 | |
Sex (male/female) | 26/4 | 26/4 | |
Age (years) | 32.5 (25.0–41.5) | 37.7 (25.8–45.2) | 0.095 |
BMI (kg/m2) | 29.5 (25.5–33.1) | 30.6 (26.7–32.9) | 0.199 |
Waist (cm) | 96.3 (90.0–105.5) | 100 (92.3–106.3) | 0.393 |
Fasting insulin (pmol/l) | 84.0 (44.8–190.0) | 38.5 (28.9–78.0) | 0.002* |
HOMA%S | 52.9 (25.5–108.9) | 117.2 (58.4–156.7) | 0.006* |
Triglycerides (mmol/l) | 1.7 (1.3–2.6) | 1.25 (0.9–1.8) | 0.028* |
Fasting glucose (mmol/l) | 4.6 (4.1–5.0) | 5.1 (4.7–5.4) | 0.026* |
Total cortisol (nmol/l) | 436.5 (370.3–510.3) | 374.5 (323.3–449.3) | 0.03* |
CBG (nmol/l) | 447.5 (349.3–580.8) | 580.5 (528.3–707.5) | 0.004* |
Free cortisol (nmol/l) | 54.4 (43.7–74.4) | 33.6 (22.0–46.7) | 0.004* |
A1C (%) | 5.7 (5.4–6.0) | 5.6 (5.3–5.8) | 0.124 |
Total cholesterol (mmol/l) | 4.9 (3.9–5.5) | 5.1 (4.4–5.8) | 0.107 |
IGF-1 (μg/l) | 197.0 (138.3–251.0) | 201 (146.3–228.0) | 0.861 |
Adiponectin (μg/l)) | 3.6 (2.5–6.2) | 5.2 (2.6–8.6) | 0.061 |
Data are median (lower quartile–upper quartile). Subjects were matched for measured variables above the line, i.e., sex, age, and BMI.
P < 0.05 are considered statistically significant.
Article Information
This work was supported by grants from the Health Research Council of New Zealand (04/141R) and the Schizophrenia Fellowship of New Zealand (to P.F.E. and N.R.P.).
We are indebted to Te Korowai, Maori Mental Health Service, and Dr. Erik Monasterio of the Canterbury District Health Board for their assistance.
References
Published ahead of print at http://care.diabetesjournals.org on 19 March 2007. DOI: 10.2337/dc06-2057.
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.