For the MDD sample, logistic regression was used to determine the association of each of a number of independent variables with the syndrome and with each metabolic syndrome criterion. Three demographic but no treatment variables, including receiving an atypical antipsychotic, were associated with an increased risk of at least 1 of the 5 criteria:
• Age 40 years or older: odds ratio (OR) = 1.73; confidence interval (CI), 1.15 - 2.60
• Female sex: OR = 1.52; CI, 1.01 - 2.27
• Latino ancestry: OR = 2.23; CI, 1.30 - 3.84
Patients 40 years or older (OR = 1.91; CI, 1.39 - 2.61) and women (OR = 4.46; CI, 3.22 - 6.18) were at increased risk for visceral adiposity. Patients 40 or older who were at double the risk (OR = 2.09; CI, 1.47 - 2.98) on the triglyceride measure, were at moderately lower risk on the high-density lipoprotein criterion (OR = 0.70; CI, 0.50 - 0.99), and were more likely to meet the glucose (OR = 2.82; CI, 1.87 - 4.24) and hypertension (OR = 13.27; CI, 7.84 - 22.47) criteria than patients younger than 40 years.
Patients at highest risk
Previous reports suggested that metabolic syndrome is common in persons with a history of depression as well as in patients with current MDD.12 Findings from our study indicate that the prevalence may be as high as that found in patients with schizophrenia.10 However, it must be noted that this study was limited to inpatients, almost all of whom were receiving an SSRI. This class of drugs has been associated with abdominal obesity (OR = 1.40), increased body mass index, and elevated cholesterol levels.13,14 Depression has also been found to be associated with insulin resistance.13-16 Any reported prevalence of metabolic syndrome is, in part, a function of the diagnostic and demographic mix of the sample. In our study, the participants in the MDD and schizophrenia groups did not differ by age (38.8 ± 12.1 vs 39.6 ± 11.9, respectively), but the proportion of women was significantly greater. This sex difference is noteworthy because women were more likely to meet the waist circumference criterion and to have at least 1 of the 5 criteria. Perhaps related is the finding that in women, but not in men, a history of MDD doubles the odds of having metabolic syndrome.15 The ages of individuals in a sample may also affect study results. For example, although the prevalence of diabetes is higher in patients with schizophrenia than in the general population, this finding holds true only for patients younger than 50 years.17,18
In our sample of patients with MDD, no medication was associated with metabolic syndrome or with the presence of any individual criterion. Earlier studies have shown that even in schizophrenia at least some metabolic syndrome criteria appear to be independent of drug exposure.8-10 While many drugs are associated with weight gain, other issues must also be considered in the pathogenesis of metabolic syndrome, including genetic predisposition and lifestyle variables.