Following a prior Kentucky clinical practice study on metabolic syndrome, serum glucose and lipid levels were used in a new sample to determine whether after correcting for confounding factors, olanzapine hyperlipidemia risk may be higher under naturalistic non-randomized treatment. Serum glucose, total cholesterol, HDL cholesterol and triglyceride levels were assessed in 360 patients with severe mental illnesses. The initial goal was to focus on olanzapine lipid profiles, but visual data inspection indicated that quetiapine needed attention as well. Patients were divided into 3 groups: 57 (16%) on olanzapine, 105 (29%) on quetiapine, and 198 (55%) on other antipsychotics (risperidone, ziprasidone, aripiprazole or typicals). HDL and glucose levels were not significantly different across the three antipsychotic groups. When compared with other antipsychotics, olanzapine patients had a borderline significantly higher mean total serum cholesterol level (178 vs. 192 mg/dl, p=0.06) and mean triglyceride level (172 vs. 202 mg/dl, p=0.06). These differences became significant (p=0.006 and 0.03) after correcting for confounders. Quetiapine appeared overprescribed in patients with metabolic syndrome complications. When compared with other antipsychotics, quetiapine patients had a significantly higher mean total serum cholesterol level (178 vs. 194 mg/dl, p=0.004) and mean triglyceride level (172 vs. 225 mg/dl, p<0.001). These differences were significant (p=0.02 and <0.001) after correcting for confounders. This study is consistent with emerging literature that suggests that some antipsychotics may have direct and immediate effects on lipid levels beyond obesity effects. The effect sizes of olanzapine and quetiapine on hyperlipidemia were about 0.40 in this naturalistic study.