Second-generation antipsychotics can cause lipid elevations at a greater rate than older typical antipsychotics. This risk may not be equivalent amongst the second-generation antipsychotics. We conducted a computerized, retrospective, nonrandomized, case-control analysis of 6331 patients receiving antipsychotics. For each patient, the first prescription for at least 60 continuous days for four antipsychotics [haloperidol (HALD), olanzapine (OLANZ), quetiapine (QUET), or risperidone (RISP)] was analyzed for total cholesterol, low-density lipoprotein, high-density lipoprotein (HDL), and triglycerides (TGL). Mean HDL was lower during OLANZ treatment than with RISP (P = 0.03) or QUET (P = 0.001). TGL were higher during OLANZ (P = 0.0007) or QUET treatment (P = 0.006) than RISP. In dichotomous analyses, odds ratios on the percentage of participants having abnormal cholesterol (P = 0.0003), low-density lipoprotein (P = 0.001), or TGL (P = 0.0001) during medication were in the order: OLANZ > QUET > RISP > HALD. For HDL, the results were less robust but the percentage of participants were in the order: OLANZ>RISP = HALD = QUET. In treatment-emergent analyses of patients without lipid abnormalities during an unmedicated baseline period, there was a greater risk of developing new HDL abnormality with OLANZ than RISP (P<0.05). In conclusion, treatment with RISP or HALD was associated with a more favorable lipid profile than with OLANZ or QUET.