We studied the effects of switching antipsychotic drug-treated patients with schizophrenia or bipolar disorder who evidenced adverse metabolic side effects as indicated by a triglyceride/high-density lipoprotein ratio (TG/HDL) ≥ 3.5 to aripiprazole (ARIP; 5-30 mg/day, n = 24) or ziprasidone (ZIP; 40-160 mg/day, n = 28). Anthropometric and metabolic measures, psychopathology, quality of life and motor adverse effects were assessed over a 52-week period with evaluations at baseline, 6, 12, 26 and 52 weeks. There were statistically significant improvements in body weight, body mass index (BMI), TG, HDL and TG/HDL which did not differ between treatments. However, numerous secondary measures including weight and BMI, and the proportion of patients who lost ≥ 7% or who no longer met criteria for obesity, favored ZIP over ARIP. Decreases in total cholesterol and increases in HDL-cholesterol also favored ZIP. On the other hand, decreases in TG/HDL ratio and reduction in HgbA1c favored ARIP. There were no significant time or group × time interaction effects for most psychopathology measures; however, Global Assessment of Functioning Scores favored ARIP at 6 and 12 months. We conclude that switching patients with evidence of metabolic side effects to either ARIP or ZIP may be beneficial for some, but not all metabolic measures, with minimal risk of worsening of psychopathology and possibly some benefit in that regard as well.