The introduction of second-generation antipsychotics (SGAs) over the past 2 decades generated considerable optimism that better antipsychotic treatments for schizophrenia and bipolar disorder were possible. SGAs offer several tolerability benefits over first-generation antipsychotics (FGAs), particularly with respect to extrapyramidal symptoms. However, SGAs can induce serious metabolic dysregulations, especially in drug-naive, first-episode, and child and adolescent populations, with olanzapine and clozapine having the highest propensity to cause these abnormalities. In this context, newer SGAs were developed to further improve the adverse effect burden of available agents. However, until now, the metabolic risk profile of the newly approved SGAs - asenapine, iloperidone, lurasidone and paliperidone (paliperidone extended release and paliperidone palmitate) - has not been compared. The objective of this systematic review and exploratory meta-analysis was to assess the effects of asenapine, iloperidone, lurasidone and paliperidone on body weight and other metabolic parameters (cholesterol, triglycerides and glucose), as this information is relevant to guide clinical decision making. A systematic literature search (1966-March 2012), using the Cochrane Central Register of Controlled Trials and MEDLINE, CINAHL and EMBASE databases, was conducted for randomized, placebo-controlled and head-to-head clinical trials of asenapine, iloperidone, lurasidone and paliperidone. Published and unpublished data on changes in body weight and glucose and lipid metabolism parameters were extracted. For placebo-controlled, short-term (≤12 weeks) and longer-term (>12 weeks) trials with available data on ≥7% weight increase compared with pre-treatment weight, or mean weight change with standard deviation, a formal meta-analysis was performed, estimating the pooled effect size (represented as relative risk [RR], numbers-needed-to-harm [NNH] and weighted mean difference [WMD]). An exploratory meta-analysis was also performed for the other metabolic variables (cholesterol, triglycerides and glucose). Data from active- and placebo-controlled studies were used for a pooled comparison of simple mean changes in weight, cholesterol, triglyceride and glucose levels. Fifty-six trials (n = 21 691) in schizophrenia (N = 49, n = 19 299) or bipolar disorder (N = 7, n = 2392) were identified (asenapine: N = 9, iloperidone: N = 11, lurasidone: N = 8, paliperidone: N = 28). Most of the trials (64.3%) were of ≤12 weeks' duration. In the short-term trials, compared with placebo, a ≥7% weight increase was statistically significantly (p < 0.05) most prevalent for asenapine (5 trials, n = 1360, RR = 4.09, 95% confidence interval [CI] 2.25, 7.43, NNH = 17), followed by iloperidone (4 trials, n = 1931, RR = 3.13, 95% CI 2.08, 4.70, NNH = 11) and paliperidone (12 trials, n = 4087, RR = 2.17, 95% CI 1.64, 2.86, NNH = 20). The effect of lurasidone on body weight (6 trials, n = 1793, RR = 1.42, 95% CI 0.87, 2.29) was not statistically significant. Short-term weight gain was statistically significantly (p < 0.001) greater than placebo with iloperidone (1 trial, n = 300, +2.50 kg, 95% CI 1.92, 3.08), paliperidone (15 trials, n = 3552, +1.24 kg, 95% CI 0.91, 1.57), asenapine (3 trials, n = 751, +1.16 kg, 95% CI 0.83, 1.49), as well as with lurasidone (5 trials, n = 999, +0.49 kg, 95% CI 0.17, 0.81, p < 0.01). Sufficient meta-analysable, longer-term, weight change data were only available for asenapine and paliperidone, showing statistically significantly (p < 0.001) greater weight gain versus placebo for both drugs (asenapine, 3 trials, n = 311, +1.30 kg, 95% CI 0.62, 1.98; paliperidone, 6 trials, n = 1174, +0.50 kg, 95% CI 0.22, 0.78). Although statistically significant, in general, no clinically meaningful differences were observed between the four newly approved SGAs and placebo regarding the mean change from baseline to endpoint in cholesterol levels in short-term trials, with the exception of iloperidone for total cholesterol (1 trial, n = 300, +11.60 mg/dL, 95% CI 4.98, 18.22, p ≤ 0.001), high-density cholesterol (1 trial, n = 300, +3.6 mg/dL, 95% CI 1.58, 5.62, p < 0.001) and low-density cholesterol (1 trial, n = 300, +10.30 mg/dL, 95% CI 4.94, 15.66, p < 0.001) and with the exception of lurasidone for high-density cholesterol (5 trials, n = 1004, +1.50 mg/dL, 95% CI 0.56, 2.44, p < 0.01). Asenapine increased total cholesterol statistically significantly (p < 0.05) during longer-term treatment (1 trial, n = 194, +6.53 mg/dL, 95% CI 1.17, 11.89). Regarding triglycerides, only short-term (3 trials, n = 1152, +1.78 mg/dL, 95% CI 0.40, 3.17, p < 0.01) and longer-term treatment with paliperidone (4 trials, n = 791, -0.20 mg/dL, 95% CI -0.40, -0.01, p < 0.05) had a statistically, but not clinically, significant effect. Statistically significant changes in glucose levels were noticed during short-term treatment with asenapine (2 trials, n = 379, -3.95 mg/dL, 95% CI -7.37, -0.53, p < 0.05) and iloperidone (1 trial, n = 300, +6.90 mg/dL, 95% CI 2.48, 11.32, p < 0.01), and during long-term treatment with paliperidone (6 trials, n = 1022, +3.39 mg/dL, 95% CI 0.42, 6.36, p < 0.05). While preliminary data suggest the lowest weight gain potential with lurasidone and potentially relevant short-term metabolic effects for asenapine and iloperidone, data are still too sparse to comprehensively evaluate the metabolic safety of the newly approved SGAs. Therefore, there is a clear need for further controlled studies to evaluate whether these agents are less problematic regarding treatment-emergent weight gain and metabolic disturbances than other currently available antipsychotics.
