The 2004 guideline recommends that selection of an antipsychotic agent be guided by the patient's past medication history, current symptoms and co-occurring conditions, other concurrent treatments, and preferences. The guideline states that second-generation agents should be considered first-line options for patients in the acute phase, mainly because of the decreased risk of extrapyramidal side effects and tardive dyskinesia, but acknowledges debate over the relative advantages, disadvantages, and cost-effectiveness of first- and second-generation agents. The guideline also states that for some patients, a first-generation agent may be an appropriate first-line option. This latter recommendation has been strengthened by the results of several recently published effectiveness studies that suggest that the first-generation antipsychotics perphenazine and molindone may be equally effective as second-generation agents. In fact, the distinction between first- and second-generation antipsychotics appears to have limited clinical utility.

Phase I of the double-blind CATIE clinical trial randomized 1,490 patients to available FDA-approved second-generation antipsychotics—risperidone, olanzapine, quetiapine, and ziprasidone—and to the first-generation antipsychotic perphenazine (4). There were few exclusion criteria, and patients were recruited from diverse programs in order to include "real-world" patients with potential co-occurring psychiatric or general medical conditions. The primary outcome measure was discontinuation from the randomized treatment, and by the end of the 18-month trial 74% of patients had switched to another antipsychotic or discontinued treatment. Olanzapine was the most effective medication, with 64% discontinuing, compared with the discontinuation rates for risperidone (74%) and quetiapine (82%). Perphenazine discontinuation rates (75%) were comparable to the other second-generation antipsychotics, including ziprasidone (79%). A similar pattern of results was found when symptom outcomes or hospitalization rates were examined. Extrapyramidal side effects were uncommon and similar across drugs, but olanzapine carried the greatest burden of metabolic side effects.

In the CUtLASS trial (5), 227 patients with schizophrenia who were judged by their treating clinician to potentially benefit from a new antipsychotic medication trial "because of inadequate response or adverse effects" were randomized to receive either a first- or a second-generation (excluding clozapine) antipsychotic. The specific antipsychotic was chosen by the treating clinician. The primary outcome measure, assessed by blind raters at 12, 26, and 56 weeks, was "quality of life," reflecting social and vocational function. Symptom changes were secondary outcomes. There was no difference in any outcome measures between groups.

In the EUFEST trial (6), 498 patients experiencing their first episode of schizophrenia were randomized to receive haloperidol, amisulpride, olanzapine, quetiapine, or ziprasidone. The study was conducted at 50 sites in 13 European countries and Israel, and treatment was not blinded. The primary outcome measure was discontinuation from treatment. At 1-year follow-up, all-cause discontinuation was higher for haloperidol (72%) than that for amisulpride (40%), olanzapine (33%), quetiapine (53%), or ziprasidone (45%). Global ratings of symptoms were least improved by treatment with quetiapine or haloperidol and most improved by treatment with amisulpride; however, there were no differences in symptom improvement as measured by the Positive and Negative Syndrome Scale or in rates of hospital admission. Extrapyramidal side effects were most severe in patients treated with haloperidol, and weight gain was most severe in patients treated with olanzapine.

In the CAFE trial (7), 400 patients early in the course of psychotic illness were randomly assigned in a double-blind manner to receive olanzapine, quetiapine, or risperidone. At 1-year follow-up, all-cause discontinuation rates were similar for all groups (68.4%–71.4%), and there was no difference in symptom severity measures. Side effects were common and in line with the known side-effect profile of these antipsychotics.

The TEOSS study (8) was a double-blind, randomized trial comparing olanzapine, risperidone, and molindone in 119 pediatric patients with early-onset schizophrenia and schizoaffective disorder. "Response" was defined as improvement on the Clinical Global Impression scale of "very much" or "much" improved, at least a 20% reduction in symptom severity as measured by the Positive and Negative Syndrome Scale, and tolerating treatment for at least 8 weeks. A significant difference in response likelihood was not found between groups (molindone 50%, olanzapine 34%, risperidone 46%). Treatment with risperidone and olanzapine was associated with significant weight gain and metabolic side effects, and patients treated with molindone were more likely to report akathisia.

Paliperidone, marketed in an extended-release (ER) formulation, is the major active metabolite (9-hydroxyrisperidone) of risperidone. It is mainly cleared by the kidneys, with negligible hepatic metabolism. Five randomized, double-blind trials, sponsored by the manufacturer, Janssen Pharmaceuticals, involving 1,647 acutely ill patients with schizophrenia, demonstrated paliperidone ER to have greater efficacy than placebo at a fixed dose over 6 weeks (9) and led to the medication receiving FDA approval in 2006. Side effects included marked prolactin elevation in men and women, a greater incidence of extrapyramidal side effects at higher doses (>6 mg/day), dose-related weight gain, and tachycardia. Comparisons with fixed-dose (10 mg/day) olanzapine (N = 1,332) showed similar efficacy and less liability for weight gain but greater liability for extrapyramidal side effects. Paliperidone ER thus appears to have a similar side-effect profile to its parent compound, risperidone. The relative advantages or disadvantages of paliperidone compared with risperidone are unknown.

As described in the 2004 guideline, weight gain and metabolic side effects are common or frequent adverse effects of the second-generation antipsychotics clozapine, olanzapine, risperidone, and quetiapine. The guideline recommends regular monitoring of weight, body mass index, serum lipids, and fasting glucose levels of all patients. When patients gain weight, it is recommended that clinicians discuss treatment options, which may include switching medications, to prevent further weight gain and encourage weight loss.

Several clinical trials have investigated pharmacological and cognitive-behavioral treatments that may attenuate or reverse antipsychotic-related weight gain and lipid, glucose, and insulin changes (10). The nonpharmacological weight management interventions are described in greater detail in the subsection "Psychosocial Interventions for Weight Management," later in this watch. There have been several pharmacological clinical trials investigating metformin (a peripheral insulin-sensitizing agent), topiramate (an anticonvulsant), reboxetine (a selective norepinephrine reuptake inhibitor), and amantadine (a dopamine agonist). Metformin has been investigated in five randomized controlled studies, with four showing some indication of benefits (11–14), and one negative trial (15). The most promising results were reported in a randomized, double-blind trial in which 128 olanzapine-treated first-episode patients received adjunctive metformin 750 mg/day, metformin 750 mg/day plus lifestyle changes, lifestyle changes plus placebo, or placebo (11). The patients who received adjunctive metformin plus lifestyle changes had the most robust weight loss, body mass index (BMI) reduction, waist circumference reduction, and fasting insulin and insulin-resistance level reduction; these outcomes were significantly better than lifestyle changes plus placebo or placebo alone. For example, BMI significantly decreased by 1.8 units on average in the metformin plus lifestyle changes group, by 1.2 units in the metformin alone group, and by 0.5 units in the lifestyle changes plus placebo group. In the placebo group, in contrast, BMI increased by an average of 0.7 units. No increase in adverse events, including nausea, occurred in the patients treated with metformin.

Two randomized, placebo-controlled trials of adjunctive topiramate have reported weight loss in overweight patients who were already receiving treatment with olanzapine (16) or with risperidone, olanzapine, quetiapine, or clozapine (17). In patients who had gained weight during olanzapine treatment, two randomized, placebo-controlled trials reported that further weight gain was less pronounced with adjunctive amantadine treatment (18,19). Two randomized, placebo-controlled trials also reported significant attenuation of weight gain when adjunctive reboxetine was initiated concomitantly with olanzapine in patients with a first episode of schizophrenia (20,21). However, the clinical utility of these adjunctive treatments is unclear, given their relatively small impact on weight as well as their cost, potential side effects, and potential interactions with other medications.