Background: Whether or not second-generation antipsychotics (SGAs) represent an advantage over first-generation antipsychotics (FGAs) in the treatment of schizophrenia is not certain. Effectiveness studies published in the past 10 years have not unequivocally confirmed the superiority of SGAs over FGAs. We aimed to compare quality of life in patients with schizophrenia on an FGA strategy with those on an SGA strategy.

Methods: In the multicentre, randomised, double-blind Neuroleptic Strategy Study (NeSSy), we recruited participants (aged 18-65 years) with schizophrenia (ICD-10: F20.X) who required treatment initiation or a change in treatment, from 14 psychiatric university hospitals and state hospitals in Germany. Double randomisation allowed for restricted selection of a treatment within each antipsychotic drug group (FGA or SGA) for an individual patient: first, patients were assigned with a random number table to two of six possible drug pairs, each pair consisting of an FGA (haloperidol [3-6 mg] or flupentixol [6-12 mg]) given orally and an SGA (aripiprazole [10-20 mg], olanzapine [10-20 mg], or quetiapine [400-800 mg]) given orally, and the investigator then selected which pair was best suited to the patient; a second, double-blind random assignment allocated either the FGA or the SGA from the investigator-chosen pair to the patient. Treatment duration was 24 weeks. Primary outcomes were change from baseline to week 24 in quality of life (SF-36) and clinical global impression (CGI-I), analysed in all randomly assigned patients who received at least one dose of the study drug. Safety was assessed in a safety set, consisting of all randomly assigned patients who received at least one dose of the study drug, coinciding with the set of the efficacy analyses. The study is registered with ClinicalTrials.gov, number NCT01164059; German Clinical Trials Register, number DRKS00000304; WHO ICTRP, number U1111-1112-9727; and EudraCT, number 2009-010966-47.

Findings: Between April 1, 2010, and May 31, 2013, 149 patients were randomly assigned, 69 to FGA treatment and 80 to SGA treatment. 136 patients received at least one dose of study drug (63 in the FGA group, 73 in the SGA group). Mean area under the curve (AUC) values of SF-36 were significantly higher in the SGA group than in the FGA group (85·1 [SD 14·7] vs 79·7 [17·3], p=0·0112). Mean AUC values for CGI-I scores decreased in both groups, but were not significantly different between the two groups (3·39 [SD 0·89] in the FGA group vs 3·26 [0·92] in the SGA group, p=0·3423). 30 (48%) of 63 patients given FGAs had at least one adverse event compared with 42 (57%) of 73 patients given an SGA (p=0·3019); the most common were nervous system disorders (18 [60%] of 30 in the FGA group vs 19 [45%] of 42 in the SGA group) and psychiatric disorders (ten [33%] vs 16 [38%]). One patient died after cessation of study drug (olanzapine), most likely as a result of an illicit drug overdose. The increase in body-mass index (BMI) was significantly higher in the SGA group than in the FGA group (p=0·0021 at week 6 and p=0·0041 at week 24).

Interpretation: Improvement of patient-reported quality of life was significantly higher in patients with schizophrenia given SGAs than in those given FGAs, when treatment selection was individualised. This advantage, however, has to be weighed against the potential metabolic adverse effects of some SGAs.

Funding: German Federal Ministry of Education and Research.