In phase 2 trials, lebrikizumab, an anti-interleukin-13 monoclonal antibody, reduced exacerbation rates and improved FEV1 in patients with uncontrolled asthma, particularly in those with high concentrations of type 2 biomarkers (eg, periostin or blood eosinophils). We undertook replicate phase 3 studies to assess the efficacy and safety of lebrikizumab in patients with uncontrolled asthma despite inhaled corticosteroids and at least one second controller medication. Adult patients with uncontrolled asthma, pre-bronchodilator FEV1 40-80% predicted, and stable background therapy were randomly assigned (1:1:1) with an interactive voice-web-based response system to receive lebrikizumab 37·5 mg or 125 mg, or placebo subcutaneously, once every 4 weeks. Randomisation was stratified by screening serum periostin concentration, history of asthma exacerbations within the last 12 months, baseline asthma medications, and country. The primary efficacy endpoint was the rate of asthma exacerbations over 52 weeks in biomarker-high patients (periostin ≥50 ng/mL or blood eosinophils ≥300 cells per μL), analysed with a Poisson regression model corrected for overdispersion with Pearson χ2 that included terms for treatment group, number of asthma exacerbations within the 12 months before study entry, baseline asthma medications, geographic region, screening periostin concentration, and blood eosinophil counts as covariates. Both trials are registered at ClinicalTrials.gov, LAVOLTA I, number NCT01867125, and LAVOLTA II, number NCT01868061. 1081 patients were treated in LAVOLTA I and 1067 patients in LAVOLTA II. Over 52 weeks, lebrikizumab reduced exacerbation rates in biomarker-high patients in the 37·5 mg dose group (rate ratio [RR] 0·49 [95% CI 0·34-0·69], p<0·0001) and in the 125 mg dose group (RR 0·70 [0·51-0·95], p=0·0232) versus placebo in LAVOLTA I. Exacerbation rates were also reduced in biomarker-high patients in both dose groups versus placebo in LAVOLTA II (37·5 mg: RR 0·74 [95% CI 0·54-1·01], p=0·0609; 125 mg: RR 0·74 [0·54-1·02], p=0·0626). Pooling both studies, the proportion of patients who experienced treatment-emergent adverse events (79% [1125 of 1432 patients] for both lebrikizumab doses vs 80% [576 of 716 patients] for placebo), serious adverse events (8% [115 patients] for both lebrikizumab doses vs 9% [65 patients] for placebo), and adverse events leading to study drug discontinuation (3% [49 patients] for both lebrikizumab doses vs 4% [31 patients] for placebo) were similar between lebrikizumab and placebo. The following serious adverse events were reported in the placebo-controlled period: one event of aplastic anaemia and five serious adverse events related to raised concentrations of eosinophils in patients treated with lebrikizumab and one event of eosinophilic pneumonia in the placebo group. Lebrikizumab did not consistently show significant reduction in asthma exacerbations in biomarker-high patients. However, it blocked interleukin-13 as evidenced by the effect on interleukin-13-related pharmacodynamic biomarkers, and clinically relevant changes could not be ruled out. F Hoffmann-La Roche.