Bivalent rLP2086 is a recombinant factor H binding protein-based vaccine approved in the USA for prevention of meningococcal serogroup B disease in 10-25-year-olds. We aimed to assess the persistence of bactericidal antibodies up to 4 years after a three-dose schedule of bivalent rLP2086. We did this randomised, single-blind, placebo-controlled, phase 2 trial at 25 sites in Australia, Poland, and Spain. In stage 1 of the study (February, 2009-May, 2010), healthy adolescents (aged 11-18 years) were randomly assigned, via an interactive voice and web-response system with computer-generated sequential random numbers, to receive either ascending doses of vaccine (60 μg, 120 μg, and 200 μg) or placebo at months 0, 2, and 6. Dispensing staff were not masked to group allocation, but allocation was concealed from principal investigators, participants and their guardians, and laboratory personnel. In stage 2 of the study (reported here), we enrolled healthy adolescents who had received three doses of 120 μg bivalent rLP2086 (the optimum dose level identified in stage 1) or saline. Immunogenicity was determined in serum bactericidal antibody assay using human complement (hSBA) by use of four meningococcal serogroup B test strains expressing vaccine-heterologous factor H binding protein variants: PMB80 (A22), PMB2001 (A56), PMB2948 (B24), and PMB2707 (B44). Immunogenicity in stage 2 was assessed at months 6, 12, 24, and 48 post-vaccination. We did analysis by intention to treat. This trial is registered as ClinicalTrials.gov number NCT00808028. Between March 17, 2010, and Feb 8, 2011, 170 participants who received 120 μg of bivalent rLP2086 and 80 participants who received placebo in stage 1 of the study were entered into stage 2; 210 participants completed stage 2 up to 48 months. 1 month after the third vaccination, 93% (n=139/149) to 100% (n=48/48) of vaccine recipients achieved protective hSBA titres equal to or greater than the lower limit of quantification to each test strain, compared with 0% (n=0/25) to 35% (n=8/23) of control recipients. Despite initial declines in seroprotective hSBA titres for all four test strains, for three test strains (A22, A56, and B24), more than 50% of bivalent rLP2086 recipients continued to achieve titres equal to or greater than the lower limit of quantification at months 6 (57% [n=93/163] to 89% [n=42/47]), 12 (54% [n=84/155] to 69% [n=33/48]), 24 (53% [n=26/49] to 54% [n=82/152]), and 48 (51% [n=24/47] to 59% [n=79/134]); corresponding values in the control group were 14% (n=11/80) to 22% (n=5/23) at month 6, 13% (n=10/78) to 29% (n=22/76) at month 12, 16% (n=12/74) to 36% (n=8/22) at month 24, and 24% (n=16/68) to 35% (n=8/23) at month 48. For test strain B44, hSBA titres equal to or greater than the lower limit of quantification were shown in 37% (n=18/49) of vaccine recipients at 6 months, in 29% (n=14/48) at 12 months, in 22% (n=11/49) at 24 months, and in 20% (n=10/49) at 48 months, compared with 0% (n=0/25) of control recipients at month 6, 4% (n=1/25) at months 12 and 24, and 12% (n=3/25) at month 48. Adverse events were reported in seven (4%) of 170 participants in the bivalent rLP2086 group and two (3%) of 80 participants in the control group; no event was deemed related to vaccine. After three doses of bivalent rLP2086, protective hSBA titres above the correlate of protection (≥1/4) were elicited up to 4 years in more than 50% of participants for three of four meningococcal serogroup B test strains representative of disease-causing meningococci expressing vaccine-heterologous antigens. Further studies will be needed to assess possible herd immunity effects with meningococcal serogroup B vaccines and the need for a booster dose to sustain individual protection against invasive meningococcal disease. Pfizer.