BACKGROUND Antibody to influenza virus hemagglutinin has been traditionally associated with protection. Questions have been raised about its use as a surrogate for vaccine efficacy, particularly with regard to an absolute titer indicating seroprotection. METHODS We examined hemagglutination-inhibition (HAI) antibody titers in subjects from a placebo-controlled trial of inactivated and live attenuated vaccines and compared titers in subjects with symptomatic influenza (cases) to those without influenza infection (noncases). RESULTS Prevaccination and postvaccination geometric mean titers were both significantly lower for cases compared with noncases in all intervention groups. Frequency of postvaccination seroconversion did not significantly differ for cases and noncases in either vaccine group. Among live attenuated vaccine and placebo recipients, cases were less likely than noncases to have postvaccination HAI titers ≥32 or 64. Nearly all recipients of inactivated vaccine had postvaccination titers of at least 64, and the small number of vaccine failures were scattered across titers ranging from 64 to 2048. CONCLUSIONS While HAI antibody is the major correlate of protection, postvaccination titers alone should not be used as a surrogate for vaccine efficacy. Vaccine failures from clinical trials need to be examined to determine why seemingly protective HAI titers may not protect. Clinical Trials Registration. NCT00538512.