Design: Phase 3, randomised, double-blind, placebo-controlled trial, stopped for futility
Population: 17,701 adults aged 60 or over with a UTI in the previous 2 years
Primary outcome: First invasive E. coli disease caused by vaccine O-serotypes
Effect: 11 vs 13 cases; vaccine efficacy 14.6% (95% CI −106.6 to 65.4)

E.mbrace enrolled adults aged 60 or over in stable health who had had a urinary tract infection in the previous two years, at 332 sites in 20 countries. They received one dose of a conjugate vaccine targeting nine E. coli O-serotypes responsible for most E. coli bloodstream infections, or placebo. The primary outcome was a first episode of invasive E. coli disease caused by those serotypes.

A prespecified futility analysis stopped the trial after 17,701 people had been vaccinated. There were 11 cases in the vaccine group and 13 with placebo, a vaccine efficacy of 14.6% (95% CI −106.6 to 65.4) — no evidence of protection. Antibody responses were strong and sustained, and safety was acceptable. Development of the vaccine has been discontinued.

The gap between good antibody levels and no clinical protection is the instructive part: for E. coli, serum antibodies to O-antigens did not translate into prevention of invasive disease. For now, preventing E. coli bacteraemia in older adults still rests on managing urinary tract infection well and limiting catheter use.

There is no vaccine against invasive E. coli disease to offer patients.
Strong antibody responses did not prevent disease, so immunogenicity alone should not reassure.
Remove urinary catheters as soon as they are no longer needed.
In older adults with recurrent UTI, review continence, retention and bladder emptying.
Use local resistance data to guide empirical therapy for urinary sepsis.

Why it matters

A vaccine that raised the right antibodies but prevented nothing shows that immunogenicity is not a stand-in for protection.

The statistics, in plain English

A vaccine efficacy of 14.6% with an interval running from −106.6% to 65.4% means the result is compatible with the vaccine doubling the risk or reducing it by two-thirds — in other words, no usable information on benefit. Only 24 events occurred, which is why the interval is so wide.