A big vaccine against invasive E. coli failed its phase 3 trial

Extraintestinal pathogenic Escherichia coli, or ExPEC, is one of the quiet workhorses of serious infection. When these gut-dwelling strains escape into the blood or other sterile sites, they cause bacteraemia and sepsis, especially in older adults, and they do so against a backdrop of rising antibiotic resistance. A vaccine that could head off those invasive infections would matter. A phase 3 trial published in The Lancet Infectious Diseases on October 12 reports that the leading candidate did not work [s1].

What the trial did

The vaccine, ExPEC9V, is a conjugate vaccine targeting nine E. coli O-serotypes that between them account for most E. coli bloodstream infections [s1]. The trial, called E.mbrace, was run at 332 clinical sites in 20 countries [s1]. It enrolled adults aged 60 years and older who were in stable health and had a history of urinary tract infection in the previous two years — a group at elevated risk of invasive disease [s1].

Participants were randomly assigned 1:1 to a single intramuscular dose of ExPEC9V or placebo, with participants, site staff and investigators all masked to the assignment [s1]. The primary endpoint was the first episode of invasive E. coli disease in blood or a sterile site caused by one of the vaccine's nine serotypes [s1]. The trial was sponsored by Johnson & Johnson and Sanofi [s1].

What it found

Enrolment ran from June 30, 2021 to February 13, 2025 [s1]. Of 18,844 people assessed for eligibility, 17,701 were vaccinated — 8,858 with ExPEC9V and 8,843 with placebo — and the per-protocol efficacy analysis included 17,637 of them [s1][s2]. There were 24 cases of invasive disease caused by the vaccine's serotypes, split 11 in the vaccine group and 13 in the placebo group [s1].

That near-even split is the whole story. It yielded a vaccine efficacy of 14.63%, with a 95% confidence interval running from −106.57% to 65.38% — a range so wide, and so firmly straddling zero, that it signals no protective effect at all [s1]. A prespecified futility analysis, triggered once at least 40% of the expected 53 disease events had accrued, found the lack of efficacy and the trial was stopped early [s1]. The vaccine's safety profile was acceptable, with more low-grade reactions than placebo, a balanced spread of serious adverse events, and no vaccine-related deaths [s1]. On the strength of these results the developers have discontinued ExPEC9V [s1].

How much weight the numbers carry

This is a clean negative result, and clean negatives are worth reading carefully. The efficacy point estimate of 14.63% is not evidence of a small benefit; with a confidence interval that reaches far into negative territory, the honest reading is that the trial could not distinguish the vaccine from placebo on its primary endpoint. The small number of events — 24 in all — is the reason the interval is so wide, and it is also why the independent futility check ended the study rather than letting it run to its planned size.

The disappointment is sharpened by what did work. The authors report that the vaccine generated robust and sustained antibody responses [s1]. In other words, the immune system responded as designed; the problem is that the antibody response did not translate into fewer invasive infections. That disconnect — good immunogenicity, no clinical protection — is one of the hardest patterns in vaccine development, because it means the marker everyone had been using to gauge progress was not predicting the outcome that counts.

Why it matters

Invasive E. coli disease is a large and growing burden, and a vaccine would have offered a way to prevent infections without adding to antibiotic pressure. The failure does not close that door, but it does say that targeting these nine O-serotypes with this conjugate design, in this population, was not enough. Reporting the result in full — including a null primary endpoint and a discontinued programme — is how the field avoids repeating the same bet, and it leaves a detailed public record of immunogenicity and safety that later candidates can be measured against.

What to watch

Whether the shortfall was about serotype coverage, the amount or type of antibody the vaccine raised, or the difficulty of protecting an older, higher-risk group; and whether other approaches to preventing invasive E. coli disease — different antigens, different formats — can clear the bar this one did not.

This article describes a single clinical trial and is informational only. It is not medical advice.