Dr. Philip McMillan,  John McMillan

On paper, it should have worked. More than 17,000 adults aged 60 and over, every one of them with a past urinary tract infection, rolled up their sleeves in 20 countries for a vaccine designed to stop one of the most common causes of bloodstream infection in older people. In the blood, the vaccine did exactly what its designers hoped. Antibody levels climbed sharply, and they stayed high. Then the infections were counted. Eleven people in the vaccine group developed invasive E. coli disease. Thirteen in the placebo group did. That works out to an efficacy of 14.6 percent, a figure so close to zero that the trial was stopped early.

The full results of the trial, known as E.mbrace, were published in The Lancet Infectious Diseases on October 12, 2026. The vaccine has been abandoned, but the question of why it failed is still wide open. Dr. McMillan, a British researcher and clinician who has been tracking ten years of hospital data on unexplained infections in England, sees the result as one piece of a wider puzzle about how immune systems have fared since the pandemic.

A Big Trial, a Small Difference

Invasive E. coli disease is no minor complaint. Most people know E. coli as the bacterium behind urinary tract infections, but certain strains can escape the bladder, reach the bloodstream and trigger sepsis. An estimated 10 million cases of invasive disease occur worldwide each year, and a growing share involve bacteria resistant to multiple antibiotics.

The candidate, developed by Janssen and Sanofi, was nine-valent, meaning it targeted the nine strains behind most invasive cases. Neither participants nor their doctors knew who received vaccine or placebo, and volunteers were followed for up to three years. This was serious science. In February 2025, Johnson & Johnson announced that an independent review had found the vaccine unlikely to work, and development has since stopped.

Strong Numbers, No Protection

What makes the failure so puzzling is how good the laboratory numbers looked. Antibody levels against the targeted strains rose between three and 22 times above their pre-vaccination levels, and they remained raised a year later. Imagine being handed those figures before the trial ended. Who would have bet against it?

The investigators offer a cautious explanation: the immunity may have been of insufficient size or quality, and nobody yet knows what antibody level actually protects against invasive E. coli. Put simply, a blood test can show that the immune system has noticed a vaccine. It cannot show that the immune system will win the fight. One more curious detail: when the vaccine was given alongside a flu shot, the antibody response was weaker.

Wrong Place, Wrong Defenses

One well-established explanation concerns where the battle takes place. An injected vaccine trains the immune system in the blood, but E. coli usually launches its attack from the lining of the bladder. Defense there depends on mucosal immunity, the cells and antibodies that guard the body’s moist inner surfaces, such as the gut, airways and urinary tract. Injectable vaccines have a poor record of training those defenses, so antibodies in the bloodstream may arrive after the bacteria have gained a foothold.

A Trial Shaped by the Pandemic

Timing raises a further question. The study ran from June 2021 to February 2025, years in which COVID was still circulating widely and older adults were receiving repeated vaccine doses. The protocol’s main nod to the pandemic was excluding anyone vaccinated against COVID within the previous 28 days. Prior infections and the total number of doses received do not appear to have been a focus. Nobody can say whether those factors made a difference, and that is exactly the problem. If a variable is not measured, it cannot be ruled out.

Infections With No Name

The hospital data in question come from England’s Hospital Episode Statistics, covering 2016 to 2026. Two diagnostic codes stand out. B99 covers other and unspecified infectious diseases, and A49 covers bacterial infections where the site is unknown. Both were edging upward before the pandemic, and B99 has climbed steeply since.

These codes describe a frustrating clinical picture. A patient is plainly fighting an infection, yet the chest X-ray is clear, the urine sample shows nothing, and no source turns up anywhere. This is an informal analysis rather than a peer-reviewed study, and changes in coding habits, an aging population and shifts in testing could all push such numbers upward. Even so, a sustained climb in infections nobody can pin down deserves a closer look.

The COVID Antibody Puzzle

The E. coli result has a familiar ring. During the pandemic, vaccinated people often recorded antibody readings far higher than those seen after natural infection, yet breakthrough infections were common. High numbers on a lab report did not reliably translate into protection.

One line of research has focused on a particular class of antibody called IgG4. Most antibodies made after infection or vaccination belong to the IgG1 and IgG3 subclasses, which are good at flagging invaders for destruction. IgG4 behaves differently. It tends to appear after repeated or prolonged exposure to the same target, and it does a poor job of rallying the rest of the immune system. That is why it is associated with tolerance; it is the antibody that rises during allergy desensitization treatment, for example. Laboratory studies have documented a shift toward IgG4 in some people after repeated mRNA COVID vaccine doses, and some researchers have proposed that this could, in effect, tell the immune system to stand down. What the change means for patients remains uncertain, and scientists disagree about its significance. Nor can it explain the E. coli result directly, since that vaccine was not an mRNA vaccine.

Suppressed or Distracted?

A common assumption is that post-pandemic immune systems might simply be suppressed: sleepy, sluggish and slow to respond. There may be a third possibility, an immune system that is oversensitive but ineffective, a state many clinicians associate with long COVID.

“It’s not that their immune system is not working, it’s just hyper-responsive and distracted,” Dr. McMillan says.

Picture a security guard who chases a cat, stares at scraps of paper and panics at a flashing red light, while the real intruder strolls straight in. Viewed this way, IgG4 looks less like suppression and more like distraction. It is a hypothesis rather than a finding, but it offers a useful way to think about patients who seem to react to everything yet fight off nothing.

What Should Happen Next

The practical lesson for future trials is modest: record each participant’s infection and vaccine history, so hidden effects can be measured rather than assumed away. The larger issue is recognition.

“How long will this go on before it is formally acknowledged there is a problem?” Dr. McMillan asks. “Because the longer it takes, the less chance you have of mitigating what is going on.”

To help close that gap, he is building a structured continuing professional development (CPD) course on long COVID, organized in four sections and beginning with the current scientific consensus. It is aimed at clinicians but open to anyone. Plenty of people roll their eyes at long COVID, yet many may have it without realizing, and the dysregulated, overreactive immune state linked to it could open the door to a wide range of problems. The course aims to give doctors a framework for patients whose illnesses refuse to fit the familiar boxes.

None of this proves that the pandemic sank the E. coli vaccine. A trial with 24 cases cannot carry that weight, and the investigators say further analysis, including of the bacteria taken from those who fell ill, is still needed. But the result leaves a question the next trial cannot afford to skip. If antibodies are rising and protection is not following, what else has changed?