A batch of contaminated food supplement given to premature babies could have claimed the life of another baby, health officials have said.
Experts are investigating a third death linked to a batch of total parental nutrition (TPN), which is supposed to deliver a variety of nutrients intravenously when a baby is unable to eat on its own.
The youngster died at Addenbrooke's Hospital in Cambridge almost five weeks after receiving the fluid.
A hospital spokesman said that despite every effort the baby died on Sunday night.
Investigators said that preliminary inquiries have shown that the equipment used to treat the youngster contained a bacteria thought to have come from the suspected batch of contaminated fluid.
The premature baby was given the fluid on May 27 but was not identified to have developed blood poisoning straight away afterwards, as seen in other cases.
The child's death has been referred to the coroner, a spokeswoman for the investigation said.
The Public Health England (PHE) and the Medicines and Healthcare products Regulatory Agency (MHRA) investigation into the blood poisoning of babies who were given the feed is now examining 23 cases, including three deaths.
The youngsters are thought to have developed septicaemia after being infected with the bacillus cereus bacteria.
A spokesman for Cambridge University Hospitals NHS Foundation Trust said: "We can confirm that a baby has died after receiving contaminated nutritional fluid (TPN).
"Our thoughts are with the family and we are supporting them during this very difficult and emotional time.
"A consultant neonatologist has spoken to all of the families on the unit.
"The babies on the unit have been closely monitored for any signs of infection since we withdrew the contaminated feed. We are confident that no other newborns have contracted bacillus cereus.
"The two other babies who were ill are stable and doing well."
A PHE spokeswoman said investigators are now in the final stages of their inquiries.
Experts have so far found that the strain of the bacteria identified in the confirmed cases has been identified in "environmental samples" located within a particular area at ITH Pharma where the feed was manufactured.
They said they have found "sufficient evidence" to confirm that the contamination was introduced to the specific supplies of TPN during manufacture on May 27.
Some unopened supplies of the feed created on this date have also been found to have the same strain of bacteria, the spokeswoman said.
Officials are investigating 19 confirmed cases and four probable cases at 11 NHS organisations in England.
One of the youngsters who died after being struck down with septicaemia after being infected with the bacteria has been named as Yousef Al-Kharboush. The nine-day-old died on June 1 while being treated at the neonatal intensive care unit at St Thomas' Hospital in London.
T hree cases have been confirmed in babies at St Thomas' Hospital, including Yousef.
Elsewhere in the capital, investigators identified four confirmed cases at the Chelsea and Westminster NHS Foundation Trust, one confirmed and one possible case at the Whittington Hospital in north London, and another possible case at private hospital the Harley Street Clinic.
Three were confirmed at Addenbrooke's Hospital, including the latest child that died.
Experts are also looking into three cases in Essex, including one possible and one confirmed case at Southend University Hospital and a possible case at Basildon University Hospital.
They are also looking into further confirmed cases including three at the Royal Sussex County Hospital in Brighton, two at Luton and Dunstable University Hospital, one at Peterborough City Hospital and another at S toke Mandeville Hospital in Buckinghamshire.
Officials have not confirmed the details of the second baby that died but it is understood that the cause of death with this child was not related to the bacillus cereus infection.
"There are still some elements of our investigation that need finalising but the main findings have all pointed towards there being a single incident that occurred on one day and was associated with the illness seen in the babies," said PHE's incident director Professor Mike Catchpole.
"We are reassured that this was a very rare occurrence as we have not seen this particular strain of bacteria in any product made since that day and there has been no further illness."
Gerald Heddell, director of inspection, enforcement and standards at the MHRA, said: " At this stage, our investigation has provided sufficient evidence to indicate that the contamination was introduced into the specific TPN supplies during manufacture in a particular sterile manufacturing area at ITH Pharma on May 27 2014.
"There is no evidence to suggest that individual ingredients, components or materials used for the manufacture of TPN on May 27, 2014, were the cause of the contamination. However, what we do know from our investigation is that the strain of bacillus cereus which infected the babies has also been identified at ITH Pharma's manufacturing facility and within some of the unopened TPN supplies manufactured on May 27, 2014.
"From our investigation to date, we continue to believe this was an isolated incident and that appropriate immediate action has been taken at ITH Pharma's facility to avoid a recurrence. Therefore we are allowing this critical product to be supplied to patients while our investigation continues."