Lucy Letby’s conviction and whole life sentence for attempting to murder the infant known as Baby L rests on a test on a single sample of his blood taken at 3.40pm on 9 April 2016, the second day of his life. A vulnerable twin who had failed to thrive in utero, he weighed just 3.2lb, and despite being put on a dextrose drip, he became dangerously hypoglycaemic. The reason, the prosecution claimed at nurse Letby’s trial, was a “deliberate act of sabotage”, for the blood test proved she had poisoned him with insulin.

Now, I can reveal, fresh evidence casts this claim in doubt. Unbeknown to the jury, the lab that tested Baby L’s blood underwent a routine assessment a few weeks later, which found it exaggerated the level of insulin in a quality control sample by almost 800%.

In recent weeks, as Letby’s 15 whole life orders for seven murders and eight attempted murders have come under mounting scrutiny, neonatal experts from Britain and around the world have criticised the medical evidence used to convict her, suggesting that none of her “victims” were ever deliberately harmed, although some received “suboptimal” care from doctors.

Next week, Letby’s barrister Mark McDonald will deliver their reports to the Criminal Cases Review Commission, asking it to refer the convictions back to the Court of Appeal. His dossier will include detailed analysis of the tests that led to her being found guilty of attempting to murder both Baby L and Baby F, whom she was found to have poisoned in the same way.

No one ever saw Letby injecting a baby with insulin, nor spiking drip bags with it, as prosecuting counsel Nick Johnson KC suggested she had. But the prosecution’s expert witnesses claimed the blood tests showed that although the two infants’ blood contained abnormally high levels of insulin, their C-Peptide levels, a substance naturally produced in tandem with insulin, were very low.

Dewi Evans, the retired paediatrician who played key roles in both the trial and the police investigation, called these results “astonishing”, saying they could have “only one explanation” — that the insulin was not natural but “exogenous”, injected to cause harm. For his part, Johnson told the jury the tests amounted to “conclusive evidence” the babies were poisoned.

This reasoning has already come under attack from experts. The prosecution took it as read that the test results were accurate. But Dr Adel Ismail, an internationally-recognised consultant in clinical biochemistry, says the Roche Elecsys immunoassay test that was used on the babies’ blood is “not reliable enough” to identify exogenous insulin. It is therefore “reckless” to use it as the basis of a criminal prosecution, while there are several possible clinical explanations for the insulin and C-Peptide test results that do not point to intentional harm.

Ismail also notes that recent submissions to the US Federal Drug Administration record “multiple falsely elevated insulin readings” from the Roche immunoassay, so “making its reliability questionable in a legal context”. It would only be safe to do so if other, more reliable tests were used to corroborate its findings — which did not happen with the samples from the Letby case babies.

However, falsely elevated readings have not occurred only in America. They have been noted at the very lab where the blood taken from babies at the Chester neonatal unit where Letby worked was sent to be analysed, at the Royal Liverpool hospital.

In the UK, clinical labs are monitored regularly by Neqas, the National External Quality Assessment Service. On 23 May 2016, the Liverpool lab tested a quality control sample from Neqas, six weeks after it analysed the sample from Baby L.

Before supplying it, Neqas had measured the sample’s insulin and C-Peptide levels using the most advanced methods available. The former was low, at 108 picomoles per litre units, and the latter high, at 873.5 units. Yet when the Liverpool lab analysed the sample using the Roche immunoassay, it claimed its insulin level was 962 units — a wildly inaccurate reading that was 788% too high. However, its C-Peptide finding was much too low at 130 units — just 14.7% of the true level of 873.5.

The prosecution experts at Letby’s trial said Baby L must have been poisoned because there was more than four times as much insulin in his blood than C-Peptide. Yet the bogus results from the quality control sample pointed to an even bigger discrepancy — an insulin level more than seven times higher. Nevertheless, the jury was told the test was totally reliable, and that the only possible explanation for the results from Babies F and L was that Letby had tried to kill them.

The Neqas assessment was mentioned at the trial, but not the bogus result. Gwen Wark, a Neqas biochemist, had made a statement to police saying there had been “no apparent issues whatsoever with the ability of the Royal Liverpool hospital’s laboratory to detect levels of both insulin and C-Peptide”. She told the court the lab was “performing very well”.

Much was made at the trial of the fact that when Letby was being cross-examined, she accepted that the babies had been given exogenous insulin, although she denied that she was the source. Of course, she did not know that the test might produce rogue results. In his summing up to the jury, the judge said they could convict even if they were not sure exactly how Letby had killed or harmed babies. He added: “They [the prosecution] say that this [evidence of insulin poisoning] assists and informs you in relation to the cases of other children who suffered sudden and unexpected collapses for which there was, at the time, no apparent medical explanation”.

The judge also told them: “There is no evidence to doubt the reliability of the test results.” Yet now, it seems, her convictions for attempting to murder Baby F and Baby L rest on an inappropriate test that was never corroborated.

“The insulin cases were absolutely crucial to the prosecution case. It is now clear that the jury heard only half the story about the tests,” says McDonald. “The emerging fresh evidence suggests the convictions are now unsafe. This matter needs to urgently be referred back to the court of appeal”.

---

David Rose is UnHerd‘s Investigations Editor.

DavidRoseUK

---