Why the Letby Insulin Science is Junk

Jamie Egan

Synopsis: The Letby "experts" used adult comparisons which do not apply to neonates to assume the synthesis and secretion of C-Peptide. Peptide and insulin may arrive in equimolar amounts, but they do not leave the body in equimolar amounts, especially in preterm neonates. The assay used does not measure insulin directly, as implied, but measures antibodies (proxies) for insulin which are notoriously cross-reactive with natural contaminants. 

The "experts" invented the notion that Letby had injected insulin into a drip bag running into the victim. Invented because no one saw Letby do anything with insulin. But the same experts in their haste to drown the witch failed to account for the amount of "exogenous insulin" that would naturally stick to the bag and tubing. So great is the propensity for insulin to stick to plastics it is likely that 3-4 vials of insulin would be needed to achieve the readings assigned to Letby. The pharmacy records did not confirm the use of this much insulin on the unit at any time.

The "experts" in the Letby trial misrepresented the physiological dynamics between the hormones proinsulin, insulin, and C-peptide protein. Hawdon et al. (1993) found that plasma immunoreactive insulin concentrations are high in newborn babies compared to older subjects, and compared with older subjects, neonatal pancreatic insulin secretion is less closely linked to circulating blood glucose concentrations.

There are important implications for interpreting studies in hypoglycaemic and hyperglycaemic neonates. The Letby experts claimed that because some neonates in their series had high serum insulin and low serum C-Peptide, Letby used exogenous insulin. While this is a genuinely astounding leap, forensically, the single tests used to convict her are insufficient. The Letby experts did not account for differing rates of C-peptide excretion in neonates, particularly after exogenous insulin administration.

The reliability of c-peptide measurements in preterm neonates is a subject of interest and research. C-peptide is a protein component of the hormone proinsulin; C-peptide + insulin = proinsulin. Specialised cells in the pancreas release the hormone proinsulin, and each proinsulin molecule has one C-peptide and one insulin. However, accurately measuring c-peptide levels, especially in preterm babies, is challenging because of their immature physiology, especially when using exogenous insulin.

There is significant variability in C-peptide secretion in preterm neonates. Since no studies have been performed in preterm or term neonates to determine C-peptide kinetics, researchers use adult data, which is a guess and not forensic evidence. This fact alone suffices to invalidate the presumptions used against Letby.

Variations in blood volume, liver metabolism, and renal function affect the reliability of these measurements. We can use C-Peptide to advise in clinical practice, but it is extraordinarily general and hopelessly removed from the tyranny of the specific needs of forensic analysis. Different laboratory methods and assays can further contribute to the variability in results.

Since none of the Letby babies had measurements for any variables applied in context with C-Peptide secretion, the readings quoted in court are essentially meaningless. In short, C-peptide and insulin may arrive in equimolar amounts, but they do not leave the body in equimolar amounts, especially in preterm neonates.

The secretion of C-Peptide is notoriously variable. And differs significantly between preterm neonates.

Adult physiology assumes a steady state based on theoretical half-lives for these proteins in the blood. The half-life is an assumed period during which the body secretes half of the protein present at the time of measurement. But even if this theoretical steady state existed in adults, it could not apply to preterm neonates because Individualised C-peptide measurement is significantly more accurate than population kinetics.A single measurement cannot define a kinetic (variable) state.

The expert witness's bold assertion that there was no doubt about her interpretation of the results was inaccurate. Worse, she failed to explain to the jury that her test does not measure actual insulin. Still, antibodies to insulin and antibodies used in these measurements often cross-react with precursors of the target protein or with its smaller, degraded fragments, creating considerable uncertainty in results.