Murder by insulin: suspected, purported and proven-a review.

Murder by insulin-whether attempted, suspected or proven-is rare. Only 66 cases worldwide could be found for this review. A conviction was secured in 31 cases and additional weapon was employed in 11. Differentiation of attempted homicide from Munchausen syndrome by proxy in the young and from 'mercy killing' in the elderly was not attempted. Most perpetrators were close relatives and most victims were alive when discovered and responded to treatment. Hypoglycaemia is the first clue to homicidal insulin use in living subjects and requires the demonstration of a plasma insulin concentration of generally more than 1000 pmol/L and undetectable plasma C-peptide concentration to establish the diagnosis. Serum glucose measurements are valueless in victims found dead. The presence near the body of insulin vials, syringes or needles, loose talk by the suspected perpetrator or their ready access to insulin may be the only clue. The demonstration of insulin in tissue around an injection site by immunohistopathology or by measuring it in an extract clinches the diagnosis. Immunoassays suitable for clinical use to detect and measure insulin and C-peptide are subject to random errors and cannot be relied upon unless special precautions including separation by gel filtration or HPLC are undertaken prior to analysis. They do not detect or measure accurately a new generation of synthetic insulin analogues. Mass spectrometry will be required to do this and to validate clinical immunoassays, upon which convictions have always had to rely in the past.