Total ketone bodies (acetone, acetoacetate, and β-hydroxybutyrate) were measured in 105 medicolegal autopsies (71 non-alcoholics, 22 chronic alcoholics, and 12 diabetics) using a coupled enzymatic head-space gas chromatographic method. Samples included vitrcous humour, pericardial fluid, and blood from the femoral vein, inferior vena cava (IVC), superior vena cava (SVC), and aorta. Vitreous ketone levels showed good correlation with blood and pericardial fluid levels, suggesting that vitreous could be used as an alternative autopsy specimen for this analysis. This opens up the possibility of using simpler clinical laboratory methodologies which cannot be applied to autopsy blood due to hemolysis. In 71 non-alcoholics (age 18 to 96, median 67) total ketones (mM/L) were: vitreous 0.19 to 3.35, median 0.49; pericardial fluid 0.02 to 1.54, median 0.35; femoral blood 0.23 to 8.08, median 1.00; aortic blood 0.25 to 9.96, median 0.90; IVC blood 0.30 to 6.49, median 1.27; SVC blood 0.32 to 6.00, median 1.07. Eleven outliers (>2.5 mM/L in femoral blood) mostly had prolonged illness prior to death. The 22 alcoholics (age 36 to 83, median 62) included four extreme outliers with femoral blood total ketone levels of 129.9 (also diabetic), 39.4 (no anatomical cause of death), 38.5 (suicidal hanging), and 18.6 (hypothermia), suggesting that while alcoholic ketoacidosis may be a previously overlooked potential cause of death, interpretation must be guarded and made within the total case context. The other 18 alcoholics had ketone levels not statistically different from non-alcoholics, suggesting that ketoacidosis is a significant factor in at most a small minority of alcoholic deaths. Three of 12 diabetics had extreme elevations of femoral blood ketone bodies: 87.5, 20.4, and 17.4 mM/L. Measurement of ketone bodies in vitreous humour or pericardial fluid using clinical laboratory methodologies is recommended in unexplained deaths in chronic alcoholics as well as diabetics.