An uninterpretable postmortem carbohydrate-deficient transferrin result cannot be classified as a normal result or evidence against chronic excessive drinking. Rainio and colleagues obtained interpretable electrophoresis results from only 41% of submitted serum specimens. Their study illustrates why analytical success must be separated from diagnostic sensitivity.
The study attempted four specimens per autopsy
The researchers collected blood from 70 corpses: an early femoral sample after mortuary arrival and left femoral, right femoral and inferior-vena-cava samples at autopsy one to six days later. The total was 280 specimens, not 280 independent deceased people.
Blood was centrifuged and serum frozen at −20°C without preservative until testing approximately two months later. All specimens underwent capillary zone electrophoresis. A subset of 76 specimens from 19 people was selected for parallel HPLC analysis.
What prevented a numerical result?
The analyte index depended on identifying transferrin isoform peaks and quantifying disialotransferrin relative to their summed peak areas. When the peaks could not be identified, the result was classified as non-interpretable. This was not a concentration below an established detection limit.
With electrophoresis, 115 of 280 specimens yielded interpretable results, reported as 41%. The authors considered hemolysis the most likely explanation for many failures, but did not experimentally establish the cause of every failed separation. Inferior-vena-cava specimens accounted for 34% of the non-interpretable electrophoresis results; that is a share of failures, not a 34% failure rate for that vessel.
HPLC had three outcome categories
Of 76 specimens intended for HPLC, 41 yielded interpretable results, 24 were non-interpretable and 11 were exhausted by previous analyses. The reported 54% analytical yield includes the specimens no longer available for testing in its denominator.
It would be misleading to describe all 35 remaining specimens as demonstrated HPLC analytical failures. Some could not be analyzed because there was insufficient material. Nor does the 54% figure establish superior diagnostic sensitivity to electrophoresis: the HPLC subset was smaller and the study was not a diagnostic-accuracy trial.
Case-level success is another denominator
At least one interpretable result was obtained in 57 of 70 autopsies, or 81%. Multiple collection sites gave more opportunities for a usable result. That case-level figure does not mean that 81% of people with independently verified heavy drinking were identified.
The investigators explicitly did not compare CDT with drinking histories or relevant autopsy findings to determine sensitivity and specificity. The paper’s analytical yield cannot be repurposed as either measure.
Method agreement did not eliminate specimen failure
Electrophoresis used ultraviolet detection at 200 nm; HPLC used anion-exchange separation with detection at 460 nm after iron saturation and lipoprotein precipitation. The study reported a significant inter-method association, with R² = 0.9128, and no systematic direction of inter-method difference in the compared results.
Agreement among measurable specimens does not supply a value for an uninterpretable one. The article does not provide a complete numerical postmortem validation dataset for detection limits, quantification limits and imprecision, and does not establish an automatic rescue procedure for failed specimens.
How the finding should be communicated
A report should distinguish a measured CDT index, an uninterpretable separation and a specimen unavailable because it was exhausted. None of these categories alone establishes acute impairment or cause of death; CDT addresses a different biological question from an ethanol concentration.
The same reporting principle appears with another analyte in unquantifiable postmortem THC results. The specific causes and analytical methods differ, but a failed measurement is not a negative exposure finding.
Primary source
Rainio J, De Paoli G, Druid H, et al. Post-mortem stability and redistribution of carbohydrate-deficient transferrin (CDT). Forensic Science International. 2008;174:161–165. doi:10.1016/j.forsciint.2007.03.020.
Leave a Reply