A therapeutic plasma range cannot be transferred automatically to postmortem femoral blood. A Finnish study drawing on 57,903 autopsy cases found that drug-specific median blood concentrations often lay outside published therapeutic plasma intervals. Its comparisons provided interpretive context, not a test proving overdose or a formula recovering the concentration during life.
Which cases and specimens contributed?
Launiainen and Ojanperä examined a national forensic laboratory database covering January 2000 through December 2010. The cases represented multiple causes of death, including poisonings, rather than a uniform drug-free control population. The headline total was the number of autopsies with comprehensive drug analysis; each drug had its own smaller number of findings.
The average death-to-autopsy interval was five days. Bodies were kept at approximately 3–5°C after arrival at the morgue. Femoral venous blood collected at autopsy contained 1% sodium fluoride and was stored at 4°C during the laboratory investigation, usually completed within 12 days after the sample arrived.
Those are cohort handling descriptions, not proof that every drug remained stable or every femoral result equaled the concentration at death. Individual premortem comparison samples were not the basis of this study.
Screening and quantitative distributions were different datasets
Urine screening used immunoassays and LC–time-of-flight MS. Blood monitoring used several chromatographic methods, with changes during the 11-year period. GC–MS and LC–MS/MS supplied confirmation and additional determinations. There was no single fixed assay with one detection limit for all 129 reported drugs.
The investigators initially selected drugs detected at least 50 times, including findings reported only as positive below the quantitation limit. Concentration statistics were calculated from numerical results. Some method-specific subsets therefore had fewer than 50 quantified cases. Positive findings and quantified observations should not be treated as identical denominators.
To limit differences caused by changing methods, certain results were drawn only from specified subsets: 53,095 cases for the stated benzodiazepine/nifedipine method period and 23,385 for selected acidic or neutral drugs. The main table gives drug-specific quantitation limits; detailed validation was assigned to supporting information. Neither the overall case count nor laboratory accreditation substitutes for analyte-specific performance.
The “blood/plasma relationship” was not a measured ratio
The authors constructed a descriptive comparison with published therapeutic plasma ranges. If a drug’s postmortem blood median fell within that range, they assigned a relationship of 1. If it fell below the range, they divided by the lower boundary; if above, they divided by the upper boundary.
Consequently, 1 did not mean equal measured blood and plasma concentrations. It was a category for any median inside an interval. Nor were the numerator and denominator paired samples from the same person. Dividing a new autopsy result by this relationship would not produce a validated premortem plasma estimate.
Of 129 drugs, 61 had an assigned relationship of 1, 22 below 1 and 46 above 1. The investigators found no marked correlation with literature volume of distribution. These descriptive results do not establish a drug-by-drug causal mechanism for the observed differences.
Amlodipine illustrates the reference-range problem
For amlodipine, 313 quantified cases had a median femoral concentration of 0.07 mg/L, with a listed quantitation limit of 0.006 mg/L. The published therapeutic plasma range used for comparison was 0.003–0.015 mg/L. Thus, the typical autopsy value in that dataset exceeded the plasma interval’s upper boundary.
That finding warned against labeling an autopsy concentration abnormal solely because it exceeded a living-person plasma reference. It did not prove that 0.07 mg/L is harmless in every death, that all 313 individuals took therapeutic doses, or that a particular driver was unimpaired. The full case remains necessary.
Upper percentiles were not lethal cutoffs
The study reported medians, means and upper percentiles. For many drugs, the means were much higher than the medians, reflecting skewed distributions. A high percentile described where a value sat among the included quantified observations; it did not mark the boundary between survival and death.
The authors separately tabulated the percentage of findings attributed as the principal drug in fatal poisonings. For 36 drugs that proportion exceeded 10%, making those distributions especially unsuitable as purely non-intoxicated reference populations. Co-detected substances, alcohol, tolerance, use patterns and selection could also influence interpretation.
The article’s conclusion contains a conflicting total of 59,903, while its abstract, methods and principal table identify 57,903. The latter is the study population used here. That inconsistency should not be converted into an additional 2,000 cases.
How the database can help
Historical postmortem distributions can identify when an inference depends on an inappropriate plasma comparator. They cannot independently establish dose, impairment, poisoning or the extent of postmortem change in an individual. The sampling site, method period, quantitation limit, numerical denominator and case circumstances need to accompany the reference value.
For complementary evidence, see paired hospital–autopsy comparisons of cardiovascular drugs and limits of ranking redistribution from volume of distribution.
Primary source
Launiainen T, Ojanperä I. Drug concentrations in post-mortem femoral blood compared with therapeutic concentrations in plasma. Drug Testing and Analysis. 2014;6:308–316. doi:10.1002/dta.1507.
Leave a Reply