A postmortem bile screen can help direct further testing, but its overall percentage agreement can conceal serious identification failures. In a 25-case study, amphetamine screening had 80% agreement with LC–MS/MS yet a negative Cohen’s kappa of −0.07. The results did not justify treating an amphetamine screen as confirmed exposure.
What specimens were compared?
Franzin and colleagues examined bile from 25 forensic cases initially investigated for suspected intoxication or overdose. This was selected casework, not a population prevalence survey or a study of 25 proven drug-caused deaths. The case table included several other recorded circumstances and many undocumented causes.
Approximately 3 mL was aspirated from the gallbladder at autopsy into tubes containing sodium fluoride and potassium oxalate. The authors recommended storage at −20°C without repeated freezing and thawing. That recommendation, based on laboratory experience, was not a prospective stability experiment within this screening comparison. The table reports a median death-to-autopsy interval of 5.5, with an interquartile range of 4–7.25, but does not label the time unit; it should not be converted into a documented uniform collection schedule.
A qualitative immunoassay and a separate confirmation method
The investigators adapted the Randox Evidence MultiStat chemiluminescence immunoassay, designed for other biological fluids, to bile. Preparation included centrifugation and 1:10 dilution in phosphate-buffered saline, followed by the assay’s further dilution steps. Its output was positive or negative relative to a cutoff, not a measured bile concentration. The listed instrumental amphetamine cutoff was 50 ng/mL; that number should remain attached to the stated preparation and instrument rather than be treated as a human toxicity threshold.
All 25 specimens underwent LC–MS/MS, including screen-negative specimens. Confirmation covered 108 drugs and metabolites, used two monitored transitions per analyte and incorporated deuterated internal standards. Bile was diluted 1:100, underwent enzymatic deconjugation and was calibrated using the diluted matrix. Results above the upper quantitation limit were handled by further dilution and calculation using the dilution factor.
The main paper describes linearity, accuracy, precision, detection and quantitation assessment, with detailed analyte-specific data assigned to supplementary tables. Those descriptions do not turn the screening device into a quantitative method. The authors expressly list inability to perform analytical validation of the screening test as a limitation.
Why 80% agreement was misleading for amphetamines
The screening table lists four amphetamine-positive specimens; confirmation lists one. Table 4 reports false-positive and false-negative screening percentages of 16% and 4%, respectively, across the 25-specimen comparison. Together, these entries indicate that the four screening positives did not identify the one confirmation-positive specimen. Agreement came from the 20 specimens negative by both methods.
Cohen’s kappa accounts for agreement expected from the positive and negative frequencies. Its −0.07 value exposed what the 80% headline obscured. The confirmed amphetamine result was reported at 111.0 ng/mL in one bile specimen; it was not a blood concentration or a demonstrated impairment level.
The authors proposed interference from putrefactive amines as an explanation, alongside possible cross-reactivity with other substances. They did not report measuring a particular interfering amine in every discordant specimen or proving the individual cause of each false result. A plausible decomposition mechanism therefore remains distinct from an experimentally identified interferent.
Other classes performed better, with important qualifications
Table 4 reports complete agreement for benzodiazepines, opiates, cocaine/benzoylecgonine, oxycodone, buprenorphine and pregabalin. Some positive groups were very small: oxycodone and buprenorphine each had one confirmed case, and pregabalin had two. Perfect agreement in a small selected set does not guarantee future sensitivity.
For barbiturates, agreement was 96% and kappa 0.65; for cannabinoids, 96% and 0.88. Methadone agreement was reported as 92.2% with kappa 0.84. The paper contains numerical and wording inconsistencies: its cannabinoid screening row totals 26 despite a 25-specimen cohort, the methadone percentages do not map exactly onto 25 specimens, and the discussion’s broad statement that all screening negatives were confirmed conflicts with the amphetamine false-negative entry. These issues limit independent reconstruction; they should not be silently corrected or summarized as flawless performance.
What can a bile finding establish?
This comparison supports using a specified bile screening workflow to guide analyte-specific confirmation, with especially poor amphetamine performance in this dataset. It did not validate conversion from bile to blood, reconstruction of the concentration at death, or a rule assigning cause of death from a positive result.
Keep the matrix, preparation, assay target, cutoff and confirmation findings together. Related examples include screening disagreements in urine and pericardial fluid and the limitations of vitreous screening.
Primary source
Franzin M, Ruoso R, Concato M, Radaelli D, D’Errico S, Addobbati R. Applicability of a Chemiluminescence Immunoassay to Screen Postmortem Bile Specimens and Its Agreement with Confirmation Analysis. International Journal of Molecular Sciences. 2024;25:3825. doi:10.3390/ijms25073825.
Leave a Reply