Category: Uncategorized

A study can find no deterioration on selected tests without establishing that a medicine carries no cognitive risk. Coghill and colleagues’ 2018 report is a useful example: its primary purpose was cognitive monitoring during prolonged lisdexamfetamine exposure, while its design could not isolate treatment benefits.

Who was followed?

The study enrolled 314 children and adolescents aged 6–17 with ADHD. All received open-label, dose-optimized oral lisdexamfetamine dimesylate, 30, 50 or 70 mg/day, for up to 104 weeks. Lisdexamfetamine is a prodrug of d-amphetamine; the reported milligram doses refer to the prodrug formulation.

Only 191 participants, 60.8%, completed the study. Withdrawal by participants and adverse events were prominent reasons for discontinuation. About 39.5% had participated in a previous lisdexamfetamine study, so the cohort was not wholly treatment-naive.

Which cognitive outcomes were measured?

Four computerized CANTAB tasks assessed delayed matching, spatial working memory, stopping responses and reaction time. The investigators prospectively defined key variables and treated group mean percentage changes exceeding 5% as potentially clinically significant.

No key variable showed potentially clinically significant deterioration under that rule. Some measures improved: spatial working-memory between-search errors declined by 22.8% at six months, while stop-signal reaction time declined by 18.9%. These are changes in test performance, not percentages of children cured or protected from harm.

Why the last assessment is not always a two-year result

The report also used the last on-treatment assessment. That observation may precede week 104 for a participant who leaves early. It should not be described as a complete two-year examination of all 314 children.

Attrition can affect which participants contribute later results. Group averages can also conceal individual deterioration, and four tasks cannot cover every cognitive, emotional or functional outcome.

Did the medicine cause improvement?

There was no placebo or untreated comparison group. Development over two years, repeated-test familiarity, changing symptoms and treatment selection could contribute to score changes. The authors specifically identified the difficulty of separating treatment effects from maturation and practice.

The study therefore supports the narrower historical conclusion that the monitored group did not show the predefined pattern of deterioration on these selected measures. It does not establish risk-free use, enhancement in people without ADHD or an individual’s fitness to drive. The work was industry sponsored, a relevant feature when evaluating the overall evidence.

Scientific source

Coghill DR, Banaschewski T, Bliss C, Robertson B, Zuddas A. Cognitive Function of Children and Adolescents with Attention-Deficit/Hyperactivity Disorder in a 2-Year Open-Label Study of Lisdexamfetamine Dimesylate. CNS Drugs. 2018;32:85–95. doi:10.1007/s40263-017-0487-z. Original commentary on the supplied paper.

Postmortem urine and bile can support cannabinoid detection, but their concentrations cannot simply be read as blood concentrations or used as a clock for impairment. Saenz and colleagues examined 11 aviation fatalities and found useful qualitative results across non-blood specimens, without consistent concentration relationships to blood. Urine hydrolysis adds another essential distinction: the measured result includes material released from conjugates.

A small, selected series of aviation fatalities

The investigators selected 11 previously cannabinoid-positive cases from 2014–2015 with blood and most desired specimens available. All were male, aged 22–69 years. Urine was available in eight cases and bile in nine. These specimen denominators must remain separate from the total of 11 fatalities.

Blood was preserved with 1% sodium fluoride and 0.2% potassium oxalate; all specimens were stored at −20°C until analysis. Postmortem intervals were unknown. Blood collection sites were often undocumented; when provided in the laboratory’s casework, chest-cavity collection was common. The study’s blood comparator therefore cannot be uniformly relabeled as femoral blood.

Urine underwent hydrolysis

The method used solid-phase extraction and LC–MS/MS to quantify THC, 11-OH-THC and THC-COOH. Urine was incubated with beta-glucuronidase for three hours at 65°C before extraction to release glucuronide-bound material. A hydrolyzed urine result is analytically different from measuring only free drug in an untreated specimen.

THC and THC-COOH glucuronide controls were included, with results within 20% of target. An 11-OH-THC glucuronide standard was unavailable, so the corresponding hydrolysis control could not be included. That limitation matters when interpreting the completeness of metabolite deconjugation.

The distributions differed by analyte and specimen

Seven of eight urine specimens contained THC, ranging 1.25–32.2 ng/mL. All eight contained 11-OH-THC and THC-COOH, with ranges of 11.7–620 ng/mL and 24.2–970 ng/mL, respectively. These were hydrolyzed urine measurements.

Five of nine bile specimens contained THC, ranging 0.78–50.4 ng/mL. All nine contained 11-OH-THC and THC-COOH, ranging 0.98–227 ng/mL and 201–3307 ng/mL. Bile had the highest THC-COOH concentration within each case where it was available. A high bile concentration therefore cannot be assumed to indicate an equally high circulating concentration.

The reporting limit was a defined analytical choice

The method used eight calibrators spanning 0.78–100 ng/mL. Detection and quantification limits were administratively set at 0.78 ng/mL because it was the lowest tested calibrator meeting the identification and accuracy requirements. Lower concentrations were not evaluated to establish a lower validated reporting limit, and case results below the adopted limit were classified as negative.

Identification required retention-time agreement, product-ion ratio agreement and minimum signal-to-noise criteria. Dilution integrity was evaluated for concentrations above the calibration range. Thus, the quantitative findings involved a defined confirmatory procedure rather than a urine immunoassay screen.

Why detection does not date impairment

No consistent distribution between blood and any non-blood specimen was identified. Dose, route, frequency and time of cannabis use were unknown. There were no matched antemortem samples to show the concentrations before death, and the unknown collection intervals complicated assessment of redistribution.

The presence of 11-OH-THC, an active metabolite, should not be confused with proof of impairment at a particular earlier time. THC-COOH is an inactive metabolite, and its high concentration in excretory specimens is not a direct measure of psychoactive effect. The study did not establish crash causation or a fatal cannabinoid threshold.

Its strongest practical contribution is support for supplemental qualitative testing with transparent specimen and method labels. For the broader issue of alternative matrices, see why a useful screening specimen does not automatically replace blood.

Primary source

Saenz SR, Lewis RJ, Angier MK, Wagner JR. Postmortem Fluid and Tissue Concentrations of THC, 11-OH-THC and THC-COOH. Journal of Analytical Toxicology. 2017;41:508–516. doi:10.1093/jat/bkx033.

Muscle can provide evidence of cannabinoids when blood is unavailable, but a muscle concentration is not an interchangeable blood concentration. Pettersen and colleagues’ 39-case autopsy study found THC and CBD in psoas and vastus lateralis muscle, with wide variation between muscle and femoral-blood results. The findings support alternative-specimen detection more strongly than a fixed conversion.

Which cases and specimens were included?

The study selected 39 cases from a larger autopsy project after routine blood testing detected THC. Specimens were collected in parallel at Oslo medicolegal autopsies during 2013–2016. The postmortem interval ranged from zero to six days. Thirty-six cases had no or slight decomposition, one had moderate decomposition and two were ungraded. Extensive burns, severe putrefaction and major blood loss were not represented as the study’s target conditions.

Peripheral blood came from the femoral vein. Blood and muscle were placed in polypropylene tubes containing potassium fluoride. Specimens were refrigerated at 4°C until analysis, generally within one to two weeks. Muscle was homogenized with water before extraction, and reported concentrations were corrected for that preparation dilution.

The anatomical muscle site mattered

THC was detected in 36 psoas specimens and 36 vastus lateralis specimens. CBD was detected in 35 psoas specimens and 33 vastus lateralis specimens, compared with 35 CBD-positive peripheral-blood specimens among the 39 THC-selected cases.

The median muscle-to-blood ratios for THC were 2.2 for psoas, ranging 0.11–17, and 5.5 for vastus lateralis, ranging 0.14–29. For CBD, corresponding medians were 2.2 and 4.2, with ranges 0.037–18 and 0.036–32. Because muscle was reported in mg/kg and blood in mg/L, these quotient values carry dimensions of L/kg; they are not dimensionless comparisons of identical materials.

Removing four vastus-lateralis THC outliers changed its median ratio from 5.5 to 4.0 and narrowed the range. That sensitivity reinforces why a published median should not be treated as a dependable divisor for an individual specimen.

Correlation did not establish interchangeability

Spearman correlations with peripheral blood were 0.57 for psoas THC and 0.45 for vastus-lateralis THC; CBD correlations were 0.64 and 0.35, respectively. These associations were statistically significant, but significance does not eliminate the wide individual ratios or validate a blood-prediction equation.

The authors found no systematic ratio differences with postmortem interval. That observation does not prove absence of redistribution. Samples were collected at one postmortem time point, and no antemortem specimens were available.

Screening and confirmation were separate

Routine peripheral-blood THC testing used UHPLC–MS/MS. Subsequent confirmation and quantitation of THC and CBD across the project specimens used GC–MS with selected-ion monitoring. The reported quantification limit was 0.0002 mg/L for both analytes. The reported long-term bias and precision data came from whole-blood controls, so they should not be represented as a complete independent muscle-validation package.

Whole-blood between-day CVs were 9% for both analytes at 0.0008 mg/L; at 0.0032 mg/L they were 9% for THC and 5% for CBD. Homogenization and matrix-specific performance remain important when evaluating an actual muscle result.

A useful alternative with a narrower inference

A confirmed muscle finding may support cannabinoid detection. This study did not establish an administered dose, time of use, impairment or cause of death from muscle concentration. Its selected, largely minimally decomposed cases also do not validate performance in every badly damaged body.

The same distinction between useful detection and quantitative substitution appears in vitreous drug screening. The appropriate matrix must be considered for the specific analyte and analytical method.

Primary source

Pettersen S, Øiestad ÅML, Rogde S, Brochmann GW, Øiestad EL, Vindenes V. Distribution of tetrahydrocannabinol and cannabidiol in several different postmortem matrices. Forensic Science International. 2021;329:111082. doi:10.1016/j.forsciint.2021.111082.

GHB can increase after collection in stored postmortem blood from people with no suspected GHB use. Beránková and colleagues demonstrated substantial, sometimes temporary changes during storage. A later measured concentration therefore cannot automatically be assigned to an administered dose or assumed to equal the concentration at autopsy.

The pooled experiment had a small biological denominator

The 2006 study’s first experiment followed material for eight months under four conditions: 4°C or −20°C, each with or without 1% weight/volume sodium fluoride. Postmortem femoral blood came from one decedent; postmortem urine was pooled from three. The antemortem comparison used serum pooled from four people and urine from one person. These were not large matched populations.

Postmortem material was collected at routine autopsy within two days after death, with the experiment beginning no later than five days after death. Repeated extracts were technical measurements of the same source material, not additional independent decedents.

Refrigerated, unpreserved blood showed a substantial rise

In the postmortem blood stored at 4°C without fluoride, the four-month result was 99.0 ± 16.6 mg/L, based on triplicate measurement and reported standard deviation. Concentrations subsequently declined; the eight-month value was 15.8 ± 6.6 mg/L. The changes were not a monotonic accumulation that could be extrapolated backward at a fixed rate.

Fluoridation and freezing generally inhibited formation, but neither should be summarized as guaranteeing a zero result. The refrigerated fluoridated blood had an initial reported value of 51.2 mg/L after one day, followed by much lower values. Frozen blood also showed occasional measurable concentrations. The authors could not determine that either freezing or fluoride was uniformly the more effective intervention across their experiments.

The individual-specimen experiment adds a different limitation

Twenty individual postmortem femoral-blood specimens were collected at autopsy approximately three days after death. The methods describe 16–27 days of refrigerated storage without preservative before the first analysis and addition of fluoride. They were then stored at 4°C for another four months.

Some specimens showed a temporary increase around the two-month measurement and later declined. By four months, the highest reported concentration was 6.5 mg/L. Adding preservative late cannot establish what happened during the earlier unpreserved interval, and the sparse later time points cannot recover an individual specimen’s complete concentration history.

Quantitation has method-specific uncertainty

The investigators used extraction, derivatization and GC–MS in selected-ion monitoring, with adipic acid as internal standard and matrix-specific calibration. The reported quantification limits were 3 mg/L for blood and serum and 0.5 mg/L for urine. Detection and quantification limits were not separately differentiated.

Intra-assay CV was 21% in postmortem blood and 22% in bovine serum at 50 mg/L, compared with 5% in urine at 20 mg/L. These are material precision limits. An unreported or below-limit value is not proof that GHB was absent, and small numerical differences should not be overinterpreted.

What the study permits an expert to conclude

The finding supports investigating storage-related formation when interpreting postmortem GHB. It does not establish that every positive result is artifactual, that known ingestion can be excluded, or that the observed concentrations define a universal endogenous/exogenous cutoff. The subjects had no suspected use; this was not independent proof of zero lifetime or recent exposure.

Obtain the collection site, interval, initial result, preservative timing, storage temperatures and analytical records. Keep specimen chemistry separate from impairment and cause-of-death attribution. The general need to distinguish detection from quantitative interpretation is also illustrated by alternative-matrix drug screening.

Primary source

Beránková K, Mutnanská K, Balíková M. Gamma-hydroxybutyric acid stability and formation in blood and urine. Forensic Science International. 2006;161:158–162. doi:10.1016/j.forsciint.2006.01.017.

More advanced decomposition does not necessarily mean more ethanol is being produced at the time a specimen is examined. In Pigaiani and colleagues’ 2024 experiment, early-putrefactive tissue preparations generated measurable ethanol, while preparations made from fresh and advanced-putrefactive tissues remained below the quantification limit. This was an in vitro comparison of stages, not proof that advanced decomposition excludes ethanol formed earlier.

How decomposition stage was tested

The study added cadaveric gastrointestinal mucosa and liver fragments to whole human blood from expired blood bags. Tubes contained no preservative, and paraffin oil created an anaerobic environment. This was a constructed microbial substrate, not blood sampled serially from the same intact bodies throughout decomposition.

The stage experiment used 30 blood preparations with glucose below 100 mg/dL and another 30 above 100 mg/dL. Within each glucose stratum, ten preparations received fresh tissues, ten early-putrefactive tissues and ten advanced-putrefactive tissues. “Fresh” meant a postmortem interval below 48 hours without putrefactive changes; early putrefaction involved approximately 48 hours to seven days with characteristic visible changes; advanced specimens came from beyond 21 days with active decay. Preparations were incubated for nine days at 20°C.

A negative stage experiment is not a negative exposure history

Under those conditions, fresh and advanced-stage preparations produced ethanol below the study’s 0.01 g/L quantification limit. Early-stage preparations produced measurable ethanol. The authors linked this pattern to differing microbial conditions and the fermentative potential of organisms associated with the stages.

Crucially, real bodies pass through stages consecutively. An advanced-stage body may contain ethanol produced during an earlier period. The experiment evaluated tissue stages separately; it did not establish that ethanol already present disappears whenever the body reaches advanced decomposition. “Below quantification” also does not mean absolute chemical absence.

Glucose and temperature changed the result

A separate comparison used ten higher-glucose and ten lower-glucose early-putrefactive preparations at 20°C for nine days. Higher-glucose preparations produced more ethanol, with a significant difference (P < 0.0001). The tabulated nine-day concentrations ranged from 0.01–0.17 g/L in the lower-glucose preparations and 0.02–0.32 g/L in the higher-glucose preparations.

Another comparison used ten higher-glucose early-putrefactive preparations at 4°C and ten at 20°C. Production was higher at 20°C after nine days (P < 0.0001). This supports temperature as a modifier under the tested conditions; it does not establish that refrigeration prevents every instance of ethanol formation.

Formation did not behave as a simple clock

In ten preparations followed over nine days, ethanol initially increased through approximately day three, then fluctuated through day eight before increasing again. The investigators proposed changing substrate availability and competing synthesis and breakdown as explanations.

The study therefore does not support a constant hourly ethanol-production rate or an equation that dates death from one ethanol concentration. Glucose, tissue source, microbial environment and temperature affected the output. The sample counts describe laboratory preparations, and the paper does not establish one independent decedent for every preparation.

What a measured concentration means

Headspace GC–FID measured the volatile compounds with two columns, replicate determinations, an internal standard and recurring quality control. Even sound analytical quantitation cannot by itself distinguish drinking from microbial synthesis in an individual decedent.

The practical inference is to investigate the whole specimen history rather than assign an origin from a decomposition label. This experiment did not establish impairment or cause of death. The complementary value of a documented predeath sample is discussed in hospital ethanol specimens in death investigations.

Primary source

Pigaiani N, Musile G, Scott KS, Dye DW, Ausania F, Davis GG, Bortolotti F. Post-mortem formation of ethanol: Is 1-propanol a reliable marker? A proof-of-concept study using an in vitro putrefactive environment setup. Journal of Forensic Sciences. 2024;69:974–985. doi:10.1111/1556-4029.15479.

1-Propanol can provide evidence relevant to microbial ethanol formation, but a laboratory correlation does not automatically reveal how much of an individual autopsy ethanol result was produced after death. Pigaiani and colleagues developed a promising 2024 model in an experimental putrefactive-blood system. The conditions and limitations of that model must travel with its equation.

A constructed substrate, not a series of reconstructed drinking histories

The investigators combined ethanol-negative whole human blood from expired blood bags with 300 mg mixtures of cadaveric gastrointestinal mucosa and liver. Samples were placed in unpreserved 6 mL tubes, with sterile paraffin oil used to create strictly anaerobic conditions. This simulated aspects of decomposition without reproducing an intact body’s circulation, anatomy or exposure history.

For the time-course and correlation phase, ten blood preparations containing early-putrefactive tissue fragments and glucose above 100 mg/dL were incubated at 20°C for nine days. Volatile compounds were measured daily. Ten preparations and repeated daily readings are not ten independently validated forensic case reconstructions.

What the regression showed

In this phase, newly formed ethanol ranged from 0.01 to 0.57 g/L, while 1-propanol ranged from 0.006 to 0.053 g/L. The reported relationship had R² = 0.91 and P < 0.0001. The fitted equation was:

Ethanol (g/L) = 9.512 × 1-propanol (g/L) + 0.0131.

The authors reported a mean percentage error of 18.4% against measured ethanol. That is the paper’s reported model-performance statistic, not a ±18.4% uncertainty interval guaranteed for a new case. The figure’s confidence curves concern the regression line; they should not be treated as individual-case prediction limits.

The analytical result is separate from the source inference

Ethanol and other volatiles were measured with headspace GC–FID. The method used two capillary columns and the mean of replicate results, tert-butanol as internal standard, and a 0.5 g/L ethanol quality-control standard every five samples. The paper identifies an ethanol quantification limit of 0.01 g/L. This is chromatographic measurement of the experimental samples, not merely a presumptive alcohol screen.

Accurately measuring both compounds does not independently validate the equation in a specimen with a different microbial population, glucose content, temperature history or pre-existing ethanol. Analytical quantitation and attribution of origin remain separate steps.

Why automatic subtraction is not supported

The starting blood was ethanol-negative. The authors identified experiments beginning above 0.0 g/L ethanol as future work. The study therefore did not establish that the same relationship remains reliable in a person who drank before death and subsequently underwent microbial ethanol formation.

Glucose did not exceed 120 mg/dL in the reported experiments. The proposed extension to higher glucose levels was an expectation, not direct validation across severe hyperglycemia. Expected production of other higher alcohols, including 1-butanol, did not occur, highlighting that the microbial environment did not capture every real-case pattern.

These limits preclude using the formula as a universal subtraction rule to calculate a “true antemortem BAC.” The model does not establish impairment, drinking time or cause of death. It supports further investigation of microbial contribution within the specific experimental framework.

How the study helps a case review

Record the specimen source, collection delay, preservation, storage, units and complete volatile profile before considering any model. Compare the experimental conditions with the case conditions and identify the evidence supporting transfer. For a different route to evidence about predeath ethanol, see hospital specimens used in death investigations.

Primary source

Pigaiani N, Musile G, Scott KS, Dye DW, Ausania F, Davis GG, Bortolotti F. Post-mortem formation of ethanol: Is 1-propanol a reliable marker? A proof-of-concept study using an in vitro putrefactive environment setup. Journal of Forensic Sciences. 2024;69:974–985. doi:10.1111/1556-4029.15479.

A cocaethylene finding and an ethanol result are different measurements. In an 81-patient clinical study, one patient in the ethanol-negative group had cocaethylene detected in blood. That observation cautions against defining exposure history solely by contemporaneous ethanol detection, but it does not establish an exact drinking time, dose or postmortem explanation.

The groups were defined by the measured ethanol result

Zucoloto and colleagues enrolled adults evaluated for suspected cocaine or crack intoxication, selected through a positive urine cocaine screen. Three of 84 initially assessed patients were excluded for exposure to other stimulants. Among the remaining 81, Group A contained 62 without positive blood ethanol, while Group B contained 19 with positive blood ethanol.

These labels did not perfectly separate people who had and had not consumed alcohol. Thirty of 62 in Group A reported alcohol association, compared with 14 of 19 in Group B. The difference in self-report proportions was not statistically significant at the prespecified 5% level (P = 0.053). An ethanol-negative classification should therefore not be paraphrased as a verified history of cocaine-only use.

Cocaethylene detection and concentration had different denominators

Cocaethylene was detected in one of 62 Group A patients and four of 19 Group B patients. The single Group A value was 0.22 μg/mL. In the four positive Group B specimens, the mean was 0.42 ± 0.38 μg/mL, with a range of 0.10–0.87 μg/mL.

Detection was more frequent in Group B (P = 0.002), but the between-group concentration difference was not significant. Comparing one positive observation with four does not establish a dependable quantitative rule for alcohol exposure. The paper discusses cocaethylene as a product of cocaine–ethanol interaction; its observed detection pattern does not independently date that interaction.

Collection timing limits interpretation

Blood and urine were collected with the initial clinical assessment. Blood samples of approximately 3 mL were taken into EDTA tubes. The report does not specify a uniform anatomical collection site or provide a complete storage timeline. Time from exposure was known for only 23 patients overall, including six in Group B. The Group B mean was 4.7 ± 4 hours, range 1–12, compared with 7.2 ± 5.4 hours in 17 Group A patients.

Those selected timing subsets cannot explain every discordant result. The design did not serially sample all patients to validate a time-since-use equation.

The analytical and statistical boundaries matter

The initial urine test was an immunoassay for benzoylecgonine, cutoff 150 ng/mL. Cocaethylene in blood was measured separately by HPLC with diode-array detection after liquid-liquid extraction. The article refers to a previously published method and does not supply a complete set of numerical blood-assay validation parameters. A urine screen is not confirmation of a blood cocaethylene concentration.

Clinical severity also requires careful reporting: mean symptom scores differed nonsignificantly between the groups in the initial analysis (P = 0.103). Removing two Group A outliers produced P = 0.008. Both analyses matter; the latter should not be described as the sole, unqualified study result or proof that cocaethylene caused greater severity.

What the finding does not establish after death

This was living-patient research, not an investigation of ethanol production during decomposition. It did not validate cocaethylene as a postmortem formation discriminator, determine a fatal concentration, or reconstruct impairment. A death investigation requires its own specimen history, analytical validation and case evidence.

The broader distinction between the time of collection and the later interpretation is addressed in hospital ethanol specimens in death investigations. The observed discordance here remains a clinical finding, with the limits of that design.

Primary source

Zucoloto AD, Eller S, de Oliveira TF, Wagner GA, Fruchtengarten LVG, de Oliveira CDR, Yonamine M. Relationship between cocaine and cocaethylene blood concentration with the severity of clinical manifestations. American Journal of Emergency Medicine. 2021;50:404–408. doi:10.1016/j.ajem.2021.08.057.

An emergency-department study cannot establish a lethal postmortem cocaine cutoff simply by reporting concentrations in symptomatic patients. Zucoloto and colleagues studied 81 living adults and found that cocaine and cocaethylene concentrations did not provide a useful concentration-based relationship with clinical severity. The design did not evaluate autopsy specimens or validate a fatal threshold.

Who was selected, and when were samples collected?

The observational study enrolled patients with suspected cocaine or crack intoxication at a São Paulo hospital between December 2015 and March 2017. A positive urine cocaine screen was required. Of 84 initially assessed patients, three were excluded for other stimulant exposure, leaving 81. Most reported intranasal use; smoked crack and other routes also occurred. There was no controlled administered dose.

Blood and urine collection accompanied the initial clinical assessment. Approximately 3 mL blood was collected into EDTA tubes. The article does not specify a uniform anatomical blood-collection site or detailed storage conditions in this report. Time since last use was estimable in only 23 patients: mean 6.6 ± 5.1 hours, range 1–24. This timing uncertainty is material, not an incidental omission.

The urine screen did not measure blood cocaine

Urine screening used an immunoassay with a 150 ng/mL benzoylecgonine cutoff. Blood cocaine and cocaethylene were assessed separately using liquid-liquid extraction and HPLC with diode-array detection. The paper refers to an earlier method publication rather than reproducing a full numerical validation package. A reader should not assume a blood detection limit from the urine cutoff.

Cocaine was detected in blood in 20 of 81 patients. In that positive subset, the mean was 0.34 ± 0.45 μg/mL, with a range of 0.11–1.79 μg/mL. Cocaethylene was found in five patients, with a mean of 0.38 ± 0.34 μg/mL. These means describe detected subsets, not the entire cohort.

Severity was measured clinically, not inferred from a number

The Stimulant Intoxication Scale assessed vital signs, orientation and mental state. Seventy-four patients had mild scores, seven had moderate scores and none occupied the scale’s severe category. The observed score range was 0–13. The study therefore had limited coverage of severe toxicity and was not a fatal-case comparison.

Patients with chest pain had higher average cocaine and cocaethylene concentrations than those without chest pain, but neither comparison was statistically significant. Likewise, blood concentration differences between ethanol-positive and ethanol-negative groups were not significant. Higher averages should not be restated as proven concentration-response rules.

Why a death investigation needs a different inference

The authors concluded that measured concentrations did not prove useful for treatment or prognosis in this sample. That finding does not mean cocaine cannot cause serious harm. It means this study did not establish a reliable concentration-based severity classification.

There were no paired predeath and postmortem specimens, no autopsy-site comparisons and no decomposition experiment. It cannot estimate redistribution, a lethal threshold, a probability of death at a specified concentration, or the cause of an individual death. Its clinical participants also cannot be treated as a representative sample of all exposed people.

For forensic interpretation, separate the laboratory finding from the proposition being offered. A concentration in a living patient is evidence about that specimen and assessment context; applying it to autopsy blood requires additional evidence. Our review of hospital specimens in death investigations illustrates why collection timing and matrix must remain explicit.

Primary source

Zucoloto AD, Eller S, de Oliveira TF, Wagner GA, Fruchtengarten LVG, de Oliveira CDR, Yonamine M. Relationship between cocaine and cocaethylene blood concentration with the severity of clinical manifestations. American Journal of Emergency Medicine. 2021;50:404–408. doi:10.1016/j.ajem.2021.08.057.

A controlled animal experiment can show that postmortem change is biologically plausible without providing a correction factor for human autopsy blood. Hilberg and colleagues’ six-rat THC experiment is a useful example: the measured direction and magnitude depended on the vessel sampled, and the protocol differed substantially from human casework.

Six animals under a specific protocol

Male Wistar rats weighing 260–300 g were fasted overnight and given 30 mg THC mixed with 1 mL water by gastric tube. This was a fixed oral dose per animal, not a 30 mg/kg dose. Ninety minutes later, the animals were anesthetized and antemortem heart blood was collected. They were then sacrificed with carbon dioxide and left at room temperature for two hours before autopsy sampling.

Postmortem heart blood was drawn after clamping the inferior vena cava above the diaphragm. Vena-cava blood was collected after clamping above the renal veins; vitreous was obtained from both eyes. These controlled sites and intervals are essential to the result.

The vessel changed the finding

The mean postmortem-heart-to-antemortem-heart concentration ratio was 1.2 ± 0.1, expressed as mean ± standard error, and was not statistically significant. The postmortem-vena-cava-to-antemortem-heart ratio was 0.6 ± 0.1, a significant decrease (Wilcoxon P < 0.05). The postmortem-heart-to-vena-cava ratio was 1.9 ± 0.4 but was not significant.

These comparisons have different denominators and must not be collapsed into “THC falls 40% after death.” The significant 0.6 ratio compared a postmortem vena-cava specimen with an earlier heart specimen. It did not follow the same vena-cava blood through time.

The vitreous-to-antemortem-heart ratio was 0.12 ± 0.04 and significantly below one. This describes a different matrix, not a blood redistribution percentage. It does not validate vitreous as an interchangeable quantitative substitute.

What was measured analytically?

THC was quantified using capillary GC–MS with selected-ion monitoring and a deuterated internal standard. The stated blood calibration range was 0.001–0.305 μmol/L, with R² ≥ 0.997. The method was unsuitable for tissue analysis, so the rat results reported blood and vitreous rather than a complete set of tissue concentrations. Calibration linearity alone should not be represented as comprehensive validation for every matrix.

Why translation to human autopsy requires restraint

The oral route, fixed dose, species, short survival interval and two-hour room-temperature postmortem period all define the experiment. The authors specifically noted that human THC exposure commonly involves inhalation, which can alter initial tissue distribution. The experiment cannot resolve effects of days of decomposition or every human sampling technique.

The same paper included two human early-to-later blood comparisons with THC ratios in opposite directions. Their sampling sites also differed, and one case’s autopsy timing differed between narrative and table. Those cases are descriptive evidence of uncertainty, not validation of either a universal increase or decrease.

Accordingly, neither dividing a human result by 0.6 nor multiplying it by 1.2 is justified by this experiment. It did not test human impairment, crash causation or a fatal THC threshold. Our discussion of human central-to-peripheral ratios addresses a separate observational design; neither design supplies a universal antemortem reconstruction.

Primary source

Hilberg T, Rogde S, Mørland J. Postmortem Drug Redistribution—Human Cases Related to Results in Experimental Animals. Journal of Forensic Sciences. 1999;44(1):3–9. doi:10.1520/JFS14404J.

A very large postmortem-to-antemortem ratio is evidence that the two measurements differ; it is not automatically the amount by which the drug redistributed after death. A chloroquine case reported by Hilberg and colleagues produced a ratio of approximately 49. The comparison involved different matrices and different collection times, making it unsuitable as a universal correction factor.

The sequence behind the number

The case concerned a 15-year-old patient who developed vomiting, unconsciousness and cardiorespiratory arrest. She received intensive care and was pronounced dead after six days with irreversible hypoxic brain damage. Serum chloroquine was 4.0 μmol/L on admission, 2.3 μmol/L eight hours later, and 1.3 μmol/L four days later.

Femoral whole blood obtained at autopsy two days after death measured 64 μmol/L. The reported ratio of 49 used 64 divided by 1.3, approximately 49.2. It did not compare autopsy blood with a sample proven to have been drawn at the exact instant of death. The case was one of ten in the paper, and the only chloroquine case in that human series.

Why matrix is central to interpretation

The numerator was femoral whole blood; the denominator was serum. The authors discussed chloroquine’s marked distribution into blood cells and tissues. A whole-blood-to-serum difference can therefore contribute to the ratio independently of a change after death.

The authors considered extensive tissue distribution and postmortem redistribution important explanations for the striking result. Nevertheless, the design cannot assign a precise percentage of the difference to redistribution, cellular partitioning, interval since the earlier sample, or other case-specific processes. Calling this “a 49-fold redistribution correction” would go beyond the comparison performed.

The collection and analytical details

The study’s femoral sampling method used a groin incision and an elevated leg, with blood drained into a scoop and transferred to a tube containing potassium fluoride. Care was taken to avoid blood from the iliac veins above. Samples were stored at −20°C.

Chloroquine was measured by gas chromatography with nitrogen-phosphorus detection after extraction, using promazin as internal standard. The described blood method reported recovery above 90% at 1 and 10 μmol/L, a detection limit of 0.2 μmol/L, calibration from 0.2–30 μmol/L, R² ≥ 0.99 and within-run CV below 6% at the two tested concentrations. The reported 64 μmol/L case result exceeds that stated calibration range. The article does not describe the specific dilution or range-extension handling for this measurement, so that detail cannot be supplied by assumption.

What one case can—and cannot—establish

This is a descriptive clinical and autopsy case, not a controlled experiment or a population estimate. There is no group statistical test establishing a typical 49-fold change. The findings demonstrate why an impressive number can remain difficult to interpret, even when collected from a peripheral vessel.

In a current case, the necessary records include the earlier specimen’s collection time and matrix, treatment history, anatomical source of the later sample, preservation, calibration and handling of results above range. The two concentrations alone do not reconstruct a dose, determine impairment, or resolve cause of death.

The same need to identify the material being measured arises in hospital specimens used in death investigations. The particular chloroquine result should remain attached to its own case and method.

Primary source

Hilberg T, Rogde S, Mørland J. Postmortem Drug Redistribution—Human Cases Related to Results in Experimental Animals. Journal of Forensic Sciences. 1999;44(1):3–9. doi:10.1520/JFS14404J.