Category: Uncategorized

A computerized test’s “ADHD Confidence Index” can identify an ADHD-like performance profile without establishing a clinical diagnosis. A 2013 study of prenatal methamphetamine exposure illustrates that distinction, as well as the difference between an odds ratio and an individual probability.

What population and test were studied?

Kiblawi and colleagues assessed 301 children at approximately 5.5 years: 153 with prenatal methamphetamine exposure and 148 comparisons. These children came from the 412-pair IDEAL cohort, with exposure identified through maternal report and/or confirmed meconium analysis.

Masked examiners administered the Conners’ Kiddie Continuous Performance Test, or K-CPT. During this 7.5-minute computerized task, children responded to pictures except a ball. The test varied presentation intervals and measured errors, response times and changes in performance over the session.

What does the confidence index represent?

The index combines performance features to describe similarity to an ADHD-related test profile. The study treated a score greater than 50% as a positive index result. That threshold is not a finding that a child has a greater-than-50% independently verified probability of ADHD in this particular population.

The analysis did not establish ADHD diagnoses for the children through a separate diagnostic reference assessment. Consequently, index positivity should not be relabeled as confirmed ADHD prevalence or as the fraction of exposed children who inevitably develop the disorder.

How many children exceeded the threshold?

Forty-one of 153 exposed children, 26.8%, exceeded the threshold, compared with 22 of 148 comparisons, 14.9%. These are the observed group proportions for the index endpoint.

The adjusted odds ratio was 3.1, with a 95% confidence interval of 1.2–7.8 and p = 0.02. That estimate concerns adjusted odds of crossing the test threshold. It is not a 3.1-fold increase in every child’s probability, a 310% diagnosis rate or proof that prenatal exposure caused the result. The broad interval also reflects uncertainty about the magnitude.

Were all attention measures different?

No. After adjustment, the groups did not differ significantly in omission errors, commission errors, overall hit reaction time or perseverations. Significant findings involved change in reaction time across blocks and change in reaction-time variability with longer presentation intervals.

Thus similar overall speed could coexist with different performance trajectories as the task progressed or its timing changed. Describing the children as uniformly slower, more error-prone or impulsive would erase the study’s null findings and the specificity of its positive results.

What did adjustment address, and what remained?

Models incorporated other prenatal substances, child characteristics, study site and several caregiver and home measures. Forty-eight home-environment scores were missing; the investigators used multiple imputation and reported similar findings in sensitivity analyses with and without those imputed values.

Parental ADHD was not included as a covariate. The authors also cautioned that the K-CPT was less extensively established in this young, higher-risk population than some instruments used in older children. Enrollment exclusions and incomplete follow-up constrain generalization. Statistical adjustment cannot rule out all shared familial or environmental explanations.

Is this independent of the inhibitory-control study?

No. The Hearts and Flowers analysis examined an overlapping IDEAL cohort at a similar age using a different task and analytic sample. The studies ask different questions but should not be counted as independent population replications.

This study supports a historical association with selected attention-test features and index positivity. It does not validate diagnosing ADHD from prenatal exposure, from a toxicology result or from the confidence index alone, and it does not determine an individual child’s later development.

Source

Kiblawi ZN et al. The Effect of Prenatal Methamphetamine Exposure on Attention as Assessed by Continuous Performance Tests: Results from the Infant Development, Environment, and Lifestyle (IDEAL) Study. Journal of Developmental & Behavioral Pediatrics. 2013;34:31–37. doi:10.1097/DBP.0b013e318277a1c5. Historical prospective observational cohort assessment.

A child can make more errors on a difficult response-conflict task without showing a significant difference in average response speed. A prenatal methamphetamine study at approximately five and a half years of age illustrates why accuracy, speed and exposure categories should be reported separately.

Which children were assessed?

Derauf and colleagues studied the longitudinal IDEAL cohort, in which prenatal exposure was identified through maternal report and/or confirmed meconium testing. Of 412 originally enrolled mother–infant pairs, 303 children received the task. Twenty-three were excluded for excessive invalid trials and thirteen for missing covariates, leaving a reported valid-data sample of 267: 137 exposed and 130 comparison children.

The resulting sample represented approximately 65% of the original cohort. Attrition and exclusion matter even when selected measured characteristics appear similar, because unmeasured differences can remain.

What did the Hearts and Flowers task require?

In the congruent condition, children pressed a key on the same side as a heart. In the incongruent condition, they pressed on the side opposite a flower. The mixed condition required applying both rules as the stimuli changed. Examiners were masked to exposure status.

Accuracy was the percentage of correct responses. Reaction time was calculated only for correct responses. These outcomes are related but not interchangeable: a response-time average omitting incorrect responses does not summarize every trial or every aspect of inhibitory control.

What counted as heavier exposure?

“Heavy” exposure meant reported use on at least three days per week across pregnancy; “some” meant use below that criterion. These are frequency categories, not measured fetal doses in milligrams. Neither meconium detection nor maternal frequency reports establish an exact absorbed dose.

Table 3 lists 25 children in the heavy category, 105 in the some-use category and 130 comparisons. Its notes identify four cases without exposure-frequency classification and three without other prenatal-exposure information. The table’s 260-person total should not be presented as identical to the broader 267-person valid-data sample.

Which results were significant?

After adjustment, heavy exposure was associated with approximately 22.69 percentage points lower accuracy in the incongruent condition, with a 95% confidence interval of −35.9 to −9.4 points. In the mixed condition, the estimate was −8.79 points, with an interval of −17.3 to −0.3.

The corresponding some-use-versus-no-use accuracy intervals included zero. The heavy-exposure congruent-condition accuracy interval also included zero, as did the reported exposure-group reaction-time comparisons. These null findings limit any claim of a universal deficit across every task condition or exposure category.

Does this establish a precise dose–response relationship?

The authors interpreted the pattern as consistent with a dose-related association. More narrowly, the analysis compared reported frequency categories, with a small heavy-exposure subgroup. It did not estimate a continuous milligram-by-milligram fetal dose–response curve or identify a safe lower exposure threshold.

Caregiver psychological symptoms, caregiver vocabulary-based cognitive measures and recorded maltreatment also related to selected outcomes. The observational design could not fully separate prenatal exposure from correlated postnatal circumstances, genetic factors or other substance exposures. Adjustment improves the comparison but does not recreate random assignment.

How does this relate to other childhood findings?

IDEAL generated several publications using different tests and overlapping children. This task study is not independent replication of each other IDEAL result, nor does a task difference diagnose ADHD or predict a particular child’s future behavior.

The study supports a specific association involving difficult response rules and heavier reported prenatal exposure. It does not establish uniform cognitive injury, prove that methamphetamine caused an individual child’s difficulty or convert a prenatal toxicology result into a measure of later functional capacity.

Source

Derauf C et al. Prenatal Methamphetamine Exposure and Inhibitory Control among Young School-Age Children. Journal of Pediatrics. 2012;161:452–459. doi:10.1016/j.jpeds.2012.02.002. Historical prospective observational cohort assessment.

Calling R-(−)-methamphetamine “inactive” removes qualifications that matter. A 2023 systematic review found limited human data, with effects differing by dose, administration route and the outcome measured. It supports a more specific account than either universal inactivity or equivalence to S-(+)-methamphetamine.

How much human evidence was available?

Barkholtz, Hadzima and Miles searched the literature in October 2022 and included thirteen publications representing ten distinct studies and 99 unique participants. Several publications reused the same participant pools. Counting papers as independent experiments would overstate the evidence base.

The review excluded animal studies and studies of R-methamphetamine formed only as a metabolite of another drug. It synthesized human administration evidence qualitatively because the doses and protocols were too heterogeneous for a meaningful pooled estimate.

What did lower oral doses show?

The review describes an eight-person crossover investigation in which oral R-methamphetamine doses up to 10 mg did not produce significant physiological or subjective differences from placebo on the reported measures. Some sessions also involved a separately identified deuterated intravenous dose for pharmacokinetic analysis.

That result concerns those doses, participants and measures. It does not demonstrate absence of every biological action, exclude uncommon adverse effects or establish safety for unrestricted use. Pharmacokinetic measurements can confirm drug exposure even when a small experiment detects no significant functional effect.

What happened with higher intravenous dosing?

A twelve-person crossover study summarized in the review compared R- and S-methamphetamine at 0.25 and 0.5 mg/kg intravenously, along with placebo and a racemic condition. The lower R dose did not produce significant physiological or subjective effects compared with placebo.

At 0.5 mg/kg, both enantiomers produced similar peak subjective effects, but the R-associated effects dissipated more rapidly, while S-associated cardiovascular and subjective effects lasted longer. Similar peaks do not establish identical time courses or interchangeable overall pharmacology.

These intravenous doses cannot be equated with inhaler administration or a fixed oral dose simply because the same molecule is involved. Route, delivered amount and absorption profile remain part of the result. The racemic condition also should not be treated as a simple arithmetic average of separate-isomer observations.

Were most studies measuring impairment?

No. Much of the literature focused on excretion and drug testing. The review reports urine measurements in nine studies involving 87 participants, with urine as the only matrix in eight studies involving 51 participants. It identified no study quantifying R-methamphetamine in whole blood within the included evidence.

These matrix categories overlap across studies and must not be added as independent participant groups. Urine concentration is not a measurement of subjective effect or task performance, and plasma results do not automatically provide whole-blood concentrations.

Why does the small evidence base matter?

Pharmacodynamic outcomes were assessed in only four studies involving 44 participants. Demographic reporting was incomplete: sex was reported for 66 of the 99 participants and age for 22. The review therefore cannot establish equally precise conclusions across ages, sexes, health conditions or patterns of use.

Historical intranasal studies also differed in formulation, inhalation instructions and dose estimation. A finding from a named historical product should not be assumed to describe every current product with similar branding. This review is a record of evidence through its search date, not a current product directory.

What is the defensible interpretation?

R-methamphetamine generally showed less or shorter activity in the reviewed comparisons, with important dose- and route-specific exceptions. The evidence does not justify a universal “no effects” label or a universal potency conversion between isomers.

For an individual toxicology interpretation, identify the enantiomer, specimen and exposure evidence separately from the proposed effect. Neither detection of R-methamphetamine nor the absence of a significant effect in one small study determines a person’s impairment at another time.

Source

Barkholtz HM, Hadzima R, Miles A. Pharmacology of R-(−)-Methamphetamine in Humans: A Systematic Review of the Literature. ACS Pharmacology & Translational Science. 2023;6:914–924. doi:10.1021/acsptsci.3c00019. Historical systematic review; individual experiments are summarized as review evidence.

Refrigeration alone did not prevent ethanol formation in the postmortem tissue homogenates studied by Lewis and colleagues. All nine specimens formed additional ethanol without storage preservative, including at 4°C. Homogenizing matched tissue portions with sodium fluoride solution largely prevented that increase under the tested conditions.

What tissue and conditions were tested?

The 2004 study examined one tissue specimen from each of nine separate aviation fatalities: four muscle specimens and five kidney specimens. Cases were deliberately selected because earlier testing suggested microbial contamination, through unusual ethanol distribution or other volatile compounds. They were not a random sample of all postmortem tissues.

Specimens came from material archived at −20°C, which the laboratory retained for up to five years. After thawing, each tissue was divided and weighed. One portion was homogenized with two parts water; the other was homogenized with two parts of a 1.00% weight/volume sodium fluoride solution. Each preparation was then split between storage at 4°C and 25°C.

The stated 1.00% describes the fluoride solution added during preparation. It should not be silently relabeled as a measured 1.00% final concentration in the complete tissue homogenate. No organisms or extra glucose were deliberately added, although homogenization and repeated sampling distinguish this experiment from an intact body.

What changed during storage?

Initial tissue ethanol concentrations ranged from 1 to 28 mg/hg, averaging 11 mg/hg. Here, mg/hg means milligrams per 100 grams of tissue. These weight-based tissue results should not be relabeled as blood concentrations in mg/dL.

Without preservative, the nine refrigerated specimens increased by 22–75 mg/hg after 96 hours, averaging 42 ± 15 mg/hg. At 25°C, the increases were 19–84 mg/hg after 48 hours, averaging 45 ± 22 mg/hg. The endpoints were different: these averages are not a same-duration temperature comparison.

For example, kidney specimen 8 started at 1 mg/hg and reached 76 mg/hg after 96 hours at 4°C. Its room-temperature counterpart reached 62 mg/hg after 48 hours. Refrigeration slowed formation but did not eliminate it in these selected materials.

The fluoride-prepared portions showed no significant increase as reported by the authors at either temperature. This does not mean every numerical result was unchanged: kidney specimen 7 went from 28 to 33 mg/hg at 4°C, which the authors considered likely experimental or instrumental variation.

How were the concentrations measured?

Ethanol was quantified by headspace gas chromatography with flame-ionization detection, using two analytical columns and a tert-butanol internal standard. Identification used retention times on both columns. The reported analytical detection and quantification limits were 0.10 and 1.00 mg/dL, respectively, for the liquid measurement method; the tissue findings were calculated and reported in mg/hg.

The analytical internal-standard solution also contained fluoride for salting out during measurement. That preparation step should not be confused with adding preservative before the experimental storage period. A specimen can be measured accurately after ethanol has already formed in storage.

What the experiment supports—and what it cannot determine

The study supports documenting tissue preparation, preservative addition and the entire temperature-and-time history. It demonstrates a mechanism for a stored tissue result to exceed its earlier value without additional drinking.

It does not estimate how often this happens across all fatalities, identify the responsible microbial species, or provide an hourly formation rate for an intact body. The tissues were selected for suspected contamination, previously frozen and homogenized. The authors also found no discernible pattern among the other measured volatiles that could resolve the ethanol question in every specimen.

These results cannot partition a particular positive result into consumed and postmortem-formed ethanol, reconstruct a blood alcohol concentration, or establish impairment or cause of death. Preservation findings also depend on the matrix and circumstances; compare the separate serial-testing case involving preserved postmortem blood.

Primary source

Lewis RJ, Johnson RD, Angier MK, Vu NT. Ethanol formation in unadulterated postmortem tissues. Forensic Science International. 2004;146(1):17–24. doi:10.1016/j.forsciint.2004.03.015.

A low average liver-to-blood ratio can support a hypothesis that hydrocodone has relatively little potential for postmortem redistribution. It cannot guarantee that every case follows that pattern or reconstruct a person’s concentration before death. A 39-case study illustrates the distinction: one high ratio persisted on repeat testing but was excluded from the reported summary ratio.

Which cases and specimens were studied?

Saitman, Fitzgerald and McIntyre examined 39 cases with hydrocodone detected in peripheral blood and sufficient central blood and liver for comparison. Cases came from the San Diego Medical Examiner’s Office between October 2013 and June 2014. Autopsy occurred within 48 hours of recorded death; the exact time of death was not established in every case.

Cases with obvious decomposition or severe trauma were excluded, as were cases with a known long delay between actual and recorded death. These restrictions matter when applying the results to a decomposed body or a substantially different postmortem interval.

Peripheral blood was collected from the common iliac veins, rather than necessarily from a femoral site. Central blood came from the heart or adjacent great vessels. Approximately 20 mL of peripheral blood was collected into standard 10 mL glass tubes, each containing 100 mg sodium fluoride and 20 mg potassium oxalate. Central blood was placed in identical tubes. Unpreserved liver came from the right upper lobe. Specimens were refrigerated at 4°C until analysis, and each case’s three matrices were tested in the same analytical run.

How was hydrocodone measured?

The study used solid-phase extraction, derivatization and GC–MS, with identification based on retention time and ion ratios. The blood calibration range was 0.02–1.0 mg/L, the detection limit 0.01 mg/L and the quantification limit 0.02 mg/L. Liver specimens were diluted to fall within the calibration range.

Analytical performance was not identical at every concentration. A nominal 0.10 mg/L blood control measured 0.087 ± 0.003 mg/L in five study runs. A 0.40 mg/kg liver control measured 0.41 ± 0.02 mg/kg in four runs. These are method-performance observations, separate from the interpretive question of redistribution.

The general pattern and its exception

Central and peripheral blood concentrations were strongly correlated, with R² = 0.965. The central/peripheral ratio averaged 1.3 ± 0.35, with a median of 1.2. Central concentrations were nevertheless statistically higher than peripheral concentrations in the paired comparison, p = 0.0021. Strong correlation therefore did not mean equality between collection sites.

After exclusion of one outlier, liver and peripheral blood correlated with R² = 0.915, and the liver/peripheral ratio averaged 3.4 ± 1.7 L/kg, with a median of 3.0 L/kg. The units matter: liver was measured in mg/kg and blood in mg/L, so their ratio has units of L/kg.

The excluded case had liver hydrocodone of 2.0 mg/kg and peripheral blood of 0.10 mg/L, a ratio of 20 L/kg. Repeat analysis produced a ratio of approximately 19 L/kg. The authors excluded it using Dixon’s Q test. Repetition did not establish why the case differed, but it means the high ratio should not simply disappear from a discussion of the study’s practical limits.

Why this is not an antemortem correction formula

The authors interpreted the generally low liver/peripheral ratios as evidence of minimal redistribution potential. However, all three specimens were postmortem. There were no paired antemortem measurements to show how much any individual’s concentration changed between life and autopsy.

A ratio below a proposed value is therefore not proof that an autopsy blood concentration equals the concentration during life. Nor can the repeated high ratio, by itself, quantify a redistribution effect. Specimen distribution, analytical variation and case circumstances need evaluation without assuming a single mechanism.

For case review, retain the exact blood collection site, both underlying concentrations, ratio units and any excluded observations. Drug detection and concentration alone do not establish dose, impairment or cause of death. See also why co-detected drugs can have different postmortem distributions.

Primary source

Saitman A, Fitzgerald RL, McIntyre IM. Evaluation and comparison of postmortem hydrocodone concentrations in peripheral blood, central blood and liver specimens: A minimal potential for redistribution. Forensic Science International. 2015;247:36–40. doi:10.1016/j.forsciint.2014.11.031.

Yes. In Elliott’s 2004 investigation, changing the internal standard used in GC–MS analysis substantially changed some reported postmortem GHB concentrations. This was an analytical comparison on the specimens, not evidence that GHB had changed in the body between measurements. An exposure threshold cannot be interpreted responsibly without knowing how the laboratory measured GHB.

What did the method comparison actually include?

The preliminary comparison included nine blood and ten urine specimens from cases 1–5, including preserved and unpreserved variants. Those 19 specimens were not 19 independent people. Samples were examined using diethylene glycol (DEG) or deuterated GHB (GHB-d6) as internal standards in a derivatization GC–MS procedure. A separate GC–FID procedure was also compared.

An internal standard helps relate the analyte signal to a reference signal through preparation and measurement. It is not simply a label attached after analysis. A compound that behaves differently from the analyte, or encounters an interference, can affect the calculated concentration.

In Table II, urine specimen 5 gave GHB results of 80 and 82 mg/L with DEG, compared with 5 mg/L using GHB-d6. The GC–FID result was 9 mg/L. Such a difference could materially alter an interpretation based on a numerical threshold.

However, DEG did not always produce a higher result: blood specimen 7 gave 5 and 6 mg/L with DEG and 20 mg/L with GHB-d6. The findings do not support a universal multiplier for correcting every DEG-based measurement.

Why did the results differ?

The paper considered possible interference involving derivatized urea near the DEG retention time, as well as differences in extraction behavior. These were possible explanations, not proof that urea caused each discrepancy. The comparison supports investigating matrix-specific performance rather than assuming all historical results from one procedure share the same error.

The GC–MS method measured derivatized GHB without first converting it to gamma-butyrolactone (GBL). The GC–FID procedure converted GHB to GBL and measured total GBL, potentially including GBL already present. Its result therefore did not necessarily represent exactly the same chemical quantity. Agreement or disagreement between these methods needs that distinction.

The reported detection limits were 1 mg/L for GC–MS and 2.5 mg/L for GC–FID. A detection limit is not an exposure threshold, and the paper did not establish that a detected low concentration proves ingestion. Routine control material also did not reproduce every discrepancy seen in authentic postmortem specimens.

What did the accompanying 40-fatality series show?

Elliott also reported a three-month series of 40 fatalities in which GHB was not implicated in the cause of death. Blood was usually femoral, with carotid or iliac specimens in some cases. Samples were stored at 4°C and analyzed within ten days of receipt. The paper’s postmortem interval ran from certification of death to analysis, rather than simply from death to autopsy.

Using the GHB-d6 GC–MS method, unpreserved blood averaged 12.3 mg/L, range 2–29, in 38 specimens; fluoride-preserved blood averaged 12.6 mg/L, range 4–25, in 17. Unpreserved urine averaged 5.5 mg/L, range 0–18, in 39 specimens; preserved urine averaged 4.8 mg/L, range 0–10, in 15. These specimen counts overlap within the 40 cases and are not separate cohorts.

The proposed interpretive values of 30 mg/L in blood and 20 mg/L in urine were tied to this specific method and case material. The paper did not establish that concentrations above those values automatically prove intake.

What to request when reviewing a result

The useful records include the internal standard, chemical conversion steps, calibration, matrix validation, chromatograms and treatment of interference. Numerical quantitation does not itself establish antemortem exposure, impairment or cause of death. The separate problem of an unusable quantitative result is discussed in unquantifiable postmortem THC results.

Primary source

Elliott S. Further Evidence for the Presence of GHB in Postmortem Biological Fluid: Implications for the Interpretation of Findings. Journal of Analytical Toxicology. 2004;28(1):20–26. doi:10.1093/jat/28.1.20.

A cardiac blood GHB concentration above 50 mg/L does not, by itself, establish ingestion before death. Kintz and colleagues found 14 such results in a 71-autopsy series selected because the case investigations did not support GHB exposure. The study supports examining additional specimens, while leaving important limits on how confidently exposure can be excluded.

How the 71 cases were investigated

The 2004 study examined cardiac whole blood from 71 autopsies performed 12–72 hours after death. Pathologists assessed the circumstances and police records and considered GHB exposure excluded. This was a case-history classification, not a controlled experiment with independently verified lifetime nonexposure.

Specimens had no preservative, were refrigerated at 4°C and were analyzed within 48 hours after autopsy. Consequently, the findings reflect both the interval before collection and the subsequent handling conditions; the study did not experimentally separate formation in the body from formation in the specimen container.

GHB was measured using derivatization and GC–MS with deuterated GHB as the internal standard. The reported linear range was 1–200 mg/L and the quantification limit was 1 mg/L. The article reports some results below that quantification limit and others above the stated linear range, without detailing the handling of every out-of-range result. Its full reported cardiac range, 0.4–409 mg/L, should be read with those analytical limitations in view.

What happened above 50 mg/L?

Fourteen cardiac specimens exceeded 50 mg/L, with reported values of 50.8–409 mg/L. The authors interpreted these elevations as compatible with postmortem formation because the investigations supplied no evidence supporting intake. That interpretation is evidence against treating 50 mg/L as an automatic exposure rule; it does not prove the biochemical origin of every molecule in every case.

Other matrices were tested when available in these high-cardiac-result cases. They were not systematically tested across all 71 autopsies. Femoral blood was available in five, vitreous humor in six and bile in seven. These small, selected comparisons cannot establish population-wide diagnostic sensitivity or specificity.

Femoral blood, vitreous and bile gave different information

In one paired case, cardiac blood measured 96.6 mg/L while femoral blood measured 17.1 mg/L. In another, cardiac blood was 50.8 mg/L, femoral blood 39.7 mg/L and vitreous humor 10.8 mg/L. These are concentrations measured in separate specimens, not interchangeable estimates of one circulating concentration.

The six vitreous results ranged from 3.9 to 21.4 mg/L. By comparison, bile ranged from 6.1 to 238 mg/L in its seven paired cases. Because bile could also contain substantial GHB, the authors did not consider it a useful discriminator in this series. Femoral blood and particularly vitreous were more informative complements to cardiac blood under the studied conditions.

Obtaining another matrix provides biological corroboration. It is not the same step as confirming a presumptive immunoassay screen: these findings came from a quantitative GC–MS procedure. Even an analytically identified and quantified substance may have an uncertain source.

What a defensible interpretation should address

The collection site should appear beside the concentration. Reviewers should also examine the death-to-autopsy interval, refrigeration, preservative, analysis timing, availability of paired specimens and independent evidence of intake. An elevated cardiac result can justify further investigation without establishing ingestion, impairment or a contribution to death.

The study’s limited paired sample sizes and history-based exposure classification prevent a universal rule that a low femoral or vitreous result excludes intake. Conversely, its findings undermine the claim that a cardiac result above 50 mg/L necessarily proves it.

For the separate issue of change after collection, see GHB increases in stored postmortem blood.

Primary source

Kintz P, Villain M, Cirimele V, Ludes B. GHB in postmortem toxicology: Discrimination between endogenous production from exposure using multiple specimens. Forensic Science International. 2004;143:177–181. doi:10.1016/j.forsciint.2004.02.036.

Brain tissue can help assess whether postmortem gamma-hydroxybutyrate (GHB) findings are compatible with intake before death, but a proposed brain threshold of 10 mg/kg is not a universal test for exposure. In a 221-autopsy study, its usefulness depended on the degree of decomposition, and only three cases had reported GHB intake.

What specimens did the study compare?

Thomsen and colleagues collected femoral venous blood and frontal-lobe cortical gray matter in 221 consecutive autopsies during an eight-month period in 2015–2016. GHB intake was not suspected in 218 cases; intake was reported in three. The absence of a reported history did not independently establish that those 218 people had never taken GHB.

Bodies were held at 5°C after arrival, with autopsy one to four days after arrival. That interval is not necessarily the time since death. Blood was collected at the start of autopsy and stored at −20°C with 1% sodium fluoride and 0.2% potassium oxalate by weight. Brain tissue was also stored at −20°C.

The investigators quantified GHB by UHPLC–MS/MS after protein precipitation. The quantification range was 2–200 mg/kg in both matrices, with dilution above the upper limit. Blood imprecision was below 14.8% CV and accuracy was 94–104%. At a brain quality-control concentration of 200 mg/kg, the reported CV was 7.3% and accuracy was 89.4%. Those high-concentration brain control results do not, by themselves, establish classification accuracy around a 10 mg/kg interpretive threshold.

Decomposition changed the comparison

Among 189 cases without suspected intake and with no or minor decomposition, femoral blood GHB ranged from 4.8 to 45.4 mg/kg, with a median of 15.3 mg/kg. Brain values ranged from not detected to 9.8 mg/kg, with a median of 4.8 mg/kg. All concentrations here, including blood concentrations, are reported in mg/kg.

In 29 cases without suspected intake but with advanced decomposition, blood ranged from 11.7 to 162 mg/kg, with a median of 21.6 mg/kg. Brain ranged from 2.2 to 49.6 mg/kg, with a median of 6.7 mg/kg. The decomposition-group differences were statistically significant in blood and brain, with p values of 9.4 × 10⁻⁶ and 4.4 × 10⁻⁶, respectively.

Thus, advanced decomposition produced brain results well above the proposed 10 mg/kg threshold in cases without a reported intake history. Brain-to-blood ratios also overlapped between decomposition groups; their medians were approximately 0.30 and 0.29. A ratio did not solve the origin question.

What did the three reported-intake cases show?

The reported-intake cases had blood/brain concentrations of 199/166, 142/78.4 and 40.3/12.7 mg/kg. The third case is especially instructive: brain exceeded 10 mg/kg although blood was within the range observed among the 189 cases without suspected intake and with limited decomposition. The detailed account attributed that death to morphine poisoning, with possible contributions from other substances including GHB.

These cases support investigating brain as an additional matrix. Three reported-intake cases cannot establish robust diagnostic sensitivity, specificity or a fatal-concentration threshold. The authors accordingly described 10 mg/kg as a tentative brain cutoff for limited decomposition.

How to apply the finding in a case review

A report should identify the sampled brain region, decomposition category, blood collection site, preservation, analytical range and the evidence supporting an intake history. Quantitative detection establishes the measured concentration in the submitted specimen; it does not establish dose, impairment or cause of death. A classification threshold is also different from an analytical detection or quantification limit.

For a separate question about changes after collection, see GHB formation in stored postmortem blood. Storage-related formation and decomposition at autopsy require their own assessment.

Primary source

Thomsen R, Rasmussen BS, Johansen SS, Linnet K. Postmortem concentrations of gamma-hydroxybutyrate (GHB) in peripheral blood and brain tissue—Differentiating between postmortem formation and antemortem intake. Forensic Science International. 2017;272:154–158. doi:10.1016/j.forsciint.2016.12.038.

A psychosis percentage from compulsory rehabilitation is not automatically the probability that anyone using methamphetamine will develop psychosis. A 2018 Shanghai study also shows why large headline samples do not eliminate the need to inspect subgroup counts and table labels.

Who supplied the 37.1% estimate?

Gan and colleagues recruited 1,430 people meeting DSM-IV methamphetamine abuse or dependence criteria from three compulsory rehabilitation centers between October 2012 and June 2014. The methods identify 530 with methamphetamine-induced psychosis and 900 without it. Dividing 530 by 1,430 produces the reported 37.1%.

This was a selected rehabilitation population, with medical exclusions, not a probability sample of all people using methamphetamine. It was also not a prospective study following initially symptom-free participants to measure incident psychosis. The estimate describes the classification within the recruited sample.

How were symptoms and exposure established?

Participants were interviewed by trained psychiatrists, with diagnostic verification by a senior psychiatrist. Drug-use history and psychosis symptoms relied on retrospective reports. The study also collected craving, depression and anxiety ratings and a much smaller cognitive-testing subset.

The investigators acknowledged recall error, other substance exposure and the lack of precise withdrawal-duration data. Participants were interviewed within a month of admission, which does not establish identical time since last use or an identical clinical state for everyone.

Why do the table headings need checking?

The methods and main results consistently describe 530 psychosis cases and 900 noncases. Tables 1 and 2 instead label the 900-person column as the psychosis group and the 530-person column as the nonpsychosis group. The 403 male participants described in the text as psychosis cases appear in the table’s 530-person column.

That internal conflict means a reader should not infer the direction of a comparison from the table heading alone. It is appropriate to flag the inconsistency and seek clarification before using the tabulated group contrasts for a consequential individual opinion.

Were duration categories available for all 530 cases?

The article reports 420 transient episodes lasting less than one month and 97 persistent episodes lasting more than one month. Those counts total 517, not 530. Their reported percentages, 81.2% and 18.8%, fit the smaller total approximately. The thirteen-case gap is not explained by those two counts.

It would therefore be inaccurate to say that the paper fully assigned every psychosis case to one of the two duration groups. The categories also leave the handling of exactly one month unclear. Neither the missing classifications nor the retrospective design supports a precise persistence probability for another population.

Did all 1,430 participants show cognitive deficits?

No such conclusion follows. The text says 56 completed cognitive testing, while the reported comparison contains 24 psychosis cases and 30 noncases, totaling 54. The discrepancy matters, and this small subset cannot be treated as the entire cohort.

The study reported differences in verbal and visual learning tasks, with p = 0.011 and p = 0.023, but six other listed task comparisons were nonsignificant. A broad statement that every measured cognitive domain was impaired would erase those null findings. Selection into testing and multiple comparisons also constrain interpretation.

Can the correlates establish causes or individual predictions?

No. The authors explicitly acknowledged that the cross-sectional design could not prove causation. Anxiety, craving, dependence and reported use patterns can be associated with psychosis without establishing temporal direction or a validated prediction rule.

The article provides historical evidence from a clinically important selected population. Its useful interpretation preserves the recruitment setting, retrospective ascertainment, actual denominators and reporting conflicts. It cannot identify a person’s mental state from drug presence, establish that methamphetamine alone caused a particular episode or assign the headline percentage to every exposed individual.

Source

Gan H et al. A Research of Methamphetamine Induced Psychosis in 1,430 Individuals With Methamphetamine Use Disorder: Clinical Features and Possible Risk Factors. Frontiers in Psychiatry. 2018;9:551. doi:10.3389/fpsyt.2018.00551. Historical retrospective, cross-sectional human evidence.

Two oral-fluid collection devices can give similar average findings without being proven interchangeable for every specimen. A controlled l-methamphetamine study illustrates how collection order, specimen volume and assay target complicate that comparison.

What was administered, and to whom?

Newmeyer and colleagues studied sixteen healthy adults, twelve men and four women, under controlled conditions. Participants used a historical Vicks VapoInhaler formulation containing l-methamphetamine: two inhalations per nostril every two hours for six doses on the first day, followed by a seventh dose the next morning.

The paper describes an estimated 0.16–0.60 mg delivered per four-inhalation dose, rather than a precisely measured absorbed dose for each participant. Specimens were collected before dosing and through 32 hours after the first dose, approximately eleven hours after the final dose. These historical formulation and administration details should not be generalized to every product currently sold under a similar name.

How were the collectors compared?

Oral fluid was collected with Quantisal and Oral-Eze, with different pad positions and processing procedures. Quantisal was always collected first. Each collection stopped when the volume indicator changed or five minutes elapsed. Low-volume specimens were analyzed without weight-based volume correction.

The study collected 432 oral-fluid specimens for each device. After dosing, there were 88 low-volume Quantisal collections and 130 low-volume Oral-Eze collections. Because order was fixed, the greater number with the second device cannot cleanly be attributed to that device alone. Prior collection may affect the fluid available for subsequent collection.

What were the positive and null results?

All sixteen participants had measurable oral-fluid l-methamphetamine at the method’s 1 µg/L quantification limit. Median maximum concentrations were 14.8 µg/L with Quantisal and 16.1 µg/L with Oral-Eze, with ranges of 5.2–380 and 4–182 µg/L, respectively. The paired maximum-concentration comparison was nonsignificant, p = 0.535.

First detection and last detection comparisons were also nonsignificant. The median last positive specimen occurred eight hours after the seventh dose for both devices. Seven participants remained positive with both at the final collection. Nonsignificant differences on these endpoints do not constitute a prespecified equivalence demonstration or establish exact agreement for an individual specimen.

Why did the on-site screen remain negative?

The DrugTest 5000 screening assay targeted d-methamphetamine, whereas the administered drug was the l-enantiomer. All screening results were negative; chiral laboratory testing detected no d-methamphetamine or d-amphetamine in oral fluid.

A negative d-targeted screen alongside a positive laboratory l-methamphetamine measurement is not automatically a false negative for the screen’s intended analyte. The target, cross-reactivity, cutoff and confirmatory identification must be examined together. The study does not establish performance for all d-methamphetamine concentrations or every real-world drug mixture.

Could oral fluid be converted to plasma concentration?

Only two participants had detectable plasma l-methamphetamine, with observed maxima of 10.0 and 6.3 µg/L. Their oral-fluid/plasma ratios varied substantially both within and between individuals. No plasma d-methamphetamine or either amphetamine enantiomer was detected.

Those observations do not supply a dependable oral-fluid-to-plasma conversion factor. Plasma must also remain distinct from whole blood. Applying a proposed analytical cutoff is a separate question from inferring a person’s concentration in another specimen or functional state.

What does the study support?

It supports broadly similar selected collector outcomes under one small controlled protocol, while documenting substantial individual variation and collection limitations. The study was not an impairment experiment, did not randomize collector order and followed only seven doses. A defensible comparison preserves specimen volume, sequence, processing, isomer and observation window instead of treating all oral-fluid results as interchangeable.

Source

Newmeyer MN et al. Oral Fluid with Three Modes of Collection and Plasma Methamphetamine and Amphetamine Enantiomer Concentrations After Controlled Intranasal l-Methamphetamine Administration. Drug Testing and Analysis. 2015;7:877–883. doi:10.1002/dta.1784. Historical controlled human administration study.