Can Postmortem Vitreous Alcohol Replace Blood BAC? What 31 Paired Samples Actually Show

Short answer: Vitreous humor is a useful separately measured specimen for investigating ethanol after death, especially when blood is missing or compromised. It is not a duplicate blood sample, however, and a correlation does not make its concentration interchangeable with a decedent’s blood alcohol concentration (BAC). In a 2020 study of 31 paired postmortem blood and vitreous specimens, Savini and colleagues reported a strong between-matrix relationship (R² = 0.9227). Vitreous alcohol exceeded blood alcohol in 23 cases, whereas blood exceeded vitreous in eight. One vitreous result was below the study’s limit of quantification despite measurable ethanol in the paired blood specimen. These findings support complementary testing—not a fixed vitreous-to-blood conversion or a conclusion about one person’s impairment when alive.

The original primary study is Savini and colleagues, “Ethanol Determination in Post-Mortem Samples: Correlation between Blood and Vitreous Humor Concentration,” Molecules (2020), DOI: 10.3390/molecules25122724. The numbers below concern that study’s particular specimens and analytical procedure, not an independently established legal standard or a prediction rule validated in another population.

What did the 31-pair study actually test?

Investigators at the Santo Spirito Hospital laboratory in Pescara, Italy, examined forensic cases during 2019 in which both blood and ocular vitreous humor were available for ethanol analysis. The authors reported 31 paired case specimens, with repeated analytical measurements used to describe the values in their original table. Blood was obtained before opening the body, principally from the femoral region; vitreous was aspirated from the eyes. This is a postmortem case series, not a controlled human-drinking trial and not a study of drivers stopped on a road.

The researchers used gas chromatography with flame-ionization detection (GC-FID) to assay ethanol in prepared samples of both matrices. They compared the numerical concentrations, summarized the direction of between-matrix differences, and fitted a linear regression. A useful interpretation must keep these three exercises separate: demonstrating an assay’s ability to quantify ethanol, observing how two sampled compartments compare, and inferring an unavailable blood concentration from a different measured compartment.

Neither specimen was a breath sample. Neither result was a mathematically estimated BAC produced by a Widmark equation. The observations were concentrations in actual collected postmortem materials, obtained with laboratory instrumentation after collection and preparation.

What is the difference between postmortem blood BAC and vitreous alcohol concentration?

A blood result reports the ethanol concentration in the submitted blood specimen. The study described its blood as postmortem whole blood, predominantly femoral, reported in grams per liter (g/L). A vitreous result reports the ethanol concentration in the gel-like fluid within the eye, also in g/L. Vitreous contains a large proportion of water and is anatomically less exposed to some postmortem changes affecting blood; those advantages explain why it is considered as an alternative or complementary specimen.

Same chemical does not mean same matrix. Ethanol can distribute between compartments, but water content, equilibration, collection site, conditions after death and the state of each specimen can affect comparisons. A 1.00 g/L vitreous measurement remains a 1.00 g/L vitreous measurement. It should not silently be reported as 1.00 g/L of whole-blood alcohol.

Blood and vitreous values may move in the same general direction across people while differing materially within an individual case. For blood collection and specimen identification issues, see the site’s published explanation of specimen collection in forensic toxicology.

How were the specimens collected, preserved and analyzed?

The authors say blood was collected before the body was opened, mainly from the femoral vein, to reduce possible contamination from neighboring fluids. Vitreous humor was collected by needle from both eyes. Samples were maintained at approximately 4 °C with sodium fluoride added at 100 mg per 10 mL of sample, and the reported analyses occurred within five days after collection. These conditions belong to the original study. They cannot be presumed for an unrelated autopsy without collection and storage records.

Gas chromatography separated the volatile components of the prepared headspace sample, and a flame-ionization detector produced analytical signals. The study used n-propanol as an added internal standard and an Agilent 7820A instrument equipped with an HP-Innowax column. The published chromatographic run lasted approximately six minutes. That run time is the instrumental portion of the test, not the entire time required for intake, labeling, standards, controls, preparation, review or chain-of-custody documentation.

Internal-standard methodology helps the laboratory compare analyte response with a known added compound. It does not independently establish where alcohol came from before collection. A method can correctly measure ethanol in a vial without answering whether a decedent drank the ethanol before death or whether circumstances contributed to its formation afterward.

How well did the GC-FID method perform analytically?

The study reported a calibrated analytical range of 0.01–10 g/L, a limit of detection (LOD) of 0.003 g/L, and a limit of quantification (LOQ) of 0.01 g/L. The reported calibration coefficient of determination was approximately 0.9981. Importantly, that high number describes how instrument response tracked analytical standards; it is not the correlation between blood and vitreous specimen concentrations.

The original method-validation table also reported within-day and between-day precision and trueness at several ethanol levels. The selected results below are summaries of the published validation data, with each metric tied to the authors’ assay rather than generalized to every laboratory.

These measurements provide evidence about one procedure’s behavior at specified quality-control concentrations. Analytical precision and bias answer different questions from uncertainty in interpreting an individual postmortem event. Nor does a strong calibration fit establish that the difference between vitreous and blood has negligible uncertainty.

What did the measured pairs show?

In the authors’ reported breakdown, 23 of 31 cases (about 74%) had higher vitreous than blood ethanol, while eight of 31 (about 26%) went the other direction. The directions are not a fixed conversion multiplier: a sample can be higher in one fluid and lower in another, and the size of the difference is case-dependent. In addition, the original case table includes one vitreous result recorded as below LOQ, not as a fully quantified concentration. That result must not be treated as an exact numerical zero.

The following selected paired observations illustrate why the actual individual numbers matter. These are four examples from the study’s original 31-case table, not a newly assembled or representative cohort. All concentrations are the authors’ GC-FID findings in g/L.

The examples refute an assumption that vitreous must always contain more ethanol than blood. They do not reveal why a particular difference occurred. The study did not assign each observation to a verified individual mechanism such as a known drinking interval, a particular microbial contaminant, or a proven equilibration phase.

What does the reported blood–vitreous regression mean?

The authors reported the relationship as vitreous ethanol (g/L) = 1.01 × blood ethanol (g/L) + 0.0479 g/L, with R² = 0.9227. Their model is expressed with vitreous on the left and blood on the right. Reversing it, or substituting a single vitreous reading to declare a precise blood value, is not an automatically validated procedure.

A coefficient of determination near 0.92 means that the fitted linear model accounts for a large share of the observed variation in the analyzed case series. It is not a 92% probability that a particular person was drinking, a 92% chance that an individual BAC estimate is correct, or an indication of 92% agreement of the two numerical measurements.

The calibration R² of approximately 0.9981 and the paired-specimen R² of 0.9227 describe different statistical relationships. A scientifically competent opinion must identify which relationship it is citing. Neither statistic replaces a matrix-specific measurement uncertainty budget or a person-level predictive interval.

Why does correlation not establish a safe individual conversion factor?

Correlation measures how values tend to covary over the study’s range. Agreement asks whether two methods or matrices yield sufficiently similar results for the decision at issue. A high correlation can coexist with a consistent offset, concentration-dependent differences, or important individual deviations. Those are separate analyses; they cannot be assumed away by reporting a large R².

For a forensic opinion, the material question is often not whether higher vitreous ethanol generally accompanies higher blood ethanol; it is how narrowly an unknown individual’s blood result could be predicted from the observed vitreous concentration. The published analysis does not supply a prospectively tested blood-from-vitreous prediction rule, its externally validated error rate, or an individual prediction interval suitable for a legal threshold decision.

The authors offer a regression description and conclude that vitreous is a useful alternative where blood is unavailable. That conclusion can reasonably justify testing a different specimen. It should not be expanded into the stronger, untested assertion that a fixed vitreous-to-blood factor is reliable for every decedent, every preservation condition, or every legal decision.

Does one below-LOQ vitreous result matter?

Yes. One case in the original table has blood ethanol reported at 0.35 g/L while vitreous ethanol is shown as <LOQ. “Below the limit of quantification” means the method did not report a reliably quantified vitreous value at that level under its stated validation criteria. It is not equivalent to saying the vitreous sample contained exactly zero ethanol.

This is also why a denominator should be specified carefully. There are 31 case pairs in the study, but only 30 pairs with fully numeric vitreous measurements in the published table. The authors’ broad 23-versus-eight directional summary uses the full 31 cases. Mathematical calculations needing two exact numbers cannot safely substitute zero for the censored observation simply to complete a spreadsheet.

The reported method LOQ was 0.01 g/L, but the original table does not explain why that specific paired vitreous result was below its quantification limit when the blood value was 0.35 g/L. An interpretation should preserve this uncertainty instead of inventing an explanation.

Could postmortem ethanol formation explain a discordant result?

It is one possibility, not a diagnosis based on a difference alone. After death, microbial activity and tissue changes may affect some biological matrices; blood may also vary with sampling site, decomposition and collection conditions. The investigators discussed vitreous as more sheltered than blood from microbial contamination and postmortem transformation. That is an important reason to retain and test it where practicable.

Yet a larger blood measurement than vitreous does not, by itself, prove microbial synthesis in that case. Likewise, a positive vitreous measurement does not, without corroboration, establish the timing, volume or circumstances of antemortem drinking. Preservation with fluoride can inhibit biochemical processes after collection; it does not retrospectively establish what happened before the specimen was secured.

Other case evidence can help evaluate competing explanations: the circumstances of death, recovery conditions, postmortem interval, degree of decomposition, documented collection site, microbiology where available, and results from other appropriately validated specimen analyses. These considerations are requirements for interpretation, not additional findings actually measured in the 31-case experiment.

Can vitreous ethanol establish whether someone was impaired before death?

Not on its own. The paper measured chemical concentrations in materials collected after death. It did not administer psychomotor tests before death, observe driving, measure reaction times, document a precise final drink, or match specimens to individualized impairment findings. A chemical detection can be relevant to alcohol exposure, but exposure is not the same as functional impairment or the cause of an event.

Even where ethanol was present before death, the time course between an earlier behavioral event and death would require information not provided by this paired-matrix validation. A postmortem concentration is not automatically the blood concentration at the moment of a collision, assault, fall or drowning. The paper supplies no person-specific elimination constant, absorption-state evidence, or uncertainty model that would make such reconstruction reliable in a particular case.

The distinction also matters if the available sample is only vitreous and a report later labels an inferred number “BAC.” The actual matrix and inferential method should stay explicit. Otherwise a modeled blood equivalent can be mistaken for an independently measured whole-blood concentration.

How should an expert evaluate a case that has blood and vitreous findings?

Start with source integrity. Identify precisely where blood was collected and whether the submitted material was whole blood, serum or another preparation. Identify which eye supplied vitreous, whether eyes were pooled, whether the sample was contaminated during collection, the container and preservative used, and documented times and temperatures from recovery to analysis. The study’s predominantly femoral collection and 4 °C storage procedure cannot be assumed where the case record is silent.

Next examine the actual laboratory package: chromatograms, method validation for each matrix, calibrator response, internal-standard behavior, blank and control data, replicate agreement, reportable range, LOD, LOQ, and reporting units. A method with a six-minute chromatographic run still depends on satisfactory pre-analytical handling and quality assurance. See the site’s published technical overview of headspace GC-FID blood alcohol testing for why the analytic result and its interpretation are distinct.

Finally examine the paired concentrations without first forcing them into a universal ratio. Determine whether both samples were quantifiable, how their uncertainties compare, and whether any alternative source of ethanol or redistribution is supported by independent evidence. A scientifically defensible report should distinguish a direct observation, an inference supported by the studied group, and a case-specific proposition that remains untested.

What can this study support in litigation—and what remains unproved?

The study supports using a properly validated vitreous assay as a valuable complementary analytical route where blood is absent, compromised or difficult to interpret. It also supports recognizing that blood and vitreous ethanol concentrations correlate strongly in the studied postmortem series while not being numerically identical in every case. Those are useful, limited propositions for forensic laboratory review and expert testimony.

It does not establish a universal vitreous-to-blood conversion factor, an individual confidence interval for a substituted BAC, a reliable threshold classification at every concentration, a proven ante-mortem drinking history, or the alcohol contribution to an identified death. For additional context on ethanol as one among several volatile analytes, see the site’s published alcohol and volatiles guide.

In testimony, precise phrasing matters: the evidence may establish “ethanol was measured in the vitreous specimen at the stated concentration.” It does not automatically establish “the decedent’s blood alcohol was that concentration at the relevant earlier time.” Scientific authority depends on preserving that distinction rather than using the word BAC for every ethanol-containing fluid.

Frequently asked questions about vitreous and blood alcohol after death

Can vitreous ethanol be measured directly?

Yes. The study used a validated GC-FID procedure to measure ethanol in submitted vitreous samples. That is a direct measurement of vitreous ethanol—not a direct measurement of ethanol in a separately unavailable blood specimen.

Were all 31 vitreous samples fully quantified?

No. The original table contains 31 paired case entries, but one vitreous result is marked below LOQ. This matters when calculating exact ratios, fitting models or interpreting how many independent paired numeric observations are available.

Does R² = 0.9227 mean vitreous and blood are interchangeable?

No. It describes the fitted between-matrix relationship in this sample. It does not quantify exact agreement or supply an externally validated person-level prediction interval, and it does not mean the two measured concentrations always match.

Does higher blood than vitreous prove postmortem fermentation?

No. It may prompt investigation of postmortem formation and other specimen-specific influences, but a direction of difference is not an experimentally confirmed explanation for a particular person’s ethanol finding.

Can the reported regression convert a vitreous result into legally established BAC?

No. The published equation models the studied relationship with vitreous as its dependent variable. It is not a separately validated individual blood-from-vitreous conversion rule, much less a universal legal reporting formula.

What do the 0.003 g/L detection limit and 0.01 g/L quantification limit mean?

The authors’ analytical procedure could identify ethanol at a lower stated level than it could reliably quantify. A result below LOQ must not be silently assigned an exact concentration or treated as an analytically established zero.

Can a postmortem vitreous ethanol result prove intoxication at the time of an incident?

Not by itself. The original study measured ethanol after death. It did not validate driving impairment, a person’s alcohol elimination history, an earlier event-time blood value or individual causation. Those propositions require separately supported evidence and uncertainty analysis.

Primary research: Savini F, Tartaglia A, Coccia L, and colleagues. Molecules. 2020;25(12):2724. Ethanol Determination in Post-Mortem Samples: Correlation between Blood and Vitreous Humor Concentration.