EVIDENCE LIBRARY
Does Serum Alcohol Equal the Concentration in the Brain? What a Three-Volunteer MRS Study Shows
Brain spectroscopy and serum gas chromatography gave different ethanol results in three volunteers. Timing, tissue composition and signal correction prevent treating the findings as a universal brain-to-blood conversion.
Serum ethanol and a spectroscopy-derived brain ethanol value are not interchangeable measurements. A three-volunteer study generally found lower brain values than serum values later in the experiment. Its methods and sample size do not support a universal conversion or an individual impairment threshold.
Two different analytical measurements
Thierauf-Emberger and colleagues studied three healthy men aged 24, 26 and 29. Each drank a calculated amount of vodka over thirty minutes. The calculation targeted a serum concentration, but actual concentrations varied substantially.
Venous blood was collected intermittently, and separated serum was analyzed by headspace gas chromatography with flame-ionization detection. These were serum ethanol results, not whole-blood results.
Brain ethanol was estimated using proton magnetic resonance spectroscopy in the occipital cortex, cerebellum, frontal cortex and putamen. The method fitted ethanol’s spectral signal and used water as an internal reference. It was not a chemical assay of removed brain tissue.
Achieved concentrations did not follow a single shared curve
At the end of drinking, serum concentrations were 0.26, 1.19 and 0.54 grams per liter in the three volunteers. One person therefore had more than four times another’s concentration at that sampling point despite the common calculation target.
Later serum values converged. Approximately 160 minutes after the end of drinking, the three values were 0.84, 0.84 and 0.85 grams per liter. These observations illustrate why a calculated target is different from an achieved measurement.
The highest reported water-normalized brain result was 0.68 grams per liter in the frontal measurement. Later serum samples generally exceeded brain values, and the brain curves appeared flatter. These descriptive results came from only three people and relatively sparse measurements.
The regions were scanned sequentially
The four brain regions were not measured simultaneously. Each spectroscopy acquisition took approximately seven minutes, with additional reference measurements and intermittent blood sampling.
Consequently, an apparent difference between regions can also include a time difference during absorption or elimination. Early comparisons were especially vulnerable to this timing mismatch. A serum sample and a nearby brain scan should not be treated as a perfectly simultaneous pair.
Water normalization did not remove every signal limitation
The researchers estimated each measurement volume’s gray-matter, white-matter and cerebrospinal-fluid fractions. Brain ethanol values were then normalized to its estimated water content. They therefore represent a different denominator from uncorrected ethanol per volume of mixed brain tissue.
No relaxation correction was performed because reliable values were unavailable for the method. Cerebrospinal fluid can contribute differently to the detectable ethanol signal, and the frontal measurement volumes contained relatively more of it. The authors identified this as a possible explanation for the higher frontal results.
The reported fitting-precision estimates also did not include all systematic errors. They should not be treated as complete uncertainty bounds for a person’s true brain concentration.
What follows for a DUI assessment?
The study confirms the importance of naming the specimen and method before comparing numbers. It does not show that a lower spectroscopy value means a person was unimpaired, or permit converting a later serum result into a precise brain concentration at driving time.
No driving, behavioral impairment or collision outcome was measured. The all-male three-person sample, sequential observations and method-dependent signal visibility leave substantial limits on individual inference.
The site’s hospital serum-alcohol analysis addresses the separate uncertainty involved in interpreting serum results. Brain spectroscopy adds another measurement problem rather than resolving that earlier one.
Primary source
Thierauf-Emberger A, Echle J, Dacko M, Lange T. Comparison of ethanol concentrations in the human brain determined by magnetic resonance spectroscopy and serum ethanol concentrations. International Journal of Legal Medicine. 2020;134:1713–1718. https://doi.org/10.1007/s00414-020-02325-w.