Retrograde alcohol estimation is often presented in litigation as elementary arithmetic: start with a later alcohol concentration, subtract or add a fixed amount per hour, and announce the “BAC at the time.” The mathematics may be simple, but the science is conditional. A defensible opinion depends on whether the person was absorbing, at peak, or eliminating alcohol; whether the drinking history is reliable; and whether the chosen elimination assumptions fit the evidence.
Lim and colleagues published a forensic application of a population pharmacokinetic model in which defendant-specific information was combined with Bayesian estimation. Rather than reporting a single deterministic answer, the authors generated a concentration-time profile with a 95% prediction interval. That approach highlights a central litigation point: uncertainty is not a defect in the science. Concealing uncertainty is.
For counsel examining an alcohol reconstruction, the first question should be whether the expert has identified the phase of the concentration curve. Back-extrapolation is most straightforward only after absorption is substantially complete. If drinking ended near the event, a later specimen may have been collected while the concentration was still rising. In that situation, a simple elimination-rate calculation can overstate confidence and may even infer the wrong directional change.
The second question is the factual foundation. What quantity was consumed? What was the alcohol concentration of the beverages? Over what period? Was food consumed? What is known about body size, sex, relevant physiology, and the specimen collection timeline? Each unsupported input widens the scientifically reasonable range even when the analytical measurement itself is excellent.
The third question is whether the expert has separated model output from observed fact. A measured breath, blood, or serum ethanol concentration is evidence. An estimated earlier concentration is an inference generated from evidence plus assumptions. That distinction should be explicit in the report and in testimony.
Human studies on gastric emptying reinforce why this matters. Alcoholic beverages can slow gastric emptying, and controlled beer studies show that gastric emptying is related to early plasma ethanol response. Those data do not allow a lawyer to calculate a case-specific absorption curve from beverage type alone. They do show why universal assumptions about rapid, complete absorption should be tested against the facts rather than treated as axioms.
A strong expert opinion should therefore disclose the model, assumptions, plausible parameter range, sensitivity to competing drinking histories, and the uncertainty interval. In cross-examination, a single-number conclusion may be less persuasive once the hidden assumptions are made visible.
Attorney-scientists are particularly well positioned to scrutinize that boundary between measurement and inference. The objective is not to make the number larger or smaller. It is to determine how much scientific confidence the evidence actually supports.
Frequently Asked Questions
Is retrograde extrapolation always unreliable?
No. It can be useful when the sampling time and drinking history are sufficiently known and the subject is reasonably established to be post-absorptive. The reliability decreases as those assumptions become uncertain.
Why use a prediction interval?
Because pharmacokinetic parameters and historical inputs vary. A prediction interval communicates the range of concentrations compatible with the model and assumptions rather than creating false precision.
Selected Research
Lim H-S, Soung JH, Bae K-S. Translational and Clinical Pharmacology. 2017. doi:10.12793/tcp.2017.25.1.5.
Kasicka-Jonderko A, et al. Journal of Gastroenterology. 2013. doi:10.1007/s00535-013-0752-y.
Stevens JE, et al. British Journal of Clinical Pharmacology. 2022. doi:10.1111/bcp.15297.