Retrograde Extrapolation of Blood Alcohol Concentration in Turkish Law: The Fixed Hourly Rate and the Evidential Limits of the Widmark Equation

Türk Hukukunda Kan Alkol Konsantrasyonunun Geriye Dönük Hesaplanması: Sabit Saatlik Oran ve Widmark Denkleminin İspat Sınırları

Av. Serkan KARA¹

Published October 20, 2026

INTRODUCTION

Drink-driving law asks a question about a moment that has passed. Turkish law states its administrative and criminal thresholds as concentrations of alcohol in the blood at the time of driving, but after an accident the measurement that reaches the file is usually made in a hospital an hour or more later. The survey of elimination rates on which the British professional guidelines base their figures opens with the same observation: 'A blood sample for forensic analysis might not be taken until several hours after an offence was committed. The courts usually want to know the suspect's blood-alcohol concentration (BAC) at some earlier time, such as the time of driving.'² The bridge between the two moments is retrograde extrapolation, the backward use of the model that Erik Widmark developed in the 1920s.

Turkish law has chosen to build that bridge with a single number. The Highway Traffic Regulation adds 0.15 per mille for every hour that passes before the hospital test, and the Twelfth Criminal Chamber of the Court of Cassation carries later results back to the time of an accident on the same average. The current professional standard of the American Academy of Forensic Sciences takes a different view. In its words, the calculation 'shall be performed using a range of elimination rates', because 'the science does not support being able to provide a single value'.³

This article asks whether a fixed hourly rate can carry the evidential weight that Turkish administrative, criminal and insurance law places on the concentration at the time of driving and, if it cannot, what a back-calculated figure is able to prove. It argues that retrograde extrapolation proves a range and not a point, that it proves the range only on the condition that the concentration was falling for the whole interval, and that the end of the range on which a decision may rest is fixed by the burden of proof of the proceedings in which the figure is used.

The method is doctrinal and interdisciplinary. The statutory provisions and six decisions of the Court of Cassation and of a regional court of appeal were read in the Ministry of Justice's UYAP legislation and case-law program, where the decisions were found by searches for the terms in which the courts describe back-calculation.⁴ The pharmacokinetic literature and the three current professional standards on alcohol calculations, from the United States, from the United Kingdom and Ireland and from England and Wales, were read from their publishers, and the Turkish forensic and legal literature from the journals' own platform. All sources were read on 20 October 2026, and translations from Turkish are the author's. The subject is ethyl alcohol in the living driver. Post-mortem cases and other substances lie outside it.

Turkish law states blood alcohol in per mille (promil), the scientific literature in milligrams per 100 millilitres (mg/dL) or grams per decilitre. The Turkish reports and decisions read hospital results on the scale on which 100 mg/dL is 1.00 per mille, and that scale is used throughout.⁵ An elimination rate of 15 mg/dL per hour is therefore 0.15 per mille per hour. Part I sets out the thresholds and the statutory bridge. Part II states the pharmacokinetic basis of the calculation and the width of its terms. Part III discusses the international and the Turkish literature and identifies the point at which this article departs from it. Part IV reads the Turkish decisions against the standards, and Part V draws the consequences for proof.

I. THE LEGAL FRAMEWORK: THRESHOLDS FIXED AT THE MOMENT OF DRIVING

A. The Administrative Thresholds and the Duty to Test

Article 48 of the Highway Traffic Law No 2918, recast in 2013 and amended in 2026, prohibits alcohol-affected drivers from driving.⁶ A driver found to have driven above 0.50 per mille is fined 25,000 Turkish liras and has the licence withdrawn for six months 'even if the act constitutes a crime', and for vehicles other than private cars the lower limit is 0.21 per mille.⁷ A candidate driver loses the candidate licence above 0.20 per mille, whatever the vehicle.⁸ The Highway Traffic Regulation states the limits as the amount of alcohol in the blood of the driver.⁹

The law ties the duty to test to the accident. A driver involved in an accident with injury or death, or in a damage accident the police attend, must be examined with the technical device. A driver who objects to the device's result or refuses the device is taken to the nearest forensic medicine institution, forensic physician's office or Ministry of Health facility, where samples of blood, saliva or urine are taken under Article 75 of the Code of Criminal Procedure, and the same applies where the driver died or was too badly injured to blow into the device.¹⁰ Outside an accident, an objection to the roadside result leads to no new measurement, and the measures taken are challenged before the courts under Article 27 of the Misdemeanours Law No 5326.¹¹ After an accident, therefore, the concentration that decides the case is regularly measured later than the moment it describes.

B. The Criminal Thresholds

Article 179(3) of the Turkish Criminal Code punishes with six months' to two years' imprisonment a person who drives 'although unable to drive safely under the influence of alcohol or drugs or for another reason'.¹² Article 48(6) of the Highway Traffic Law adds that drivers found to be above 1.00 per mille are in addition subject to that provision, and Article 48(7) that a driver of a private car above 0.50 per mille, or of another vehicle above 0.20, who causes an accident is subject to the relevant provisions of the Criminal Code.¹³ The Regulation hands a driver above 1.00 per mille to the police for criminal proceedings, and a driver below that figure only where a record establishes that alcohol made safe driving impossible.¹⁴

In fatal and injury cases the threshold enters the assessment of fault. Article 22(3) of the Criminal Code defines conscious negligence as the case where the result the offender foresaw, but did not want, occurs, and increases the punishment for the negligent offence by one third to one half.¹⁵ The Twelfth Criminal Chamber of the Court of Cassation, relying on the views of the Forensic Medicine Institution and on its own settled practice, treats a concentration above one per mille at the time of the accident as removing the ability to drive safely and as a ground of conscious negligence, and a lower concentration as insufficient on its own.¹⁶ The criminal consequence of a fatal accident can therefore turn on whether the concentration at the moment of the accident lay on one side of one per mille or on the other.

C. The Insurance Questions

The same question reaches the civil courts. The general conditions of motor own-damage insurance, as the Court of Cassation has quoted them, exclude damage that occurs while the vehicle is driven by persons who drank more than the amount the Highway Traffic Law prohibits, and those of compulsory liability insurance, as a regional court summarised them, give the insurer a right of recourse where the vehicle was driven above the level the legislation sets.¹⁷ Under Article 1409(2) of the Turkish Commercial Code the insurer bears the burden of proving that a risk falls outside the cover, and the courts have added the requirement that the accident occurred exclusively under the effect of alcohol.¹⁸ Whatever the weight of that further requirement, which Part III discusses, the first fact the insurer must prove is again a concentration at a moment that nobody measured.

D. The Statutory Bridge: A Fixed Hourly Rate

Article 97 of the Highway Traffic Regulation closes the interval for the administrative sanction. Where the driver objected to the device and the determination was made at a health facility, the period between the two determinations is taken into account, and where the driver refused the device, the period between the accident and the hospital determination. The traffic unit that made the first measurement then determines the alcohol level 'by adding 0.15 per mille for each hour' to the hospital result, and takes action on that figure.¹⁹

The rule is written for the administrative sanction, but the criminal courts work with the same figure. The Twelfth Criminal Chamber has recorded the settled reports of the Fifth Specialized Board of the Forensic Medicine Institution as stating that the alcohol level in the body falls by an average of 0.15 per mille every hour,²⁰ and in the case discussed in Part IV it reached the concentration at the time of the accident by 'adding 0.15 per mille for each hour, backward'.²¹ The rule contains no term for the drinking history, no exception for the period after the last drink, no lower limit and no reference to the kind of sample. Whether a single figure of this kind can carry the weight that the thresholds place on it is a question the pharmacokinetic evidence answers.

II. THE PHARMACOKINETIC BASIS OF RETROGRADE EXTRAPOLATION

A. Widmark's Model and Its Terms

Alcohol calculations rest on the work Widmark carried out in the 1920s and published in 1932.²² In its forward form the model gives the concentration at a given time as the mass of alcohol absorbed divided by the body weight and by a distribution factor, less the alcohol eliminated since drinking began, which is the elimination rate multiplied by the time.²³ The distribution factor, now read as the apparent volume of distribution of ethanol in the body water, averages about 0.70 L/kg in healthy men and about 0.60 L/kg in women.²⁴ The American standard nonetheless holds that 'the use of a single fixed Vd is inappropriate' and gives 95 per cent ranges of 0.58 to 0.83 L/kg for men and 0.43 to 0.73 L/kg for women.²⁵

The backward form is simpler. The concentration at the time of the incident equals the measured concentration plus the elimination rate multiplied by the time between the incident and the test.²⁶ It contains neither the distribution factor nor the body weight, which is why it can be applied to a measured result without any drinking history. It needs three things: the elimination rate, the interval, and the condition that the concentration fell throughout the interval. The distribution factor returns when the drinking history matters, for example where the contribution of a drink that may not yet have been absorbed has to be estimated and subtracted.²⁷

B. The Elimination Rate and Its Distribution

Once the first drinks have saturated the liver enzyme, alcohol leaves the blood at a roughly constant amount per hour, and Jones's evidence-based survey found that 'zero-order kinetics is adequate to describe the declining phase of the BAC profile in most forensic situations (BAC>20mg/100mL)'. The survey placed the rate at 10 to 15 mg/dL per hour after drinking on an empty stomach, 15 to 20 in non-fasted subjects and 25 to 35 in alcoholics during detoxification, put the physiological range at 10 to 35, and concluded that 'in moderate drinkers 15 mg/100mL/h remains a good average value for the population, whereas in apprehended drivers 19 mg/100mL/h is more appropriate, since many of these individuals are binge drinkers or alcoholics'.²⁸ The guidelines of the United Kingdom and Ireland Association of Forensic Toxicologists, drawing on 1,090 individuals from casework with a mean of 19.1 mg/dL per hour, adopt a range of 9 to 29 mg/dL per hour with 19 as the most likely rate. The range is a 95 per cent interval, so that '1 in 20 individuals may fall outside of this range'.²⁹ The American standard requires at least the range of 0.010 to 0.025 g/dL per hour, which 'encompasses the majority of the population regardless of age, sex, ethnicity, and drinking experience'.³⁰ German forensic practice, when it estimates a minimum concentration in the subject's favour, applies a low rate of 0.1 per mille per hour, which is 'legally accepted', and a recent re-analysis of four controlled drinking trials with 87 subjects found an overall decline of 0.156 per mille per hour across the full elimination curve.³¹

Read in per mille, the published ranges are 0.10 to 0.25 per hour on the American standard, 0.09 to 0.29 on the British guidelines and 0.10 to 0.35 at the physiological limits, with 0.15 as the average of moderate drinkers and 0.19 as that of apprehended drivers. The Regulation's figure is therefore the average of moderate drinkers. It is lower than the average that Jones and the British guidelines give for drivers who have been stopped, and it lies inside ranges whose upper ends are two and a half times their lower ends or more. The uncertainty of a calculated concentration follows the same logic. Searle, re-working the published examples by error propagation, showed that the coefficient of variation 'is not some fixed percentage but must be calculated in each case', and that a general figure is 'far from correct, especially when elimination has played a large part'.³² In a back-calculation over several hours, elimination is most of the result.

C. Absorption and the Post-Absorptive Condition

The constant fall begins only after absorption. The maximum concentration 'is usually reached between 10 and 60 min postdosing',³³ but the American standard records that 'it can take up to 2 hours to reach the post absorptive phase after the last drink', that the time needed to reach the peak 'is not the same as the time to reach the post absorptive phase', and that 'the linear elimination rate only applies when the subject is in the post absorptive phase'.³⁴ The guidance of the Forensic Science Regulator for England and Wales draws the consequence. The process 'is, however, based on the assumption that over the period between the time of interest and the time the sample was taken the alcohol concentration in the person's blood is decreasing. If this is not true then the result of the back calculation may be inaccurate. It could significantly overestimate the alcohol concentration.' It follows that 'back calculation should not be performed if it is likely the alcohol concentration was not decreasing for the whole calculation period', and an expert who performs one must explain why the assumption was valid in the circumstances of the case.³⁵

The standards translate the condition into time. The British guidelines advise that a standard back-calculation should not normally be performed where alcohol was consumed within one hour of the incident, and that caution applies within two hours of the last drink where a meal was eaten.³⁶ The American standard treats a subject as post-absorptive where the incident occurred more than two hours after drinking ended, requires the impact of potentially unabsorbed alcohol to be addressed, and holds that 'when the drinking history is unknown, it is not reasonable to assume that the subject is post absorptive'. Its own example is a driver who drank a last shot as she left a bar and crashed while leaving its car park. Her result of 0.082 g/dL, taken ninety minutes later, gives 0.097 to 0.120 g/dL at the time of the crash on the assumption that absorption was complete, but 0.057 to 0.090 g/dL if the last drink was not yet absorbed, so that she may have been below the legal limit of 0.08 g/dL.³⁷ Even the two-hour window is a population figure. Food, certain medications and medical conditions can prolong absorption beyond two hours, and a review of 36 back-calculation reports of one state laboratory in the United States found that every one of them assumed absorption complete within two hours.³⁸

D. The Lower Limit of Linear Elimination

The constant rate also ends at low concentrations. The American standard records that below 0.020 g/dL 'the elimination rate may not be linear as zero order kinetics may no longer apply' and forbids retrograde extrapolation from concentrations below that figure, and the British guidelines allow a standard back-calculation only from 20 mg/dL, below which Michaelis-Menten kinetics predominate.³⁹ On the Turkish scale that floor is 0.20 per mille. German expert practice typically refrains from back-calculation below 0.15 per mille, and the 2026 re-analysis that proposes a corrected rule for low results would still omit the calculation entirely below 0.06 per mille. The same study notes that endogenous ethanol in human blood is reported at about 0.01 to 0.05 per mille.⁴⁰ A result in that region says little about the interval before it, and a calculation that starts from it consists almost entirely of the assumed rate multiplied by the time.

E. The Matrix of the Sample

Statutory limits refer to whole blood, while hospital laboratories may measure serum or plasma, which hold more water. Jones and Tilson, reviewing the distribution ratios, note that statutory limits 'are defined as the concentration of ethanol in WB and not in plasma, serum or red-blood cells', that the mean serum or plasma to whole blood ratio is about 1.15, and recommend that for legal purposes a hospital's plasma or serum result be divided by 1.2.⁴¹ The American standard requires serum and plasma results to be converted to a whole-blood equivalent, with a ratio of 1.13 to 1.19, before any other calculation, and forbids retrograde extrapolation from urine results, 'even those converted to a whole blood equivalent'.⁴² Turkish law sets its limits on the amount of alcohol in the blood, yet Article 48(3) of the Highway Traffic Law allows urine samples to be taken.

III. THE STATE OF THE LITERATURE

A. The International Debate

Jones's survey states the position plainly: making back calculations of blood alcohol in criminal cases 'has many proponents and critics'.⁴³ Montgomery and Reasor gave the case for the method. Retrograde extrapolation, they wrote, may be applied 'with reliability when reasonable assumptions are made concerning absorption rates, elimination rates, and patterns of alcohol consumption'. After peak absorption a relatively narrow range of results follows from elimination rates of 0.015 to 0.020 g/dL per hour, and a wider range of 0.01 to 0.03 has 'little practical effect', while before the peak 'a wider range of predicted BAC values will result'. They added that reported drinking volumes are 'notoriously inaccurate'.⁴⁴ Gullberg and McElroy, comparing roadside and later breath tests of 161 arrested drivers, found that extrapolation at an assumed rate produced a small but significant overestimate of the roadside result, and concluded that evidential analyses within two hours of driving 'will provide good estimates and certainly not overestimates' and that 'extrapolation may be unwarranted in these cases'.⁴⁵ Posey and Mozayani observed that many practitioners are reluctant to perform the calculation, that 'some jurisdictions expressly prohibit them', and that improvements to Widmark's formula are frequently ignored by those who do.⁴⁶

The recent literature has moved from whether the calculation can be made to the conditions on which it is made. Two letters in the Journal of Forensic Sciences called for a scientific consensus on the application of retrograde extrapolation in drink-driving cases, and a third, in 2024, set out an answer.⁴⁷ Saskőy and co-authors found that the guidance issued by the British toxicologists and by the American Organization of Scientific Area Committees is 'not fully agreed', and proposed to widen the range of elimination rates to 8 to 29 mg/100 mL per hour and to adopt a two-hour limit after the last drink, which in their view would better meet the standard of proof beyond reasonable doubt.⁴⁸ Maskell and co-authors argued that the percentile range used for the volume of distribution and the elimination rate should depend on the case type, recommending the 25th to 75th percentiles in civil cases and the 0.5th to 99.5th or 2.5th to 97.5th percentiles in criminal cases.⁴⁹ Olson, on the absorption side, asked toxicologists to disclose 'that population averages regarding the absorption of alcohol may not apply to individual cases'.⁵⁰ The professional standards of 2020 to 2024 converge on a range of elimination rates, the post-absorptive condition, a lower limit and the conversion of serum results, and they differ on the width of the range and on the length of the absorption window.

B. The Turkish Literature

The Turkish forensic literature reports the same averages. In their review of the forensic problems of alcohol, Baduroğlu and Durak give the elimination rate as 15 mg/100 mL per hour in moderate drinkers and 19 in heavy drinkers, and note that the amount lost to metabolism since drinking began must be considered whenever a concentration is carried back to the time of an event.⁵¹ They also report the scheme of the Fifth Specialized Board of the Forensic Medicine Institution: at 0 to 30 mg/dL the ability to drive safely is taken as unimpaired, from 31 to 100 mg/dL it cannot be determined from the level alone, and from 101 mg/dL it is taken as lost, so that drivers in the middle band are judged on the police records and the accounts of witnesses.⁵²

The Turkish legal literature has addressed the threshold rather than the calculation. Seven, writing on the effect of the 2013 amendment on the insurer's burden of proof, records that the Court of Cassation had required proof that the accident occurred exclusively under the effect of alcohol and had held that alcohol limits could not be left to a regulation. Because the recast Article 48 now fixes the limits in the statute itself and no longer speaks of a loss of the ability to drive safely, Seven concludes that the insurer meets its burden under Article 1409(2) of the Commercial Code by proving that the driver was above the statutory limits, without proving that the accident occurred exclusively under the effect of alcohol.⁵³ The case law has not followed that view in every instance. In 2026 the Court of Cassation upheld a judgment whose reasoning, as the Court summarised it, still required that the accident have occurred exclusively under the effect of alcohol.⁵⁴

C. The Point of Departure

The Turkish literature takes the elimination rate as an average to be applied, and the legal literature treats the concentration at the time of the accident as a fact the insurer must prove, without asking how a later sample proves it. The international literature treats the range, the absorption condition, the lower limit and the matrix as scientific questions. This article reads them as legal questions. Which concentration the evidence proves depends on which end of the range the applicable burden of proof requires, and the conditions of the calculation are conditions of its value as proof of that concentration. On Seven's view, under which a concentration above the statutory limit is the whole of the insurer's burden, the question of how a later sample proves that concentration becomes the whole of the case.

IV. THE TURKISH DECISIONS READ AGAINST THE STANDARDS

A. A Criminal Case Decided at 0.97 Per Mille

In a decision of 28 December 2022 the Twelfth Criminal Chamber considered a driver who had struck a car parked on the emergency lane and injured its occupant. A test 1 hour 48 minutes after the accident showed 0.70 per mille. Adding 0.15 per mille for each hour gave 0.97 per mille at the time of the accident. Because 0.97 is below one per mille, and because the file contained no finding on the driver's loss of ability or outward behaviour and no evidence that alcohol caused the accident, the Chamber held that the conditions of conscious negligence were not met and reversed the judgment.⁵⁵

On the published ranges the same result gives 0.88 to 1.15 per mille on the American standard and 0.86 to 1.22 per mille on the British guidelines, and at the British most likely rate of 0.19 per mille per hour it gives 1.04 per mille. The finding that the driver was below one per mille was therefore a property of the chosen average. The result of the decision stands on a sounder ground: the lower end of every published range lies below one per mille, so it was not proved that the driver was above it. The fixed figure can as easily cut the other way. A result of 0.75 per mille two hours after an accident becomes 1.05 per mille on the average of 0.15, above the criminal threshold, while the range of the American standard runs from 0.95 to 1.25 per mille and its lower end lies below that threshold. A result of 0.30 per mille ninety minutes after a refusal becomes 0.525 per mille under the Regulation, above the 0.50 limit, while the lower end of the range is 0.45 per mille. A single number turns a range that straddles a threshold into a finding on one side of it.

B. Back-Calculation from Results Near Zero

Turkish decisions show both readings of results near zero. In 2012 the former Seventeenth Civil Chamber of the Court of Cassation upheld the dismissal of an own-damage claim that rested on an expert's calculation. The expert had carried 0.06 per mille, measured at 03:36, back to 0.86 per mille at the time of an accident at 22:15 the evening before, at 0.15 per mille per hour. Two members dissented. So small an amount, they wrote, could have come from a medicine or from fruit, the insurer bore the burden of proof, and carrying 0.06 per mille back to the time of the accident by Widmark's formula did not accord with the file.⁵⁶ Of the 0.86 per mille, 0.80 was the assumed rate multiplied by five hours and twenty-one minutes, and 0.06 was the measurement. In 2026 the Fourth Civil Chamber upheld a judgment whose expert report, as the appellant's grounds describe it, carried 7.3 mg/dL, found in blood taken three hours after a crash, back to 0.64 per mille, without addressing the calculation.⁵⁷ Both starting values lie below the floor of 0.20 per mille that the American and British standards set, and the first lies inside the range reported for endogenous ethanol.

The Forensic Medicine Institution has taken the other view, and in the decisions read here the courts followed it. In an own-damage case decided by the Ankara Regional Court of Appeal in 2025, an expert had carried a hospital result of 0.04 per mille back to 0.67 per mille at the time of the accident. The Institution's Fifth Specialized Board answered that a result below the device's limit of detection and quantification means that the driver was alcohol-free when the blood was drawn, that the available data could not show whether alcohol had been eliminated before the draw or had never been present, and that no backward calculation can be made from a value accepted as zero. The court preferred the Board's report to that of a panel with a single neurologist and held that the insurer had not proved the driver to be over the limit.⁵⁸ The Board's position is the position of the standards.

C. Two Results That Cannot Both Be True

In 2025 the Twelfth Criminal Chamber, on an appeal in the interest of the law, acquitted a driver whose breath test read 1.92 per mille at 20:19 and whose blood, drawn at 00:57, showed 0.7 mg/dL. The Board had found it medically impossible for 1.92 per mille to fall to that level in 278 minutes, had read the hospital result as alcohol-free at the draw, and had refused to calculate back from zero. The Chamber held that there was no evidence sufficient for a conviction, free of every doubt.⁵⁹ The pharmacokinetic evidence bears the Board out. Even at the fastest rate of Jones's physiological range, 0.35 per mille per hour, 278 minutes remove about 1.6 per mille, so that a driver at 1.92 per mille at 20:19 would still have had about 0.30 per mille at 00:57. The forward use of the range thus shows that the two results cannot both describe the driver that evening, and nothing in the file allowed the court to prefer one of them.

D. The Matrix in the Case Law

The hospital value in the Ankara case was recorded as 'ethanol (per mille) serum 0.04', and the expert carried it back without conversion.⁶⁰ On the ratio of about 1.15 that Jones and Tilson report, a serum value used as a blood value overstates the blood concentration by about a seventh, and the American standard requires the conversion before any calculation. The point mattered little in that case, because the value lay below the device's limit. It matters near the thresholds, where a seventh of the result can decide on which side of 0.50 or 1.00 per mille it falls. A urine result, which the Highway Traffic Law allows to be taken, cannot carry a back-calculation at all.

V. CONSEQUENCES FOR PROOF

A. The Fixed Rate as a Convention of Administration

For the administrative sanction the Regulation adopts a convention. It applies the average of moderate drinkers to every driver, so that it errs against the driver who eliminates more slowly and in favour of the driver who eliminates faster, and it presumes that every driver was past absorption for the whole interval. A convention of this kind may serve the administration of a sanction that is open to judicial review, but it is not a measurement of the driver. It cannot be carried into criminal proceedings as proof of the concentration at the time of the accident, because there the question is not what the administration may presume but what the evidence proves.

B. The End of the Range the Burden of Proof Requires

In criminal proceedings the Court of Cassation's own formula answers the question. A conviction requires evidence sufficient for it and free of every doubt, and the acquittal of 2025 rested on exactly that requirement.⁶¹ The concentration that a back-calculation proves to that standard is the lower end of its range, and only where the post-absorptive condition is itself proved. Where the lower end lies below the threshold, the threshold is not proved, whatever the average gives. In the insurance cases Article 1409(2) of the Commercial Code places the burden on the insurer, and the same logic applies to the insurer that relies on a concentration to exclude cover. Maskell and co-authors show that the science expects the choice of range to follow the case type. Which percentile corresponds to the Turkish civil standard of proof is a question of law for the court and not a choice for the expert, and the British guidelines add that 'when a calculated range is close to the prescribed limit or the charging threshold (prosecution limit) extra care should be taken in wording statements and any uncertainties clearly expressed to avoid possible miscarriages of justice'.⁶²

C. The Content of the Expert Report

Article 67(1) of the Code of Criminal Procedure requires the expert to explain 'the operations performed and the conclusions reached'.⁶³ The professional standards give that duty its content for a back-calculation. The Forensic Science Regulator requires the report to give the court sufficient information to determine whether the method is reliable, to explain why it was appropriate and to explain why the assumption of a falling concentration was valid in the case. The British guidelines require the minimum, maximum and most likely values, and the American standard requires the calculations to be documented and the assumptions clearly stated.⁶⁴ Read together with the Turkish thresholds, these requirements mean that a report states the time of driving, of each test and of the last drink as far as the file shows them, with the source of each, and the matrix of each result with its conversion to whole blood. It states the limit of detection and quantification of the instrument, and makes no calculation from a result below it or below the concentration at which elimination ceases to be linear. It states the range of elimination rates with its source and the minimum, most likely and maximum concentration that range gives at the time of driving, and the evidence on which the driver is taken to have been past absorption. A report that gives one figure at one rate does not tell the court what the evidence proves.

CONCLUSION

The research question was whether a fixed hourly rate can carry the evidential weight that Turkish law places on the concentration of alcohol at the time of driving. It cannot. Widmark's equation can carry a later result back to the moment of driving, but what it carries back is a range whose width is set by the measured variation of elimination rates, and it does so only if the concentration was falling throughout the interval. The Regulation's 0.15 per mille for every hour, and the same average in the criminal courts, is the average the literature gives for moderate drinkers and below the average it gives for apprehended drivers. It is a convention of administration and not a measurement of the driver. Where a back-calculated figure lies close to 0.50 or 1.00 per mille, the end of the range required by the burden of proof decides. Where the result lies at or near zero, nothing can be carried back. Where the sample was serum, it must first be converted, and where it was urine, it cannot be carried back at all. The Forensic Medicine Institution's refusal to calculate back from zero, and the Twelfth Criminal Chamber's acquittal where two results could not both be true, follow the science. The remaining step lies with the expert report and the court: a report that states the range, its condition and its matrix, and a court that asks which end of the range its standard of proof requires, leave the equation to prove what it can prove.

ABBREVIATIONS

• ASB: Academy Standards Board of the American Academy of Forensic Sciences
• BAC: blood alcohol concentration
• CMK: Ceza Muhakemesi Kanunu (Code of Criminal Procedure No 5271)
• E: esas (docket number of a Turkish court)
• FSR: Forensic Science Regulator (England and Wales)
• K: karar (decision number of a Turkish court)
• KTK: Karayolları Trafik Kanunu (Highway Traffic Law No 2918)
• mg/dL: milligrams per decilitre (per 100 millilitres)
• OG: Official Gazette (Resmi Gazete)
• TCK: Türk Ceza Kanunu (Turkish Criminal Code No 5237)
• TTK: Türk Ticaret Kanunu (Turkish Commercial Code No 6102)
• UKIAFT: United Kingdom and Ireland Association of Forensic Toxicologists
• UYAP: Ulusal Yargı Ağı Projesi (the Turkish judiciary's information system)
• Vd: volume of distribution
• WB: whole blood

BIBLIOGRAPHY

Books and Articles

Baduroğlu E and Durak D, 'Alkol ile İlgili Adli Tıp Sorunları' (2010) 36(2) Uludağ Üniversitesi Tıp Fakültesi Dergisi 65.

Gullberg RG and McElroy AJ, 'Comparing Roadside with Subsequent Breath Alcohol Analyses and Their Relevance to the Issue of Retrograde Extrapolation' (1992) 57 Forensic Science International 193.

Jones AW, 'Evidence-Based Survey of the Elimination Rates of Ethanol from Blood with Applications in Forensic Casework' (2010) 200 Forensic Science International 1.

Jones AW, 'Alcohol, Its Absorption, Distribution, Metabolism, and Excretion in the Body and Pharmacokinetic Calculations' (2019) 1(5) WIREs Forensic Science e1340.

Jones AW and Tilson C, 'Distribution Ratios of Ethanol and Water between Whole Blood, Plasma, Serum, and Erythrocytes: Recommendations for Interpreting Clinical Laboratory Results in a Legal Context' (2023) 68 Journal of Forensic Sciences 9.

Labay L and Logan B, 'Letter to the Editor: Call for a Scientific Consensus Regarding the Application of Retrograde Extrapolation to Determine Blood Alcohol Content in DUI Cases' (2018) 63 Journal of Forensic Sciences 1602.

LeBeau MA and Limoges JF, 'Letter to the Editor: Call for a Scientific Consensus Regarding the Application of Retrograde Extrapolation to Determine Blood Alcohol Content in DUI Cases' (2019) 64 Journal of Forensic Sciences 322.

LeBeau MA and Limoges JF, 'Letter to the Editor: Answering the Call for a Scientific Consensus Regarding the Application of Retrograde Extrapolation to Determine Blood Alcohol Content in DUI Cases' (2024) 69 Journal of Forensic Sciences 1935.

Maskell PD, Fitsanakis VA, LeBeau MA, Castillo LV and Wilson AL, 'Assessing Uncertainty in Forensic Alcohol Calculations: The Case for Using Percentile Ranges' (2026) 71 Journal of Forensic Sciences 1793.

Mihan ATR, Toennes S, Paulke A and Verhoff MA, 'On the Possibility of Retrograde Calculation of Blood Alcohol Concentrations below 0.15‰' (2026) 140 International Journal of Legal Medicine 2833.

Montgomery MR and Reasor MJ, 'Retrograde Extrapolation of Blood Alcohol Data: An Applied Approach' (1992) 36 Journal of Toxicology and Environmental Health 281.

Olson A, 'Extended Absorption, Implications: Rethinking Alcohol Pharmacokinetics in Forensic Calculations' (2026) 94 Medico-Legal Journal 24.

Posey D and Mozayani A, 'The Estimation of Blood Alcohol Concentration: Widmark Revisited' (2007) 3 Forensic Science, Medicine and Pathology 33.

Saskőy L, Taylor L, Rooney B and Trotter G, 'Back to the Future: Retrograde Alcohol Calculations an Uncertain Science' (2024) 64 Science & Justice 455.

Searle J, 'Alcohol Calculations and Their Uncertainty' (2015) 55 Medicine, Science and the Law 58.

Seven V, 'Alkollü Araç Kullanma ile İlgili Karayolları Trafik Kanunu'nda Yapılan Değişikliğin Sigorta Şirketinin İspat Yüküne Etkisi' (2016) 22(3) Marmara Üniversitesi Hukuk Fakültesi Hukuk Araştırmaları Dergisi 2521.

Widmark EMP, Die theoretischen Grundlagen und die praktische Verwendbarkeit der gerichtlich-medizinischen Alkoholbestimmung (Urban & Schwarzenberg 1932).

Standards and Guidelines

AAFS Standards Board, ANSI/ASB Best Practice Recommendation 122: Best Practice Recommendation for Performing Alcohol Calculations in Forensic Toxicology (1st edn, 2024), aafs.org.

Forensic Science Regulator, Guidance: Alcohol Back Calculation for Road Traffic Investigations, FSR-G-220 (Issue 2, 2020), assets.publishing.service.gov.uk.

United Kingdom and Ireland Association of Forensic Toxicologists, Guidelines for Alcohol Calculations (Version 4.4, 2024), ukiaft.org.

Legislation

Ceza Muhakemesi Kanunu (Code of Criminal Procedure) No 5271, OG 17 December 2004, No 25673.

Karayolları Trafik Kanunu (Highway Traffic Law) No 2918, OG 18 October 1983, No 18195.

Karayolları Trafik Yönetmeliği (Highway Traffic Regulation), OG 18 July 1997, No 23053 (repeated).

Türk Ceza Kanunu (Turkish Criminal Code) No 5237, OG 12 October 2004, No 25611.

Türk Ticaret Kanunu (Turkish Commercial Code) No 6102, OG 14 February 2011, No 27846.

Cases

Ankara Bölge Adliye Mahkemesi 21. Hukuk Dairesi (Ankara Regional Court of Appeal, 21st Civil Chamber), E 2023/1732, K 2025/1631, 12 December 2025.

Yargıtay 4. Hukuk Dairesi (Court of Cassation, 4th Civil Chamber), E 2023/1051, K 2026/1268, 10 February 2026.

Yargıtay 12. Ceza Dairesi (Court of Cassation, 12th Criminal Chamber), E 2022/1654, K 2022/10915, 28 December 2022.

Yargıtay 12. Ceza Dairesi (Court of Cassation, 12th Criminal Chamber), E 2025/2343, K 2025/8840, 15 December 2025.

Yargıtay 12. Ceza Dairesi (Court of Cassation, 12th Criminal Chamber), E 2021/4697, K 2025/8889, 16 December 2025.

Yargıtay 17. Hukuk Dairesi, closed (Court of Cassation, former 17th Civil Chamber), E 2012/8427, K 2012/9997, 25 September 2012.