A pecularity of alcohol (ethanol) is the comparatively stable relationship between dose and effects. With increasing blood alcohol concentrations (BACs), the magnitude of impairment increases [1, 2]. Nevertheless, with regard to road traffic, interindividual differences at similar BACs have repeatedly been described [3, 4].

Whether alcohol-related impairments are observable or not regularly depends on the complexity of the driving task [1]. In general, driving with a BAC of around 0.50 g/l increases the relative probability for single vehicle accidents by factor 2 [3]. Performance differences in comparable situations are often explainable by different states of alcohol habituation.

Besides habituation, several variables can directly or indirectly influence the driving-related effects of alcohol resp. the BAC. For example, the driver’s age is associated with different crash risks, which is likely linked to more increased alcohol-impaired crash avoidance skills in young drivers and risk-taking personality traits [5]. Similar BACs may provoke different levels of impairment at different times of the day as the circadian rhythm interferes [6, 7]. The ethanol elimination rate is usually (slightly) higher in women than in men [8, 9]. The state of alcohol hangover may also be associated with significantly impaired performances even if the BAC has returned to zero [10, 11].

A comparable BAC supposedly leads to more signs of impairment in the phase of alcohol resorption than in the phase of alcohol elimination. In other words, the BAC supposedly lags behind the alcohol effect. This observation is called “Mellanby effect”, since Sir Edward Mellanby was the first to describe it [12]. Mellanby carried out his experiments on dogs at the time, and he could therefore only assess gross motor abnormalities. The evidence for Mellanby effect is still under discussion, and driving safety skills might be more affected in the phase of alcohol elimination [13].

The presented work should clarify the question whether a Mellanby effect is evident in alcoholised e-scooter drivers. Results of a previously published study [14] were therefore re-analysed. As repeated e-scooter runs at different BACs were performed, several test persons completed the course with comparable BACs in the phase of alcohol resorption and alcohol elimination.

Material and methods (for details, see [14]).

Sixty-three healthy subjects, who were experienced in driving an e-scooter participated in the study (31 females, 32 males). Six subjects (3 females, 3 males) remained sober during the whole trial and served as control group. Two female test persons dropped out during the trial.

The test persons’ state of alcohol and drug soberness were checked at the beginning of each test day by breath-alcohol analyses and immunochemical screening of the urine.

Questionnaires were used to assess alcohol experience (AUDIT) and e-scooter experience.

For these re-analyses, matched pairs (e-scooter runs with comparable BACs with a maximum difference of 0.20 g/kg between the phase of alcohol resorption and elimination) were found in 16 alcohol-consuming test persons (9 females, 7 males) in the age range of 18 to 47 (median 24 years) (see below “Statistical analyses to evaluate a possible “Mellanby effect””).

Each of the 4 test days started at 10 a.m. and lasted until approx. 8 p.m.

The course was built on a non-public area (52 m × 18 m).

After accommodation to course and e-scooters (Tier, Modell ES 400B, Tier Mobility AG, Berlin), the sober run, which served as baseline, was performed.

Afterwards, alcohol consumption started (duration between approx. 2 h 22 min and 4 h 41 min). The amount of alcohol to be consumed was calculated in advance using the Widmark formula [15] in such a way that the BAC should reach a maximum value of approx. 1.30 g/kg.

Under the influence of alcohol, 3–4 runs were completed by each test person. All runs were videotaped for the purpose of later evaluation.

After each run, blood samples were taken.

Nine obstacles had to be passed that can be seen in Fig. 1 (in chronologic order: narrowing track (45-m length); gate passage (spaced at 1.30 m); gravel bed of 6.90-m length; driving in circles counterclockwise 3.5 times; three turns with timely indication of the direction (left-right-left); three thresholds; alley drive (width: 1.05 m; length 5.55 m); slalom ride with decreased spacing (2 x 4 m; 2 x 3 m; 2 x 2 m; 1 x 1,5 m); speed track (17.7 m resp. 16.5 m)).

Course I: narrowing track; II: gate passage; III: gravel bed; IV: driving in circles counterclockwise; V: three turns with timely directional indication; VI: three thresholds; VII.i.: alley drive; VIII: slalom ride; IX: speed track) [14]

Two elements were slightly modified after the trial had already started: The speed track was shortened from 17.70 to 16.50 m, and a light signal was installed at the alley drive, which signaled the lane to be used. Furthermore, some subjects were asked to pass 6 of the abovementioned 9 elements a second time on each trip to better assess concentration losses or exhaustion effects during longer trips. Only the first passage was taken into account for further analyses.

A run through the 9 obstacles lasted on average about 1.5 min.

Demerits were allocated for distinctive driving (coordinative or cognitive) features. Driving and medical features were evaluated separately.

More demerits were allocated for features that were considered more relevant to road traffic. For example, touching a boundary line (e.g. when driving straight ahead or during driving in circles) was allocated 1 demerit, crossing the line with one wheel was allocated 2 demerits and skipping a gate during gate passage was allocated 3 demerits.

BAC differences were set to a maximum of 0.20 g/kg between a test person’s run in the phase of alcohol resorption and alcohol elimination.

Persons were supposed to reach a maximum BAC of 1.30 g/kg. The alcohol to be consumed was calculated by the formula of Widmark [15].

To fulfill the demand of alcohol resorption, test persons were either still drinking alcohol or the termination of alcohol consumption was not longer than 60 min ago.

To fulfill the demand of alcohol elimination, test persons had finished alcohol consumption at least 90 min ago.

Basically, the added demerits per run were evaluated.

The added demerits per obstacle formed the “error score”.

The added demerits of all unchanged obstacles (I, II, III, IV, V, VI, VIII) formed the “absolute score”. If obstacles had to be passed twice, only the first passage of each obstacle during each run was included.

The “individual score” was calculated by comparing the results of the respective run under the influence of alcohol with the results of the sober run (with all test persons as control group). Subjects with an absolute score of zero points received one point to enable the calculations of all individual scores.

An “influence score” (relative driving performance in different states of alcohol influence) was determined by dividing the error score of the alcohol elimination phase by the error score of the phase of alcohol resorption.

The result of a statistical test was considered significant if the p value was less than 0.05.