707 of the 740 cases involved male subjects, only 33 involved female subjects. The overall median age was 28.6 years with a range from 15 to 62 years. A majority of the cases (n = 569) involved individuals under the age of 34. No significant variation in the frequency of incidents was observed across different weekdays.

In most cases (n = 730) the type of vehicle was specified (or plausibly assumed to be) a car. 10 cases concerned motorbikes (N = 4) or e-scooters (N = 6).

In view of the newly established THC legal threshold in Germany (3.5 ng/mL, replacing the previous 1 ng/mL limit), we examined its potential impact on legal classifications. 182 of the 546 cases (33,3%) classified as suspected violations of § 24a (excluding the cases with a co-consumption of alcohol) showed a THC serum concentration under 3.5 ng/ml (with one of them having a THC serum concentration under 1 ng/ml).

For a better understanding of the effects of cannabis on driving, we first analysed the 695 DUIC cases’ distribution of age, THC serum concentration and CIF value, excluding cases in which cannabis in combination with alcohol was found (n = 45).

As illustrated in Fig. 1, the age distribution was similar between the administrative offense group (§ 24a) and the criminal offense group (§ 316), although slight differences were observed. The overall median age in the dataset was 27 years, with a mean age of approximately 28.5 years. The § 24a group had a median age of 28 years (interquartile range (IQR): 23–33 years), whereas the § 316 group showed a median of 24 years (IQR: 21–31 years).

Boxplot for age distribution of exclusively THC positive cases (cases with co-consumption of alcohol excluded); 541 cases being medicolegally classified as administrative offense (§ 24a) and 154 cases as potential criminal offenses (§ 316). The black line indicates the median, the box includes 50% of all values. The circles indicate outliers

The group with an exclusive influence of cannabis with specific evidence for criminal offenses (§ 316) from a medicolegal point of view showed higher mean THC serum concentrations and CIF values as illustrated in Figs. 2 and 3:

Comparison of the distribution of the THC serum concentration between the groups medicolegally classified as administrative offenses (§ 24a) and potential criminal offenses (§ 316)

Comparison of CIF value distribution in both groups. The case number differs slightly, as in some cases not all metabolites were measurable and CIF, therefore, not calculable

The median THC serum concentration in the § 24a group was 5.4 ng/ml (interquartile range [IQR]: 2.7–11 ng/ml), with a mean value of approximately 8.39 ng/ml. The § 316 group exhibited a higher median concentration of 6.8 ng/ml (IQR: approximately 3.63–14 ng/ml), and a higher mean value of approximately 11.53 ng/ml.

To assess the distribution of the Cannabis Influence Factor (CIF), values were capped at a maximum of 30, in line with established methodological recommendations. The median CIF value in the § 24a group was approximately 13 (IQR: 3–19; mean value: 12), while the § 316 group showed a higher median CIF of approximately 15.5 (IQR: 8–21; mean value: 14). Notably, in 61 cases (11.3%) from the § 24a group and 28 cases (18.2%) from the § 316 group, the calculated CIF exceeded the upper threshold of 30.

Given the higher mean/median CIF values observed in cases with specific evidence of criminal offenses (§ 316 group), we further examined THC-COOH serum concentrations in both groups to explore a potential link to regular cannabis use and its contribution to the CIF. As shown in Fig. 4, no clearly differential pattern (e.g. clusters of high-CIF/low-THC-COOH in the § 316-group or low-CIF/high-THC-COOH in the § 24a-group) can be observed between the two groups.

Comparison of the coherent CIF and THC-COOH (ng/ml) values of each case in both groups with a line of best fit

In order to identify driving and behavioural impairments that are solely due to cannabis consumption, we excluded cases with an additional consumption of alcohol (analysed n = 695 cases).

A total of 125 cases with driving impairments and 219 cases (double mentioning possible) with behavioural impairments were found.

Regarding the driving impairments, 70 of the 125 (56%) cases were presenting “hard criteria” for a criminal offense (e.g. loss of vehicle control, swerving, lane departure or use of multiple lanes, red light violations, single-vehicle accidents, collisions with vehicles, rear-end collisions, violations of right-of-ways) with most of them being abnormal speed, followed by accidents and lane-keeping violations.

Of the 219 behavioural impairments, 107 (49%) were classified as “hard criteria” for a criminal offense, most of which were found in the cognition, consciousness, mood and gait/posture categories.

Despite these findings, no statistically significant relationship could be established between either THC serum concentration or CIF value and the presence of driving impairments. The same was true for behavioural impairments, despite a significant difference between the two groups in the category “consciousness” with significantly more subjects described as “confused” in the § 316 group.

To further explore the relationship between THC serum concentration and impairments according to a criminal classification under § 316, we analysed all 194 cases with specific evidence of criminal offenses—including those involving co-consumption of alcohol.

As shown in Fig. 5, the majority of these cases (n = 134) had THC serum concentrations below 10 ng/mL, although outliers reached values exceeding 70 ng/ml. The mean THC serum concentration in this group was approximately 10 ng/mL, while the median value was 6.1 ng/ml.

All cases (including co-consumption of alcohol) with specific evidence for criminal offenses (§ 316) in correlation to the measured THC serum concentration (n = 194)

Using t-tests, cases with specific evidence of criminal offenses exhibited significantly higher THC serum concentrations compared to cases classified as administrative offenses (p < 0.01). A similar, statistically significant difference was observed for CIF values (p = 0.045).

A total of 45 cases showed positive blood alcohol and THC serum concentrations. BACs varied from 0.4 g/kg to 2.41 g/kg with a median of 1.31 g/kg. Six cases were classified as administrative offenses, whereas 39 cases showed specific evidence for criminal offenses.

The cases involving co-consumption presented a total of 162 driving impairments (100 impairments or 62% of those being classified as “hard criteria”) with most of them in the categories “speeding violation”, “lane keeping violation” and “accident”, showing a significant correlation for lane keeping violations and accidents.

As illustrated in Fig. 6, the majority of co-consumption cases showed specific medicolegal evidence for criminal offenses (41 out of 45 cases). In most cases, BAC alone already explains the criminal offense (“per se”-law). Of the four cases classified as putative administrative offenses, most showed no documented driving or behavioural impairments. In the one case where impairments were noted, further toxicological analyses were recommended. These four cases had BACs ranging from 0.4 g/kg to 0.67 g/kg. In contrast, all cases within the § 316 group demonstrated at least one documented driving and/or behavioural impairment. However, no statistically significant interaction was observed between THC serum concentration and BAC with regard to the likelihood or severity of impairments.

Cases with co-consumption of cannabis and alcohol in correlation with the measured blood alcohol concentration (BAC) and THC serum concentration (n = 45). Cases classified as putative administrative offenses were illustrated as triangles (n = 4), whereas cases with specific medicolegal evidence for criminal offenses were illustrated as dots (n = 41). The horizontal red line marks a BAC of 1.1 g/kg, the vertical red lines mark THC serum concentration of 1 ng/ml resp. 3.5 ng/ml

Only one case — presenting no compelling signs for a criminal offense — had a BAC below 0.5 g/kg with a THC serum concentration of 0.61 ng/ml. All other cases with co-consumption had a BAC above 0.5 g/kg, with 32 cases having a BAC above 1.1 g/kg. The THC serum concentration ranged between 0.34 ng/ml and 26 ng/ml with only 8 cases showing a THC serum concentration below 1 ng/ml (and 22 cases showing a THC blood concentration of less than 3.5 ng/ml).