Hair samples were collected between 2021 and 2022 among students attending university courses in three Northern Italian regions. The study was submitted and approved by the Bioethics Committee of the University of Bologna (Prot. n. 76,007 of 26/03/2021) and has been performed in accordance with the ethical principles of the Declaration of Helsinki. Students from the Universities of Bologna, Parma (Emilia-Romagna region), Pavia (Lombardy region), and Turin (Piedmont region) were asked to participate. Inclusion criteria were the attendance of an academic degree course, a minimum age of 18 years, and the presence of head hair. Hair from other body districts was not considered. Exclusion criteria were relevant neurological, psychiatric, cardiovascular, pulmonary, endocrinological, or neoplastic diseases. Dyed and bleached hair and samples with a hair length lower than 3 cm were excluded from the study. Participants provided their informed consent before participating in the study. They were then asked to anonymously complete a general questionnaire, which included information on gender, height, weight, degree course, year of study, and their tobacco and coffee consumption habits. Ethanol consumption data was collected through self-reports (see Supplementary Information), where participants indicated the type of beverages typically consumed (such as wine, beer, plain spirits, cocktails, and aperitifs) and the frequency of consumption.

Methanol, acetonitrile, and formic acid (all LC–MS grade) were purchased from Merck (Darmstadt, Germany), while ammonium acetate was obtained from Sigma-Aldrich (Saint Louis, MO). Strata-X-A solid-phase extraction (SPE) cartridges (60 mg, 3 ml) were acquired from Phenomenex (California, USA). Water was purified by PURELAB Chorus ELGA Veolia (High Wycombe, UK).

The reference materials, ethyl-β-D-glucuronide and ethyl-β-D-glucuronide-D5 (IS), were purchased from Sigma-Aldrich (Saint Louis, MO). Working solutions were prepared at a concentration of 100 ng/ml in methanol and stored at – 20 °C. Mobile phases consisted of 20 mM ammonium acetate at pH 6 in water (mobile phase A) and acetonitrile (mobile phase B). Flow rate was set at 0.3 ml/min.

Analyses were performed by LC–MS/MS (Xevo TQD, Waters, Milford, USA) via an electrospray ion source (ESI) operating in negative ion mode. A zwitterionic HILIC LC column (Poroshell 120, 2.1 × 100 mm, 2.7 μm), maintained at 30 °C was used for the separation. Gradient elution was as follows: 90% B for 1 min, 80% B from 1 to 7 min, held for 3 min, and equilibration for 5 min. Injection volume was 8 µl. The analytes of interest were detected in multiple reaction monitoring (MRM) mode, monitoring the following transitions: m/z 221 → 75, 221 → 85, and 221 → 113 and ETG-d5 (m/z 226 → 75 and 226 → 85). Cone voltage was 30 V for all, and collision energy was 18 V except for m/z 221 → 113 (collision energy: 12 V). Autosampler was kept at 10 °C. Data analysis was performed by MassLynx software (Waters, Milford, USA).

Subjects were invited to participate in the study after a brief presentation of the project’s objectives and methodology, which was given prior to the attended lesson. Hair samples were obtained by cutting them as close to the scalp as possible from the vertex region of the head. These samples were then stored in paper envelopes at room temperature until analysis, for a maximum period of 2 months. The analysis focused on the proximal 3 cm of each hair sample. For calibration curves and validation, human hair samples were collected from children aged 3 and 4 years and prepared by creating a homogenized pool of hair. Samples and calibrators/controls were processed by adapting the procedure described in ref. [15]. A lock of hair was washed with 10 ml dichloromethane and 10 ml of methanol for 10 min each. Samples were left to dry at room temperature overnight and then cut into small pieces (1–2 mm) with scissors. For the extraction, 100 mg of hair was weighted, spiked with 30 µl of IS, and soaked in 1 ml of deionized water. Calibrators and QC were added to a proper amount of EtG. Samples were incubated overnight at room temperature, followed by ultrasound extraction for 2 h, at 50 °C. After centrifugation, 1 ml of the supernatant was submitted to solid-phase extraction (SPE). Cartridges were conditioned with 2 ml of methanol and 2 ml of deionized water. After sample loading, cartridges were rinsed with 1 ml of 5% NH4OH and 2 ml methanol. To remove all residual liquid, a strong vacuum was applied for 15 min. Elution was performed by 2 ml of 2% formic acid in methanol. The eluate was evaporated to dryness under a stream of nitrogen at 50 °C and then reconstituted in 150 μl of mobile phases A/B, (10: 90 v/v).

Selectivity, linearity, sensitivity, precision, accuracy, matrix effect, and stability were considered for validation. Selectivity was assessed by analyzing ten blank samples without IS and three blank samples with IS to check for interfering signals. Linearity was evaluated in the 5–45 pg/mg range of concentrations (5, 7, 10, 20, 30, 45 pg/mg). Four calibration batches were analyzed on 4 non-consecutive days. Accuracy, precision, matrix effect, and stability were calculated on quality controls (QC) prepared at three different concentrations (5, 15, and 25 pg/mg). QC were analyzed in two replicates for each concentration per day (intra-day precision) and on 4 non-consecutive days (inter-day precision). Precision was calculated as relative standard deviation (RSD) both intra-day and inter-day. Accuracy was calculated as percentage bias. The extraction procedure was tested on samples from proficiency tests (n.6) with authentic hair material. Limit of detection (LOD) and limit of quantification (LOQ) were determined by a signal-to-noise ratio of 3 and 10, respectively, and experimentally verified by spiking the calculated amount. Matrix effect was identified by comparing the ratio of peak areas of samples pre- and post-extraction and expressed as a percent. EtG stability was assessed in processed samples after 24 h at 10 °C by calculating the percent deviation on freshly prepared samples.

We gathered data on age, biological sex, weight, height, degree program, academic year, smoking habits (cigarettes per day), daily coffee consumption, and alcohol-related behaviors through questionnaires. The degree programs were re-categorized into bachelor’s and master’s degrees. The data analysis involved classifying the collected information into predefined intervals. BMI was classified as BMI < 25 (normal) or BMI > 25 (overweight); smoking habits were classified as non-smokers, smokers of 1–9 cigarettes/day, and smokers over 10 cigarettes/day. Coffee drinking was classified into no consumption, up to 5 coffees/day (up to 400 mg caffeine) [16], and over 6 coffees/day. The self-reported frequency of ethanol consumption was pre-categorized into four groups based on the number of drinking occasions per week: no alcohol consumption, alcohol consumption on 1–2 days/week, alcohol consumption on 3–5 days/week, and alcohol consumption on all 7 days of the week. Hair EtG concentrations were classified according to the guidelines outlined in the Society of Hair Testing Consensus document. In addition, data were also grouped for negative samples (hEtG < LOD) and LOD < hEtG ≤ LOQ (5 pg/mg). No toxicological data exceeded the upper limit of 30 pg/mg. Congruence between self-reporting of alcohol consumption and hEtG levels was evaluated.

Normality and non-parametric statistics were assessed by Sktest (p > 0.05). Descriptive statistics was provided for all data, by the mean and standard deviation (SD) and/or by the median and interquartile range (IQ), for non-parametric variables. hEtG levels were compared among males and females and bachelor’s/master’s degrees by means of a non-parametric t-test (Mann–Whitney test). A similar comparison was performed on the basis of the reported frequency of ethanol consumption, by non-parametric ANOVA (Kruskal–Wallis test).

Statistical associations were attempted between hEtG data, divided into the 3 predefined concentration intervals (hEtG negative, hEtG ≤ 5 pg/mg, 5 < hEtG ≤ 30 pg/mg) and the following parameters, by means of chi-square analysis: biological sex, BMI category, degree course, year of course, smoking habits, coffee consumption, and frequency of ethanol consumption.

To evaluate participants’ awareness of their consumption patterns, we conducted a Spearman correlation analysis between hEtG levels and the self-reported frequency of ethanol consumption among questionnaire respondents. Furthermore, we calculated the congruence between the reported frequency of ethanol consumption and the interpretation of hEtG levels based on the Society of Hair Testing (SoHT) cut-off values.

For all analyses, a p < 0.05 was set for significance. Statistics was performed by Stata (StataCorp LP, version 14.0, Texas, USA) and images by Prism (GraphPad Software, LLC, version 9.3.0).