The clinical study has been described in detail earlier (Landmesser et al. 2019). Briefly, 25 healthy adult males belonging to three groups participated in the study: 5 regular smokers (smoking non-filter cigarettes spiked with 13.4 mg 13C3-PG, 13.6 mg 13C3-G, and 2.4 mg D7-nicotine dissolved in 100 µL ethanol per cigarette), 10 regular vapers (vaping-labeled e-liquid at 10 W) and 10 regular vapers (vaping-labeled e-liquid at 18 W). In the applied e-liquid containing PG and G 50/50% (m/m) and 1.2% nicotine (m/m), 10% of each PG, G and nicotine was replaced by 13C3-PG, 13C3-G and D7-nicotine, respectively. The e-liquid was tested with respect to the presence of the (un)labeled carbonyls formaldehyde, acetaldehyde, acrolein, acetone, crotonaldehyde, methacrolein, and propionaldehyde. No carbonyls were detectable in the e-liquid. The labeled compounds were purchased from AptoChem, Montreal, Canada. All three substances were characterized for their identity, chromatographic purity, water content, and residual solvents. 13C3-PG, 13C3-G, and D7-nicotine showed a purity of 99.2%, 100%, and 99.7%, respectively. The subjects stayed in the clinic for 84 h (evening of day-1 until morning of Day 4). Vaping or smoking took place only on day 1 during 10 sessions, each comprising the consumption of 10 controlled puffs (two puffs per minute, 4 s puff duration) by the 20 vapers or one cigarette by the 5 smokers. Before and after each sampling in the clinical study, the tanks were weighed to determine the amount of e-liquid consumed per session. The amount of e-liquid consumed was used to normalize the results. All urine voids were collected from the morning of Day 1 prior to the first vaping/smoking session until the morning of day 4.
Chemicals, standards and stock solution
Acetaldehyde-dinitrophenylhydrazone (DNPH) (99.9% purity), acetone-DNPH (99.7%) acrolein-DNPH (99.8%), formaldehyde-DNPH (99.9%), crotonaldehyde-DNPH (99.6%), methacrolein-DNPH (96.7%) and propionaldehyde-DNPH (98.3%) were purchased from Neochema (Bodenheim, Germany). 3,5,6-D3-Acetaldehyde-DNPH (99%), 3,5,6-D3-acetone-DNPH (99%), 3,5,6-D3-acrolein-DNPH (99%), 3,5,6-D3-formaldehyde-DNPH (98.8%), 3,5,6-D3-crotonaldehyde-DNPH (99%) and 3,5,6-D3-propionaldehyde-DNPH (98.6%) were obtained from CDN Isotopes Inc. (Quebec, Canada). 2,4-Dinitrophenylhydrazine for HPLC derivatization (> 99%), perchloric acid (70%) and pyridine (anhydrous, 99.8%) were purchased from Sigma (Taufkirchen, Germany). Acetonitrile (ULC/MS grade) was obtained from Biosolve BV (Valkenswaad, Netherlands). Ultrapure water was prepared using the arium® pro ultrapure water system (Sartorius, Göttingen, Germany).
Analysis of carbonyls in aerosol and smoke
The carbonyls (formaldehyde, acetaldehyde, acrolein, acetone, crotonaldehyde, methacrolein, propionaldehyde) are highly volatile and reactive and therefore they were derivatized with 2,4-dinitrophenylhydrazine upon trapping, followed by dilution. 10 puffs of the EC or a single cigarette (approx. 7–8 puffs) were drawn through two glass impingers in sequence each containing 20 mL of an acidic dinitrophenylhydrazine derivatization solution (2.5 mM in acetonitrile) according to Miller et al. (2010). The puffing regime was set according to CORESTA recommended method CRM no. 81 (CORESTA 2015) with minor modifications: puff duration: 4 s (instead of 3 s according to CRM no. 81), puff interval: 30 s, puff volume: 55 mL. Immediately after the trapping procedure, 200 μL pyridine was added to stop the derivatization reaction. The trapping solution was diluted 1:10 (1:100 for the cigarette) with acetonitrile. Prior to the UPLC-MS/MS analysis, 10 μL of the internal standard mix of the carbonyl-DNPHs was added to 100 μL of the diluted sample, 5 μL of which was injected for analysis.
Liquid chromatography was performed with a Shimadzu Nexera X2 UPLC system consisting of a binary pump, an auto-sampler, a degaser, and a column oven (Shimadzu Corp., Kyoto, Japan). A triple quadrupole mass spectrometer QTRAP® 6500 + equipped with a Turbo V ion spray source, operating in negative ESI mode, was used for detection (AB Sciex, Darmstadt,Germany). High purity nitrogen was produced by a nitrogen generator NGM 22-LC/MS (cmc Instruments, Eschborn, Germany). Chromatographic separation was achieved on a Kinetex® 5 μm EVO C18 column (150 × 2.1 mm, 5 μm, Phenomenex, Aschaffenburg, Germany) with water (eluent A) and acetonitrile (eluent B) applying the following gradient: 0–5.0 min: 35–55% B; 5.0–7.0 min: 70% B; 7.0–7.1 min: 70–35% B; 7.1 – 10.0 min: 35% B. The column was kept at 50 °C with a flow rate of 0.7 mL/min. Labeled and unlabeled analytes as well as their corresponding internal standards were monitored in the multiple reaction monitoring mode (MRM, Supplementary Information Table S1).
All determinations of the carbonyls were repeated eight times for EC aerosol at low (10 W) and high (18 W) wattage and ten times for cigarette mainstream smoke, respectively. The presented data were corrected for isotope overlap according to Scherer et al. (2010). The lower limit of quantification (LLOQ) was 0.2 ng/puff for EC aerosol and 0.25 ng/puff for cigarette mainstream smoke. The upper limit of quantification (ULOQ) was 100 ng/puff for EC aerosol and 125 ng/puff for cigarette mainstream smoke.
Analysis of epoxides in aerosol and mainstream smoke
Various trapping agents and chromatographic conditions were tested for the determination of the epoxides propylene oxide (PO) and glycidol (GLY) in EC aerosol and mainstream smoke. The analytical method showing the best performance with regard to the sensitivity and recovery is presented in the Supplementary Information. However, the sensitivity of the final method with an LLOQ of 0.05 µg/mL was still not sufficient for the quantification of PO and GLY in EC aerosols or smoke in our study.
Analysis of the biomarkers for formaldehyde and acetaldehyde
The specific biomarkers TCA and TCG for FA as well as MTCA and MTCG for AA were determined according to the fully validated method published by Landmesser et al. (2020). Briefly, the biomarkers were cleaned up from major matrix components by solid-phase extraction followed by derivatization under alkaline conditions using propyl chloroformate. The LC–MS/MS analysis was performed using a Shimadzu Nexera X2 UPLC system consisting of a binary pump, an autosampler, a degaser and a column oven (Shimadzu Corp., Kyoto, Japan) combined with a triple quadrupole mass spectrometer QTRAP® 6500 + equipped with a Turbo V ion spray source (AB Sciex, Darmstadt, Germany), operated in positive ESI (ESI+) mode. The presented data are corrected for isotope overlap according to Scherer et al. (2010). LLOQs and ULOQs were 0.5 ng/mL and 200 ng/mL for (M)TCA and 1.0 ng/mL and 400 ng/mL for (M)TCG in accordance with Landmesser et al. (2020).
Analysis of mercapturic acids
The following mercapturic acids were determined in 24 h urine samples according to Pluym et al. (2015) with modifications: 3-HPMA (biomarker for acrolein), HMPMA (crotonaldehyde), 2-HPMA (propylene oxide), and DHPMA (glycidol). DHPMA, which was not implemented in the initial method (Pluym et al. 2015) was included for the purpose of our study into the analytical method for 2-/3-HPMA. LLOQs (ng/mL) of the newly integrated analytes (DHPMA and all labeled mercapturic acids) were as follows: [13C4]-HMPMA: 5.0, [13C3]-2-HPMA: 0.5, [13C3]-3-HPMA: 0.5, DHPMA: 10; [13C3]-DHPMA: 0.8. ULOQs (ng/mL) of the newly integrated analytes were as follows: [13C4]-HMPMA: 2,500, [13C3]-2-HPMA: 2,000, [13C3]-3-HPMA: 10,000, DHPMA: 2,000; [13C3]-DHPMA: 2,000. The applied mass transitions for the labeled analytes were selected according to the published fragmentation pathways of mercapturic acids (summarized in Supplementary Information Table S2).
Biomarker analysis was performed separately for urine fractions collected over 48 h, beginning with the first fraction voided after start of the first vaping/smoking session on Day 1. The presented data are corrected for isotope overlap according to Scherer et al. (2010).