The commercial ENDS “Endura T20-S” (Innokin) was chosen as the portable dosing device for participants’ exposure to DEHP. ENDS consists of a battery-powered heated coil that functions as a thermal atomizer, a mouthpiece, and an e-liquid tank. Generally, the e-liquid is mainly composed of nicotine, flavoring agents, and ethanol in a base of propylene glycol (PG) and vegetable glycerol (VG). Base composition can vary from a PG/VG ratio of 20/80 to 70/30 and is typically 50/50. When the user activates the heated coil by pressing a button on the ENDS, the e-liquid is vaporized, releasing a flavored nicotine-containing aerosol called a “puff.” The user inhales this aerosol through the mouthpiece.

Aerosols generated by ENDS are complex mixtures containing volatile and semivolatile gases (mainly PG and glycerol) and nonvolatile particles (mainly glycerol, PG, and nicotine) [44,45]. The e-liquid formulated in this study is composed of glycerol (purity 99.5%, Hänseler Swiss Pharma), PG (PG 0 mg/mL nicotine, Revolute), ethanol (absolute for analysis EMSURE, Supelco, Merck), and DEHP. Ring-deuterated DEHP (DEHP-d4, CAS# 93951-87-2, Sigma Aldrich) is used to differentiate exposures generated in our experiment to environmental and laboratory DEHP contaminations.

To adapt the ENDS to generate DEHP aerosols at controlled concentrations, we used a vaping machine that was developed in the Unisanté laboratory in collaboration with the University Centre of Legal Medicine of Lausanne (CURML) [46]. The purpose of the vaping machine was to replicate the puffing behavior of an individual using an ENDS according to the method n°81 recommended by the Cooperation Centre for Scientific Research Relative to Tobacco (CORESTA) [47]. This method describes the requirements for the generation and collection of e-cigarette aerosols for analytical testing purposes and led to an ISO (International Organization for Standardization) standard (20768:2018). Our vaping machine included 3 ports, each for an ENDS head (e-liquid tank, atomizer, and mouthpiece), a power supply, a piston syringe, a pinch-valve system, and silicon tubing (Figure 2). The machine was equipped with its own power supply, which replaced the individual ENDS batteries. This provided a steady-state power of 12 W throughout all experiments. The replacement of the ENDS battery by the machine’s power supply precisely controlled puff duration during the experiments. The piston, in combination with the pinch valve system, controlled the puff volume, that is, the volume leaving the ENDS. Puff volumes were trapped by n-hexane in subsequent glass impingers (with P1 porosity fritted nozzle tip, joint connection NS 29/32, Bistabil, Sigma Aldrich) for analytical analysis. Puff duration (3 seconds per puff), volume (55 mL per puff), and frequency (30 second between puffs) were set up as per CORESTA criteria. Tests were carried out in triplicates. Control tests served to verify that the entire setup was not contaminated by DEHP-d4. A total of 3 ENDS with 0.8-ohm atomizers were used for all the experiments. New atomizers were used in each experiment.

A first series of tests was carried out on the vaping machine to optimize the e-liquid formulation for generating stable aerosol concentrations as well as trapping conditions. We prepared several formulations with different PG/VG ratios (from 0/100 to 50/50), different ethanol amounts (0%-10%), and different DEHP amounts (0.002%-0.3%, w/w) to determine e-liquid formulations that allowed DEHP to be dissolved in the e-liquid and aerosolized by the atomizer. For each ENDS, DEHP-d4 was quantified in the n-hexane of all impingers. The number of impingers (2 and 3), impinger volume (20 and 100 mL), and n-hexane volume in each impinger (13, 50, 75, and 80 mL) were tested for optimal DEHP-d4 trapping conditions.

We carried out a second series of experiments with the vaping machine to determine mass balance and confirm DEHP-d4 concentration in the aerosols. DEHP-d4 was quantified in the n-hexane of all impingers, in the e-liquid before and after the experiment, and in an acetone/ethanol (50/50 v/v) mixture (acetone, suitable for high-pressure liquid chromatography [HPLC] ≥99.9%, Sigma Aldrich) that was used to extract DEHP-d4 from disassembled ENDS parts at the end of the experiment by the Central Environmental Laboratory of the Ecole Polytechnique Fédérale of Lausanne (EPFL).

The stability of the DEHP-d4 concentration in e-liquid was an important control step for the participants’ safety and to avoid any exposure concentration variability between the participants. The stability test for DEHP-d4 in the e-liquid was performed on the e-liquid stored at room temperature for 1 week and frozen at –20 °C for 2 weeks.

A portable optical particle counter (aerosol spectrometer, model 1.109, GRIMM) was used to determine the mass distribution of e-liquid particles in the aerosols and assess deposition in the respiratory system [48,49]. Aerosols were generated by equipping the vaping machine with ENDS. The e-liquid used for these tests was composed of a PG/VG ratio of 50/50 with 1% ethanol and was not spiked with DEHP-d4. A total of 6 puffs of aerosol (55 mL×6=330 mL) were diluted in ambient air in a 3 L sampling bag (Tedlar sampling bag, Nutech Instruments Inc). The dilution in ambient air simulated the aerosol dilution in inhaled air by an individual puffing, assuming a 55 mL puff and a 500 mL tidal volume (330 mL/3000 mL=55 mL/500 mL). The particle size distribution profile was then measured. A control test was performed with ambient air.

The concentration of DEHP-d4 in the e-liquid (percentage by weight) was set at a value that produced DEHP-d4 concentrations for participant exposures (C_exp, mg/m³) at or below the Swiss occupational exposure limit (OEL) for DEHP (2 mg/m³) [50]. The OEL is set at concentrations that, based on available scientific data, should not produce health effects in workers exposed to a substance in the air every day for a workweek throughout their working lifetime (in Switzerland: 8 hours per day, 5 days a week, for 40 years). C_exp was calculated based on the following equation:

where M_puff was the mass of DEHP-d4 in the aerosols generated in a puff by ENDS coupled with the vaping machine (µg per puff), p was the number of puffs per minute that an ENDS user would consume per minute (puff per minute), V_tidal was the human mean tidal volume (0.5 L of air per inhalation [51,52]), and IE was the number of inhalations/exhalations per minute.

A maximum of 10 participants (N=10) will be recruited for repeated exposure to DEHP. Participants who are interested in the study will be checked for eligibility through a short general health questionnaire. These questions are about smoking status, age, BMI, alcohol consumption, general health status, and current job (to exclude the possibility of exposure to phthalates at work). The eligibility criteria are described in Textbox 1.

The flowchart in Figure 3 shows the steps from participant screening and visits through exposure periods until data analysis. Briefly, an initial contact with the participant will be established by email to explain the study and obtain informed consent. Then, a phone call will be made to check if the inclusion and exclusion criteria are met and to get answers for a short general health questionnaire. If the inclusion criteria are met, the participant will be invited for a first visit. At this visit, the participant will sign the written informed consent. Women of reproductive age will also undergo a pregnancy test. Then, the participant will be given an aerosol-forming device and, after being instructed on how to use the device, will undergo the first inhalation exposure under the supervision of a research team member (see “Participant Exposure” section). A blood and urine sample will be collected before the first exposure (baseline) and will serve as baseline concentrations (control). Blood will be drawn again 30 minutes after the first exposure. Then, participants will follow the expected vaping schedule with DEHP-d4-spiked e-liquid and collect spot urine samples 4 times throughout the day (morning, midday, evening, and night) for the 4 days of exposure. Blood and urine collecting times are shown in Multimedia Appendix 1.

The Laboratory of Clinical, Forensic, and Environmental Toxicology, University Hospital of Liege, will determine DEHP-d4 and its metabolites concentrations in blood. The quantified metabolites will be deuterated mono(2-ethylhexyl)phthalate (MEHP-d4), deuterated mono(2-ethyl-5-hydroxyhexyl)phthalate (5OH-MEHP-d4), and deuterated mono(2-ethyl-5-oxohexyl)phthalate (5oxo-MEHP-d4). The Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr-Universität Bochum (IPA), will quantify DEHP-d4 metabolites in urine, as described in Koch et al [53]. The metabolites quantified by the IPA laboratory will be MEHP, 5OH-MEHP, 5oxo-MEHP, and mono(2-ethyl-5-carboxypentyl)phthalate (5cx-MEPP).

Participants will use the portable aerosol-forming device themselves, first in our facilities (day 1) and then at home (days 1-4). The e-liquid will have a PG/VG ratio of 50/50 with 1% (w/w) ethanol and 0.02% (w/w) DEHP-d4. The participants enrolled in the study will press a button on the ENDS battery. The coil will heat the e-liquid, generating an aerosol with DEHP concentrations less than or equal to the OEL.

The e-liquid will be prepared in advance for DEHP-d4 quantification analysis and will be stored at room temperature during the exposure week. A fraction of the e-liquid will be sent to the analytical laboratory of EPFL to confirm the DEHP-d4 concentration in the e-liquid, and the rest will be frozen at –20 °C until use. On the day of the first exposure to DEHP, the e-liquid will be defrosted. A vial with a weighted amount of e-liquid sufficient for 1-week exposure (ie, reserve) will be prepared for each participant. This vial will be stored at room temperature at the participant’s place, where they will use it to refill the ENDS e-liquid tank when needed. Therefore, a stability test is required to verify DEHP-d4 concentration after several days of storage at room temperature. This step is also important for the relevance of the study. On day 5, the participant will return the ENDS and the e-liquid vial, which will be weighed to calculate the exact amount of e-liquid puffed by the participant, hence the amount of puffed DEHP-d4.

Standard descriptive statistics (ie, scatter plots, box plots, means, and SD) will be performed to describe the within-participant evolution of parent compound and metabolite distribution and elimination over time. Absorption and elimination rates will be estimated for each of the participants using standard nonlinear regression methods (hockey plots, piecewise linear regression models, or exponential models). These rates will be compared while accounting for potential effect modifiers (eg, age and gender) using mixed models.

The optimized e-liquid formulation had a PG/VG ratio of 50/50 with 1% (w/w) ethanol and 0.02% (w/w) DEHP-d4. Ethanol was necessary to dissolve DEHP-d4 and then create a homogeneous solution with PG/VG. The mean recovery in impingers was 89% (SD 5%) of the mass of aerosolized DEHP-d4 (n=5). The optimized number of impingers, impinger volume, and n-hexane volume in each impinger were 3 mL, 125 mL, and 80 mL, respectively. The number of puffs per experiment (n=36) was set to produce 2 dm³ of aerosol, a volume that generated quantifiable amounts of DEHP-d4, and that was a submultiple of 1 m³ (the volume used to define the OEL), which simplified calculations.

Total DEHP-d4 mass mean recovery for e-liquids, ENDS heads, and impingers was 100% (SD 22%; n=4). A DEHP-d4 concentration of 0.02% (w/w) in the e-liquid was confirmed by chemical analysis and produced aerosols containing 25.4 (SD 3.4) mg/m³ DEHP-d4 (median 26 mg/m³), corresponding to a mean of 1.4 (SD 0.2) μg DEHP-d4 per puff (n=6).

The results of the stability test (Figure 4) showed a decrease in DEHP concentration of about 10% when the e-liquid was stored at room temperature. This loss may be explained by slight adsorption of the DEHP to the recipient wall or analytical error (up to 10% is considered in the error range of the analytical method). Therefore, the concentration of DEHP was stable when the e-liquid was stored in the freezer at –20 °C.

The aerosol particle size distributions are shown in Figure 5. Table 1 lists the particle means expressed in terms of environmental (PM10, PM2.5, and PM1) and occupational (inhalable, thoracic, and respirable) mean mass concentrations. Most of the particles in the diluted ENDS-generated aerosols were at the respirable mass fraction level, with a diameter of 2.5 μm or smaller.

We observed that on average, an individual puffs 4 times per minute and inhales and exhales 12 times per minute (data not shown). Therefore, a 0.02 % (w/w) DEHP-d4 concentration in the e-liquid generating an aerosol with 1.4 µg per puff of DEHP-d4 led to an estimated exposure concentration C_exp of 0.9 mg/m³ over a 10-minute exposure, which was below the OEL. Assuming that a participant puffed 40 times on average over the 10-minute exposure duration twice a day (80 puffs per day), the predicted DEHP-d4 dose for a week of exposure was 0.45 mg.

Our first participant could not complete the exposure week because some e-liquid leaked out of the ENDS, preventing our control of the puffed dose. Therefore, results for this participant are shown in Figure 6 as proof of concept. Our results from our first participant showed quantifiable amounts of DEHP-d4 in this participant’s urine. DEHP-d4 was quantified in the e-liquid by chemical analysis before participant exposure (0.019%, w/w). The participant was asked to puff 40 times over 10 minutes twice a day. After five 10-minute exposure periods over 3 days and minor leaking, 2.52 g of e-liquid were used, corresponding to 0.47 mg DEHP-d4. This amount was in line with the DEHP-d4 predicted dose and led to exposure concentrations lower than the OEL but still sufficiently high to quantify DEHP in urine and to elucidate the urinary elimination kinetics. However, C_exp led to concentrations below our limit of quantification (LOQ) for blood.