Determination of Benzoic Acid and Benzyl Alcohol in E-Liquids (JUULTM Pods) by Isotopic Dilution High-Performance Liquid Chromatography and Tandem Mass Spectrometry
Pubblicato online: 02 gen 2022
Pagine: 212 - 220
Ricevuto: 30 mar 2021
Accettato: 14 ott 2021
DOI: https://doi.org/10.2478/cttr-2021-0016
© 2021 Roberto Bravo Cardenas et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Although modern electronic cigarettes (e-cigarettes) were introduced over a decade ago (2006 in Europe and 2007 in the United States), the concept even then was not new (1, 2, 3). An earlier attempt in 1965 to develop e-cigarettes in the United States was not profitable despite being advertised as a safe and harmless way to smoke or quit smoking (4). The first product to show some success on the market was developed in China in 2003 by HON LIK (5) and marketed with the help of the China Inventors Program. HON's e-cigarette gained some favor among Chinese smokers as a potential cessation device or an alternative cigarette product.
The appearance of e-cigarettes has varied widely, and include those that look like regular cigarettes, to pens, personal vaporizers, tank systems, and computer memory sticks. The devices share similar design components, including a battery that powers an atomizer or heating element, a reservoir to hold a liquid which generally contains nicotine and flavoring agents solubilized in propylene glycol (PG) and/or glycerin (Gly), and a mouthpiece through which the user inhales the aerosol of the electronic juice. The two popular formats of e-cigarettes are open systems and closed (“pod”) systems. Open systems allow the consumer to adjust the temperature and use different e-liquids individually. Closed pod-based systems such as JUULTM, Vuse AltoTM, and mybluTM do not allow the user to make changes to the device and/or e-liquid. However, the distinction between open and closed systems for pod-based devices has become muddled as third parties offer refillable replacement carts (6).
Pod e-cigarettes have become extremely popular, especially the JUULTM brand because they are easy and discreet to use, often resembling a computer memory stick, and generate less exhaled vapor (7, 8, 9, 10). These products are also available in multiple flavors that go beyond the standard tobacco and menthol offerings. Pod e-cigarette companies target their advertising to groups such as youth and military personnel (10, 11, 12, 13, 14), successfully using social media platforms such as YouTube, Facebook, and Instagram (15). According to the National Youth Tobacco Survey about 19.6% of high schoolers and 4.7% of middle schoolers reported e-cigarette use in 2020. Much higher nicotine concentrations can be made available in pod e-cigarettes due to the use of e-juices. The e-juices contain high concentrations of nicotine salts, which allow greater amounts of nicotine to be consumed per puff of aerosol produced. Traditional e-liquids, by contrast, are based on free nicotine, which is volatile and produces higher absorption rates in the body. Free nicotine, like the smoke of cigars that also have high free-nicotine levels, is associated with harshness that makes inhalation difficult due to irritation (16) and limits the nicotine concentration. The nicotine salt e-liquids used in pod e-cigarettes such as JUULTM are prepared using weak acids, such as benzoic acid, which serves to reduce the pH of the e-liquid, allowing the administration of higher doses of protonated nicotine with reduced throat irritation. Some publications suggest that nicotine salt e-liquids can produce satisfaction levels similar to those of traditional cigarettes to quench nicotine desire (14, 17, 18, 19, 20).
Although there is a global consensus that the use of e-cigarettes lowers exposure to many harmful compounds compared to combustible tobacco cigarettes (21), the use of e-cigarettes remains controversial due to the high concentrations of their e-liquid components. Gly, PG, flavoring additives and nicotine salts are present in high concentrations in some of these e-liquids and could contribute to health problems. High concentrations of nicotine, for example, have been reported to adversely affect brain development (22, 23, 24) and could exacerbate cardiovascular issues and atherogenesis via prolonged nicotine exposure especially in individuals with underlying disease such as chronic kidney disease, hypertension, cardiovascular diseases, etc. (25, 26). Gly, PG, and flavorings can lead to the formation of highly carcinogenic compounds such as carbonyl groups and benzene during the vaping process when the coil temperature is sufficiently high. Additionally, the use of benzoic acid in nicotine salt liquids, such as those used in JUULTM devices, can increase benzene formation via decarboxylation of benzoic acid, which can occur during vaping as previously postulated by P
All solvents used were analytical and/or HPLC grade with purity greater than 99.5%. Methanol Plus (MeOH), benzoic acid, and benzyl alcohol certified reference materials (CRM) were purchased from Sigma-Aldrich (Milwaukee, MO, USA). Formic acid 0.1% in water and acetonitrile (Optima) were obtained from Fisher (Fair Lawn, NJ, USA). Benzoic acid (992 μg/mL) and benzyl alcohol (998 μg/mL) reference materials solutions were acquired from SPEX CertiPrep (Metuchen, NJ, USA). Benzoic acid (Ring-13C6, 99%) and benzyl alcohol (Ring-13C6, ≥ 98%) were procured from Cambridge Isotope Laboratories, Inc. (Andover, MA, USA).
JUULTM cartridges (5% nicotine, pods) (Pax Inc., San Francisco, CA, USA) were acquired in five different flavors (Fruit Medley, Mango, Cool Mint, Virginia Tobacco, and Creme Brulee) along with two U.S. e-liquid products through The Lab Depot, Inc. (Dawsonville, GA, USA) in June 2019. The U.S. e-liquid products selected are not necessarily representative of the U.S. market; they constitute a “convenience” sample, purchased solely to test validity of the method using commercial e-liquids. Samples were labeled and stored at room temperature until they were analyzed. JUULTM refill cartridges were uncapped and used only one time for each analysis.
Separate isotopically labeled internal stock solutions were prepared by dissolving 1000.00 mg of benzoic acid-13C6 in 1 mL of methanol and 100.00 mg of benzyl alcohol-13C6 in 5 mL of methanol. Aliquots of 800 μL of both labeled internal standard stock solutions were mixed into a 10-mL volumetric flask and diluted with methanol. The resulting internal standard solution (ISTD) of 80 mg/mL benzoic acid-13C6 and 1.6 mg/mL benzyl alcohol-13C6 was divided into 2-mL aliquots and stored at −20 ºC until used. An aliquot of ISTD was added to the dilution solution which is needed to dilute each sample (2.5 mL dilution solution containing ISTD/sample) to yield a final concentration of 0.63 μg/μL of benzoic acid-13C6 and 12,64 ng/μL of 13C6-benzyl alcohol.
Standards were prepared from CRMs and covered the full range of the analytes concentration observed during preliminary investigation of the e-liquids, specifically the JUULTM samples. A native standard stock solution of benzoic acid was prepared by weighing 2.5 g into a 25-mL volumetric flask and dissolving it with methanol to yield a stock solution with a concentration of 100 μg/μL. A native stock solution of benzyl alcohol was prepared by weighing 25.00 mg into a 25-mL volumetric flask and dissolving it with methanol to generate a stock solution with a concentration of 1.00 μg/μL. Aliquots of the two analytes were diluted with MeOH to generate nine primary calibration standards with concentrations ranging from 0.45 to 1.82 μg/μL for benzoic acid and from 0.50 to 27.20 ng/μL for benzyl alcohol.
A 250-mL non-commercial matrix solution was prepared by mixing Gly and PG (1:1, v/v) and dividing it into three pools. The first pool (low concentration quality control, QCL) was prepared by spiking the corresponding amount of benzyl alcohol and benzoic acid (CRMs) to yield concentrations of 1.3 ng/μL and 0.50 μg/μL, respectively. The second pool (high concentration quality control, QCH) was spiked with a corresponding amount of benzyl alcohol and benzoic acid (CRMs) to yield concentrations of 13.1 ng/μL and 1.20 μg/μL, respectively. The third pool was not spiked. After being examined for the presence of possible endogenous and/or cross contamination of the two analytes, the third pool was used as matrix material for blank samples, accuracy, and matrix-effect validation steps. QCL and QCH pools were characterized to determine the mean, 95th, and 99th control limits by consecutive analysis of at least 2 samples/day from each QC pool for 20 days. After establishing the control limits of the pools, QC high and low samples were analyzed with each analytical run and evaluated for validity using a modification of the Westgard rules (29). These values were used to certify the performance and long-term analytical stability of the method. A solvent, blank, and QCs samples were analyzed to monitor background levels and guard against contamination from sample carryover.
All e-liquid samples, QC materials, reagents, and calibration standards were brought to room temperature. To prepare the QCs and the unknown samples, a 100 μL aliquot was weighed on an analytical balance with a precision of ± 0.00001 g into an amber 16-mL vial.
Samples were diluted with 2.5 mL of MeOH/H2O (1:1) containing the internal standards and agitated for 60 min at 160 rpm on a Barnstead Lab-line E-class orbital shaker (Dubuque, IA, USA).
The diluted extract was analyzed using a Dionex Ultimate 3000 HPLC coupled to a Thermo Scientific TSQ Endura mass spectrometer® (ThermoScientific, San Jose, CA, USA). The HPLC separation was carried out on a Luna® PFP (2.0 mm × 150 mm, 3.0 μm particle size) column (Phenomenex, Torrance, CA, USA). The column temperature was kept at 45 °C and the injection volume was 10 μL. The mobile phase consisted of (A) 0.1% formic acetic acid in water and (B) methanol:acetonitrile (1:1). The HPLC gradient conditions applied are listed in Table 1.
HPLC gradient conditions for benzoic acid and benzyl alcohol analysis.
| Run time | Flow rate | Mobile phase A | Mobile phase B |
|---|---|---|---|
| (min) | (mL/min) | Composition (%) | Composition (%) |
| 0.0 | 0.225 | 80 | 20 |
| 2.0 | 0.225 | 30 | 70 |
| 11.0 | 0.225 | 30 | 70 |
| 12.0 | 0.250 | 80 | 20 |
To prevent excessive buildup of sample residue inside the mass spectrometer ion source, a diverter to waste container valve was used for the first and last segment of the separation. The HPLC eluents were ionized by heated electro-spray ionization probe (H-ESI) in positive-ion mode using ultra-high purity nitrogen as the sheath and auxiliary gases at pressures of 30 psi and 10 psi, respectively.
The vaporizer and ion transfer tube temperatures were kept at 300 and 270 ºC, respectively. Ultra-high purity argon (Airgas, Atlanta, GA, USA) was used as the collision gas at a pressure of 1 mTorr. The mass spectrometer was operated in the selected reaction monitoring mode as shown in Table 2.
Segments used in the mass spectrometer during analysis.
| Segment | Segment time (min) | Compound | Scan type | Ions monitored (m/z) | Collision energy (eV) |
|---|---|---|---|---|---|
| 1 | 0.00 − 6.00 | NA | Full Scan | NA | |
| 2 | 6.00 − 9.50 | BOH_Q | SRM | 91.0 → 65.0 | 20.0 |
| BOH_C1 | SRM | 91.0 → 91.0 | 10.0 | ||
| BOH_C2 | SRM | 91.0 → 39.1 | 34.0 | ||
| BOH_L | SRM | 97.0 → 69.0 | 20.5 | ||
| 3 | 9.55 − 13.50 | BCOOH_Q | SRM | 123.0 → 79.0 | 10.2 |
| BCOOH_C1 | SRM | 123.0 → 77.0 | 22.1 | ||
| BCOOH_C2 | SRM | 123.0 → 105.0 | 13.6 | ||
| BCOOH_L | SRM | 129.0 → 85.0 | 10.0 | ||
| 4 | 13.55 − 20.0 | NA | Full Scan | NA |
BOH: benzyl alcohol; BCOOH: benzoic acid; _Q: quantitation; _C: confirmation; _L: Labeled; NA: Does not apply, SRM: selected reaction monitoring
The main objective in our group is to develop analytical techniques to monitor and measure chemical components that impact the release of nicotine in tobacco products. In this project, we focus on the analysis of benzoic acid and benzyl alcohol, chemicals that are considered potentially human harmful that are present in new and/or popular tobacco products. JUULTM is a popular style of e-cigarettes in the United States, and at one point accounted for more than half of the market share in tracked retail channels and becoming so popular that its use has led to a new verb: “JUULing” (15). We developed and fully validated a method to ensure reliable results in detecting and quantifying benzyl alcohol and benzoic acid in e-liquids. Accuracy, dynamic range, linearity, detection limit, precision, and robustness (including stability) were assessed. In addition, matrix effects were also determined in order to confirm the use of solvent-based calibrators was appropriate.
A calibration curve was constructed for each analytical run using the response factors of ten calibrators covering the linear dynamic range (LDR) from 0.50 to 27.20 ng/μL and 0.45 to 1.82 μg/μL for benzyl alcohol and benzoic acid, respectively. The LDR was selected such that the lowest standard concentration was near the limit of detection (LOD) and the highest calibrator concentration was higher than the concentrations measured for some e-liquids in domestic products. Analysis of the calibration curves (n = 7) indicated that a linear regression with (1/×) weighting resulted in an optimal distribution of residuals, and calibration curves displayed a coefficient of determination (R2) greater than 0.995, indicating appropriate linearity for the analysis of both compounds.
We estimated LODs by evaluating the signal-to-noise (S/N) ratio for a low concentration standard of 0.50 ng/μL and 0.45 μg/μL for benzyl alcohol and benzoic acid, respectively, injected into the instrument over the course of twenty days (n = 20). The LOD was extrapolated for an S/N value of 3 based on the mean (S/N) value of the 20 measurements. Table 3 shows the LODs calculated using this approach.
Limits of detection (LOD), accuracy, and precision of the method.
| Analyte | LOD (ng/μL) | Accuracy (%) | % RSD intermediate precision | |||
|---|---|---|---|---|---|---|
| Low | Medium | High | QCL | QCH | ||
| Benzyl alcohol | 0.11 | 96.6 ± 1.9 | 95.4 ± 1.4 | 98.9 ± 2.3 | 5.3 | 4.4 |
| Benzoic acid | 9.05 | 92.2 ± 1.7 | 91.7 ± 4.1 | 93.8 ± 2.5 | 4.8 | 4.6 |
RSD: relative standard deviation; QCL: quality control samples at low concentration; QCH: quality control samples at high concentration
The absolute accuracy of the method was not determined because no reference e-liquids were available for evaluation. Accuracy was assessed by spiking a known amount of each analyte to the blank e-liquid samples at three different concentrations in triplicate (low, medium, and high) spanning the range of expected values in commercial samples. The averaged calculated percent at low, medium, and high recovery samples are reported in Table 3.
To guarantee the method's long-term analytical stability and reproducibility of the results, a blank sample and two QC samples at low and high concentrations (QCL and QCH, respectively) were analyzed with each sample set consisting of twenty-five unknown samples and ten standards. Analysis of intermediate precision yielded relative standard deviations that are reported in Table 3. The mean value and limits of each QC pool remained constant throughout this study. No carryover was observed for any blank sample.
We needed to determine the effects of the matrix due to internal CDC guidelines. The matrix effects between e-liquid calibrators and solvent-based calibrators were evaluated by comparing the slopes of two sets of calibrators prepared in Gly and PG solution (1:1,
Figure 1
Representative chromatograms of benzyl alcohol (BOH) and benzoic acid (BCOOH) quantification transitions for (A) quality control sample low concentration, (B) HAUS Cool Ice TM, (C) JUUL Cool Mint TM, (D) labeled internal standard.
We assessed sample stability using two different e-liquid samples (low and high concentration) under two different conditions: room temperature (24 ± 2) ºC and in a dark freezer at (−20 ± 2) ºC.
To avoid confounding issues resulting from simultaneous degradation of the standard and internal standard, the concentrated internal standard stock solutions were kept separate at (−20 ± 2) ºC. On days of analysis, a single vial of internal standard and vials corresponding to each sample from each environment were equilibrated to room temperature. Then, each vial was spiked with internal standard solution and vortexed. Results following 30 days under the specified conditions were determined as a percentage of the original response for the sample. After 30 days all samples exhibited less than a 10% change in apparent concentration. Samples were stored in the dark at −20 °C if deemed necessary. Samples were typically analyzed the day they were generated, but if storage was required, they were stored for no longer than 30 days.
The robustness of the method was assessed by changing five method parameters using three different conditions including the final conditions. All the analyses were performed in commercial samples at two different levels. The five method parameters evaluated were column temperature, sample extraction composition, injection volume, sample weight and extraction time. For each parameter tested, all other variables were set to their final method conditions. Column temperature, sample extraction composition, injection volume, sample weight and extraction time varied from ± 0.1% to ± 4.7% with respect to the final method conditions (Table 4), demonstrating that the technique is robust. Based on CDC internal parameters, a maximum variability in the robustness test of ± 5% with respect to the final method is acceptable for the intended purpose of the method.
Results and conditions for the determination of method robustness.
| Test parameters | Analyte | Method settings | Low level settings | High level settings | |||||
|---|---|---|---|---|---|---|---|---|---|
| Method level | Amount (μg) a | Low level | Amount (μg) a | % Diff. | High level | Amount (μg) a | % Diff. | ||
| Column temperature | Benzyl alcohol | 50 °C | 3.76 ± 0.12 | 45 °C | 3.69 ± 0.13 | 1.9 | 55 °C | 3.66 ± 0.03 | 2.7 |
| 61.65 ± 0.08 | 62.90 ± 1.34 | 2.0 | 61.54 ± 0.82 | 4.7 | |||||
| Benzoic acid | 3279 ± 50 | 3241 ± 40 | 1.2 | 3171 ± 35 | 3.3 | ||||
| 3158 ± 103 | 3038 ± 84 | 3.8 | 3053 ± 55 | 3.3 | |||||
| Sample dilution composition | Benzyl alcohol | MeOH:H2O 50:50 | 2.95 ± 0.01 | MeOH:H2O 40:60 | 3.06 ± 0.06 | 3.7 | MeOH:H2O 60:40 | 2.96 ± 0.04 | 0.3 |
| 52.12 ± 4.8 | 52.21 ± 0.59 | 0.2 | 52.42 ± 1.32 | 0.6 | |||||
| Benzoic acid | 3344 ± 72 | 3324 ± 90 | 0.6 | 3415 ± 77 | 2.1 | ||||
| 2711 ± 201 | 2831 ± 111 | 4.4 | 2813 ± 60 | 3.8 | |||||
| Injection volume | Benzyl alcohol | 10 μL | 3.00 ± 0.30 | 7 μL | 3.11 ± 0.32 | 3.7 | 13 μL | 3.01 ± 0.41 | 0.3 |
| 61.26 ± 6.95 | 61.91 ± 7.09 | 1.1 | 61.17 ± 6.76 | 0.1 | |||||
| Benzoic acid | 3332 ± 396 | 3419 ± 437 | 2.6 | 3434 ± 386 | 3.1 | ||||
| 3077 ± 27 | 3137 ± 52 | 1.9 | 3147 ± 29 | 2.3 | |||||
| Sample weight | Benzyl alcohol | 100 mg | 3.03 ± 0.11 | 80 mg | 3.09 ± 0.12 | 2.0 | 130 mg | 3.05 ± 0.20 | 0.7 |
| 53.54 ± 4.25 | 53.19 ± 3.05 | 0.7 | 54.15 ± 2.25 | 1.1 | |||||
| Benzoic acid | 3360 ± 68 | 3484 ± 58 | 3.7 | 3308 ± 78 | 1.5 | ||||
| 2806 ± 210 | 2909 ± 90 | 3.7 | 2856 ± 110 | 1.8 | |||||
| Extraction time | Benzyl alcohol | 60 min | 3.01 ± 0.07 | 42 min | 2.98 ± 0.04 | 1.0 | 78 min | 2.98 ± 0.06 | 1.0 |
| 51.64 ± 4.04 | 51.75 ± 1.49 | 0.2 | 52.12 ± 0.96 | 0.9 | |||||
| Benzoic acid | 3345 ± 65 | 3293 ± 27 | 1.6 | 3327 ± 20 | 0.5 | ||||
| 2773 ± 235 | 2714 ± 21 | 2.1 | 2721 ± 23 | 1.9 | |||||
% Diff.: Percentage of difference between the final method settings and the new settings used for the robustness study.
Average amount in sample with standard deviation. All samples were prepared in triplicates.
The suitability of the method was evaluated in the analysis of five JUULTM flavors (5% nicotine), two commercially available e-liquids, and two quality control samples (QCL and QCH) for the quantification of benzyl alcohol and benzoic acid. The two commercially available brands were randomly selected and are not representative of the U.S. market. The samples were analyzed in triplicate seven times (n = 21). Additionally, JUULTM samples were analyzed for nicotine following a previously reported method (30, 31) to establish the ratio correlation between nicotine and benzoic acid.
Table 5 presents a summary of the measured results. Two of the five JUULTM samples tested yielded detectable values for benzyl alcohol (BOH); these concentrations are low and agree with data previously reported (32). However, the concentration of BOH in the Cool Mint JUULTM flavor was the highest of all the samples analyzed in the study, including the two-regular e-liquids and the QCH samples. The concentration of BOH in Cool Mint JUULTM was 22 and 2 times higher than BOH concentration of the Mango JUULTM flavor and QCH sample, respectively. It should be noted that JUULTM Lab Corporate withdrew all flavors from the U.S. market (except for the menthol and tobacco flavors). This was in response to the January 2020 U.S. Food and Drug Administration regulation that prohibited the sale of flavors in non-disposable e-cigarettes except for menthol and tobacco flavored varieties (31). Of note, the amount of benzoic acid did not vary significantly between different flavors of the initial pods (32.8 ± 2.8 mg/g; 3.3 ± 0.3%, w/w). The average results for benzoic acid and nicotine in JUULTM are lower than results previously reported in the literature (20).
Levels of benzyl alcohol (n = 21), benzoic acid (n = 21), and nicotine (n = 7) in JUULTM liquids and two commercially available e-liquids.
| Sample type | Benzyl alcohol Total (mg/g) | Benzoic acid (BA) Total (mg/g) | Benzoic acid Total (%,W/W) | Nicotine (N) Total (mg/g) | Molar ratio | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Average | STD | CV | Average | STD | CVs | Average | Average | STD | CV | N:BA | |
| QCL | 0.0340 | 0.0009 | 2.5 | 13.03 | 0.67 | 5.2 | 1.30 | NP | — | — | NA |
| JUULTM - Fruit Medley | ND | NA | NA | 32.58 | 2.06 | 6.3 | 3.26 | 49.4 | 0.84 | 1.7 | 1.14 |
| JUULTM - Mango | 0.0283 | 0.0045 | 16.0 | 33.64 | 3.74 | 11.1 | 3.36 | 50.6 | 0.92 | 1.8 | 1.13 |
| JUULTM - Cool Mint | 0.6420 | 0.0485 | 7.6 | 33.18 | 2.84 | 8.5 | 3.32 | 49.3 | 0.75 | 1.5 | 1.12 |
| JUULTM - Virginia Tobacco | ND | NA | NA | 32.52 | 2.54 | 7.8 | 3.25 | 50.3 | 0.88 | 1.7 | 1.16 |
| JUULTM - Creme Brulee | ND | NA | NA | 32.20 | 2.62 | 8.2 | 3.22 | 50.2 | 1.08 | 2.2 | 1.17 |
| HAUSTM - Cool Ice | 0.0487 | 0.0021 | 4.3 | ND | NA | NA | ND | NP | — | — | NA |
| Pacha Mama - Mango | 0.0972 | 0.0065 | 6.7 | ND | NA | NA | ND | NP | — | — | NA |
| QCH | 0.3317 | 0.0118 | 3.6 | 29.87 | 1.27 | 4.3 | 2.99 | NP | — | — | NA |
ND: Non-detectable; NA: Does not apply; % CV: Coefficient of variation; STD: Standard deviation; NP: Non-performed; QCL: quality control sample-low concentration; QCH: quality control sample-high concentration; N: nicotine; BA: benzoic acid
The possible discrepancy in the results may be due to the fact that the previous results were expressed in mg/mL, so for comparison purposes, a weighted density of the e-liquid was used based on a theoretical composition of the JUULTM e-liquids of 30% PG and 70% Gly (0.30 × 1.04 g/mL + 0.70 × 1.26 g/mL = 1.19 g/mL). The molar ratio of nicotine:benzoic acid (1 : 0.9) for all JUUL TM samples suggests that 90% of the nicotine present in JUULTM samples is in protonated form as nicotine benzoate. The next step in our research will be to quantify benzoic acid, benzyl alcohol, benzene, total nicotine, and freebase nicotine captured in the aerosol particulate using a standard puffing regimen. Another interesting path could be to quantify the amount of benzene in the same phase due to the carcinogenic nature of this analyte (33) and see if it correlates with the benzoic acid content. We have also considered quantifying other carboxylic acids such as lactic, levulinic, salicylic, and tartaric acids in e-liquids that can be found in other commercial brands that use these types of acids.
We developed and validated an analytical method to determine the concentration of benzoic acid and benzyl alcohol in e-liquids using a dilution and mechanical extraction sample preparation step coupled with isotope-dilution HPLC/MS-MS. The new method was applied to the analysis of five JUULTM flavors and two other U.S. commercial e-cigarette brands. This method is characterized by its straightforward sample preparation, sensitivity, selectivity, and precision. The stability and precision of the measurement system over several weeks demonstrated the robustness of the method. The method is suitable for the routine analysis of benzoic acid and benzyl alcohol in e-liquids.