Urinary Cyanoethyl Mercapturic Acid, a Biomarker of the Smoke Toxicant Acrylonitrile, Clearly Distinguishes Smokers From Nonsmokers

Abstract

Introduction

Cyanoethyl mercapturic acid (CEMA) is a urinary metabolite of acrylonitrile, a toxicant found in substantial quantities in cigarette smoke, but not in non-combusted products such as e-cigarettes or smokeless tobacco and rarely in the diet or in the general human environment. Thus, we hypothesized that CEMA is an excellent biomarker of combusted tobacco product use.

Aims and Methods

We tested this hypothesis by analyzing CEMA in the urine of 1259 cigarette smokers (urinary cotinine ≥25 ng/mL) and 1191 nonsmokers. The analyses of CEMA and cotinine were performed by validated liquid chromatography–tandem mass spectrometry methods. Logistic regression was fit for log-transformed CEMA to construct the receiver operating characteristic curve.

Results

We found that a CEMA cutpoint of 27 pmol/mL urine differentiated cigarette smokers from nonsmokers with sensitivity and specificity greater than 99%. The use of different cotinine cutpoints to define smokers (10–30 ng/mL) had little effect on the results.

Conclusions

CEMA is a highly reliable urinary biomarker to identify users of combusted tobacco products such as cigarettes as opposed to users of non-combusted products, medicinal nicotine, or nonusers of tobacco products.

Implications

CEMA can be used to distinguish users of combusted tobacco products from non-combusted products such as e-cigarettes, smokeless tobacco, and medicinal nicotine. Levels of CEMA in the urine of people who use these non-combusted products are extremely low, in contrast to cotinine.

Introduction

Biochemical verification of tobacco product use is critical in the conduct of clinical trials to avoid misclassification of users of combusted products such as cigarettes as opposed to smokeless tobacco users, e-cigarette and medicinal nicotine users, and nonsmokers. A recent review discusses the strengths and weaknesses of various indicators of smoking status including carbon monoxide in exhaled breath, serum and urine nicotine metabolites, minor tobacco alkaloids, and metabolites of tobacco-specific nitrosamines.¹ In this report, we show that cyanoethyl mercapturic acid (CEMA, Figure 1), a urinary metabolite of acrylonitrile, is a highly effective biomarker for distinguishing cigarette smokers from nonsmokers.

Figure 1.

Structures of cyanoethyl mercapturic acid and acrylonitrile.

Acrylonitrile is classified as “reasonably anticipated to be a human carcinogen” by the National Toxicology Program and “possibly carcinogenic to humans” by the International Agency for Research on Cancer.^2,3 Cigarette smoking is by far the most common source of relatively high exposures to acrylonitrile. Its mean concentration in mainstream cigarette smoke has been reported as 28.4 µg per cigarette in a market survey of U.S. brand cigarettes analyzed using the Health Canada smoking regimen.⁴ The formation of acrylonitrile during cigarette smoking from tobacco nitrate and nitrite has been postulated, while other sources could include pyrolysis of nicotine and anatabine.⁵ There are no other common environmental or dietary sources of human exposure to comparable levels of acrylonitrile, and there is no evidence that it is formed endogenously. The general population may be exposed when using products made with acrylonitrile polymers, but these exposures are extremely low.² Some foods may contain low levels of acrylonitrile but the FDA’s Total Diet Study found no acrylonitrile residue in any foods tested from 1991 to 2004.² Occupational exposures may occur during the production of acrylonitrile and in factories where it is used as a monomer.² The only other source of exposure to levels of acrylonitrile comparable to those of tobacco smoke is marijuana smoke, while waterpipe smokers have levels of urinary CEMA significantly lower than cigarette smokers.^6–8

A quantitatively important pathway of acrylonitrile metabolism is a reaction with glutathione followed by normal metabolic processing of the glutathione conjugate resulting in excretion of CEMA in the urine.⁹ Urinary CEMA can be readily quantified by validated liquid chromatography–tandem mass spectrometry (LC-MS/MS) methods.^10–15 Multiple studies have demonstrated substantially higher levels of CEMA in the urine of cigarette smokers than nonsmokers consistent with the exposure data noted above.^{7,10–12,14,16,17} Furthermore, CEMA decreases by approximately 90% 8 days after cessation of cigarette smoking.¹¹ However, none of these studies has investigated whether urinary CEMA level can be reliably utilized as an indicator of combusted tobacco product use. We quantified CEMA in the urine of 1259 cigarette smokers and 1191 nonsmokers and report here that it reliably distinguishes smokers from nonsmokers.

Methods

Urine samples from subjects who reported being smokers were obtained at the week 0 timepoint of a 10-site randomized clinical trial that studied the effect of immediate versus gradual reduction in nicotine content of cigarettes on smoking-related behavior and biomarkers of exposure. Smoking status was confirmed by the analysis of urinary cotinine. The smokers were using their usual brand of cigarettes. The study protocol was approved by the University of Minnesota Institutional Review Board. Written consent was obtained from all subjects. Details of the trial have been described.¹⁸

Nonsmokers were selected from the Hawaii participants in the Multiethnic Cohort.¹⁹ The study protocol was approved by the University of Hawaii and the University of Southern California Institutional Review Boards. A urine sample was collected from the subjects, stored on ice overnight, and subsequently at −80°C. Equal numbers of participants were selected among Native Hawaiians, Japanese Americans, and whites. Participants were selected from those linked to Medicare rolls, in order to allow for use of the samples for nested case–control studies of relevant endpoints available from Medicare data.

Analysis of Urine

Urinary CEMA and cotinine were analyzed as described, using validated LC-MS/MS methods.^15,20 The limits of quantitation (LOQ) were 0.07 pmol/mL and 0.07 ng/mL, respectively.

Statistical Analysis

We used a urine cotinine concentration of at least 25 ng/mL as the standard for defining smokers and less than 25 ng/mL for nonsmokers. Summary statistics for CEMA were tabulated for the smoker and nonsmoker groups. CEMA levels that were lower than the LOQ were replaced with LOQ/√2 or LOQ/2 (as a sensitivity analysis). A histogram was presented to illustrate the distribution of CEMA in the two groups. Logistic regression was fit for the log-transformed CEMA to construct the receiver operating characteristic curve. Optimal cutoff points were determined both as that which equalized sensitivity and specificity and that which maximized the Youden index (ie, sensitivity + specificity − 1). SAS 9.4 (SAS Institute) was used for analyses and the R 3.6.1 (R Core Team) package ggplot2²¹ was used for plotting.

Results and Discussion

Demographics and smoking history of the subjects from the reduced nicotine content cigarette study have been described previously.¹⁸ Briefly, the median age was 45 years, they were 44% females, 61% whites, and 30% blacks, and they smoked a mean of 17 cigarettes/day for a mean of 27 years. The median age of the nonsmokers was 71 years, they were 57% females and 33.3% Native Hawaiians, 33.3% Japanese Americans, and 33.3% whites.

The results are summarized by smoking status in Table 1 and Figure 2. The mean and median levels of CEMA in smokers (N = 1259) were 412 and 1364 times as great, respectively, as those in nonsmokers (N = 1191). There was a clear separation of smoker and nonsmoker levels of CEMA. The distribution of urinary CEMA levels is illustrated in Figure 2.

Table 1.

Summary Statistics for Urinary CEMA Levels in Nonsmokers (Urine Cotinine <25 ng/mL; 142 pmol/mL) and Smokers (Urine Cotinine ≥25 ng/mL)

	Nonsmoker	Smoker
Sample size, N	1191	1259
Number of LOQ,1  N	173	0
Mean (SD), pmol/mL	2.42 (20.37)	994 (961)
Median, pmol/mL	0.54	737
Interquartile range, pmol/mL	0.14–1.83	390–1257
Range, pmol/mL	0.02–675	0.10–10 363
Geometric mean, pmol/mL	0.53	677

CEMA = cyanoethyl mercapturic acid; LOQ = limit of quantitation.

¹LOQ = 0.07 pmol/mL. For calculating mean and other summary statistics, LOQ/√2 was imputed for CEMA levels <LOQ. Sensitivity analysis was performed with CEMA <LOQ being replaced with LOQ/2; and only the mean (2.42–2.41 pmol/mL) and geometric mean (0.53–0.50 pmol/mL) of the nonsmokers slightly changed.

Figure 2.

Levels of urinary cyanoethyl mercapturic acid (CEMA) in nonsmokers (red) and smokers (green). The vertical lines indicate the geometric mean of CEMA for each group (nonsmoker = 0.53 pmol/mL; smoker = 677 pmol/mL); limit of detection (LOQ) = 0.07 pmol/mL.

Mean urinary cotinine values in smokers and nonsmokers were 3546 ng/mL (SD = 2520 ng/mL; geometric mean = 2737 ng/mL) and 0.45 ng/mL (SD = 1.44 ng/mL; geometric mean = 0.17 ng/mL), respectively. Mean urinary cotinine values in smokers and nonsmokers expressed as pmol/mL were 20 148 and 2.55, respectively.

The receiver operating characteristic analysis result when using urine cotinine of at least 25 ng/mL to define active smoking is summarized in Table 2. Both the sensitivity and the specificity were greater than 99% under both the equalized sensitivity and specificity criteria and the maximized Youden criterion, while the optimal cutpoints for CEMA were 27 and 23 pmol/mL, respectively, under the two criteria. We also examined varied cutpoints for urine cotinine for defining active smoking (ranging from 10 to 30 ng/mL). These did not impact the sensitivity (all >99%), specificity (all >99%), Youden (all >98%), or AUC (all >0.99), but slightly impacted the optimal cutpoint for CEMA (ranging from 19 to 30 pmol/mL). Note that all results given in Table 2 were unchanged when we replaced CEMA values that were lower than the detection limit with LOQ/2.

Table 2.

Cutpoints and Related Parameters for Distinguishing Smokers and Nonsmokers

Cotinine values to define smokers	AUC	Criterion	Cutpoint for smokers (CEMA ≥pmol/ mL)	Sensitivity	Specificity	Youden	No. of correctly predicted smokers	No. of correctly predicted nonsmokers	No. of nonsmokers predicted as smokers	No. of smokers predicted as nonsmokers
Cotinine ≥10 ng/mL	0.997	Equality1	19.13	0.991	0.992	0.983	1255	1174	10	11
		Max Youden	23.12	0.991	0.994	0.985	1255	1177	7	11
Cotinine ≥15 ng/mL	0.997	Equality	22.90	0.993	0.993	0.986	1255	1178	8	9
		Max Youden	23.12	0.993	0.994	0.987	1255	1179	7	9
Cotinine ≥20 ng/mL	0.998	Equality	27.00	0.994	0.994	0.989	1252	1184	7	7
		Max Youden	23.12	0.996	0.993	0.989	1254	1183	8	5
Cotinine ≥25 ng/mL	0.998	Equality	27.00	0.994	0.994	0.989	1252	1184	7	7
		Max Youden	23.12	0.996	0.993	0.989	1254	1183	8	5
Cotinine ≥30 ng/mL	0.998	Equality	30.00	0.994	0.994	0.989	1251	1185	7	7
		Max Youden	27.00	0.995	0.994	0.989	1252	1185	7	6

CEMA = cyanoethyl mercapturic acid; LOQ = limits of quantitation.

When CEMA levels <LOQ were replaced with LOQ/2, the receiver operating characteristic analysis result was unchanged.

¹Equalized sensitivity and specificity.

Our results demonstrate that urinary CEMA, which is readily quantified by LC-MS/MS, is an excellent biomarker for distinguishing cigarette smokers from nonsmokers, with sensitivity and specificity greater than 99%. Since acrylonitrile is present only in combusted tobacco products, not in e-cigarettes, medicinal nicotine, or smokeless tobacco products, the same characteristics apply. The mean level of CEMA reported in the PATH study of exclusive smokeless tobacco users was 8.3 pmol/mg creatinine²² while in e-cigarette users it was 18 pmol/mg creatinine,²³ both well below the CEMA cutpoint of 27 pmol/mL (based on 1.18 mg creatinine/mL).²⁴ Cotinine and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol cannot be used to distinguish cigarette smokers from smokeless tobacco users, and cotinine is also not useful for distinguishing smokers from e-cigarette or medicinal nicotine users. One potential problem with CEMA is that its urinary levels in marijuana smokers have been reported as 71 pmol/mL, which could lead to misclassification in the absence of other data; exclusive marijuana users would not have urinary cotinine but would have urinary cannabinoid metabolites.⁸

This study has certain limitations. The nonsmoker group was recruited in Hawaii and was comprised of an equal distribution of three racial/ethnic groups—Native Hawaiians, Japanese Americans, and whites—while the smoker group was recruited in the mainland United States and was comprised of 61% whites and 30% blacks. The median age of the nonsmokers was 71 years while that of the smokers was 45 years. While these are clear differences, there is no reason to believe that they should influence the levels in cigarette smoke of acrylonitrile, a tobacco combustion product, or the consequent levels of CEMA in the urine.

Thus, urinary CEMA should be added to the armamentarium of biomarkers for biochemical verification of combusted tobacco use. A user of combusted tobacco products will have high levels of CEMA whereas an e-cigarette, smokeless tobacco, or medicinal nicotine user, or nonuser of these products, will have extremely low levels of this urinary metabolite.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

This study was supported by the National Institute on Drug Abuse and Food and Drug Administration grant U54DA-031659 and National Cancer Institute grants CA-138338 and CA-164973. Mass spectrometry was carried out in the Analytical Biochemistry Shared Resource of the Masonic Cancer Center, supported in part by the National Cancer Institute Cancer Center Support grant CA-077598.

Declaration of Interests

None declared.