Plasma nicotine and its metabolite as biomarkers of tobacco exposure and their relevance to pulmonary nodule

Abstract

Aim: Explore the optimal cut-off values for plasma nicotine and its metabolites in assessing smoking status and quantify the association between individual tobacco exposure and pulmonary nodules (PNs).

Materials & methods: A total of 2245 plasma samples were included for the determination of nicotine (Nic), cotinine (Cot) and trans-3′-hydroxycotinine (OHCot) concentrations. The receiver operating characteristic curve was used to determine the optimal biomarkers reflecting smoking status. Binary logistic regression, restricted cubic spline and generalized linear model were used to analyze the association of nicotine and its metabolites with PNs. Quantile g-computation was used to investigate the mixed effects between them.

Results: Cot was found to be the best biomarker of self-reported active-passive smoking, with optimal thresholds of 9.06 and 1.26 ng/ml, respectively. Except for OHCot, increased concentrations of Cot, Nic, total nicotine equivalent (TNE2) and TNE3 were significantly positively associated with the risk of PNs, whereas nicotine metabolite ratio presented a negative association. The mixed effects of OHCot, Cot and Nic were associated with PNs, with an odds ratio of 1.17 and a 95% CI of 1.05–1.30.

Conclusion: Nicotine and its metabolites as potential biomarkers of tobacco exposure were significantly associated with PNs.

Graphical Abstract

Plain language summary

Article highlights.

Introduction

• To quantify individual tobacco exposure is a public health challenge.

• Pulmonary nodules (PNs) detection rates have risen in recent years, but there is a paucity of research on the association between smoking and PNs.

Methods

• Nic and its metabolite concentrations in plasma samples were measured using liquid chromatography-triple quadrupole mass spectrometry (LC–MS/MS).

• Identify the optimum biomarker for self-reported smoking status by obtaining the area under the curve from the receiver operator characteristic curve.

• The binary logistic regression and restricted cubic spline, generalized linear model were used to analyze the association between nicotine and its metabolites and PNs.

• The Quantile g-Computation model was used to explore the mixed effects of nicotine (Nic), cotinine (Cot) and trans-3′-hydroxycotinine (OHCot) on PNs.

Results

• Cotinine (Cot) is the optimal biomarker for self-reported smoking status, with an area under the curve of 0.958.

• The risk of PNs increased significantly with increasing concentrations of Cot, Nic, TNE2 and TNE3.

• There was a significant positive effect of the mixed effect of OHCot, Cot and Nic on PNs in the Quantile g-Computation model.

Discussion & conclusion

• Nicotine and its metabolites as biomarkers of tobacco exposure are highly associated with the development of PNs, but the relationship between them needs to be explored in greater depth.

1. Introduction

Global tobacco control has been a priority of the WHO for decades, but smoking remains a major risk factor for cardiovascular disease, respiratory disease and more than 20 cancers [1]. Tobacco smoke contains more than 4000 compounds, including tar, carbon monoxide and irritating oxidizing gases, of which more than 100 are harmful to human health, including approximately 50–60 carcinogens, several mutagens and many irritating or toxic substances [2]. Nicotine (Nic), an alkaloid and sympathetic-like stimulant contained in tobacco, is a major driver of addiction [3]. After Nic enters the body, about 70–80% will be converted into cotinine (Cot). Cot is eventually excreted in urine, with the remainder converted to metabolites, primarily trans-3′-hydroxy Cotinine (OHCot) (33–40%) [4]. Although not a carcinogen, Nic is one of the key substances leading to smoking and ongoing exposure to many carcinogens in tobacco [5], which has also become a leading cause of premature death in developed countries [4].

At present, many studies have confirmed the association between Nic and its metabolites and lung cancer. Nic is not a carcinogen, but the intake of Nic and the metabolism of Nic are mainly influenced by smoking behavior. Serum Cot is associated with the risk of lung cancer in large epidemiological studies [6]. In a nested case–control study of lung cancer involving 1741 lung cancer cases and 1741 controls in Norway, results showed that elevated serum Cot levels were associated with a significantly increased risk of lung cancer [7].

However, pulmonary nodules (PNs), which are often found incidentally on examination for other reasons, are overwhelmingly benign, and a small proportion of them may develop into early-stage lung cancers with potential to be cured [8]. A study of PNs in northern China found a prevalence of 26.32% [9]. A total of 2.27% of sporadic PNs developed into lung cancer during a 2-year follow-up period [10]. Therefore, early screening of lung cancer high-risk groups and detection of PNs have a positive impact on the diagnosis, treatment and prognosis of lung cancer patients.

Previous studies have found that smoking is a possible factor affecting PNs. One study found that 57% of 5444 patients with incidental PNs were smokers [11]. A domestic study of PNs found that smoking was associated with positive PNs [12]. One study assessed risk factors for the development of PNs and found that heavy smokers were more common in the positive PNs group [13]. In addition, studies on the relationship between passive smoking and PNs (which may refer to lung-related diseases) are relatively scarce, and a Japanese study found that women who lived with smokers (especially their husbands) had a higher risk of developing lung cancer. Also, even if women are not smokers themselves, prolonged passive smoking in environments such as the workplace increases their risk of lung cancer [14]. Based on these findings, it can be hypothesized that passive smoking may also affect the occurrence of PNs, which needs to be confirmed by more studies.

Nowadays, the current status of active and passive smoking is usually collected on the basis of standardized questionnaires, which are a quick and inexpensive way of assessing an individual's exposure to tobacco smoke and have been widely used in relevant studies. However, not all self-reported active and passive smoking statuses reflect the “true” status of an individual's exposure to tobacco smoke [15]. The underestimation of exposure to tobacco smoke based on self-reported active and passive smoking status can be due to the omission of personal information, recall bias and reporting bias in questionnaire completion, which is a limitation of the results of tobacco exposure and health studies [16]. Therefore, objective measurement of the levels of biomarkers and their metabolites associated with exposure to tobacco smoke in humans is essential for reflecting the smoking status of the subjects in the research. Studies have demonstrated that self-reported active and passive smoking can be verified by in vivo levels of Nic and its metabolites [17].

Nevertheless, epidemiologic studies on the association between Nic and its metabolites and PNs are still lacking. As key biomarkers of tobacco smoke exposure, Nic and its metabolites play an important role in revealing potential associations between tobacco smoke exposure and PNs. Therefore, to deeply investigate the relationship between them, we conducted this study to examine the plasma concentrations of Nic, Cot and OHCot in high-risk individuals (50–74 years old) from the lung cancer screening cohort in Ma'anshan City, Anhui Province, combining with the self-reported smoking status of the population, to determine the optimal biomarkers as well as the cutoff values. And further explored the relationship between Nic and its metabolites and PNs, to provide scientific clues for the prevention and control of PNs.

2. Materials and methods

2.1. Study population and design

From June to November 2020, the community lung cancer screening cohort in Ma'anshan City [18], Anhui Province adopted the cluster sampling method, and selected 10,038 permanent residents aged 50–74 years old (local residence ≥3 years) from Huashan District of Ma'anshan City for baseline questionnaire survey. The exclusion criteria of the selected subjects were: patients with previous lung cancer; newly diagnosed cancers within the last 5 years (except localized prostate cancer, cervical carcinoma in situ, nonmelanic skin cancer); have potential symptoms of lung disease, such as coughing up blood, hemoptysis, etc.; serious diseases of cardiovascular and cerebrovascular systems, gastrointestinal tract, liver and kidney systems, or serious mental diseases (e.g., mania, panic disorder and other severe anxiety disorders).

Based on the questionnaire survey, 3376 people at high risk of lung cancer were evaluated [19], and a total of 2289 people completed low-dose computed tomography (LDCT) examination [20], 839 people were detected PNs, of which 290 were positive PNs. From November 2022 to May 2023, we collected plasma samples from 2289 individuals who completed LDCT examination in high-risk groups and tested for concentrations of Nic and its metabolites, after excluding 44 loss of plasma samples and abnormal determination values (e.g., height of 180 m or weight of 500 kg), a total of 2245 individuals had plasma samples included in the analysis (Supplementary Figure S1). This study was approved by the Ethics Committee of Ma'anshan Center for Disease Control and Prevention (Approval No. 2020001), and all participants were required to provide written informed consent for admission.

2.2. Covariate collection and definitions

A community health risk assessment questionnaire was used to conduct an interview survey on the study subjects. We collected General demographic information (sex and age), lifestyle habits (cooking frequently, drinking status, tea drinking, physical exercise and thurification which means burning incense at least once a month in the household [21], occupational exposure (occupational history of metal exposure and occupational exposure to other harmful substances), disease (history of lung-related diseases, hypertension and diabetes) and mental depression. Physical measurements, including height, weight, systolic blood pressure and diastolic blood pressure, were taken by trained investigators using professional instruments. Total nicotine equivalent (TNE2) represents the sum of plasma Nic and Cot concentrations, and TNE3 represents the sum of plasma Nic, Cot and OHCot concentrations. Nicotine metabolite ratio (NMR) refers to the ratio of OHCot to Cot and is used to predict the clearance of Nic and its metabolites in the body [22].

At Ma'anshan Central Hospital, free LDCT examinations are conducted for high-risk populations assessed through questionnaire surveys. The diagnosis of LDCT results is mainly carried out by senior doctors in the hospital, and another senior physician reviews the results and issues an examination report. On the morning of the LDCT examination, 5 ml of venous blood was drawn on an empty stomach, and anticoagulant vessels containing EDTA-K2 were collected. Fasting blood glucose was measured after collection, and the remaining parts were immediately stored away from light and sent to the Ma'anshan Center for Disease Control and Prevention laboratory in a low-temperature specimen transport box for centrifugation (4000 rpm for 10 min). The upper plasma was then extracted in a 5 ml EP (Eppendorf) tube and quickly stored in an ultra-low temperature refrigerator at -80°C for further testing.

2.3. Definition of PNs

The PNs were defined as an approximately rounded opacity with a diameter of less than 3 cm [10]. The positive PNs were defined as solid or partially solid nodules with a diameter ≥5 mm; or nonsolid nodules ≥8 mm in diameter; or found suspicious lesions of trachea and bronchus; or single, multiple nodules or masses of lung cancer diagnosed by LDCT [23]. If the nodule diameter was smaller than this standard and no tracheal nodule was detected, it was defined as a negative pulmonary nodule.

2.4. Laboratory measurements

The instrument reagents used in this study include liquid chromatograph-triple quadrupole mass spectrometry (LC–MS–MS, Altis Plus, Thermo Fisher), analytical balance, centrifuge, ultrasonic instrument, ultrapure water instrument, vortex mixing instrument, fume hood, nitrogen blowing instrument, Nic, OHCot, Cot-d3 (USA, Good Laboratory Practice Bioscience) and Cot (Toronto Research Chemicals Incorporation, Canada).

The detection of Nic and its metabolites in plasma using the standard curve method [24]. During sample preparation, plasma samples were defrosted according to a certain procedure and then fully vortexed. About 100 μl of plasma was collected into a 2 ml EP tube, and 10 μl of external standard, internal standard and 50 mg/ml DTT solution were added, then swirled for 15 s. Add 1 ml extractant into EP tube from fume hood and shake in a vortex mixer for 5 min. After the shock, centrifuge with 13,800 × g centrifugal force for 10 min, remove 900 μl from the fume hood and rinse it into a 2 ml EP tube, then blow nitrogen on the nitrogen blower until it is completely dry. Add 200 μl compound solution into the EP tube after nitrogen blowing, swirl for 30 s, ultrasound for 7 min, centrifuge at 4°C with 13,800 × g centrifugal force for 20 min, take 100 μl supernatant into the sample vial, and then put it into LC–MS–MS for detection.

2.5. Statistical analyses

The distributions of age and BMI were described by means and standard deviations, and differences were compared using t-tests. Categorical variables were represented by n (%) and Chi-square tests were used to compare differences between groups. The concentrations of Nic and its metabolites were more than 70% above the detection limit (LOD), and the parts below the detection limit were replaced by half of LOD [25], and the quartile was used to describe their distribution. The specific LOD and limit of quantification in this study are shown in Supplementary Table S1. The receiver operator characteristic curve (ROC) was used to obtain the area under the curve (AUC). The accuracy of the assessment of self-reported active and passive smoking status based on plasma Nic and its metabolite concentration was analyzed, and the optimal cutoff value of plasma Nic and its metabolite to distinguish active and passive smoking [26,27] was obtained according to the Youden’s index. In addition, the sensitivity and specificity of the self-reported and assessed by the Nic or its metabolites active/passive smoking were calculated, and the consistency between the two was assessed using Kappa values.

In this study, six metabolites or indicators of Nic and its metabolites (Nic, Cot, OHCot, TNE2, TNE3 and NMR) were quartiled and divided into four groups, Q1 (Quartile), Q2, Q3 and Q4. With Q1 as a reference, binary logistic regression was used to explore the influence of the other three quantiles on the occurrence of PNs and positive PNs, and the odds ratio (OR) and 95% confidence interval (CI) were expressed. Model 1 adjusts for sex and age; based on Model 1, Model 2 adjusts for occupational history of metal exposure, drinking status, tea drinking, diabetes and thurification; based on Model 2, Model 3 adjusts for BMI, physical exercise, cooking frequently, mental depression, occupational history of exposure to other harmful substances, hypertension, and history of lung-related diseases.

The restricted cubic spline model was used to analyze the dose–response relationship between six metabolites or indicators and PNs. The dose-response relationship between six metabolites or indicators and PNs was introduced as continuous variables after logarithmic transformation, and three nodes were automatically generated by the system, namely P₁₀, P₅₀ and P₉₀. And the adjustment of covariables is the same as the above model 3. In this study, the generalized linear model (GLM) was used to further explore the linear association between 6 metabolites or indicators and PNs and 3 models were also used to explore. We used binary logistic regression to analyze the interaction between Nic and its metabolites. Each metabolite is divided into “Low” (Q1 + Q2) and “High” (Q3 + Q4) and each of the two metabolites is divided into four categories (low/low, low/high, high/low and high/high), using the “Low/low” group as a reference. To explore the impact of the mixed effects of Nic and its metabolites (Nic, Cot and OHCot) on the occurrence of PNs and positive PNs, the Quantile g-Computation (QgC) model was used in this study.

To further explore the association of Nic and its metabolites with PNs in specific populations, different population analyses were performed in this study, including the male population, the elderly population, and the population with passive smoking status. EpiData 3.1 software was used for data entry, SPSS 23.0 software and R 4.1.1 software for statistical analysis, the test level α = 0.05. p < 0.05 was considered to indicate a statistically significant difference.

3. Results

3.1. Identify the optimal cutoff value and assess the consistency of the results

According to ROC curve results (Figure 1), in the self-reported smoking status reflected by Nic or its metabolite levels, the AUC of Cot, TNE2, TNE3, OHCot and Nic were 0.958 (95% CI: 0.949–0.968), 0.959 (95% CI: 0.951–0.968), 0.959 (95% CI: 0.950–0.968), 0.942 (95% CI: 0.932–0.953) and 0.926 (95% CI: 0.914–0.939), respectively. Although the AUC values of TNE2 and TNE3 were slightly higher than Cot, considering factors such as cost–effectiveness, we ultimately chose Cot levels to reflect self-reported smoking status and determined that the cut-off value used to distinguish smokers from nonsmokers was 9.06 ng/ml. Similarly, we determined that Cot concentration reflected self-reported passive smoking status and determined that the cut-off value for the difference between passive smoking and passive smoking was 1.26 ng/ml.

Figure 1.

Receiver operator characteristic curve for assessing self-reported active and passive smoking status based on nicotine or its metabolite concentration. (A) is the ROC curve for evaluating self-reported smoking status based on the concentration of nicotine or its metabolites, and (B) is the ROC curve for evaluating self-reported passive smoking status based on the concentration of nicotine or its metabolites.

AUC: Area under curve; Cot: Cotinine; Nic: Nicotine; OHCot: Trans-3′-hydroxycotinine; ROC: Receiver operator characteristic curve; TNE: Total nicotine equivalent; TNE2: The sum of nicotine and cotinine concentrations; TNE3: The sum of nicotine, cotinine and trans-3′-hydroxycotinine concentrations.

Supplementary Tables S2 & S3 show the results of consistent evaluation of active and passive smoking status based on self-report and cot levels, respectively. When the optimal cutoff value of Cot was 9.06 ng/ml, with sensitivity of 91.5% and specificity of 95.0% for distinguishing smokers from nonsmokers, it showed favorable agreement with self-reported information (Kappa = 0.867, p < 0.001). Among 1237 self-reported nonsmokers, the sensitivity and specificity of Cot levels for passive smoking status were 37.0% and 82.9%, respectively, with lower agreement with self-reported passive smoking status (Kappa = 0.215, p < 0.001). The results above indicate that the ability to differentiate between smokers and nonsmokers based on plasma Nic levels is satisfactory, but differentiating passive smokers from self-reported nonsmokers is less effective.

3.2. Basic characteristics of the study population and distribution of Nic and its metabolite levels

To improve the representation of subjects in the study, we compared and analyzed the consistency between self-reported smoking status and laboratory cotinine test results based on the data in Supplementary Table S2, excluding 148 subjects with inconsistent assessment results, and finally including 2097 subjects, of whom 765 were PNs and 1332 were non-PNs. There were statistically significant differences in occupational history of metal exposure and diabetes between the PNs and non-PNs groups. There were 266 positive PNs and 499 negative PNs in the PNs group. There were statistically significant differences in age, sex, drinking status, tea drinking and thurification between the positive and negative PNs group (Table 1). The median concentrations of Nic and its metabolites in the positive PNs group were higher than those in the negative PNs group, and the median NMR values were lower than those in the negative PNs group (Supplementary Table S4).

Table 1.

Study the basic characteristics of the population (n = 2097).

Variable	Non-PNs (n = 1332)	PNs (n = 765)	p-value	Negative PNs (n = 499)	Positive PNs (n = 266)	p-value
Cotinine-based smoking status			0.354			0.120
No	740 (55.56)	409 (53.46)		277 (55.51)	132 (49.62)
Yes	592 (44.44)	356 (46.54)		222 (44.49)	134 (50.38)
Age ( $\overline{\text{x}}$  ± s, year)	62.93 ± 6.81	63.51 ± 6.64	0.055	63.14 ± 6.79	64.21 ± 6.30	0.034
BMI ( $\overline{\text{x}}$  ± s, kg/m2)	23.99 ± 3.01	24.07 ± 2.95	0.533	24.07 ± 3.02	24.08 ± 2.82	0.957
Sex			0.911			0.016
Male	904 (67.87)	521 (68.10)		325 (65.13)	196 (73.68)
Female	428 (32.13)	244 (31.90)		174 (34.87)	70 (26.32)
Drinking status			0.625			0.028
Never	717 (53.83)	411 (53.73)		284 (56.91)	127 (47.74)
Drinking	499 (37.46)	296 (38.69)		176 (35.27)	120 (45.11)
Abstinence	116 (8.71)	58 (7.58)		39 (7.82)	19 (7.14)
Tea drinking			0.365			0.026
Barely drink	613 (46.02)	331 (43.27)		233 (46.69)	98 (36.84)
Occasionally	74 (5.56)	39 (5.10)		22 (4.41)	17 (6.39)
Often	645 (48.42)	395 (51.63)		244 (48.90)	151 (56.77)
Physical exercise			0.052			0.890
Never	572 (42.94)	289 (37.78)		186 (37.27)	103 (38.72)
Occasionally	157 (11.79)	88 (11.50)		59 (11.82)	29 (10.90)
Often	603 (45.27)	388 (50.72)		254 (50.90)	134 (50.38)
Thurification			0.107			0.026
No	1216 (91.29)	682 (89.15)		454 (90.98)	228 (85.71)
Yes	116 (8.71)	83 (10.85)		45 (9.02)	38 (14.29)
Cooking frequently			0.800			0.364
No	425 (31.91)	240 (31.37)		151 (30.26)	89 (33.46)
Yes	907 (68.09)	525 (68.63)		348 (69.74)	177 (66.54)
Mental depression			0.950			0.811
No	1230 (92.34)	707 (92.42)		462 (92.59)	245 (92.11)
Yes	102 (7.66)	58 (7.58)		37 (7.41)	21 (7.89)
Occupational history of metal exposure			0.023			0.276
No	1009 (75.75)	545 (71.24)		349 (69.94)	196 (73.68)
Yes	323 (24.25)	220 (28.76)		150 (30.06)	70 (26.32)
Occupational history of exposure to other harmful substances			0.353			0.583
No	585 (43.92)	352 (46.01)		226 (45.29)	126 (47.37)
Yes	747 (56.08)	413 (53.99)		273 (54.71)	140 (52.63)
Hypertension			0.319			0.481
No	629 (47.22)	344 (44.97)		229 (45.89)	115 (43.23)
Yes	703 (52.78)	421 (55.03)		270 (54.11)	151 (56.77)
Diabetes			0.014			0.900
No	1106 (83.03)	602 (78.69)		392 (78.56)	210 (78.95)
Yes	226 (16.97)	163 (21.31)		107 (21.44)	56 (21.05)
History of lung-related diseases			0.168			0.986
No	1188 (89.19)	667 (87.19)		435 (87.17)	232 (87.22)
Yes	144 (10.81)	98 (12.81)		64 (12.83)	34 (12.78)

BMI: Body mass index.

3.3. Association of nicotine and its metabolites with PNs in a multifactor model

This study explored the association between OHCot, Cot, Nic, TNE2, TNE3 and NMR with the occurrence of PNs after quartile placement. Although different confounder factors were adjusted in each of the three models, the results were similar (Table 2). Compared with the first quantile, Cot, TNE2 and TNE3 of the other three quantiles increased the risk of PNs occurrence, among which the OR values of the fourth quantile in Model 3 were 1.55 (95% CI: 1.14–2.09), 1.46 (95% CI: 1.08–1.97) and 1.47 (95% CI: 1.08–1.98); NMR of the remaining three quantiles all reduced the risk of PNs, with the fourth quantile of Model 3 having an OR value of 0.73 (95% CI: 0.56–0.95). This study also investigated the association of OHCot, Cot, Nic, TNE2, TNE3 and NMR quartile with the occurrence of positive PNs. Compared with the first quantile, the OR values of OHCot, Cot, Nic, TNE2 and TNE3 of the remaining quantiles were greater than 1, but the differences were not statistically significant (Supplementary Table S5).

Table 2.

Association between pulmonary nodules group and nonpulmonary nodules group in a multifactor model (n = 2097).

	Model 1	Model 2	Model 3
	OR (95% CI)	OR (95% CI)	OR (95% CI)
OHCot
Quartile 1	1 (ref)	1 (ref)	1 (ref)
Quartile 2	1.06 (0.82–1.37)	1.07 (0.83–1.39)	1.04 (0.80–1.35)
Quartile 3	1.41 (1.08–1.84)	1.39 (1.06–1.82)	1.41 (1.07–1.85)
Quartile 4	1.19 (0.90–1.58)	1.17 (0.88–1.56)	1.22 (0.91–1.64)
Cot
Quartile 1	1 (ref)	1 (ref)	1 (ref)
Quartile 2	1.35 (1.04–1.74)	1.33 (1.03–1.71)	1.32 (1.02–1.71)
Quartile 3	1.39 (1.06–1.84)	1.36 (1.03–1.80)	1.42 (1.06–1.88)
Quartile 4	1.50 (1.12–2.01)	1.46 (1.09–1.97)	1.55 (1.14–2.09)
Nic
Quartile 1	1 (ref)	1 (ref)	1 (ref)
Quartile 2	1.23 (0.95–1.59)	1.24 (0.95–1.60)	1.22 (0.94–1.58)
Quartile 3	1.41 (1.08–1.84)	1.40 (1.06–1.83)	1.42 (1.08–1.87)
Quartile 4	1.50 (1.13–2.00)	1.49 (1.11–1.98)	1.55 (1.15–2.08)
TNE2
Quartile 1	1 (ref)	1 (ref)	1 (ref)
Quartile 2	1.33 (1.03–1.72)	1.31 (1.01–1.69)	1.29 (1.00–1.67)
Quartile 3	1.37 (1.04–1.81)	1.34 (1.01–1.77)	1.38 (1.04–1.83)
Quartile 4	1.42 (1.06–1.91)	1.39 (1.03–1.87)	1.46 (1.08–1.97)
TNE3
Quartile 1	1 (ref)	1 (ref)	1 (ref)
Quartile 2	1.34 (1.04–1.73)	1.33 (1.03–1.72)	1.31 (1.01–1.69)
Quartile 3	1.43 (1.08–1.89)	1.40 (1.06–1.85)	1.44 (1.08–1.91)
Quartile 4	1.44 (1.07–1.92)	1.40 (1.04–1.88)	1.47 (1.08–1.98)
NMR
Quartile 1	1 (ref)	1 (ref)	1 (ref)
Quartile 2	0.70 (0.54–0.90)	0.70 (0.54–0.90)	0.69 (0.54–0.90)
Quartile 3	0.69 (0.54–0.89)	0.68 (0.53–0.88)	0.67 (0.51–0.86)
Quartile 4	0.74 (0.57–0.97)	0.76 (0.58–0.99)	0.73 (0.56–0.95)

Model 1 adjusts for sex and age; On the basis of Model 1, Model 2 adjusted occupational history of metal exposure, drinking status, tea drinking, diabetes, thurification; On the basis of Model 2, Model 3 adjusts for BMI, physical exercise, cooking frequently, mental depression, occupational history of exposure to other harmful substances, hypertension and history of lung-related diseases.

95% CI: 95% confidence interval; Cot: Cotinine; OHCot: Trans-3′-hydroxycotinine; Nic: Nicotine; NMR: Nicotine metabolite ratio; OR: Odds ratio; TNE: Total nicotine equivalent; TNE2: The sum of nicotine and cotinine concentrations; TNE3: The sum of nicotine, cotinine and trans-3′-hydroxycotinine concentration.

3.4. Linear association of nicotine and its metabolites with PNs

We conducted a dose-response relationship study to further investigate the association between nicotine and its metabolite concentrations and PNs. As shown in Figure 2, with the increase of the concentrations of OHCot, Cot, Nic, TNE2 and TNE3, the risk of PNs occurrence is also increasing, and at a certain concentration, the OR value is greater than 1. But NMR is the opposite. As the above P_Total was less than 0.05 and P_non-line was greater than 0.05, it indicates that there is no nonlinear dose-response relationship between Nic and its metabolites with PNs.

Figure 2.

Dose–response relationship between nicotine and its metabolite concentrations and pulmonary nodules (PNs).The solid red line indicates the OR of each nicotine and its metabolites in relation to PNs, the red shading indicates its 95% CI, and the black dashed line indicates OR=1. Adjusted for sex, age, occupational history of metal exposure, drinking status, tea drinking, diabetes, thurification, BMI, physical exercise, cooking frequently, mental depression, occupational history of exposure to other harmful substances, hypertension and history of lung-related diseases.

95% CI: 95% Confidence interval; Cot: Cotinine; Nic: Nicotine; NMR: Nicotine metabolite ratio; OHCot: Trans-3′-hydroxycotinine; OR: Odds ratio; TNE: Total nicotine equivalent; TNE2: The sum of nicotine and cotinine concentrations; TNE3: The sum of nicotine, cotinine and trans-3′-hydroxycotinine concentration.

Based on the above study, we continued to investigate the linear association of nicotine and its metabolites with PNs using the GLM, and the results were similar in the three models. As shown in Figure 3, Cot, Nic, TNE2 and TNE3 increase the risk of PNs, while NMR reduces the risk of PNs. In Model 3, for each unit increase in the concentration of Cot, Nic, TNE2 and TNE3 (log-converted), the OR values of the risk of PNs increased by 1.14 (95% CI: 1.04–1.24), 1.19 (95% CI: 1.06–1.34), 1.15 (95% CI: 1.05–1.27) and 1.14 (95% CI: 1.03–1.27), respectively. The OR value of the risk of PNs changed by 0.76 (95% CI: 0.64–0.91) for every one-unit decrease in NMR (log-converted).

Figure 3.

Association of nicotine and its metabolites with pulmonary nodules based on the generalized linear model. Solid dot represents OR value and the length of a straight line represents the 95% CI. Model 1 adjusted for sex and age; On the basis of Model 1, Model 2 adjusted occupational history of metal exposure, drinking status, tea drinking, diabetes, thurification; On the basis of Model 2, Model 3 adjusted for BMI, physical exercise, cooking frequently, mental depression, occupational history of exposure to other harmful substances, hypertension and history of lung-related diseases.

Cot: Cotinine; Nic: Nicotine; NMR: Nicotine metabolite ratio; OHCot: Trans-3′-hydroxycotinine; TNE: Total nicotine equivalent; TNE2: The sum of nicotine and cotinine concentrations; TNE3: The sum of nicotine, cotinine and trans-3′-hydroxycotinine concentrations.

3.5. Association of nicotine and its metabolites with PNs in different populations

Due to the large number of male smokers, we analyzed the male population, and the results were consistent with the analysis of the general population (Supplementary Table S6 & Supplementary Figures S2 & S4A). However, the dose-response curves of OHCot showed an inverted “U” shape, and the OR values did not exceed 1. Similarly, the results in the older population were similar to those in the general population (Supplementary Table S7 & Supplementary Figures S3 & S4B). However, results from the GLM of the passive smoking status population showed that neither nicotine nor its metabolites had a statistically significant effect on PNs (Supplementary Figure S4C).

3.6. Association of mixed effects of nicotine and its metabolites with PNs

We further explored the interaction of OHCot, Cot and Nic among PNs. As shown in Supplementary Table S8, there is an interaction between Nic and Cot, and P_interaction is 0.047. Quantile g-computation was used to evaluate the association of mixed effects of nicotine and its metabolites with PNs. As shown in Table 3, the mixed effects of OHCot, Cot and Nic had a significant positive correlation with the occurrence of PNs, with an OR value of 1.17 (95% CI: 1.05–1.30). The same results were obtained in the male population, the elderly population and the passive smoking population. However, the mixed effects of OHCot, Cot and Nic did not have a significant association with the occurrence of positive PNs, and the same results were obtained in sensitivity analysis (Table 3). In the general population (Figure 4), the negative correlation between OHCot and PNs has the greatest weight, while the positive correlation between Cot has the greatest weight. The same results were found for men and older people. In the passive smoking status population, the weight of the positive association of Nic is the largest.

Table 3.

The association of mixed effects of nicotine and its metabolites with pulmonary nodules by using quantile g-computation.

	Non-PNs → PNs			Negative PNs → Positive PNs
	N	OR (95% CI)	p-value	N	OR (95% CI)	p-value
Totala	2097	1.17 (1.05–1.30)	0.005	765	1.11 (0.92–1.34)	0.277
Maleb	1425	1.14 (1.01–1.28)	0.027	521	1.03 (0.85–1.26)	0.740
The elderlyc	1401	1.14 (1.01–1.30)	0.040	527	1.12 (0.91–1.39)	0.289
Passive smoking statusa	1237	1.16 (1.01–1.33)	0.042	439	1.10 (0.86–1.41)	0.451

^aAdjusted for sex, age, occupational history of metal exposure, drinking status, tea drinking, diabetes, thurification, BMI, physical exercise, cooking frequently, mental depression, occupational history of exposure to other harmful substances, hypertension and history of lung-related diseases.

^bReduced adjustment for sex.

^cReduced adjustment for age.

95% CI: 95% Confidence interval; OR: Odds ratio; PNs: Pulmonary nodule.

Figure 4.

Weight of nicotine and its metabolites in positive or negative associations between pulmonary nodule and non-pulmonary nodule groups in Quantile g-Computation model. (A) is the general population, (B) is the male, (C) is the elderly and (D) is the passive smoking population. (A & D) were adjusted for sex, age, occupational history of metal exposure, drinking status, tea drinking, diabetes, thurification, BMI, physical exercise, cooking frequently, mental depression, occupational history of exposure to other harmful substances, hypertension and history of lung-related diseases. (B) reduced adjustment for sex, (C) reduced adjustment for age.

Cot: Cotinine; Nic: Nicotine; OHCot: Trans-3′-hydroxycotinine.

4. Discussion

This study mainly investigated the ability of plasma Nic and its metabolite concentrations to assess the self-reported status of active and passive smoking in a population at high risk of lung cancer and identified plasma Cot as the optimal biomarker reflecting active and passive smoking. Based on the levels of Cot metabolites and self-reported information, we evaluated tobacco smoke exposure to ultimately include the study population with consistent evaluation results. Subsequently, we investigated the association of six indicators (OHCot, Cot, Nic, TNE2, TNE3 and NMR) with PNs, and the results suggested that there was no nonlinear dose-response relationship between them. GLM results showed that in addition to OHCot, the risk of PNs increased significantly with the increase of Cot, Nic, TNE2 and TNE3 concentrations, while the risk of PNs decreased significantly with the increase of NMR. In the QgC model, the mixed effects of OHCot, Cot and Nic had significant positive effects on PNs.

In this study, a plasma Cot level of 9.06 ng/ml was used as the best cutoff value (sensitivity 91.5%, specificity 95.0%) to evaluate smoking status. Cot level has been proposed by many studies as a biomarker to distinguish whether smoking or not. Due to the differences in metabolic rate of the population in different studies, the cut-off value is slightly different. The serum cotinine concentrations used to distinguish smoking among non-Hispanic blacks, non-Hispanic whites, and Mexican Americans were 5.92, 4.85 and 0.84 ng/ml, respectively [15]. One study in pregnant women used urine Cot level as a biomarker to distinguish between smoking and non-smoking, with a cutoff value of 74.1 ng/ml (sensitivity 96.7%, specificity 98.0%) [16].

An optimal cutoff value of 1.26 ng/ml (sensitivity 37.0%, specificity 82.9%) for passive smoking plasma Cot was found in this study, with a high misclassification rate of 33.71%, which is similar to the results of several studies [16,28]. This suggests that nonsmokers do not report passive smoking very accurately. However, all participants in this study were voluntary, and community workers were familiar with these groups, so most of the groups did not provide incorrect answers. This study based on Cot and self-reported smoking status analysis found that its sensitivity and specificity are very high, more can be demonstrated from the side. Therefore, the possible reasons are that the age of the study object is 50–74 years old, and there are blurred memory and unclear expression; In daily life, passive smoking is not a spontaneous behavior and may be less sensitive to the increase and decrease of the number of smokers around, leading to the misestimation of passive smoking [28]. The situation of passive smoking inhalation is also greatly related to the air circulation and the interval distance, which will lead to a large gap between the Cot content in the body and the self-reported.

A Chinese employee physical examination study consisting of 96,476 subjects found that smoking was associated with positive PNs [12], while this study found no significant association between nicotine and its metabolites and positive PNs. This may be due to the small sample size of positive and negative PNs in this study, and further studies are needed to confirm this. Studies have shown that smokers have a higher incidence of PNs than non-smokers [29]. However, there is currently little literature on the association between smoking dose and PNs, which can be explained by the possible mechanism of the effect of tobacco smoke exposure on the lungs. Possible mechanisms include: First, cigarette smoke induces oxidative stress in airway epithelial cells, which may lead to continued aggregation of immune cells, excessive mucus secretion and loss of ciliary beats on the surface of airway epithelium, resulting in airflow restriction [30]. Second, cigarette smoke alters the structure and function of mitochondria, a key link in the pathogenesis of chronic lung disease [31]. Third, smoking also promotes the activation of neutrophils by macrophages and immune cells, thereby altering the respiratory microbiome primarily associated with Haemophilus, leading to neutrophilic airway inflammation [32]. Therefore, the higher the dose of smoking, the more frequently these mechanisms occur, and the more likely PNs are to occur.

Studies have shown that TNE3 is more suitable as a biomarker of tobacco smoke exposure than multiple nicotine and its metabolites [33] and more reflective of individual nicotine intake and internal smoking dose [34]. According to the ROC of this study, it can also be found that the AUC values of TNE2 and TNE3 are the highest, but only 0.001 higher than that of Cot. Therefore, considering the cost–effectiveness and other factors, we finally chose to reflect the smoking status only with Cot. To more fully explore the association between tobacco smoke exposure and PNs, TNE2 and TNE3 were also included in subsequent analyses. Similar to previous findings, TNE is associated with a higher risk of lung cancer [34,35].

Studies have shown a positive dose-response relationship between NMR levels and cytochrome P450 2A6 (CYP2A6) [36], so some studies use NMR values to reflect the activity of CYP2A6 [35]. Over 75% of nicotine metabolism in vivo is due to oxidation catalyzed by CYP2A6. It follows that nicotine metabolic capacity directly affects exposure to various poisons and tobacco carcinogens. In addition, despite the lack of literature on the association of NMR with PNs, NMR was found to be significantly associated with long-term smoking behavior and may serve as a biomarker for lung cancer [22,34–38]. Regarding the specific mechanisms involved in the results of this study, one possible explanation is that a fast metabolism of nicotine means that nicotine is cleared from the body quickly, reducing the chance of it binding to nicotinic receptors in lung cells, and thus decreasing the potential damage to lung cells [38]. And more relevant studies are needed in the future to explore the association and mechanisms between them further.

In this study, the linear association between OHCot and PNs was not statistically significant, and it was negatively weighted with PNs in the QgC model, which was inconsistent with the results of previous studies. In a cohort of workers, it was found that for every one-quantile increase in OHCot after the third percentile, Club cells secreted 3.30 ng/ml less protein, whose decreased concentration has been associated with lung disease [39]. The inconsistent results mentioned above may be due to the following reasons: first, the study needs to be combined with measurements of nicotine and its metabolites in urine and saliva; and second, some of the PNs were not caused by tobacco smoke, instead, due to scarring or inflammation left after lung infection.

The effects of smoking on different groups vary. A study has shown that adult smoking males have lower levels of Cot than females [40]. In this study, a subgroup analysis of male smokers was conducted and it was found that there was a negative correlation between OHCot and PNs. Another study found that female CYP2A6 diplotypes have a greater effect on tobacco metabolism [41], thus OHCot alone could not reflect true exposure to tobacco smoke. In addition, this study found that older smokers had similar findings to those of the general population and that the mixed effect of multiple tobacco substances in passive smoking had a positive effect on PNs, which are findings that have important implications for the development of prevention and treatment strategies.

There are some limitations of this study: first, the results are based on a single plasma sample, which may affect accuracy due to mishandling during sampling, and multiple samples are needed to ensure reliable results. Second, as a baseline cross-sectional study of a lung cancer screening cohort, it only revealed the association between smoking and PNs and did not address causality. Third, the study only measured three nicotine and its metabolites in plasma, which is not comprehensive enough to explore the exposure to tobacco smoke, multiple related substances in blood could be measured subsequently.

5. Conclusion

Results of this study indicate that the assessment of active smoking status based on actual plasma Cot levels is in better agreement with self-reported information, while passive smoking status assessment based on Cot was less consistent with self-report, which provides a reference for subsequent studies based on data related to this cohort. In addition, we used three measured substances (Nic, OHCot and Cot) and three combined metrics (TNE2, TNE3 and NMR) to reflect the true smoking dose of an individual and to explore the association with PNs. The results found that smoking was positively associated with the risk of developing PNs and that there was a significant positive effect of the mixed effect of OHCot, Cot and Nic in the QgC model on PNs. Our study provides clues to the relationship between tobacco smoke exposure and PNs, the mechanism of which needs to be further investigated.

Funding Statement

This study was supported by the project of Special Foundation for Science and Technology Development of Central Government Guiding Locals (202007d07050008) and the Scientific Research Project for Health Commission of Anhui Province (AHWJ2021a026).

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/17520363.2024.2422809

Financial disclosure

This study was supported by the project of Special Foundation for Science and Technology Development of Central Government Guiding Locals (202007d07050008) and the Scientific Research Project for Health Commission of Anhui Province (AHWJ2021a026). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Competing interests disclosure

The authors have no competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, stock ownership or options and expert testimony.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research

The authors state that they have obtained appropriate institutional review board approval (Ma'anshan Center for Disease Control and Prevention [Approval No. 2020001]) and/or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.