Effect of Second-Hand Smoke Exposure on Establishing Urinary Cotinine-Based Optimal Cut-Off Values for Smoking Status Classification in Korean Adults

Regulations for banning smoking in indoor public places and workplaces have increased worldwide in recent years. A consecutive Korean National Health and Nutrition Examination Survey (KNHANES) between 2008 and 2018 showed a trend toward significant decreases in self-reported tobacco smoke exposure and measured urinary cotinine concentrations. We established and compared each optimal cut-off value for assessing the effect of second-hand smoke (SHS) exposure on establishing urinary cotinine-based cut-off values for smoking status classification in a population setting controlled for racial and cultural diversity, using four KNHANES datasets consisting of the 2008, 2011, 2014, and 2018 surveys. A total of 18,229 Korean participants aged >19 years with measured urinary cotinine concentrations were enrolled. Self-reports of current smoking status showed that the prevalence of current smokers decreased from 22.9% to 18.2% between 2008 and 2018. During this period, the median value of urinary cotinine in nonsmokers decreased from 5.86 µg/L to 0.48 µg/L, whereas the median value showed no remarkable decrease in current smokers. The AUC-based optimal cut-off values of urinary cotinine concentration for distinguishing current smokers from nonsmokers decreased from 86.5 µg/L to 11.5 µg/L. Our study showed that decreased SHS exposure would result in decreased optimal cut-off values for distinguishing current smokers from nonsmokers. In addition, the study suggests that the range of urinary cotinine concentration to define SHS exposure for the trend monitoring of populationof SHS exposure is appropriate between 0.30 µg/L and 100 µg/L. In addition, our study showed the importance of determination of cotinine concentration, which would have allowed us to avoid mistakes in qualification to the study group in an increased use of e-cigarette setting.


Introduction
Tobacco smoking is a common public health issue and a cause of preventable morbidity and mortality worldwide [1]. Smoking or exposure to second-hand smoke (SHS) is associated with the risk of asthma, chronic obstructive lung disease, respiratory tract infection, and various cancers [2][3][4][5]. Regulations and social consensus for banning smoking in indoor public places and workplaces are becoming ubiquitous in recent years [6][7][8][9]. The consecutive Korean National Health and Nutrition Examination Survey (KNHANES) between 2008 and 2018 showed a trend toward significant decreases in self-reported tobacco smoke exposure and measured urinary cotinine concentrations [10][11][12]. Cotinine is a major metabolite of nicotine, and it is used as a biomarker for tobacco smoke exposure [12]. Urinary cotinine is a noninvasive biomarker, and it has comparable diagnostic performance for tobacco smoke exposure to serum cotinine [13,14]. Therefore, urinary cotinine has been used in KNHANES for many years to validate reported smoking status and to monitor population exposure to tobacco over time [15]. cotinine has been used in KNHANES for many years to validate reported smoking status and to monitor population exposure to tobacco over time [15].
The cut-off values of urinary cotinine concentration, established for smoking status classification, might be affected by various factors such as sex, age, pregnancy, nicotinecontaining food intake, ethnic variation of cotinine half-life and nicotine metabolism, and SHS exposure [16][17][18][19]. Therefore, cut-off values of urinary cotinine concentration established by previous studies varied from 31.5 µg/L to 550 µg/L [18,[20][21][22][23][24][25]. Although recent studies have reported that the cut-off values of cotinine have decreased over the past 20 years [25], the generalization of these trends is still limited without controlling the heterogeneity of racial and cultural factors related to nicotine absorption and metabolism.
In the current study, we established and compared each area under the curve (AUC)based optimal cut-off value of urinary cotinine concentration using four KNHANES datasets from 2008, 2011, 2014, and 2018. We aimed to assess the effect of a declining trend in SHS exposure on establishing urinary cotinine-based cut-off values for smoking status classification during the recent decade in a controlled population setting with racial and cultural diversity.  Table S1), 7595 participants aged <19 years were excluded. Of 26,209 adult participants, we excluded 7980 for whom urinary cotinine concentrations were not measured. In total, 18,229 Korean participants with measured urinary cotinine concentrations, aged ≥19 years, were enrolled in the current study. This study was approved by the Institutional Review Board of Wonkwang University Hospital (IRB file no. 2022-03-027-001). A waiver of consent was obtained given the nature of the project, which aimed to establish a cut-off value using a public dataset.

Self-Report for Smoking Status
In the current study, no response to smoking status was defined as a participant who did not participate in the self-report for smoking status or did not answer the selfreport among all enrolled participants. Current smokers were defined as participants who reported a history of smoking 100 cigarettes in their lifetime and reported currently smoking cigarettes. A daily smoker was defined as a participant who reported, "Yes, I smoke at least one cigarette a day". among current smokers. Non-daily smokers were defined as participants who reported "Yes, I smoke, but not every day". among current smokers. Nonsmokers were defined as participants who did not meet the current smoking definition among participants for self-reported smoking status. The questionnaires for smoking status and data processing in the current study are described in Supplementary  Tables S2 and S3. 2.3. Self-Report for SHS Exposure In the current study, SHS exposure was defined as a participant who reported a history of SHS exposure at home or in the workplace among nonsmokers. An unclear response to SHS exposure was defined as a participant who is not included in SHS exposure and who reported an unclear history of SHS exposure, such as no response, unknown response, and non-defined response, at home or in the workplace among nonsmokers. No SHS exposure was defined as a participant who definitively reported no history of SHS exposure at home and in the workplace among nonsmokers. The questionnaires for SHS exposure and data processing in the current study are described in Supplementary Tables S4 and S5.

Self-Report for the Use of e-Cigarettes
In the current study, a current e-cigarette user was defined as a participant who reported a history of e-cigarette smoking within 30 days. A non-e-cigarette user was defined as a participant who definitively reported no history of e-cigarette smoking within 30 days. The questionnaires for the use of e-cigarettes and data processing in the current study are described in Supplementary Table S6.

Measurement of Urinary Cotinine Concentration
Measurement methods of urinary cotinine concentration and the diagnostic performance of the methods, including the limit of detection (LoD) and the limit of quantitation (LoQ), are described in Supplementary Table S7. Gas chromatography-tandem mass spectrometry (GC-MS/MS) with a Perkin Elmer Clarus 600T instrument (Perkin Elmer, Turku, Finland) was used in 2008, 2011, and 2014 to measure urinary cotinine concentrations. LoD of the GC-MS/MS method varied from 0.25 µg/L to 0.27 µg/L, and LoQ varied from 0.75 µg/L to 0.82 µg/L, between 2008 and 2014. In 2018, a high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) system comprising an 1100 HPLC system (Agilent, Santa Clara, CA, USA) to API 4000 (AB Sciex, Redwood City, CA, USA) was used to measure urinary cotinine concentrations. LoD and LoQ of HPLC-MS/MS method were 0.27 µg/L and 0.31 µg/L, respectively.

Statistical Analysis
To reduce the impact of methodological differences on LoD in the current study, values equal to or lower than 0.30 µg/L were converted to 0.30 µg/L. Histograms of urinary cotinine concentration are expressed as log10 urinary cotinine (µg/L) and percentage frequency. Data with normal distribution are expressed as mean ± SD, while skewed data are expressed as median (interquartile range, IQR). However, some urinary cotinine concentration statistics associated with smoking status are intentionally expressed as mean ± SD because tailed portions in the distribution are important to establish cut-off values of urinary cotinine concentration for smoking status classification. The AUC-receiver operating characteristic (ROC) curve analysis with the maximum value of Youden's index was used to establish optimal cut-off values of urinary cotinine concentration for smoking status classification. Fisher's exact test or the chi-square test was used to analyze categorical data. The Mann-Whitney U test or Kruskal-Wallis test was used to analyze continuous data. The Statistical Package for the Social Sciences version 25.0 (IBM Corporation, Armonk, NY, USA) and the GraphPad Prism (version 9.1.2.; GraphPad Software, La Jolla, CA, USA) were used for statistical analyses and graphs. A p value of less than 0.05 was considered significant.

The Characteristics of Study Participants
The characteristics of the study participants are described in Table 1

The Result of the Self-Reported Smoking Status
The results of the self-reported smoking status in the current study are described in Tables 1 and 2

The Result of the Self-Reported SHS Exposure
The results of the self-reported SHS exposure in the current study are described in Table 3

The Result of the Self-Reported Use of e-Cigarettes
The results of the self-reported use of e-cigarettes in the current study are described in Tables 1 and 4 Table 4. The comparison between self-report for use of e-cigarettes and urinary cotinine concentration in self-report participants.

The Distributions of Urinary Cotinine Concentration from All Participants
The distribution of urinary cotinine concentration in all participants is illustrated in Figure 2. Trimodal distribution of urinary cotinine concentration was observed based on a visual inspection of the histogram. The first subgroup was located at 0.30 µg/L. This subgroup was prominent in 2008 and 2018, whereas it decreased in 2011 and 2014. The second subgroup was located at various positions from 0.30 µg/L to around 100 µg/L in each survey, and this subgroup shifted toward 0.30 µg/L during the last decade. In 2018, the second subgroup showed a merged pattern with the first subgroup. The third subgroup started at approximately 100 µg/L and was most frequently observed at approximately 1500 µg/L. The third subgroup did not show markedly different locations or patterns among the four surveys.

The Distributions of Urinary Cotinine Concentration in Nonsmokers
The distribution of urinary cotinine concentration in nonsmokers is illustrated in

The Distributions of Urinary Cotinine Concentration in Nonsmokers
The distribution of urinary cotinine concentration in nonsmokers is illustrated in

The Established Optimal Cut-Off Values of Urinary Cotinine Concentration for Smoking Status Classification
The established optimal cut-off values of urinary cotinine concentration for smoking status classification are described in Table 5  Only the optimal cut-off value of urinary cotinine concentration for distinguishing current smokers from nonsmokers was established in the 2008 KNHANES dataset. The diagnostic performance for each optimal cut-off varied from 93.00% to 98.85% for sensitivity and 94.06% to 96.29% for specificity.

The Established Optimal Cut-Off Values of Urinary Cotinine Concentration for Smoking Status Classification
The established optimal cut-off values of urinary cotinine concentration for smoking status classification are described in Table 5 and illustrated in in Figure 2. According to smoking status, two cut-off values were established to distinguish current smokers from nonsmokers and to distinguish daily smokers from non-daily smokers. The opti- Only the optimal cut-off value of urinary cotinine concentration for distinguishing current smokers from nonsmokers was established in the 2008 KNHANES dataset. The diagnostic performance for each optimal cut-off varied from 93.00% to 98.85% for sensitivity and 94.06% to 96.29% for specificity.

Comparison of Diagnostic Performance According to Various Urinary Cotinine Cut-Off Values for Smoking Status Classification
A comparison of diagnostic performance according to various cut-off values of urinary cotinine concentration for smoking status classification is presented in Table 6 and illustrated in Figure 4. When using 100 µg/L, 50 µg/L, 25µg/L, and 12 µg/L as cut-off values for distinguishing current smokers from nonsmokers, sensitivity and negative predictive value (NPV) were relatively similar among the four datasets. As the cut-off value, 12 µg/L showed the best performance for sensitivity (96.06% to 98.91%) and NPV (98.54% to 99.71%), whereas 100 µg/L showed the worst performance for sensitivity (93.22% to 96.85%) and NPV (97.71% to 99.28%). In contrast, the specificity and positive predictive value (PPV) were significantly different according to the survey. As the cut-off value, 12 µg/L showed the worst performance for specificity and PPV, whereas 100 µg/L showed the best performance. According to a visual inspection of the scatter plot of urinary cotinine concentration, the distribution patterns in current smokers were relatively similar among the four datasets. Most current smokers showed values greater than 100 µg/L of urinary cotinine, and only 2.0% (20/1047, in 2018) to 2.9% (13/457, in 2011) of current smokers showed values ranging from 12 µg/L to 100 µg/L in all datasets (Table 7). However, the distribution of nonsmokers showed a markedly decreased pattern during the recent decade. The percent of nonsmokers with values in the 12 µg/L to 100 µg/L range decreased from 31.5%    According to a visual inspection of the scatter plot of urinary cotinine concentration, the distribution patterns in current smokers were relatively similar among the four datasets. Most current smokers showed values greater than 100 µ g/L of urinary cotinine, and only 2.0% (20/1047, in 2018) to 2.9% (13/457, in 2011) of current smokers showed values ranging from 12 µ g/L to 100 µ g/L in all datasets (Table 7). However, the distribution of nonsmokers showed a markedly decreased pattern during the recent decade. The percent

Comparison between Self-Reported SHS Exposure and Urinary Cotinine Concentrations in Nonsmokers
A comparison between self-reported SHS exposure and urinary cotinine concentrations in nonsmokers is presented in Table 3

Discussion
To define SHS exposure, the Centers for Disease Control and Prevention in the United States have used the range of serum cotinine between 0.05 µg/L and 10 µg/L [26]. A serum cotinine value of 0.05 µg/L is equal to the previous LoD of the LC-MS/MS method, and 10 µg/L is widely accepted as the cut-off value for defining smokers [26]. However, there is no consensus on the cut-off value of urinary cotinine concentration to define SHS exposure. A previous study using KNHANES datasets between 2007 and 2010 asserted that urinary cotinine concentration did not distinguish SHS exposure from nonsmokers because the half-life of cotinine is less than 24 h [15]. However, this conclusion might be misleading due to limitations in self-reporting for SHS exposure. Self-report of SHS exposure is subjective and less correlated with the quantitative values of urinary cotinine concentration. Different proportions of self-reported responses to SHS exposure and the proportion of unclear SHS exposure between the surveys may bias the correlation between self-reported SHS exposure and urinary cotinine values in nonsmokers. Our data clearly showed the limitations of selfreporting SHS exposure (Table 3). For example, the no-SHS-exposure participants among the participants with equal to or less than 0.30 µg/L in 2008 were 69.6% (780/1121). This is similar to the 71.2% (79/111) of the no SHS exposure participants among the participants with equal to or less than 0.30 µg/L in 2011, although the proportion of the participants with equal to or less than 0.30 µg/L differed significantly (25.8% [1121/4340] vs. 8.1% [111/1375], p < 0.0001). Furthermore, 28.5% (320/1121) of participants with equal to or less than 0.30 µg/L in 2008 reported SHS exposure, whereas most participants with more than 0.30 µg/L in 2018 reported no SHS exposure. Above all, the median value of urinary cotinine concentration from no-SHS-exposure participants in 2008 was significantly higher than that of SHS exposure participants, in 2011, 2014, and 2018 KNHANES. Therefore, the cut-off value of urinary cotinine concentration to define SHS exposure should be established based on the distribution of urinary cotinine concentrations rather than self-reports. Our study indicated that a 0.30 µg/L urinary cotinine value, which is close to the current LoD of GC-MS/MS or HPLC MS/MS, is the cut-off value for distinguishing SHS exposure from nonsmokers in the Korean adult population.
During the last decade, the use of novel forms of smoking, such as electronic cigarettes, heated tobacco products, little cigars, and cigarillos, has rapidly increased among the public. However, the urinary cotinine concentration distribution in current smokers does not appear to have changed markedly during this period in the current study (Table 7 and Figure 4). A study of urinary cotinine concentrations in e-cigarette users reported that e-cigarette users are exposed to the same nicotine levels as cigarette smokers [27], and the results of our study are in line with those of a previous study (Table 4, Supplementary  Figures S1 and S2). Therefore, the optimal cut-off values for distinguishing daily smokers from non-daily smokers did not change markedly at approximately 100 µg/L in each survey. The range of urinary cotinine concentration between 0.30 µg/L and 100 µg/L appears to be suitable for monitoring the trend of SHS exposure. These cut-off points, consisting of the LoD value for urinary cotinine concentration measurement and the cut-off value to define daily smokers, are similar to the characteristics of cut-point values of serum cotinine for SHS exposure, as defined by the Centers for Disease Control and Prevention in the United States.
Our study also showed that the optimal cut-off value for distinguishing current smokers from nonsmokers changed mainly because of the continuous quantitative decline of urinary cotinine concentration in participants with SHS exposure. Among the nonsmokers, the number of participants with the range of urinary cotinine concentration between 12 µg/L and 100 µg/L significantly decreased from 31.5% (1369/4340) in 2008 to 0.7% (27/4692) in 2018, whereas the current smokers with that range were not different, with 2.2% (26/1287) in 2008 and 2.0% (20/1047) in 2018 ( Table 7). As a result, the optimal cut-off value for differentiating non-daily smokers and participants with SHS exposure decreased from 86.48 µg/L in 2008 to 11.50 µg/L in 2018. The optimal cut-off value of 11.50 µg/L for distinguishing current smokers from nonsmokers for the 2018 dataset showed equal or superior diagnostic performance compared to the optimal cut-off value of 86.48 µg/L for 2008 dataset (sensitivity 98.85% vs. 97.20%, specificity 95.80% vs. 94.06%; Table 5).
A transference of the cut-off value is generally unacceptable when there are significant geographic or demographic differences between the populations that are known to cause a difference in the cut-off value [28]. Furthermore, our study showed that the optimal cut-off value for distinguishing current smokers from nonsmokers could change according to the quantitative alteration of urinary cotinine concentration in nonsmokers, suggesting the intensity or frequency of SHS exposure, even in a controlled population with racial and cultural diversity. Therefore, it is ideal to investigate and analyze the pattern and characteristics of SHS exposure in different populations and then establish the population's own cut-off value rather than a transference of the cut-off value. If transference of the cut-off value is the only option without a survey for SHS exposure, 100 µg/L might be recommended rather than other lower values of urinary cotinine concentration as the cut-off value for distinguishing current smokers from nonsmokers. Although improved sensitivity and NPV would be helpful for the individual management of smoking-related issues, cut-off values lower than 100 µg/L of urinary cotinine concentration may cause significantly lower specificities and PPVs in a high-SHS-exposure setting such as the 2008 KNHANES (Table 6).
Our study has some limitations. First, the smoking status verification was not performed by determination of cotinine concentration in serum, which would have allowed us to avoid mistakes in qualification to the study group. Next, several smoking statuses were ignored in the current study, such as former smokers, users of novel forms of smoking or nicotine alternatives, and participants undergoing nicotine replacement therapy. Of current e-cigarette users, 22% (14/65, in 2014) and 12% (19/163, in 2018) had self-reported as nonsmokers, and 0.5% (17/3754, in 2014) and 0.4% (19/4692, in 2018) of misclassified nonsmokers were identified, respectively. However, between 3.6% and 5.5% of suspected nonsmokers with 100 µg/L or higher value of urinary cotinine were not excluded in each survey (Table 7), because questionnaires for use of novel forms of smoking or nicotine alternatives were only included in two surveys, of the total surveys. These limitations might lead to a minor bias for statistics and established cut-off values, although they could not affect trends toward significant declines in measured urinary cotinine in nonsmokers and a decline in the optimal cut-off value for distinguishing current smokers from nonsmokers.

Conclusions
In conclusion, our study showed that decreased SHS exposure mainly resulted in decreased optimal cut-off values for distinguishing current smokers from nonsmokers in a controlled Korean population with racial and cultural diversity. Moreover, the current study suggests that the range of urinary cotinine concentration to define SHS exposure for population-monitoring the trend of SHS exposure is appropriate between 0.30 µg/L and 100 µg/L. In addition, our study showed the importance of determination of cotinine concentration, which would have allowed us to avoid mistakes in qualification to the study group in an increased use of e-cigarette setting.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board of Wonkwang University Hospital (IRB file no. 2022-03-027-001).

Informed Consent Statement:
A waiver of consent was obtained given the nature of the project, which aimed to establish a cut-off value using a public dataset.

Data Availability Statement:
The datasets used in this study were obtained from the Korean National Health and Nutrition Examination Survey (KNHANES) between 2008 and 2018. The dataset recruitment process is detailed in the Supplementary Table S1.

Conflicts of Interest:
The authors declare no conflict of interest.