The association between urine cotinine level and hemoglobin and hematocrit levels: a cross-sectional study using the Korea National Health and Nutrition Examination Survey VIII (2019–2021)
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Smoking is a global health risk known to elevate hemoglobin (HB) levels through its effects on hematopoiesis. Urine cotinine, a metabolite strongly correlated with serum cotinine, serves as an effective biomarker for assessing smoking status. This study aimed to explore the relationship between urinary cotinine levels and both HB and hematocrit (HCT) levels in a Korean population.
Methods
The study analyzed 4,454 healthy participants, categorized into three groups based on urine cotinine tertiles. Steiger’s Z tests were used to assess correlations between HB, HCT levels, and urine cotinine. After adjusting for clinical variables, multiple linear regression was employed to evaluate the relationship between urine cotinine levels and HB/HCT. Receiver operating characteristic curves helped determine the cut-off values for urine cotinine in relation to HB and HCT levels.
Results
After adjusting for covariates, a positive correlation was found between urine cotinine and both HB and HCT levels. HB levels were 18% higher in the second tertile and 23% higher in the highest tertile than the lowest. Similarly, HCT levels increased by 44% in the second tertile and 50% in the highest tertile. The highest tertiles of HB and HCT had values of 504.650 and 202.950, respectively, with area under the curve values of 0.634 for HB and 0.616 for HCT.
Conclusion
This study demonstrates a significant correlation between urine cotinine levels and elevated HB and HCT levels in a representative Korean population. Clinicians should consider urine cotinine levels when assessing anemia in smokers or individuals exposed to secondhand smoke. Further research is needed to validate these findings.
Smoking remains a leading cause of death worldwide, with the global smoking rate still high and smoking-related mortality continuing to rise. In Korea, there were 67,982 deaths attributed to smoking in 2019 [1]. As smoking persists as a significant public health challenge, accurately assessing its somatic effects is crucial.
Smoking status is commonly assessed using questionnaires and urinary cotinine measurements. However, self-reported smoking may not reliably reflect an individual’s actual smoking status, especially among populations likely to omit information or those susceptible to memory-related biases. In contrast, cotinine, a metabolite of nicotine [2], offers greater reliability in assessing tobacco exposure, including secondhand smoke, compared to questionnaires. Additionally, urinary cotinine levels strongly correlate with serum cotinine levels, making it a suitable, noninvasive measure for evaluating smoking status, including smoking intensity [3].
Previous studies have shown that smoking can elevate hemoglobin (HB) levels through its effects on hematopoiesis [4-9]. The carbon monoxide in cigarette smoke binds more strongly to HB than oxygen, leading to hypoxemia [10,11]. As a compensatory mechanism, smokers typically exhibit higher HB levels than nonsmokers [8,12]. Consequently, anemia in smokers may be masked due to this adaptive increase in HB levels, and smoking status should be adjusted for and considered when evaluating anemia [13]
Accurately assessing smoking status is essential for improving clinical management and developing effective disease prevention strategies. However, few studies have explored the relationship between HB, hematocrit (HCT), and urinary cotinine, which can accurately measure smoking status. This study, therefore, investigates the relationship between urinary cotinine levels and HB and HCT levels using the Korea National Health and Nutrition Examination Survey (KNHANES) data, which provides a representative sample of Korean adults.
Methods
Data collection
Data for this study were gathered from the 8th KNHANES (KNHANES VIII), conducted between 2019 and 2021. This survey employed a complex, stratified, multistage probability sampling design, accounting for age, sex, and region to accurately represent the noninstitutionalized civilian Korean population. KNHANES VIII used a rolling sample survey method. A technical investigation team, including a nurse, a nutritionist, and a health science specialist, ensured reliable and consistent performance, minimizing bias in surveys and interviews.
Study population
This cross-sectional study utilized data from the 2019‒2021 KNHANES. KNHANES is a national survey that undergoes annual review and approval by the Research Ethics Deliberation Committee of the Korea Centers for Disease Control and Prevention (KCDC). Initially, 25,559 participants were included in the study, and individuals with a history of hematologic malignancy (n=10), a history of kidney disease (n=1,241), a creatinine level greater than 1.5 (n=2,508), anemia (n=434), and missing urine cotinine data (n=13,912) were excluded. Ultimately, 4,454 subjects were analyzed in this study (Figure 1). According to the World Health Organization diagnostic criteria, anemia was defined as HB <13 g/dL or HCT <40% in men and HB <12 g/dL or HCT <37% in women.
The study was performed in accordance with the Declaration of Helsinki, and all participants provided written informed consent. The Institutional Review Board (IRB) of Severance Hospital approved the study protocol (IRB no., 4-2024-0863).
Clinical and anthropometric data
Physical measurements of the participants were conducted by trained staff from the Division of Chronic Disease Surveillance under the KCDC and the Korean Ministry of Health and Welfare. Additionally, the survey collected information on exercise, smoking, and drinking habits. A current smoker was defined as a participant who had smoked more than 100 cigarettes in their lifetime and still smoked all types of cigarettes, including cigarette-type e-cigarettes and liquid-type e-cigarettes. A current drinker was defined as a participant who consumed alcohol once or more per month in the past year. Regular exercise was defined as ≥2.5 hours per week of moderate-intensity physical activity (moderate shortness of breath or moderate elevation of heart rate), ≥1 hour and 15 minutes of vigorous-intensity physical activity per week (rapid breathing and substantial elevation of heart rate), or mixed-intensity physical activity for the equivalent amount of time, with 1 minute of vigorous-intensity physical activity being equal to 2 minutes of moderate-intensity physical activity.
A standard mercury sphygmomanometer (Baumanometer) was used to measure blood pressure. Systolic and diastolic blood pressures were measured twice at 5-minute intervals, with the average values used for analysis. A bioelectrical impedance analyzer measured body weight and body composition (Inbody 720; Biospace). Each participant’s body mass index (BMI) was calculated by dividing their weight by the square of their height. Waist circumference (WC) was measured in a standing position at the midpoint between the lowest rib and the iliac crest. One trained professional obtained all measurements.
Biochemical analyses
Blood samples were collected from the antecubital vein of each participant after fasting for more than 8 hours and analyzed within 24 hours of collection. Serum levels of cholesterol, triglycerides (TG), high-density lipoprotein, low-density lipoprotein, and hemoglobin A1c were measured using a Hitachi Automatic Analyzer 7600 (Hitachi) with enzymatic methods and commercially available kits (Daiichi). Serum levels of white blood cells (WBC), HB, HCT, and platelets (PLT) were measured using a sodium lauryl sulphate HB (no cyanide) method with an XE-2100D instrument (Sysmex). Spot urine samples were collected from participants through a midstream void in the morning. Urine cotinine levels were determined by gas chromatography-mass spectrometry using the PerkinElmer Clarus 600T (PerkinElmer).
Statistical analysis
Continuous variables were described using means and standard deviations, while categorical variables were expressed as frequencies and percentages. Clinical characteristics across urine cotinine tertiles were compared using analysis of variance. Pearson chi-square test was used to compare proportions. The relationship between urine cotinine levels and HB and HCT was evaluated using Pearson correlation.
Multiple linear regression models were constructed to analyze the relationship between urine cotinine and clinical variables such as HB and HCT, with confounding factors selected based on clinical expertise. The final model was chosen using the forward selection method, in which predictors were sequentially added to a baseline model, starting with the most significant variables and continuing until no further improvements were observed. The linear regression results were presented as beta coefficients, 95% confidence intervals (CIs), and P-values.
Receiver operating characteristic (ROC) curves were used to calculate the area under the curve (AUC) and to identify the urine cotinine cut-off value for HB and HCT dominance. The optimal cut-off value was determined by selecting the point on the ROC curve that maximized both sensitivity and specificity, known as the Youden index. The standard error and 95% CI of the AUC were calculated using DeLong’s method.
Statistical significance was defined as a two-sided P-value of less than 0.05. All statistical analyses were conducted using R ver. 4.3.0 (R Foundation).
Results
Clinical characteristics of the participants
Table 1 shows the clinical characteristics of the participants, stratified by urine cotinine levels. Among the 4,454 participants, 59.1% had a smoking history, and 71.5% were men. The mean age of the participants was 46.3±16.7 years. As urine cotinine levels increased across tertiles, participants exhibited significantly higher WC (P<0.01), diastolic blood pressure (P<0.01), serum cholesterol levels (P<0.01), TG (P<0.01), LDL levels (P=0.03), and WBC counts (P<0.01). Additionally, the proportion of current smokers and alcohol consumers was significantly higher in groups with higher urinary cotinine tertiles (P<0.01).
Relationship between hemoglobin, hematocrit, and urine cotinine
The relationship between urine cotinine levels and various hematologic parameters was evaluated across different tertiles of urine cotinine concentration. As urinary cotinine levels increased, significant changes were observed in several hematologic parameters. Higher tertiles of urinary cotinine were associated with increased WBC count, HB levels, and HCT, all of which showed statistically significant associations with P-values less than 0.001. However, the association with PLT count was not statistically significant. (Table 2, Supplement 1).
Multiple linear regression analyses were used to assess the independent relationship between urine cotinine levels and HB and HCT, adjusting for various covariates, including age, sex, BMI, alcohol intake, WBC count, and hypertension. The results showed that higher urine cotinine levels were independently associated with increased HB and HCT levels after adjusting for confounding variables.
Specifically, compared to tertile 1, HB levels increased by 18% in tertile 2 and 23% in tertile 3 (Table 3). Similarly, HCT levels increased by 44% in tertile 2 and 50% in tertile 3 (Table 4). These findings demonstrate a positive relationship between urine cotinine levels and both HB and HCT (Figure 2).
Cut-off value of urine cotinine for high hemoglobin and hematocrit
Figure 3 illustrates the ROC curve analysis conducted to determine the optimal urine cotinine cut-off levels for predicting the highest tertiles of HB and HCT. The AUC for urine cotinine in predicting the highest tertile of HB was 0.634 (standard error [SE], 0.01; 95% CI, 0.616–0.653), while for HCT, the AUC was 0.616 (SE, 0.01; 95% CI, 0.596–0.635). The analysis identified an optimal urine cotinine cut-off value of 504.650 for predicting elevated HB levels, with a sensitivity of 74.5% and a specificity of 46.6%. For predicting higher HCT levels, the optimal cut-off value was 202.950, with a sensitivity of 82.8% and a specificity of 36.4%. These findings suggest that urine cotinine levels could serve as a predictive marker for identifying individuals with elevated HB and HCT levels.
Discussion
In this study, we analyzed HB and HCT levels based on the tertiles of urinary cotinine levels. A significant positive correlation was observed between urinary cotinine levels and both HB and HCT levels, with a progressive increase in both parameters across higher tertiles of urinary cotinine. After adjusting for covariates, individuals in the highest tertile of urinary cotinine levels had HB levels that were 23% higher and HCT levels that were 50% higher compared to those in the lowest tertile. These findings suggest that higher urinary cotinine levels are associated with increased HB and HCT levels. Additionally, the AUC value for HB was 0.634, and for HCT, it was 0.616, indicating a moderate discriminatory ability for predicting elevated HB and HCT levels.
The relationship between smoking and HB levels is well-documented, with substantial evidence indicating that smoking leads to an increase in HB level [9]. Previous studies suggest that to more accurately assess anemia in smokers during health or nutritional surveys, it is recommended to adjust HB levels according to the number of cigarettes smoked daily [9]. Specifically, an increase of 3 g/L should be considered for individuals who smoke 10 to 19 cigarettes per day, 5 g/L for those who smoke 20 to 39 cigarettes per day, and 7 g/L for those who smoke more than 40 cigarettes per day [9]. This adjustment is crucial, as anemia, which may be indicative of inflammatory or chronic diseases, malignant conditions, or iron deficiency due to poor nutritional status, could be masked by the elevated HB and HCT levels associated with smoking [14]. Therefore, it is essential to assess anemia while individually considering the smoking status of each patient.
A study by Nordenberg et al. [9] found that smokers had higher HB levels than never-smokers, with HB levels increasing with the number of cigarettes consumed per day. However, most previous studies on anemia in smokers have relied on smoking questionnaires to assess smoking status, which may not accurately reflect an individual’s true smoking habits and fails to account for long-term exposure to secondhand smoke. To address this limitation, our study used urine cotinine as a more accurate measure of tobacco smoke exposure, including secondhand smoke [15]. Urine cotinine is a more sensitive and reliable marker due to its higher concentration compared to plasma or saliva, making it easier to detect smoking exposure. Additionally, urine collection is both accurate and noninvasive [16,17].
Building on these findings, our study demonstrated that as urine cotinine levels increased, HB and HCT levels were significantly higher in the Korean population. This underscores the importance of using urine cotinine as a measure of tobacco smoke exposure, enabling more accurate adjustments in anemia assessments.
While the exact mechanism remains unclear, this association can be explained by several possible factors. Firstly, tobacco use is a major cause of chronic obstructive pulmonary disease (COPD), and secondhand smoke is also recognized as a risk factor for COPD [18]. COPD leads to progressive airflow limitation and destruction of the pulmonary capillary bed, resulting in hypoxemia due to ventilation/perfusion mismatch, which can cause compensatory secondary polycythemia [19,20]. However, considering that the mean age of participants in this study was 46 years and COPD is typically diagnosed in older individuals [21], another mechanism may be more relevant. Smoking introduces carbon monoxide into the body, which forms carboxyhemoglobin, reducing oxygen transport capacity and contributing to hypoxemia [22]. This hypoxemia may trigger increased red blood cell production as a compensatory response [23].
This study has some limitations. As a cross-sectional study, it is challenging to establish a definitive causal relationship between HB, HCT, and urinary cotinine levels. Moreover, lifestyle factors such as alcohol consumption and exercise were assessed using self-reported questionnaires, which may introduce individual biases or subjectivity. Additionally, the study did not consider the use of electronic cigarettes, which has become an increasingly important aspect of tobacco exposure. Lastly, the relatively small proportion of women in the participant pool (approximately 28.5%) limited our ability to analyze the data separately by gender. Despite these limitations, our study has several strengths. We recruited a large number of participants and adjusted for various variables, which allowed us to create a sample that was representative of the Korean population. Urine cotinine serves as a more sensitive and reliable marker of tobacco smoke exposure compared to plasma or saliva due to its higher concentration, making it easier to detect smoking exposure, including secondhand smoke [16,17]. Furthermore, urine collection is accurate, noninvasive, and straightforward. Finally, this is the first study to investigate and establish a correlation between HB, HCT, and urinary cotinine levels, significantly contributing to the fields of public health and epidemiology.
In conclusion, this study found a significant correlation between urinary cotinine levels and elevated HB and HCT levels in a Korean population. Clinicians should be cautious when evaluating anemia in smokers or individuals exposed to secondhand smoke, as urine cotinine levels may necessitate individualized adjustments to anemia criteria. However, further studies are needed to validate these findings and explore the long-term implications of tobacco exposure on hematologic health.
Notes
Conflict of interest
No potential conflict of interest relevant to this article was reported.
Funding
None.
Data availability
Data of this research are available from the corresponding author upon reasonable request.
Flow chart for enrollment. Spearman correlation was used to assess the relationship between the urine cotinine and clinical variables.
Figure. 2.
Relationship between urine cotinine and hemoglobin (A) and hematocrit (B). Steiger’s Z tests were applied to evaluate the differences in absolute correlation coefficients with hemoglobin and hematocrit levels and urine cotinine levels.
Figure. 3.
Receiver operating characteristic (ROC) curve analysis for predicting hemoglobin (A) and hematocrit (B) levels using urine cotinine. ROC-curve was performed for predicting hemoglobin and hematocrit levels using urine cotinine, assuming null hypothesis of no difference in area under the curve (AUC) at 0.5.
Table 1.
Clinical characteristics of participants according to urine cotinine level
Characteristic
Urine cotinine
Overall (n=4,454)
Tertile 1 (n=1,424)
Tertile 2 (n=1,477)
Tertile 3 (n=1,553)
P-value
Urine cotinine range
759.2 (0.5–4,910)
1.6 (0.5–187)
710.2 (187–1,160)
1,740 (1,160–4,910)
Age (y)
46.3±16.7
46.5±19.3
46.2±17.0
46.3±13.7
0.825
Sex
<0.001
Male
3,185 (71.5)
731 (51.3)
1,094 (74.1)
1,360 (87.6)
Female
1,269 (28.5)
693 (48.7)
383 (25.9)
193 (12.4)
Body mass index (kg/m2)
24.5±3.9
24.5±4.0
24.6±4.0
24.3±3.7
0.107
Waist circumference (cm)
86.0±10.8
84.9±11.6
86.6±10.8
86.4±10.1
<0.001
Hypertension (yes)
2,477 (58.6)
762 (59.5)
842 (59.0)
873 (57.5)
0.534
Systolic blood pressure (mm Hg)
119.1±14.9
119.3±15.4
119.1±14.8
119.1±14.4
0.933
Diastolic blood pressure (mm Hg)
76.6±10.1
75.5±9.9
76.6±10.2
77.6±10.2
<0.001
Exercise (yes)
1,767 (43.9)
560 (45.8)
611 (44.8)
596 (41.3)
0.045
Cholesterol (mg/dL)
191.4±38.1
188.6±37.2
190.7±37.8
194.8±39.1
<0.001
Triglyceride (mg/dL)
154.5±136.2
132.3±108.4
154.5±116.2
175.0±169.6
<0.001
High density lipoprotein (mg/dL)
50.6±12.7
52.5±12.6
50.6±12.6
48.8±12.6
<0.001
Low density lipoprotein (mg/dL)
113.1±34.1
111.4±32.5
112.2±34.5
115.5±35.1
0.003
Glycated hemoglobin (%)
5.8±0.9
5.8±0.9
5.8±0.9
5.8±0.8
0.374
Diabetes (yes)
2,363 (57.4)
712 (56.4)
777 (56.3)
874 (59.3)
0.197
Current smoker
2,510 (59.1)
156 (12.1)
1,050 (73.5)
1,304 (85.6)
<0.001
Amount of smoke (packday)
0.6±0.4
0.2±0.2
0.5±0.3
0.8±0.4
<0.001
Current drinker
3,008 (68.2)
790 (56.4)
1,065 (72.6)
1,153 (74.8)
<0.001
Urine cotinine (ng/dL)
900.5±879.1
18.2±40.3
699.0±280.5
1,901.1±621.5
<0.001
Values are presented as median (range), mean±standard deviation, or number (%).
Table 2.
Blood cell count based on urine cotinine tertile
Variable
Urine cotinine
Overall (n=4,454)
Tertile 1 (n=1,424)
Tertile 2 (n=1,477)
Tertile 3 (n=1,553)
P-value
WBC (103/μL)
6.8±1.9
6.2±1.6
6.9±1.8
7.3±2.0
<0.001
Hemoglobin (g/dL)
14.8±1.3
14.3±1.2
14.9±1.3
15.3±1.2
<0.001
Hematocrit (%)
44.6±3.6
43.2±3.4
44.9±3.6
45.7±3.3
<0.001
Platelet (103/μL)
261.8±60.2
259.7±59.5
261.6±60.9
264.0±60.0
0.151
Values are presented as mean±standard deviation.
WBC, white blood cell.
Table 3.
Multiple linear regression analysis to determine the relationship between urine cotinine and hemoglobin levels after adjusting clinical variables
Variable
Model 1 (reduced model)
Model 2
Model 3
Model 4
Model 5 (final model)
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Urine cotinine
Tertile 1
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Tertile 2
0.21 (0.14 to 0.29)
<0.001
0.22 (0.15 to 0.30)
<0.001
0.17 (0.10 to 0.24)
<0.001
0.17 (0.10 to 0.25)
<0.001
0.18 (0.10 to 0.25)
<0.001
Tertile 3
0.27 (0.19 to 0.34)
<0.001
0.30 (0.22 to 0.38)
<0.001
0.22 (0.14 to 0.30)
<0.001
0.23 (0.15 to 0.31)
<0.001
0.23 (0.16 to 0.31)
<0.001
Age (y)
–0.01 (–0.01 to –0.01)
<0.001
–0.01 (–0.01 to –0.01)
<0.001
–0.01 (–0.01 to –0.01)
<0.001
–0.01 (–0.01 to –0.01)
<0.001
–0.01 (–0.01 to –0.01)
<0.001
Female sex
–1.83 (–1.90 to –1.77)
<0.001
–1.79 (–1.85 to –1.72)
<0.001
–1.77 (–1.84 to –1.70)
<0.001
–1.75 (–1.82 to –1.68)
<0.001
–1.76 (–1.83 to –1.69)
<0.001
BMI
0.04 (0.03 to 0.05)
<0.001
0.04 (0.03 to 0.04)
<0.001
0.03 (0.02 to 0.04)
<0.001
0.03 (0.02 to 0.04)
<0.001
WBC
0.07 (0.06 to 0.09)
<0.001
0.07 (0.05 to 0.09)
<0.001
0.07 (0.05 to 0.09)
<0.001
Hypertension (yes)
0.15 (0.08 to 0.21)
<0.001
0.15 (0.09 to 0.21)
<0.001
Current drinker
–0.06 (–0.12 to 0.01)
0.084
CI, confidence interval; Ref, reference; BMI, body mass index, WBC, white blood cell.
Table 4.
Multiple linear regression analysis to determine the relationship between urine cotinine and hematocrit levels after adjusting clinical variables
Variable
Model 1 (reduced model)
Model 2
Model 3
Model 4
Model 5 (final model)
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Urine cotinine
Tertile 1
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Tertile 2
0.55 (0.33 to 0.76)
<0.001
0.35 (0.14 to 0.57)
0.001
0.41 (0.20 to 0.62)
<0.001
0.42 (0.20 to 0.63)
<0.001
0.44 (0.22 to 0.65)
<0.001
Tertile 3
0.62 (0.40 to 0.84)
<0.001
0.33 (0.11 to 0.56)
0.004
0.46 (0.24 to 0.68)
<0.001
0.48 (0.25 to 0.70)
<0.001
0.50 (0.27 to 0.72)
<0.001
Age (y)
–0.03 (–0.03 to –0.02)
<0.001
–0.03 (–0.03 to –0.02)
<0.001
–0.03 (–0.03 to –0.02)
<0.001
–0.03 (–0.04 to –0.02)
<0.001
–0.03 (–0.04 to –0.03)
<0.001
Female sex
–4.84 (–5.04 to –4.64)
<0.001
–4.76 (–4.95 to –4.56)
<0.001
–4.65 (–4.84 to –4.45)
<0.001
–4.60 (–4.79 to –4.40)
<0.001
–4.64 (–4.84 to –4.44)
<0.001
BMI
0.28 (0.23 to 0.33)
<0.001
0.24 (0.19 to 0.29)
<0.001
0.24 (0.19 to 0.28)
<0.001
0.23 (0.19 to 0.28)
<0.001
WBC
0.11 (0.08 to 0.13)
<0.001
0.10 (0.07 to 0.12)
<0.001
0.10 (0.07 to 0.12)
<0.001
Hypertension (yes)
0.31 (0.13 to 0.49)
<0.001
0.33 (0.15 to 0.52)
<0.001
Current drinker
–0.24 (–0.43 to –0.05)
0.013
CI, confidence interval; Ref, reference; BMI, body mass index, WBC, white blood cell.
References
1. Yang YS, Jung KJ, Kimm H, Lee S, Jee SH. Smoking-attributable mortality among Korean adults in 2019. Epidemiol Health 2024;46:e2024011.
3. Muranaka H, Higashi E, Itani S, Shimizu Y. Evaluation of nicotine, cotinine, thiocyanate, carboxyhemoglobin, and expired carbon monoxide as biochemical tobacco smoke uptake parameters. Int Arch Occup Environ Health 1988;60:37-41.
7. Stewart RD, Baretta ED, Platte LR, Stewart EB, Kalbfleisch JH, Van Yserloo B, et al. Carboxyhemoglobin levels in American blood donors. JAMA 1974;229:1187-95.
13. Braverman Bronstein A, Lomelin Gascon J, Eugenio Gonzalez CI, Barrientos-Gutierrez T. Environmental tobacco exposure and urinary cotinine levels in smoking and nonsmoking adolescents. Nicotine Tob Res 2018;20:523-6.
15. Jarvis MJ, Tunstall-Pedoe H, Feyerabend C, Vesey C, Saloojee Y. Comparison of tests used to distinguish smokers from nonsmokers. Am J Public Health 1987;77:1435-8.
16. Benowitz NL, Dains KM, Dempsey D, Herrera B, Yu L, Jacob P 3rd. Urine nicotine metabolite concentrations in relation to plasma cotinine during low-level nicotine exposure. Nicotine Tob Res 2009;11:954-60.
17. Thomas CE, Wang R, Adams-Haduch J, Murphy SE, Ueland PM, Midttun O, et al. Urinary cotinine is as good a biomarker as serum cotinine for cigarette smoking exposure and lung cancer risk prediction. Cancer Epidemiol Biomarkers Prev 2020;29:127-32.
18. Laniado-Laborin R. Smoking and chronic obstructive pulmonary disease (COPD): parallel epidemics of the 21 century. Int J Environ Res Public Health 2009;6:209-24.
19. Kent BD, Mitchell PD, McNicholas WT. Hypoxemia in patients with COPD: cause, effects, and disease progression. Int J Chron Obstruct Pulmon Dis 2011;6:199-208.
20. Sharma AJ, Addo OY, Mei Z, Suchdev PS. Reexamination of hemoglobin adjustments to define anemia: altitude and smoking. Ann N Y Acad Sci 2019;1450:190-203.
21. Holm KE, Plaufcan MR, Ford DW, Sandhaus RA, Strand M, Strange C, et al. The impact of age on outcomes in chronic obstructive pulmonary disease differs by relationship status. J Behav Med 2014;37:654-63.
The association between urine cotinine level and hemoglobin and hematocrit levels: a cross-sectional study using the Korea National Health and Nutrition Examination Survey VIII (2019–2021)
Figure. 1. Flow chart for enrollment. Spearman correlation was used to assess the relationship between the urine cotinine and clinical variables.
Figure. 2. Relationship between urine cotinine and hemoglobin (A) and hematocrit (B). Steiger’s Z tests were applied to evaluate the differences in absolute correlation coefficients with hemoglobin and hematocrit levels and urine cotinine levels.
Figure. 3. Receiver operating characteristic (ROC) curve analysis for predicting hemoglobin (A) and hematocrit (B) levels using urine cotinine. ROC-curve was performed for predicting hemoglobin and hematocrit levels using urine cotinine, assuming null hypothesis of no difference in area under the curve (AUC) at 0.5.
Graphical abstract
Figure. 1.
Figure. 2.
Figure. 3.
Graphical abstract
The association between urine cotinine level and hemoglobin and hematocrit levels: a cross-sectional study using the Korea National Health and Nutrition Examination Survey VIII (2019–2021)
Characteristic
Urine cotinine
Overall (n=4,454)
Tertile 1 (n=1,424)
Tertile 2 (n=1,477)
Tertile 3 (n=1,553)
P-value
Urine cotinine range
759.2 (0.5–4,910)
1.6 (0.5–187)
710.2 (187–1,160)
1,740 (1,160–4,910)
Age (y)
46.3±16.7
46.5±19.3
46.2±17.0
46.3±13.7
0.825
Sex
<0.001
Male
3,185 (71.5)
731 (51.3)
1,094 (74.1)
1,360 (87.6)
Female
1,269 (28.5)
693 (48.7)
383 (25.9)
193 (12.4)
Body mass index (kg/m2)
24.5±3.9
24.5±4.0
24.6±4.0
24.3±3.7
0.107
Waist circumference (cm)
86.0±10.8
84.9±11.6
86.6±10.8
86.4±10.1
<0.001
Hypertension (yes)
2,477 (58.6)
762 (59.5)
842 (59.0)
873 (57.5)
0.534
Systolic blood pressure (mm Hg)
119.1±14.9
119.3±15.4
119.1±14.8
119.1±14.4
0.933
Diastolic blood pressure (mm Hg)
76.6±10.1
75.5±9.9
76.6±10.2
77.6±10.2
<0.001
Exercise (yes)
1,767 (43.9)
560 (45.8)
611 (44.8)
596 (41.3)
0.045
Cholesterol (mg/dL)
191.4±38.1
188.6±37.2
190.7±37.8
194.8±39.1
<0.001
Triglyceride (mg/dL)
154.5±136.2
132.3±108.4
154.5±116.2
175.0±169.6
<0.001
High density lipoprotein (mg/dL)
50.6±12.7
52.5±12.6
50.6±12.6
48.8±12.6
<0.001
Low density lipoprotein (mg/dL)
113.1±34.1
111.4±32.5
112.2±34.5
115.5±35.1
0.003
Glycated hemoglobin (%)
5.8±0.9
5.8±0.9
5.8±0.9
5.8±0.8
0.374
Diabetes (yes)
2,363 (57.4)
712 (56.4)
777 (56.3)
874 (59.3)
0.197
Current smoker
2,510 (59.1)
156 (12.1)
1,050 (73.5)
1,304 (85.6)
<0.001
Amount of smoke (packday)
0.6±0.4
0.2±0.2
0.5±0.3
0.8±0.4
<0.001
Current drinker
3,008 (68.2)
790 (56.4)
1,065 (72.6)
1,153 (74.8)
<0.001
Urine cotinine (ng/dL)
900.5±879.1
18.2±40.3
699.0±280.5
1,901.1±621.5
<0.001
Variable
Urine cotinine
Overall (n=4,454)
Tertile 1 (n=1,424)
Tertile 2 (n=1,477)
Tertile 3 (n=1,553)
P-value
WBC (103/μL)
6.8±1.9
6.2±1.6
6.9±1.8
7.3±2.0
<0.001
Hemoglobin (g/dL)
14.8±1.3
14.3±1.2
14.9±1.3
15.3±1.2
<0.001
Hematocrit (%)
44.6±3.6
43.2±3.4
44.9±3.6
45.7±3.3
<0.001
Platelet (103/μL)
261.8±60.2
259.7±59.5
261.6±60.9
264.0±60.0
0.151
Variable
Model 1 (reduced model)
Model 2
Model 3
Model 4
Model 5 (final model)
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Urine cotinine
Tertile 1
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Tertile 2
0.21 (0.14 to 0.29)
<0.001
0.22 (0.15 to 0.30)
<0.001
0.17 (0.10 to 0.24)
<0.001
0.17 (0.10 to 0.25)
<0.001
0.18 (0.10 to 0.25)
<0.001
Tertile 3
0.27 (0.19 to 0.34)
<0.001
0.30 (0.22 to 0.38)
<0.001
0.22 (0.14 to 0.30)
<0.001
0.23 (0.15 to 0.31)
<0.001
0.23 (0.16 to 0.31)
<0.001
Age (y)
–0.01 (–0.01 to –0.01)
<0.001
–0.01 (–0.01 to –0.01)
<0.001
–0.01 (–0.01 to –0.01)
<0.001
–0.01 (–0.01 to –0.01)
<0.001
–0.01 (–0.01 to –0.01)
<0.001
Female sex
–1.83 (–1.90 to –1.77)
<0.001
–1.79 (–1.85 to –1.72)
<0.001
–1.77 (–1.84 to –1.70)
<0.001
–1.75 (–1.82 to –1.68)
<0.001
–1.76 (–1.83 to –1.69)
<0.001
BMI
0.04 (0.03 to 0.05)
<0.001
0.04 (0.03 to 0.04)
<0.001
0.03 (0.02 to 0.04)
<0.001
0.03 (0.02 to 0.04)
<0.001
WBC
0.07 (0.06 to 0.09)
<0.001
0.07 (0.05 to 0.09)
<0.001
0.07 (0.05 to 0.09)
<0.001
Hypertension (yes)
0.15 (0.08 to 0.21)
<0.001
0.15 (0.09 to 0.21)
<0.001
Current drinker
–0.06 (–0.12 to 0.01)
0.084
Variable
Model 1 (reduced model)
Model 2
Model 3
Model 4
Model 5 (final model)
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Estimate (95% CI)
P-value
Urine cotinine
Tertile 1
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Ref
Tertile 2
0.55 (0.33 to 0.76)
<0.001
0.35 (0.14 to 0.57)
0.001
0.41 (0.20 to 0.62)
<0.001
0.42 (0.20 to 0.63)
<0.001
0.44 (0.22 to 0.65)
<0.001
Tertile 3
0.62 (0.40 to 0.84)
<0.001
0.33 (0.11 to 0.56)
0.004
0.46 (0.24 to 0.68)
<0.001
0.48 (0.25 to 0.70)
<0.001
0.50 (0.27 to 0.72)
<0.001
Age (y)
–0.03 (–0.03 to –0.02)
<0.001
–0.03 (–0.03 to –0.02)
<0.001
–0.03 (–0.03 to –0.02)
<0.001
–0.03 (–0.04 to –0.02)
<0.001
–0.03 (–0.04 to –0.03)
<0.001
Female sex
–4.84 (–5.04 to –4.64)
<0.001
–4.76 (–4.95 to –4.56)
<0.001
–4.65 (–4.84 to –4.45)
<0.001
–4.60 (–4.79 to –4.40)
<0.001
–4.64 (–4.84 to –4.44)
<0.001
BMI
0.28 (0.23 to 0.33)
<0.001
0.24 (0.19 to 0.29)
<0.001
0.24 (0.19 to 0.28)
<0.001
0.23 (0.19 to 0.28)
<0.001
WBC
0.11 (0.08 to 0.13)
<0.001
0.10 (0.07 to 0.12)
<0.001
0.10 (0.07 to 0.12)
<0.001
Hypertension (yes)
0.31 (0.13 to 0.49)
<0.001
0.33 (0.15 to 0.52)
<0.001
Current drinker
–0.24 (–0.43 to –0.05)
0.013
Table 1. Clinical characteristics of participants according to urine cotinine level
Values are presented as median (range), mean±standard deviation, or number (%).
Table 2. Blood cell count based on urine cotinine tertile
Values are presented as mean±standard deviation.
WBC, white blood cell.
Table 3. Multiple linear regression analysis to determine the relationship between urine cotinine and hemoglobin levels after adjusting clinical variables
CI, confidence interval; Ref, reference; BMI, body mass index, WBC, white blood cell.
Table 4. Multiple linear regression analysis to determine the relationship between urine cotinine and hematocrit levels after adjusting clinical variables
CI, confidence interval; Ref, reference; BMI, body mass index, WBC, white blood cell.
