Urine cotinine is used widely as a noninvasive marker of nicotine exposure in smokers as well as in nonsmokers who are exposed to SHS (N. L. Benowitz, 1996
; U.S. Department of Health and Human Services, 2006
). Urine measurement has an advantage over plasma cotinine in that concentrations in urine are four- to fivefold higher and easier to measure in most laboratories, and urine collection is less invasive than blood sampling.
We simulated nicotine exposure from SHS by administering deuterium-labeled nicotine orally in low doses at regular intervals throughout the day. Nicotine taken orally undergoes hepatic first-pass metabolism, so that the ratio of nicotine to cotinine in plasma is lower than when nicotine is inhaled in tobacco smoke. In contrast to nicotine, cotinine undergoes little or no first-pass metabolism (Zevin, Jacob, Geppetti, & Benowitz, 2000
). Therefore, cotinine and cotinine metabolite measurements would not be expected to differ after oral administration compared with after inhalation of nicotine.
We found that, on average (across all dosing groups), approximately 1 ng/ml cotinine in plasma at steady state corresponds to a dose of 100 μg nicotine per day. Because considerable individual variability was observed in plasma cotinine levels at any particular dose of nicotine (see ), the use of plasma cotinine to estimate daily nicotine intake for individuals is only a rough estimate. Urine cotinine concentrations were highly correlated with plasma cotinine concentrations. The extent of correlation was improved substantially by a correction of urine cotinine concentration for urine creatinine. This finding differs from those of Thompson et al. (1990)
in smokers. They found no improvement in the correlation between urine cotinine and plasma cotinine concentrations with simple creatinine correction, although they did improve the correlation when they corrected for creatinine using a regression equation.
The strength of the correlation between urine and plasma cotinine concentration varied somewhat with time of day. Correlations tended to be stronger using the morning collection (8 a.m. to noon) compared with other times of day and strongest using the 24-hr pooled collection.
Cotinine in plasma is eliminated unchanged in the urine as well as via metabolism. The major metabolites of cotinine include cotinine glucuronide, 3HC, and 3HC glucuronide, with minor routes of metabolism to cotinine-N-oxide, 5′-hydroxycotinine, and 5′-hydroxynorcotinine (Hukkanen et al., 2005
). It might be expected that the sum of cotinine plus cotinine glucuronide would better estimate plasma cotinine. However, this was not the case. Urine cotinine alone was more highly correlated with plasma cotinine compared with the sum of cotinine plus cotinine glucuronide.
3HC is the most prevalent metabolite of cotinine in the urine (Hukkanen et al., 2005
). Therefore, measurement of the sum of cotinine plus 3HC would substantially increase the analytical sensitivity of urine measurements with low-level cotinine exposure. We found that the sum of cotinine plus 3HC with or without their glucuronides did not correlate as well with plasma cotinine as did urine cotinine alone.
Renal clearance of cotinine averaged about 9 ml/min and was not affected by the dose of nicotine. This represents 15%–20% of total cotinine clearance, similar to that reported in prior studies (Hukkanen et al., 2005
). Men had higher renal clearance of cotinine than women, but this difference disappeared when corrected for body weight. This finding is consistent with the fact that men have higher glomerular filtration rates and greater kidney mass than women (Schwartz, 2007
Our data on the relationship between plasma cotinine and the daily dose of nicotine indicates that plasma cotinine can be used to quantitatively estimate daily exposure to nicotine when an individual is exposed to SHS. Each 1 ng/ml plasma cotinine represents on average exposure to 100 μg nicotine per day, although considerable individual variability exists. The relationship can be used to estimate daily nicotine exposure in population studies but is less accurate for individuals. Our data on the ratio of urine to plasma cotinine (on average, 4 to 5) allow the use of urine cotinine to estimate plasma cotinine and, in turn, the daily exposure to nicotine from tobacco smoke.
We provide compelling evidence that correction of urine cotinine for creatinine concentration improves the correlation between urine and plasma cotinine, particularly for urine sampled at particular times of day (that is, spot urines). We therefore recommend creatinine correction for future studies. Since correlations between urine cotinine and plasma cotinine are highest in morning urine collections, we recommend samples at that time of day, although correlations remain robust throughout the day. Men and women differ in their daily creatinine excretion due to differences in muscle mass. However, sex correction for urine creatinine did not improve the correlation with plasma cotinine and, therefore, is not necessary.
Although there is reason to believe that measuring multiple cotinine metabolites in urine might enhance the correlation between urine measurements and plasma cotinine, our data did not support this hypothesis. Measurement of urine cotinine corrected for creatinine appears to be the best predictor of plasma cotinine concentration. However, in some situations, urine cotinine concentrations may be below the limit of quantitation. In these cases, the measurement of cotinine plus its glucuronide with or without the measurement of 3HC would enhance sensitivity and would allow a reasonable estimate of plasma cotinine concentration.
This paper focused on predictors of plasma cotinine, which has been the most widely used biomarker and is considered by many researchers to be the gold standard for estimating nicotine exposure from SHS in epidemiology and treatment studies. The optimal urine metabolites to predict daily nicotine intake might differ from those that best predict plasma cotinine concentration. An analysis of the relationship between urine metabolites and nicotine intake will be the subject of a future paper.