Abstract
Cigarette smoking is the foremost modifiable risk factor for adverse pregnancy outcomes. Nicotine is a suspected fetal neuroteratogen. There is concern that nicotine may achieve toxic levels during pregnancy if nicotine replacement therapies are prescribed at doses used in the nonpregnant state. Ten healthy, volunteer, pregnant smokers received infusions of deuterium-labeled nicotine and cotinine during pregnancy and again postpartum. From blood and urine measurements, the following were determined: clearance (renal and nonrenal) of nicotine and cotinine, clearance of nicotine via the cotinine pathway (an indicator of CYP2A6 activity), and daily intake of nicotine from smoking. The clearance of nicotine and cotinine was significantly higher (60 and 140%, respectively), and the half-life of cotinine was much shorter (8.8 versus 16.6 h, P < 0.01) during pregnancy. Although plasma levels of cotinine were lower during pregnancy (119 versus 202 ng/ml, P < 0.05), daily intake of nicotine from smoking was similar during pregnancy and postpartum. For a given level of intake, the pharmacologic and toxicologic effects of nicotine during pregnancy are anticipated to be less than expected from nicotine metabolism data in nonpregnant women. Our data indicate that no downward dose adjustment needs to be made for nicotine replacement therapy during pregnancy. Conversely, higher than usual doses of nicotine may be necessary to optimize efficacy. Lower cotinine levels observed during pregnancy do not necessarily reflect less smoke exposure, and cut-off levels used to classify nonsmokers, passive smokers, and active smokers need to be established for pregnancy.
Cigarette smoking is the foremost modifiable risk factor for adverse pregnancy outcomes, affecting the mother, fetus, and newborn (Ebrahim et al., 2000). In the United States, approximately 24% of women 18 years and older are regular cigarette smokers and many continue to smoke after becoming pregnant (Ershoff et al., 1990; Fingerhut et al., 1990; Mayer et al., 1990; Floyd et al., 1993; Ebrahim et al., 2000). Counseling pregnant smokers does aid smoking cessation to some extent, but quit rates in smoking cessation programs remain low (Windsor et al., 1985; Mayer et al., 1990; Floyd et al., 1993). The urgency of developing safe and effective treatment to aid smoking cessation during pregnancy is apparent and is an important public health priority.
Nicotine replacement therapies have been extensively studied in nonpregnant adults (Fiore et al., 1994; Henningfield, 1995). Nicotine treatment has been shown to double smoking cessation rates and has gained wide usage in nonpregnant smokers (Fiore et al., 1994;Henningfield, 1995). One efficacy study (Wisborg et al., 2000) has been conducted during pregnancy and a few small short-term safety studies of nicotine replacement therapy have also been published (Oncken et al., 1996, 1997; Wright et al., 1997; Ogburn et al., 1999). The efficacy study provided 15-mg nicotine patches to women during the second trimester of pregnancy, but found no effect of nicotine compared with placebo on cessation rates (Wisborg et al., 2000). A recent study found that 92% of obstetrical providers in the northeastern United States believe that nicotine replacement therapy is likely to be effective for smoking cessation during pregnancy but most do not prescribe it because of safety concerns (Oncken et al., 2000).
There are two major barriers to studying the efficacy of nicotine replacement therapy during pregnancy. First, animal studies suggest that nicotine is a neuroteratogen (Slotkin, 1998). In utero exposure of rats to nicotine results in altered brain neuronal maturation and in behavioral disturbances in pups. Second, there are few pharmacokinetic data for nicotine during pregnancy upon which to make nicotine dosing recommendations (Oncken et al., 1996, 1997; Wright et al., 1997; Ogburn et al., 1999). Pregnancy is well known for affecting the metabolism of some drugs and may result in higher or lower clearances compared with the nonpregnant state (Loebstein et al., 1997). If nicotine metabolism was slower during pregnancy, then at usual adult nicotine replacement therapy doses nicotine plasma levels might rise to toxic levels in the pregnant woman. Nicotine replacement therapy might deliver higher, potentially more toxic, levels of nicotine to the fetus than would be acquired from the usual smoking rate of the mother. Conversely, if nicotine metabolism was faster during pregnancy, higher than usual doses of nicotine might be required to achieve efficacy.
Smokers regulate their cigarette consumption to maintain similar levels of nicotine in the body from day to day (Benowitz, 1996). It has been reported that women reduce their smoking during pregnancy (Sexton and Hebel, 1984; Windsor et al., 1985). Slower metabolism of nicotine during pregnancy could be an explanation for such behavior. Thus, one objective of our study was to determine the effect of pregnancy on the metabolism and disposition kinetics of nicotine. Such information could serve as a basis for optimizing medicinal nicotine dosing during pregnancy and to better understand changes in cigarette consumption that are observed in smokers during pregnancy.
Cotinine, the major proximate metabolite of nicotine, is widely used as a biomarker of nicotine exposure from tobacco (Benowitz, 1999). Plasma or saliva cotinine levels are reported to be lower in pregnant smokers compared with nonpregnant women (Mathai et al., 1990; Rebagliato et al., 1998; Ogburn et al., 1999). The use of a biomarker to quantitate exposures requires an understanding of its disposition kinetics in different conditions (Benowitz, 1999). A second objective of our study was to determine the disposition kinetics of cotinine as a basis for its application as a biomarker for nicotine and tobacco smoke exposure during pregnancy.
Materials and Methods
Participants.
Pregnant women were recruited through newspaper advertisements and flyers posted in the prenatal clinic at San Francisco General Hospital. Ten pregnant women completed the study. Eligibility criteria included: 1) active smoker with a plasma cotinine level above 25 ng/ml; 2) over 17 years of age; 3) receiving regular prenatal care without any complications of pregnancy (gestational diabetes, hypertension, intrauterine growth retardation, preterm labor, etc.); 4) good health on the basis of prenatal care records, history, physical examination, electrocardiogram, and blood chemistries; and 5) negative urine substance abuse toxicology screen and no current illicit drug use or alcoholism. The study was approved by the University of California, San Francisco, Committee on Human Research, and was carried out in accordance with the Declaration of Helsinki.
Six women were Caucasian and four were African-American. They smoked an average of 11 cigarettes per day (range 5–20). Their average age was 26.5 years (range 19–36). The mean body weight at the time of the first drug infusion was 79.1 kg (range 65–103, S.D. 11.1), whereas postpartum it was 82.1 kg (range 63–110, S.D. 16.5). Dating of the gestational age of the pregnancy was done by the prenatal care provider based on the date of the last monthly period and/or sonogram. Prenatal records were supplied to us by prenatal care providers. All pregnancies except one delivered at term, and all birth weights were consistent with the gestational age at birth. The mean gestational age at the time of the first drug infusion was 25.2 weeks (range 16–37, S.D. 7). All women delivered healthy babies at term without complications.
Ten women completed the study. Five women had two nicotine infusions during pregnancy, and five had one infusion during pregnancy. All had one infusion postpartum, occurring at least 12 weeks after the birth. The 15 infusions given during pregnancy were administered during the following gestational weeks (five subjects who had two infusions during pregnancy are designated A, B, C, D, and E after the gestational age): 16, 18, 19, 21A, 23B, 27, 27C, 30D, 34A, 34E, 36B, 37, 38E, 38C, and 40D (for example, subject A had two infusions, one during the 21st week of gestation and one during the 34th week of gestation).
Experimental Procedure.
Subjects came to the General Clinical Research Center at San Francisco General Hospital by 7 AM. Subjects were asked to abstain from cigarette smoking from 10 PM the previous night. Venous catheters were placed in both forearms. Subjects received a simultaneous 30-min infusion of deuterium-labeled nicotine-d2 (3′,3′-dideuteronicotine) and cotinine-d4(2,4,5,6-tetradeuterocotinine), each at a dose of 1.0 or 1.5 μg/kg/min (calculated as the free base) up to a maximum dosing weight of 90 kg. The dose of nicotine was typically equivalent to that obtained from smoking two cigarettes and, therefore, was judged not to pose a significant additional risk to that daily experienced by a regular smoker. The syntheses of these deuterium-labeled compounds have been described previously (Jacob et al., 1988; Jacob and Benowitz, 1993). Blood samples for measurement of plasma nicotine and cotinine levels were collected at 0, 10, 20, 30, 45, 60, 90, 120, 240, 360, and 480 min, then 24, 48, 72, and 96 h after the infusion. Urine was collected for 8 h after the start of the infusion.
Subjects were monitored by continuous electrocardiographic monitoring and frequent blood pressure measurement. The fetal heart rate was monitored using an obstetrical doppler stethoscope. All fetuses were active, with heart rates in the normal range and exhibiting normal variability. One infusion was stopped at 17 min because of nausea and vomiting; otherwise, all infusions were uneventful.
Analysis of Nicotine and Metabolites in Biological Fluids.
Nicotine and metabolite concentrations were determined by gas chromatography-mass spectrometry. Nicotine, nicotine-d2, cotinine, cotinine-d2, cotinine-d4(cotinine-2,4,5,6-d4),trans-3′-hydroxycotinine,trans-3′-hydroxycotinine-4′,4′-d2, andtrans-3′-hydroxycotinine-2,4,5,6-d4were determined by published methods (Jacob et al., 1991, 1992).
Glucuronide-conjugated nicotine, cotinine, andtrans-3′-hydroxycotinine in urine were measured as the difference in the total concentration of each analyte before and after incubation with β-glucuronidase, as described previously (Benowitz et al., 1994). Enzymatic hydrolysis was performed using 6000 ς units of β-glucuronidase (EC 3.2.1.3 from Helix Pomita; Fluka Chemical AG, Milwaukee, WI).
Pharmacokinetic Analyses.
Pharmacokinetic parameters were estimated from blood concentration and urinary excretion data using model-independent methods as described previously (Benowitz and Jacob, 1994; Benowitz et al., 1994; Perez-Stable et al., 1998). Among women who had two infusions during pregnancy, data from the first infusion were used for statistical comparison with the postpartum data. Data from all 10 women were included in the analyses; none were excluded or treated as an outlier. The daily intake of nicotine from tobacco was estimated using pharmacokinetic data determined from the infusion study and from measurement of plasma cotinine concentration derived from ad libitum smoking as follows: Dnic = Ccot × CLcot/fNIC→COT, as described previously (Benowitz and Jacob, 1994; Perez-Stable et al., 1998). Dnic is the daily dose of nicotine from smoking, Ccot is the plasma level of cotinine determined during ad libitum cigarette smoking as measured at the screening visit, CLcot is the clearance of cotinine determined from the cotinine-d4 infusion, and fNIC→COT is the fractional conversion fof nicotine to cotinine.
All data presented are normalized for body weight because the weights of the subjects fluctuated over the course of the study. Urine metabolite concentrations were expressed as a fraction of the total recovered nicotine-d2 plus its metabolites.
Statistical Analysis.
Pregnant and nonpregnant data were compared by a paired t test. Data from all 15 infusions done during pregnancy were examined to evaluate a possible effect of gestational age upon pharmacokinetic parameters using NONMEM, a mixed model statistical test (NONMEM Project Group, 1992).
Results
Comparisons of pharmacokinetic data during pregnancy and postpartum for the 10 subjects are presented in Table1 and Figs.1 and 2. The total plasma clearances of nicotine and cotinine were significantly higher during pregnancy compared with those postpartum. There was one woman with a very large difference in cotinine clearance during pregnancy and postpartum; if this subject was excluded, the difference was still highly significant (P < 0.001). The clearance of nicotine increased on average by 60% during pregnancy, whereas the clearance of cotinine increased by 140%. There was a 54% increase in the metabolic clearance of nicotine via the cotinine pathway during pregnancy. The renal clearance of nicotine tended to be lower during pregnancy, but this difference was not significant. The renal clearance of cotinine was similar during and after pregnancy. The half-lives of nicotine and cotinine were shorter during pregnancy. Cotinine elimination was nearly twice as rapid during pregnancy than postpartum. The changes in nicotine and cotinine clearance during pregnancy could be accounted for by changes in nonrenal (metabolic) clearance. No trend was found for any pharmacokinetic parameter with advancing gestational age. When comparing pharmacokinetic data from the first and second infusions done during pregnancy in the same women, no difference was found.
The average plasma cotinine concentration from ad libitum smoking was 119 ng/ml (S.D. 75) during pregnancy and 202 ng/ml (S.D. 77,P < 0.05) postpartum. The daily intake of nicotine from tobacco during pregnancy averaged 17.4 mg (S.D. 9.1) and 16.7 mg (S.D. 5.1) postpartum (P = 0.9). Urine metabolite data for the 8-h urine collection are presented in Table2. During pregnancy, a lower percentage of the dose of infused nicotine was recovered as nicotine, although there was an increase in the percentage recovered as nicotine glucuronide, cotinine glucuronide, and 3′-hydroxycotinine.
Discussion
Our study presents several novel findings. First, we found that the metabolic clearance of nicotine was substantially increased during pregnancy. Second, the metabolic clearance of cotinine was markedly accelerated during pregnancy, resulting in a substantial decrease in the half-life of cotinine. Third, the profile of nicotine and its metabolites in urine was altered during pregnancy. The excretion of nicotine was substantially decreased, with a small increase in the excretion of nicotine glucuronide and a substantial increase in the excretion of cotinine glucuronide and 3′-hydroxycotinine. Fourth, despite large differences in plasma cotinine concentration during ad libitum smoking, there was no difference between the daily dose of nicotine absorbed from cigarette smoking during and after pregnancy. Our study was relatively small but, considering the within-subject design and the large effect of pregnancy, the findings are unambiguous.
It is reported that many pregnant women reduce the number of cigarettes smoked per day (Sexton and Hebel, 1984; Windsor et al., 1985). We had originally hypothesized that the clearance of nicotine would be reduced during pregnancy—such that nicotine would remain in the body longer. This would then be expected to result in fewer cigarettes smoked per day. We found the opposite. The clearance of nicotine was increased and the half-life was shorter during pregnancy. Despite these changes, the daily intake of nicotine was unchanged during pregnancy. Thus, the intake of nicotine from cigarette smoking seems not to be influenced by the rate of nicotine metabolism during pregnancy. Our study was conducted between 16 and 40 weeks gestation. Further study is needed to identify when during pregnancy the increase in nicotine and cotinine clearance occurs.
Our findings have important clinical implications regarding the use of nicotine replacement products during pregnancy. There has been concern that nicotine plasma levels could rise to toxic levels with regular nicotine replacement therapy because the unconscious control of nicotine levels in the body associated with smoking would be lost. Our data indicate that with nicotine replacement therapy, nicotine plasma levels will not accumulate to a greater extent in the body during pregnancy compared with the nonpregnant state. The findings of no benefit in a recent nicotine patch study during pregnancy might be explained by inadequate levels of nicotine due to faster metabolism (Wisborg et al., 2000). Our data indicate that efficacy trials should consider changes in nicotine metabolism when planning doses of nicotine therapy.
Cotinine, the major proximate metabolite of nicotine, is important clinically because of its widespread use as a biomarker for nicotine exposure from smoking (Benowitz, 1999). Among pregnant smokers, maternal levels of cotinine correlate better with outcome measures such as birth weight than the number of cigarettes smoked per day (Haddow et al., 1987; Mathai et al., 1990; Bardy et al., 1993; Li et al., 1993;Ellard et al., 1996; Klebanoff et al., 1998; Secker-Walker et al., 1998). Recently, a study of women during pregnancy and again postpartum found that during pregnancy the median saliva cotinine concentration per cigarette was 3.5 ng/ml versus 9.9 ng/ml when not pregnant (Rebagliato et al., 1998). Our data explain lower cotinine levels reported during pregnancy. For any given intake of nicotine from smoking, the increased metabolic clearance of cotinine will result in lower steady-state plasma, saliva, or urine concentrations of cotinine.
The increased clearance and decreased half-life for cotinine will affect the interpretation of cotinine levels used in clinical trials or epidemiology studies during pregnancy. In nonpregnant adults, the average half-life of cotinine is approximately 17 h (Benowitz and Jacob, 1994), whereas during pregnancy it is a little less than 9 h. The faster clearance and shortened half-life during pregnancy have consequences for the use of cotinine as a biomarker of nicotine exposure. The cotinine levels that are used to classify nonsmokers, passive smokers, and active smokers will be lower, and cut-off levels need to be established for pregnancy. In addition, the time of day that a sample is collected will have a much greater effect during pregnancy because there will be a greater decline in cotinine levels during periods of nonsmoking, such as after sleeping overnight (Benowitz and Jacob, 1994).
In nonpregnant adults, 70 to 80% of nicotine is metabolized to cotinine (Benowitz and Jacob, 1994), primarily by liver cytochrome P450 CYP2A6 (Messina et al., 1997). Cotinine is for the most part metabolized to 3′-trans-hydroxycotinine, primarily by the same CYP2A6 enzyme (Nakajima et al., 1996; Messina et al., 1997). Both nicotine and cotinine undergo N-glucuronidation, whereas 3′-hydroxycotinine undergoes O-glucuronidation (Jacob and Benowitz, 1991; Benowitz and Jacob, 1994; Benowitz et al., 1994, 1999).
Pregnancy has a variable and unpredictable effect upon the metabolic clearance of drugs (Loebstein et al., 1997). Drugs with increased clearance during pregnancy include methadone (Pond et al., 1985), phenytoin, carbamazepine, penicillin, ampicillin, piperacillin, and imipenem (Loebstein et al., 1997). The mechanism for the increase in metabolic clearance of nicotine and cotinine during pregnancy is not known. Our data suggest that nicotine and cotinine clearances are accelerated by faster oxidation via CYP2A6 and faster glucuronide formation. Although nicotine and cotinine share the same metabolizing enzymes, their increased clearances may occur by different physiologic mechanisms. Nicotine is a rapidly cleared drug with a high affinity for CYP2A6, and the rate of clearance is primarily controlled by liver blood flow (Lee et al., 1989; Nakajima et al., 1996). There is a substantial increase in cardiac output and blood volume during pregnancy, which would be expected to be associated with increased liver blood flow. Increased liver blood flow could account for the increase in nicotine metabolic clearance. However, one study indicates that there is no increase in liver blood flow during pregnancy (Robson et al., 1990). Cotinine is a slowly metabolized chemical, with a low affinity for CYP2A6 relative to nicotine. The rate of cotinine metabolism is primarily determined by the level of metabolizing enzymes in the liver (Nakajima et al., 1996). Our study suggests that the levels of enzymes responsible for cotinine metabolism, presumably CYP2A6, are markedly increased during pregnancy. It is possible that extrahepatic sites of drug metabolism, such as the placenta, may be involved in the increased clearance of nicotine and cotinine. However, in vitro studies indicate that there is very little CYP2A6 activity in the placenta (D. L. Kroetz, personal communication) (Tutka et al., 2000).
There was a substantial increase in the percentage of nicotine and cotinine excreted as their glucuronide conjugates. There was no increase in the percentage of 3′-hydroxycotinine excreted as a glucuronide. These data suggest an acceleration of nicotine and cotinine metabolism via the N-glucuronidation pathway, but no effect on hydroxycotinine metabolism by theO-glucuronidation pathway. The mechanism of up-regulation of the N-glucuronidation pathway during pregnancy remains to be determined.
In summary, the clearance of nicotine is increased during pregnancy. For a given level of nicotine intake from smoking, the pharmacologic and toxicologic effects on the fetus will be less than expected from nicotine clearance based on data from nonpregnant women. Our data indicate that no downward dose adjustment needs to be made for nicotine replacement therapy during pregnancy. In fact, our data suggest the opposite—that higher doses of nicotine replacement therapy may be necessary during pregnancy compared with the nonpregnant state. The metabolic clearance of cotinine is markedly accelerated during pregnancy, resulting in a half-life nearly 50% shorter than in the nonpregnant state. This observation has implications for the use of cotinine as a biomarker for cigarette smoking during pregnancy. Our data do not affect the validity of cotinine levels as a predictor of pregnancy outcomes, which are based on empirical observations (Mathai et al., 1990; Bardy et al., 1993; Li et al., 1993; Ellard et al., 1996). But our study does indicate that the lower cotinine levels observed in smokers during pregnancy compared with the same individuals before or after pregnancy do not necessarily reflect less smoke exposure. We found a similar level of nicotine intake from smoking during pregnancy and postpartum, despite the 2-fold differences in cotinine levels. Finally, our observations on the marked induction of CYP2A6 activity and N-glucuronidation during pregnancy add to the limited body of data concerning the effect of pregnancy upon drug metabolism.
Acknowledgments
We thank Patricia Buley, Sandra Tinetti, and the staff of the General Clinic Research Center at San Francisco General Hospital for assistance in conducting the clinical study. We thank Lisa Yu for performing the analytical chemistry, Gunnard Modin for statistical analysis, and Kaye Welch for editorial assistance.
Footnotes
-
Supported by U.S. Public Health Service Grants DA09761, DA02277, and DA12393 from the National Institute on Drug Abuse, National Institutes of Health. Carried out in part at the General Clinical Research Center at San Francisco General Hospital Medical Center with support of the Division of Research Resources, National Institutes of Health (RR-00083).
- Received September 6, 2001.
- Accepted January 14, 2002.
- The American Society for Pharmacology and Experimental Therapeutics