Hum. Reprod. Advance Access originally published online on October 19, 2006
Human Reproduction 2007 22(2):337-345; doi:10.1093/humrep/del406
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Maternal exposure to second-hand tobacco smoke and pregnancy outcome among couples undergoing assisted reproduction
1 Department of Environmental Health Sciences, University of Michigan, Ann Arbor, MI 2 Obstetrics, Gynecology and Reproductive Science, Brigham and Women’s Hospital, Harvard Medical School 3 Department of Environmental Health, Harvard School of Public Health and 4 Vincent Memorial Obstetrics and Gynecology Service, Andrology Laboratory and In Vitro Fertilization Unit, Massachusetts General Hospital, Boston, MA, USA
5 To whom correspondence should be addressed at: Department of Environmental Health Sciences, University of Michigan School of Public Health, M6226 SPH II, 109 S. Observatory St., Ann Arbor, MI 48109, USA. E-mail: meekerj{at}umich.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
BACKGROUND: Exposure to second-hand tobacco smoke is preventable, yet common. This study assessed relationships between maternal exposure to second-hand tobacco smoke and adverse pregnancy outcomes. METHODS: We measured cotinine (a biomarker of tobacco smoke) in urine from 921 women undergoing assisted reproductive technologies (ARTs) between 1994 and 1998. We also collected information on self-reported exposure to second-hand smoke at home or at work, in addition to parental smoking during the women’s childhood. RESULTS: In crude analysis, creatinine-adjusted cotinine levels were associated with a slight decrease in implantation rate among non-smoking women (11.1% in the lowest cotinine quintile versus 8.2% in the highest cotinine quintile; P = 0.13). However, in multivariate logistic regression, cotinine levels above the median were not associated with failed fertilization, failed implantation or spontaneous abortion, nor was there evidence of a dose–response relationship among cotinine quintiles. After excluding women in couples diagnosed with male factor infertility, there were increased odds of having a spontaneous abortion among non-smoking women who reported that both parents smoked while they were children growing up compared with women reporting that neither parent smoked [adjusted odds ratio (OR) = 4.35; 95% confidence interval (CI) = 1.04–18.1]. CONCLUSIONS: Female exposure to second-hand smoke as a child or in utero may be associated with an increased risk of spontaneous abortion in adulthood. However, this may be a chance finding due to multiple comparisons. Similar associations should be explored in additional studies with more refined estimates of childhood and in utero exposure to tobacco smoke.
Key words: ETS/IVF/pregnancy/second-hand smoke/spontaneous abortion
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
There is growing concern surrounding potential adverse reproductive health effects and pregnancy outcomes resulting from exposure to second-hand tobacco smoke [also known as environmental tobacco smoke (ETS)]. Although exposure to second-hand tobacco smoke is preventable, it remains prevalent. The Second and Third National Reports on Human Exposure to Environmental Chemicals, part of the National Health and Nutrition Examination Survey (NHANES 1999–2000 and 2001–02, respectively), recently reported measurable levels of serum cotinine, a biomarker of tobacco smoke, in
50% of the non-smoking US population (CDC, 2003
, 2005a; Pirkle et al., 2006). The majority of second-hand tobacco smoke is in the form of sidestream smoke generated from the burning end of a lighted cigarette, whereas the remainder is composed of mainstream smoke exhaled by individuals actively smoking. Both mainstream and sidestream smoke contain thousands of compounds—many of them are harmful to humans. Mainstream and sidestream smoke are produced at differing temperatures and oxygen conditions, and harmful constituents exist in varying proportions between the two types of smoke. For example, sidestream smoke contains more CO and less CO2, and higher levels of combustion products formed by nitrozation and amination than mainstream smoke (Woodward and al-Delaimy, 1999). In addition to CO, second-hand tobacco smoke contains other chemicals that are known or suspected reproductive toxicants—for example, benzene, cadmium, ethylbenzene, formaldehyde, hydrazine, lead, limonene, methylamine, methylene chloride, nicotine, pyridine, toluene and radioactive polonium-210 (Lindbohm et al., 2002).
Approximately 15% of all couples have difficulty achieving pregnancy. Only 50–60% of all conceptions advance beyond 20 weeks of gestation (Wilcox et al., 1988), and up to 75% of the lost pregnancies are a result of blastocyst implantation failure and are never clinically recognized as pregnancies (Norwitz et al., 2001). Thus, these early losses may manifest clinically as female infertility. For the most part, factors involved in failed implantation remain unidentified. The ability of a blastocyst to implant in the uterus may be associated with uterine receptivity, oocyte quality or delayed implantation, although oocyte quality is likely the most important factor (Norwitz et al., 2001). Oocyte quality can be significantly affected by its surrounding environment (Homburg and Shelef, 1995), and the follicular microenvironment is influenced by a complex balance of many physiological, biochemical and potentially environmental factors. Active smoking is associated with female subfertility resulting from a number of potential mechanisms (Hull, 1995; Augood et al., 1998; Shiverick and Salafia, 1999).
Approximately one-third of pregnancies are lost after implantation (Wilcox et al., 1988). Causes of post-implantation loss are also poorly understood. A large percentage of cleaved embryos have chromosome aberrations (Carrera and Veiga, 1998), mostly as a result of abnormalities during oogenesis leading to abnormal gametes, which likely contribute to a portion of the high rate of embryonic arrest (Simpson, 1980). Active smoking is associated with spontaneous abortion, abnormal placentation, intrauterine growth retardation, preterm delivery, perinatal mortality, congenital malformations (Werler, 1997) and lower rates of implantation and IVF success (Van Voorhis et al., 1996; Lintsen et al., 2005). Less is known about these outcomes among passive smokers, but metabolites of cigarette smoke have been measured in fetal blood at higher concentrations than in that of the mother (Jauniaux et al., 1999; Whyatt et al., 2001; Perera et al., 2004), suggesting that nicotine and other compounds in second-hand smoke concentrate in the fetus. In addition, fetal serum and amniotic fluid metabolite levels from passive smokers reached 30–44% of the corresponding levels from active smokers, respectively (Jauniaux et al., 1999), higher than what would be expected considering passive smokers typically have serum metabolite concentrations at <1–10% of those found in active smokers (Pirkle et al., 1996). In hamsters, exposure to sidestream cigarette smoke at concentrations that reflect second-hand smoke exposure in humans was found to delay the rate of embryo transport through the oviduct, and sidestream smoke actually exhibited greater toxicity than mainstream smoke concentrations that were representative of active smokers (DiCarlantonio and Talbot, 1999). Thus, the potential reproductive and developmental effects associated with second-hand tobacco smoke exposure are of concern.
Whereas more data are available on reproductive health and pregnancy outcomes related to active smoking than among women exposed to second-hand tobacco smoke, several recent studies have found associations between maternal exposure to second-hand tobacco smoke and adverse reproductive effects or pregnancy outcomes. Most similar in design to the present study was a recent report among 225 women undergoing IVF or ICSI in a Canadian reproductive clinic in the years 2003–04 (Neal et al., 2005), where significantly lower implantation rates and pregnancy rates were found among both active smokers and passive smokers compared with non-smokers. However, passive smoking (second-hand tobacco smoke exposure) among the women in the study was measured only by self-report. Another recent study investigated associations between paternal smoking and pregnancy loss among 526 non-smoking Chinese female textile workers and reported increased odds of early pregnancy loss among women with husbands who smoked more than 20 cigarettes per day (Venners et al., 2004). Results associated with paternal smoking may reflect either effects related to maternal exposure to second-hand tobacco smoke or sperm damage associated with active smoking in the male partner, or possibly a combination of both factors. A third study of non-smokers among 3000 California women enrolled in a 1992 prenatal screening programme reported associations between maternal serum cotinine levels and increased odds for fetal death, preterm delivery and term-low birthweight (Kharrazi et al., 2004). An earlier study from California also showed an increased risk of spontaneous abortion among mothers exposed to second-hand tobacco smoke for 1 h or more per day (Windham et al., 1992). Finally, a study of fertile women found that the risk of experiencing delayed conception for at least 6 months was significantly elevated among women who reported second-hand tobacco smoke exposure, and the risk estimate was similar in magnitude to that for women who actively smoked (Hull et al., 2000).
The present study was designed to assess relationships between maternal exposure to second-hand tobacco smoke, estimated through urinary cotinine concentrations and self-report, and adverse pregnancy outcomes. The pregnancy outcomes explored were the various points of failure observable during treatment involving assisted reproductive technologies (ARTs): primarily failure of fertilization, implantation failure and developmental failures, which include spontaneous abortion.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Between August 1994 and March 1998, couples undergoing IVF or ICSI were recruited through three clinics in the Boston area (Boston IVF, the Brigham and Women’s IVF Program and the Reproductive Science Center of Boston). Study protocols were approved by the Brigham and Women’s Hospital Human Research Committee. Approximately 65% of couples who were approached agreed to participate in the study, and 1250 couples were enrolled. Couples who required either donor oocytes or donor semen were excluded, as were couples who were gestational carriers and those who underwent gamete intra-Fallopian transfer. Of the 1250 couples enrolled, 105 were excluded for these reasons. An additional 224 women did not have a urine sample analysed for cotinine concentration, leaving 921 couples with data on female tobacco smoke exposure and ART outcomes. The study used self-administered questionnaires to obtain information about medical history and lifestyle factors, and details about IVF treatment and outcome were abstracted from clinical records. The self-administrated questionnaire included medical and lifestyle variables, such as demographics, ages of both male and female partner, medical and reproductive history, smoking history and duration of infertility. Information abstracted from clinical records included diagnosed causes of infertility, type of GnRH agonist treatment, gonadotrophin treatment variables, estradiol levels, follicle counts, number and quality of oocytes retrieved, day of embryo transfer and number of embryos transferred. Other clinical information collected included chemical pregnancies, clinical pregnancies and follow-up information on whether women with a clinical pregnancy had at least one liveborn pregnancy.
All IVF outcome variables are defined from data abstracted from the clinic medical record. Couples who ended their first IVF cycle after starting ovulation induction treatment but before oocyte retrieval are defined as such. Unsuccessful oocyte retrieval was determined when a retrieval procedure was undertaken, but no oocytes were successfully aspirated or the oocytes that were aspirated were of such poor quality that insemination was not attempted. Failure of fertilization (poor semen quality) was defined by no insemination despite oocyte retrieval. Embryo creation of such poor quality that an embryo transfer was not attempted was defined accordingly. When at least one embryo was transferred but hCG levels never reached 5.0 mIU/ml, the cycle outcome was defined as a failure of implantation. A chemical pregnancy was defined by a luteal hCG measurement of 5.0 mIU/ml or greater with no further evidence (gestational sac and fetal heartbeat) of a continued pregnancy. Clinical pregnancy was determined by ultrasound visualization of a gestational sac and a fetal heartbeat. Among clinically recognized pregnancies, outcomes included an ectopic pregnancy (gestation outside of the uterus), a molar pregnancy, a spontaneous abortion (fetal demise before 20 weeks of gestation), stillbirth (fetal demise at 20 gestational weeks or later) or live birth of at least one infant. Numbers for some of these outcomes lacked power to investigate independently.
Before their first IVF treatment cycles, a first morning void urine sample was collected from the female partner of each couple. Urine samples were collected in a sterile wide-mouthed 1-l plastic container, taken back to the laboratory and aliquoted and then frozen for cotinine analysis. Cotinine concentrations in urine were determined by competitive radioimmunoassay techniques described elsewhere (Langone et al., 1973; Van Vunakis et al., 1993). Concentrations were reported in nanograms per millilitre (ng/ml). The procedure had a lower reporting limit of 0.1 ng/ml, with inter-assay and intra-assay variations of 5%. Self-reported second-hand tobacco smoke exposure at home (by either a spouse or other household member) or at work (in either a smoke-regulated environment or unregulated) was obtained through the self-administered questionnaire. Questionnaire information was also collected on male partner’s present smoking status. In addition, participants’ childhood exposure to second-hand tobacco smoke was estimated through questions on parental smoking while growing up.
Statistical analysis
Preliminary exploratory data analysis was used to evaluate variable distributions and to assess bivariate relationships among key covariates. All primary statistical analyses were performed after excluding active smokers. Information on smoking from the questionnaire was used to identify women who actively smoked during pregnancy. Previously published urinary cotinine threshold values for active smoking (Haufroid and Lison, 1998) were also tested to identify women who were smokers who may have purposely responded incorrectly on the questionnaire because smoking during pregnancy is strongly discouraged. To account for inter-individual variation in hydration level and urine sample dilution, we used cotinine levels adjusted by urinary creatinine in the primary statistical analyses (Boeniger et al., 1993). Associations between second-hand tobacco smoke exposure and outcomes were examined when exposure was categorized as a dichotomous variable (above or below the median), then again when categorizing cotinine levels into quintiles. Associations were also explored using self-reported second-hand tobacco smoke exposure. To compare the present data with the results recently reported by Neal et al. (2005), we calculated implantation rates (number of fetal sacs with a positive heartbeat divided by the total number of embryos transferred) and explored differences between cotinine level groups using non-parametric methods (Kruskal–Wallis test).
Conditional logistic regression analyses were then performed by fitting separate multiple logistic regressions for the various outcome measures to estimate the odds of a specific failure on a subset of subjects who have not experienced a failure up to that point, instead of among all subjects. For example, a woman/couple who experienced a failed implantation would not be included in the analysis for spontaneous abortion. Thus, the sample size (i.e. number of subjects at risk) decreased with each point of failure along the progressing stages of pregnancy. The primary statistical analysis was limited to couples not diagnosed with male factor infertility. Pregnancies and live births (dichotomous yes/no variables), oocyte fertilization (yes/no), implantation (yes/no) and spontaneous abortions (yes/no) were used as the outcome variables. Variables considered as potential confounders were age, race, body mass index, months spent attempting to get pregnant, year of IVF treatment, clinic/site of IVF treatment, ovarian stimulation method, ampules of gonadotrophin, number of oocytes transferred, assisted hatching and day of embryo transfer. Inclusion of covariates in the final models was based on biological and statistical considerations (Hosmer and Lemeshow, 1989). Bivariate relationships between each covariate and the exposure and/or outcome variables were first assessed to identify covariates to include in the multivariate model. The original multivariate model was constructed using forward selection. Non-significant covariates were added to the model and individually retained if inclusion resulted in a change of >10% in the main parameter estimate of interest. The covariates that appear in the final models were significant in some, but not all, models. The same covariates were included in each of the models to maintain consistency, and they were considered to be biologically or clinically important in models in which they were not statistically significant.
In secondary analyses, couples with male factor infertility were retained in the models. Secondary analysis also involved restricting the data set to women with tubal factor or unexplained infertility, followed by further restriction to include only women with tubal factor infertility. A stratified analysis including only women diagnosed with tubal factor infertility was performed to increase generalizability of the results, as women with tubal factor infertility are two times more likely to reach a 12-week pregnancy following IVF/embryo transfer than women with any other classification of infertility (Joesbury et al., 1998) and are likely to more closely represent women in the general population. However, this analysis substantially reduced sample size and statistical power.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Demographic data of women in this study are presented in Table I. Subjects had a mean (SD) age of 35.3 (4.2) years and were primarily white (92%), and the majority had never actively smoked (60%). Eight per cent of women were self-reported active smokers. A male factor was the primary infertility diagnosis in nearly one-third of the couples (32%), tubal factors were diagnosed in nearly one-quarter of the couples (24%) and infertility remained unexplained in 12% of couples. Table II lists the outcome of first-cycle IVF treatment for the 921 couples. A successful live birth was reported for 18% of the couples, and failure of implantation was the most common treatment outcome (52%). Year of treatment was associated with suggestive but non-significant decreasing trends in certain IVF failures from 1995 to 1997, where failed implantations decreased from 54 to 48% of all IVF treatments and spontaneous abortions decreased from 3.9 to 2.6% of treatments (P-value >0.05).
|
|
Cotinine concentrations in all 921 maternal urine samples ranged from 0.1 to 15 000 ng/ml, with an unadjusted median of 48.9 ng/ml. Thirty-two of the women did not have creatinine data and were excluded from analyses involving creatinine-adjusted data. Distributions did not differ greatly when couples with male factor infertility were excluded or when the study population was limited to couples with a primary diagnosis of tubal factor infertility (data not shown). Among all non-smokers (n = 847), the median cotinine concentration was 44.6 ng/ml unadjusted and 57.1 ng/g when adjusted for creatinine (Table III). These values were significantly lower than median levels among self-reported active smokers (2334 ng/ml unadjusted and 3334 ng/g creatinine). Creatinine-adjusted cotinine levels did not differ between non-smoking women who reported ETS exposure at home or at work (median = 56.4 ng/g) compared with those who reported no ETS exposure (median = 57.3 ng/g), representing disagreement between self-reported exposure and biomarker data.
|
Although this study spanned only 3 years, cotinine concentrations were associated with year of treatment. The median (25th percentile and 75th percentile) creatinine-adjusted level among non-smokers decreased from 73.2 ng/g (45.7, 127 ng/g) in 1995 to 47.5 ng/g (29.0, 92.2 ng/g) in 1996 to 30.1 ng/g (9.57, 67.1 ng/g) in 1997 (P-value < 0.05). Preliminary analysis showed a slight decreasing trend in implantation rate among non-smokers with increasing creatinine-adjusted cotinine quintiles. Mean implantation rates for the lowest to highest exposure quintiles were as follows: 11.1, 9.4, 8.8, 8.8 and 8.2%, respectively. However, differences between groups and the decreasing trend were not statistically significant (P = 0.13 and P = 0.3, respectively).
When exploring associations between exposure to tobacco smoke and pregnancy outcomes among non-smokers using multivariate logistic regression, analyses were performed among all non-smoking women and also when stratified by infertility diagnosis (male factor excluded, tubal factor and unexplained only, tubal factor only). Table IV presents results from both before and after exclusions based on infertility diagnosis and shows no evidence of associations between creatinine-adjusted cotinine levels above the median and an increased risk of failed fertilization, failed implantation, spontaneous abortion or failure of development. Odds ratio (OR) estimates for spontaneous abortion and failure of development were inconsistent and sensitive to fertility diagnosis restriction, whereas the analysis for failed fertilization was limited because of extremely low numbers of cases when stratified by infertility diagnosis. When additional exposure groups were used (subjects categorized by creatinine-adjusted cotinine quintile), there remained a lack of association for all outcomes with no evidence of a dose–response trend and no difference in the odds of adverse cycle outcomes between the highest and lowest cotinine quintile (data not shown). OR estimates for quintiles two through five did not show a trend, with some above and some below unity in comparison with the lowest quintile, and all 95% confidence intervals (CIs) included 1.0 (P-value >0.05). Results did not differ when identifying and excluding active smokers by cotinine concentration (>500 ng/ml) instead of by self-report (data not shown). The overall results also did not change when the homogeneity of the comparison group was increased by limiting the comparison group to only women who experienced a live birth (data not shown).
|
57 ng/g creatinine)Results of multivariate logistic regression from self-reported exposure obtained through questionnaire, along with other self-reported variables related to second-hand tobacco smoke exposure, are presented in Table V. When all non-smoking women were included in the analysis regardless of infertility diagnosis, there was a suggestive increase in the odds of failed implantation among women reporting exposure to second-hand tobacco smoke either at home or at work compared with women reporting no second-hand smoke exposure (OR = 1.55; 95% CI = 0.93–2.60). This self-reported exposure variable was formed by including all women who answered yes to at least one of the four questions about exposure at home or at work. None of the four individual variables on self-reported second-hand smoke exposure at home or work were associated with the IVF outcomes (data not shown). Variables associated with the women’s childhood exposure to second-hand tobacco smoke were also explored through questions on whether their parents had smoked. When women in couples diagnosed with male factor infertility were excluded from the analysis, there were increased odds of spontaneous abortion among women reporting that both their parents smoked while they were children growing up compared with women who reported that neither parent smoked (OR = 4.35; 95% CI = 1.04–18.1). Results in Table V were similar when recent smokers (subjects who reported that they quit smoking within 1 year before their ART cycle) were also excluded. The most notable change was a slight increase in the odds of failed implantation among all women reporting any second-hand exposure (OR = 1.70; 95% CI = 0.98–2.96) and among women in couples not diagnosed with male factor infertility who reported any ETS exposure (OR = 1.72; 95% CI = 0.88–3.36).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
In this study, we found no association between exposure to second-hand tobacco smoke, estimated by urinary cotinine levels and by self-reported exposure, and an increased risk of adverse pregnancy outcomes (failed fertilization, failed implantation, failure of development or spontaneous abortion) among couples undergoing IVF. However, we did find increased odds of having a spontaneous abortion among non-smoking women who reported having two parents who smoked while growing up compared with women who reported that neither parent smoked.
Consistent with the recent study by Neal et al. (2005), our preliminary bivariate results showed a slight declining trend in implantation rate as cotinine level increased. However, the differences in implantation rate between non-smokers exposed to tobacco smoke and those unexposed were much smaller in the present study (8 and 11% among the highest and lowest cotinine groups, respectively) compared with those in the Neal study (12 and 25% among passive smokers and non-smokers, respectively). Moreover, when more sophisticated (multivariate) statistical analyses were utilized in the present study, we found no associations between second-hand tobacco smoke exposure and failed implantation. In addition to a relatively small study size (n = 225 couples), another limitation of the study by Neal et al. (2005) was that exposure groups were assigned based solely on self-reported tobacco smoke exposure among study subjects. Biological markers (known as biomarkers) of tobacco smoke provide a more valid, quantitative measure of exposure in epidemiology studies compared with self-report measures (Benowitz, 1999). The main advantage of using biomarkers in exposure assessment for epidemiology is that they provide a measure of absorbed dose as opposed to the potential dose (exposure) in the external environment (Haufroid and Lison, 1998), and self-reports of time spent exposed to second-hand smoke are imprecise because of variations in the number of cigarettes smoked, proximity of non-smokers to smokers, room ventilation and other environmental characteristics.
Cotinine is a major metabolite of nicotine and is currently regarded as the best biomarker in active smokers and in non-smokers exposed to tobacco smoke (Haufroid and Lison, 1998; Benowitz, 1999; CDC, 2003). Measuring cotinine is preferred to measuring nicotine, because cotinine persists longer in the body (20–24 h, compared with 2 h for nicotine). Cotinine can be measured in saliva, hair or more commonly serum or urine. Urine is often the preferred medium due to the ease of sample collection and because cotinine is present at higher concentrations in urine than in serum (Benowitz, 1999). The primary analysis in the present study used the median cotinine level among non-smokers to classify subjects as exposed or unexposed to second-hand tobacco smoke. Owing to the observed median level in the present study (44 ng/ml unadjusted and 57 ng/g creatinine-adjusted), this approach also satisfied the recommended method of separating exposed and unexposed non-smokers by using a urinary cotinine threshold level of 50 ng/ml (Apseloff et al., 1994; Haufroid and Lison, 1998). The discrepancy between the prevalence of cotinine measured in urine and self-reported second-hand tobacco smoke exposure suggests that most people may have been unaware of exposure or that exposure to second-hand tobacco smoke may have occurred in places other than the locations included on the questionnaire. The questionnaire only asked about exposure at home or at work, whereas exposure could have occurred in other places where subjects spent time, as in other people’s homes or in public places such as restaurants.
The questionnaire also did not ask about the timing of ETS exposure, which may also account for some of the discrepancy between self-reported exposure and biomarker measurements due to the rapidity at which cotinine is excreted from the body. For this reason, although often imprecise, self-report is still a useful supplement to biomarkers for the estimation of longer-term intermittent ETS exposure. A limitation of using cotinine concentrations as a biomarker for second-hand tobacco smoke exposure, and thus a limitation of the present study, is its relatively short half-life in the body. Thus, cotinine level reflects exposure on the order of several days, and the use of a single urine sample may lead to exposure misclassification if attempting to estimate second-hand tobacco smoke exposure over several months (i.e. pregnancy) in an epidemiology study. However, if sources and activities related to an individual’s exposure to second-hand smoke are relatively stable over time, then a single sample may adequately estimate long-term exposure. For example, in a study with repeat measures in children, exposure ranks based on urine cotinine concentrations were found to be stable over the course of 1 month (Henderson et al., 1989). More research defining the temporal variability of biomarkers of tobacco smoke exposure among non-smokers is needed, and this topic is a future goal of our work.
In this study, we observed decreased cotinine levels in the years 1995–97 among women undergoing IVF treatment. This temporal decline is consistent with national data, where serum cotinine levels among non-smoking subjects participating in NHANES were found to decrease by 70% from years 1988–2002 (Pirkle et al., in press). The decrease in cotinine levels over time likely reflects increased smoking restrictions or bans in the workplace and in other public places along with overall declining smoking rates (Farrelly et al., 1999; Fichtenberg and Glantz, 2002; CDC, 2005b).
When using cotinine levels to estimate exposure to second-hand tobacco smoke in the present study, we did not find any associations with pregnancy outcomes. Statistical power was limited when cotinine levels were divided into quintiles, especially when the analysis was limited to subsets of subjects according to infertility diagnosis, which resulted in wide CIs. Thus, the results focus on the analysis using above or below median cotinine levels. In that analysis, most of the ORs were unexpectedly below unity. Although this may reflect random error because all CIs included 1.0, another possible explanation relates to the dose–response fallacy in reproductive studies of toxic exposures where there are consecutive dependent potential outcomes (Selevan and Lemasters, 1987). This phenomenon may occur if the exposure of interest is related to an earlier adverse outcome (e.g. infertility), leading to apparent lower risk of later outcomes (e.g. implantation) with respect to exposure. Another factor in the dose–response fallacy is level of exposure, where more toxic levels of exposure are associated with earlier adverse outcomes and lower exposure levels are associated with later effects (e.g. low birthweight). Thus, in the present study, it is possible that the dose–response fallacy may be present for associations between exposure and outcome before those we assessed (e.g. infertility, for which we did not have an appropriate comparison group), or perhaps, the early pregnancy outcomes we explored were related to lower exposure levels that we were not able to observe with adequate resolution within our exposure range. That is, it is possible that a threshold of exposure for the outcome of interest exists below the upper threshold for the lowest exposure group, which would result in effect estimates that are below unity for the higher exposure groups.
We found significantly increased odds of having a spontaneous abortion among non-smoking women who reported having two parents who smoked while growing up compared with women who reported that neither parent smoked (P-value = 0.04). Although this finding may be due to chance because the present study made a number of comparisons (i.e. there were 24 separate analyses in Table V alone) and there were a small number of spontaneous abortion cases, it may also be plausible that it indicates second-hand tobacco smoke exposure as a child or in utero may have reproductive implications later in life. Evidence for this possibility was illustrated in a recent study where a mother’s own prenatal exposure to tobacco smoke modified the association between her smoking status and having children with reduced birthweight (Misra et al., 2005). An earlier study also found that women with prenatal exposure to cigarette smoking had reduced fecundability (Weinberg et al., 1989). In the present study, the fact that we did not also see an increased risk of spontaneous abortion among women who reported that only their mothers smoked while growing up further suggests that our findings of an increased risk when both parents smoked may be due to chance. However, a woman who smokes may be more likely to continue smoking during pregnancy if her husband also smokes, potentially leading to an increased fetal exposure to both mainstream (direct exposure from mother actively smoking) and sidestream (passive exposure from father smoking) tobacco smoke constituents. In addition, when both mother and father (or female and male partner) smoke, they may be more likely to smoke inside the home or car than if only one of them smokes (Gilpin et al., 1999; Okah et al., 2002), which in this case would lead to an increased childhood exposure to second-hand tobacco smoke (and an increased fetal exposure to sidestream smoke constituents if the mother did not smoke during the pregnancy). This might be especially true several decades ago when the women in this study were themselves children and/or in utero, because at that time much less was known about the adverse health effects from smoking and exposure to tobacco smoke and very few public smoking restrictions were in place. Future studies on second-generation reproductive health, with more detailed information on first-generation smoking and the related nature and timing of second-generation exposure, are needed to support or refute our findings.
Because the present study was limited to couples undergoing ART, the ability to generalize our findings may be limited. Some potential reasons for reduced generalizability may include differences in the sensitivity of gametes from infertility patients to second-hand tobacco smoke, medical risk factors for infertility increasing an individual’s sensitivity to tobacco smoke exposure or IVF media/culture conditions altering early embryonic development in response to tobacco smoke exposure. Because several oocytes are harvested and only the best embryos are selected for transfer, this process does not fully represent what occurs in a natural pregnancy. Thus, some important factors may differ from the time leading up to fertilization through implantation among ART couples compared with couples conceiving naturally. If these conditions are associated with a differential response to tobacco smoke exposure, then the results would not be generalizable. Demographic characteristics of an ART study population could also potentially limit generalizability. For example, smoking rates and exposure to second-hand smoke may vary by socioeconomic status. Because ART is costly, couples seeking treatment tend to be of higher socioeconomic standing. Thus, the results from a study of tobacco smoke exposure among ART couples may not be generalizable to all socioeconomic groups if the groups have differing biological responses to tobacco smoke exposure.
Although results from the present study may only be generalizable to similar populations, the study also has several strengths. First, the study of ART couples allows for monitoring prepregnancy and early pregnancy stages that are unobservable in other populations. Second, the study had a high participation rate which serves to reduce the likelihood of selection bias. In the present study, 65% of couples who were approached about the study agreed to participate. This is likely higher than might be expected for a similarly invasive pregnancy study among the general population. Although low participation may not lead to selection bias in cohort studies of birth outcomes among pregnant women (Nohr et al., 2006), it remains problematic in fecundity studies among the general population, e.g. time to pregnancy studies (Muller et al., 2004; Bonde et al., 2006). Finally, information bias was minimized in the present study through the use of exposure biomarkers and by administering the questionnaire on current and historical tobacco smoke exposure before treatment and knowledge of pregnancy outcomes.
In conclusion, we did not find evidence of an association between exposure to second-hand tobacco smoke, measured through urinary cotinine, and adverse IVF outcomes. It is possible that our inability to find associations between urinary cotinine levels and pregnancy outcomes in the present study was in part due to exposure misclassification resulting from the collection of only a single urine sample. In future work, we plan to estimate exposure to second-hand smoke in these couples by measuring cotinine in follicular fluid samples which may better reflect dose at or near the site of interest and may better represent longer-term exposure. We also plan to retest our finding of an increased risk of spontaneous abortion among women who had parents who smoked in a larger, more recent, cohort.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
The authors thank Ms Allison Vitonis for data management. This work was supported by the Flight Attendant Medical Research Institute and Grant R01-HD32153 NICHD, NIH.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Apseloff G, Ashton HM, Friedman H, Gerber N. (1994) The importance of measuring cotinine levels to identify smokers in clinical trials. Clin Pharmacol Ther 56:460–462.[ISI][Medline]
Augood C, Duckitt K, Templeton AA. (1998) Smoking and female infertility: a systematic review and meta-analysis. Hum Reprod 13:1532–1539.
Benowitz NL. (1999) Biomarkers of environmental tobacco smoke exposure. Environ Health Perspect 107:Suppl 2, 349–355.[ISI][Medline]
Boeniger MF, Lowry LK, Rosenberg J. (1993) Interpretation of urine results used to assess chemical exposure with emphasis on creatinine adjustments: a review. Am Ind Hyg Assoc J 54:615–627.[ISI][Medline]
Bonde JP, Joffe M, Sallmen M, Kristensen P, Olsen J, Roeleveld N, Wilcox A. (2006) Validity issues relating to time-to-pregnancy studies of fertility. Epidemiology 17:347–349.[CrossRef][ISI][Medline]
Carrera M and Veiga A. (1998) Chromosome analysis of human preimplantation embryos. Croat Med J 39:121–127.[Medline]
CDC. (2003) Second National Report on Human Exposure to Environmental Chemicals. (Centers for Disease Control and Prevention, Atlanta, GA).
CDC. (2005a) Third National Report on Human Exposure to Environmental Chemicals. (Centers for Disease Control and Prevention, Atlanta, GA) (http://www.cdc.gov/exposurereport/3rd/default.htm) (accessed 27 May 2006).
CDC. (2005b) Health, United States, 2005 with chartbook on trends in the health of Americans. (Centers for Disease Control and Prevention, National Center for Health Statistics, Hyattsville, MD)254–255.
DiCarlantonio G and Talbot P. (1999) Inhalation of mainstream and sidestream cigarette smoke retards embryo transport and slows muscle contraction in oviducts of hamsters (Mesocricetus auratus). Biol Reprod 61:651–656.
Farrelly MC, Evans WN, Sfekas AE. (1999) The impact of workplace smoking bans: results from a national survey. Tob Control 8:272–277.
Fichtenberg CM and Glantz SA. (2002) Effect of smoke-free workplaces on smoking behaviour: systematic review. BMJ 325:188–194.
Gilpin EA, White MM, Farkas AJ, Pierce JP. (1999) Home smoking restrictions: which smokers have them and how they are associated with smoking behavior. Nicotine Tob Res 1:153–162.[CrossRef][Medline]
Haufroid V and Lison D. (1998) Urinary cotinine as a tobacco-smoke exposure index: a minireview. Int Arch Occup Environ Health 71:162–168.[CrossRef][ISI][Medline]
Henderson FW, Reid HF, Morris R, Wang OL, Hu PC, Helms RW, Forehand L, Mumford J, Lewtas J, Haley NJ, et al. (1989) Home air nicotine levels and urinary cotinine excretion in preschool children. Am Rev Respir Dis 140:197–201.[ISI][Medline]
Homburg R and Shelef M. (1995) Factors affecting oocyte quality. In Grudzinskas JG and Yovich JL (eds). Gametes – The Oocyte(Cambridge University Press, Cambridge) pp. 277–291.
Hosmer DW Jr and Lemeshow S. (1989) Model building strategies and methods for logistic regression. Applied Logistic Regression(John Wiley & Sons, New York, USA) pp. 82–134.
Hull MG. (1995) Clinical disorders affecting ovulation and oocyte quality. In Grudzinskas JG and Yovich JL (Eds.). Gametes – The Oocyte(Cambridge University Press, Cambridge) pp. 292–328.
Hull MG, North K, Taylor H, Farrow A, Ford WC. (2000) Delayed conception and active and passive smoking. The Avon Longitudinal Study of Pregnancy and Childhood Study Team. Fertil Steril 74:725–733.[CrossRef][ISI][Medline]
Jauniaux E, Gulbis B, Acharya G, Thiry P, Rodeck C. (1999) Maternal tobacco exposure and cotinine levels in fetal fluids in the first half of pregnancy. Obstet Gynecol 93:25–29.
Joesbury KA, Edirisinghe WR, Phillips MR, Yovich JL. (1998) Evidence that male smoking affects the likelihood of a pregnancy following IVF treatment: application of the modified cumulative embryo score. Hum Reprod 13:1506–1513.
Kharrazi M, DeLorenze GN, Kaufman FL, Eskenazi B, Bernert JT Jr, Graham S, Pearl M, Pirkle J. (2004) Environmental tobacco smoke and pregnancy outcome. Epidemiology 15:660–670.[CrossRef][ISI][Medline]
Langone JJ, Gjika HB, Van Vunakis H. (1973) Nicotine and its metabolites. Radioimmunoassays for nicotine and cotinine. Biochemistry 12:5025–5030.[CrossRef][Medline]
Lindbohm ML, Sallmen M, Taskinen H. (2002) Effects of exposure to environmental tobacco smoke on reproductive health. Scand J Work Environ Health 28:Suppl 2, 84–96.
Lintsen AM, Pasker-de Jong PC, de Boer EJ, Burger CW, Jansen CA, Braat DD, van Leeuwen FE. (2005) Effects of subfertility cause, smoking and body weight on the success rate of IVF. Hum Reprod 20:1867–1875.
Misra DP, Astone N, Lynch CD. (2005) Maternal smoking and birth weight: interaction with parity and mother’s own in utero exposure to smoking. Epidemiology 16:288–293.[CrossRef][ISI][Medline]
Muller A, De La Rochebrochard E, Labbe-Decleves C, Jouannet P, Bujan L, Mieusset R, Le Lannou D, Guerin JF, Benchaib M, Slama R, et al. (2004) Selection bias in semen studies due to self-selection of volunteers. Hum Reprod 19:2838–2844.
Neal MS, Hughes EG, Holloway AC, Foster WG. (2005) Sidestream smoking is equally as damaging as mainstream smoking on IVF outcomes. Hum Reprod 20:2531–2535.
Nohr EA, Frydenberg M, Henriksen TB, Olsen J. (2006) Does low participation in cohort studies induce bias? Epidemiology 17:413–418.[CrossRef][ISI][Medline]
Norwitz ER, Schust DJ, Fisher SJ. (2001) Implantation and the survival of early pregnancy. N Engl J Med 345:1400–1408.
Okah FA, Choi WS, Okuyemi KS, Ahluwalia JS. (2002) Effect of children on home smoking restriction by inner-city smokers. Pediatrics 109:244–249.
Perera FP, Tang D, Tu YH, Cruz LA, Borjas M, Bernert T, Whyatt RM. (2004) Biomarkers in maternal and newborn blood indicate heightened fetal susceptibility to procarcinogenic DNA damage. Environ Health Perspect 112:1133–1136.[ISI][Medline]
Pirkle JL, Flegal KM, Bernert JT, Brody DJ, Etzel RA, Maurer KR. (1996) Exposure of the US population to environmental tobacco smoke: the Third National Health and Nutrition Examination Survey, 1988–91. JAMA 275:1233–1240.[Abstract]
Pirkle JL, Bernert JT, Caudill SP, Sosnoff CS, Pechacek TF. (2006) Trends in the exposure of nonsmokers in the U.S. population to secondhand smoke: 1988–2002. Environ Health Perspect 114:853–858.[ISI][Medline]
Selevan SG and Lemasters GK. (1987) The dose-response fallacy in human reproductive studies of toxic exposures. J Occup Med 29:451–454.[ISI][Medline]
Shiverick KT and Salafia C. (1999) Cigarette smoking and pregnancy I: ovarian, uterine and placental effects. Placenta 20:265–272.[CrossRef][ISI][Medline]
Simpson JL. (1980) Genes, chromosomes, and reproductive failure. Fertil Steril 33:107–116.[ISI][Medline]
Van Voorhis BJ, Dawson JD, Stovall DW, Sparks AE, Syrop CH. (1996) The effects of smoking on ovarian function and fertility during assisted reproduction cycles. Obstet Gynecol 88:785–791.[Abstract]
Van Vunakis H, Gijka HB, Langone JJ. (1993) Radioimmunoassay for nicotine cotinine. In Seifert B, van de Wiel HJ, Dodet B, O’Neill IK (Eds.). Environmental carcinogens—methods of analysis and exposure measurement(International Agency for Research on Cancer, Lyon, France) Vol. 12: pp. 293–299.
Venners SA, Wang X, Chen C, Wang L, Chen D, Guang W, Huang A, Ryan L, O’Connor J, Lasley B, et al. (2004) Paternal smoking and pregnancy loss: a prospective study using a biomarker of pregnancy. Am J Epidemiol 159:993–1001.
Weinberg CR, Wilcox AJ, Baird DD. (1989) Reduced fecundability in women with prenatal exposure to cigarette smoking. Am J Epidemiol 129:1072–1078.
Werler MM. (1997) Teratogen update: smoking and reproductive outcomes. Teratology 55:382–388.[CrossRef][ISI][Medline]
Whyatt RM, Jedrychowski W, Hemminki K, Santella RM, Tsai WY, Yang K, Perera FP. (2001) Biomarkers of polycyclic aromatic hydrocarbon-DNA damage and cigarette smoke exposures in paired maternal and newborn blood samples as a measure of differential susceptibility. Cancer Epidemiol Biomarkers Prev 10:581–588.
Wilcox AJ, Weinberg CR, O’Connor JF, Baird DD, Schlatterer JP, Canfield RE, Armstrong EG, Nisula BC. (1988) Incidence of early loss of pregnancy. N Engl J Med 319:189–194.[Abstract]
Windham GC, Swan SH, Fenster L. (1992) Parental cigarette smoking and the risk of spontaneous abortion. Am J Epidemiol 135:1394–1403.
Woodward A and al-Delaimy W. (1999) Measures of exposure to environmental tobacco smoke. Validity, precision, and relevance. Ann N Y Acad Sci 895:156–172.
Submitted on June 12, 2006; resubmitted on August 1, 2006; resubmitted on September 13, 2006; accepted on September 20, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. D. Meeker, S. A. Missmer, A. F. Vitonis, D. W. Cramer, and R. Hauser Risk of Spontaneous Abortion in Women with Childhood Exposure to Parental Cigarette Smoke Am. J. Epidemiol., September 1, 2007; 166(5): 571 - 575. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||