Click on the links below to access all the ArticlePlus for this article.
Please note that ArticlePlus files may launch a viewer application outside of your web browser.
Small-for-gestational-age (SGA) is defined as birth weight below the 10th percentile according to gestational age and sex based on national standards.1 Maternal smoking during pregnancy is an established risk factor for SGA.2 Although the specific mechanism explaining the impact of smoking on fetal growth is not understood, many plausible ones have been suggested. First, uteroplacental insufficiency would result from a nicotine-mediated reduction in uteroplacental blood flow or from structural changes in the placenta (eg, placental infarcts).3,4 Second, maternal smoking may induce fetal hypoxia through diffusion of carbon monoxide, nicotine, and thiocyanate across the placenta.3,4 Third, changes in the transplacental transport of amino acids or zinc could explain the impact of cigarette smoking on fetal growth.3,5 Finally, chemical components within tobacco smoke could exert direct effects on fetal and placental cells4; these toxic agents, including carcinogen and free radical-forming substances, may cause extensive damage to DNA, proteins, and lipids, as well as to organelles such as mitochondria. Based on these possible mechanisms, the role of genetic susceptibility, defined as polymorphisms in genes involved in the metabolism of tobacco smoke, is relevant but has rarely been studied. Only a few studies have considered the direct role of polymorphic xenobiotic-metabolizing genes in fetal growth6 or their interaction with maternal smoking.7–9 Most of the previous studies have been limited in size, and none determined polymorphisms in both mothers and newborns. In addition, because none was family-based, these studies could not evaluate the excess transmission of the variant alleles to SGA newborns, nor could they assess the role of maternally mediated genetic effects (ie, the effects of genetic variants acting through their influence on the maternal uterine environment as distinct from the maternal transmission of the variant allele). Finally, most studies considered only one or 2 polymorphisms in relevant genes.
Numerous polymorphic genes are involved in the metabolism of tobacco smoke.10 Cytochrome P450 (CYP) proteins represent the first line of defense against toxic lipophilic chemicals11; they participate in the activation of these chemicals (phase I) to allow the attachment of a conjugate by phase II enzymes such as glutathione S-transferases (GST). With the latter reaction, the original substrate becomes more hydrophilic, thus facilitating its excretion from the cell. Through this pathway, most carcinogens are metabolically activated during phase I, resulting in the formation of chemically reactive electrophiles that can then act as DNA-damaging agents and form adducts.12 CYP1A1, a polymorphic gene in the CYP family, is considered to play an important role in the activation of polycyclic aromatic hydrocarbons (PAHs),13,14 which are present in cigarette smoke. Low birth weight has been associated with PAH-DNA adducts in the placenta,15,16 suggesting that placental damage by toxic metabolites may contribute to adverse effects of maternal smoking.
Another mechanism through which CYP1A1 could lead to toxicity is the production of reactive oxygen species.17 These induce cellular damage by oxidation of macromolecules. Evidence for increased oxidative stress is seen in animal studies in which mice are exposed to intrauterine tobacco and alcohol exposure18 and in human SGA and preeclampsia pregnancies.19,20 The mechanism through which reactive oxygen species can negatively act on pregnancy is not known, but oxidant radicals produced by inhaled or ingested chemicals can cause an inflammatory response and increase plasma viscosity, compromising circulation.21 In the placenta, SGA is characterized by thrombosis and vascular insufficiency.
Interindividual differences in expression of the CYPs and GSTs may be due to genetic polymorphisms. Maternal cigarette smoking induces placental CYP1A1 and the isoform has been identified in fetal liver in the early days of development.22,23 GSTM1 and GSTT1, in the GST family, are both involved in the biotransformation of a wide range of reactive toxic and mutagenic compounds, including ROS oxygen species and components of tobacco smoke.10,24,25 GST enzymes are present in large amounts in the placenta in early pregnancy and are expressed early in embryonic development.26,27
Finally, cigarette smoke can generate reactive oxygen species, which are capable of inducing double-strand breaks in DNA. A large number of genes are involved in DNA repair to maintain the integrity of the genetic code, and many are polymorphic.28 XRCC3 is one of the base excision repair system genes that play a pivotal role in the homologous recombination repair machinery.29 Polymorphisms in these genes may have great impact on the level and quality of DNA-damage repair in response to environmental insults and subsequently on induction of effects associated with reduced DNA repair capacity. In a previous study among healthy individuals, the relationship between DNA damage, as measured by 32P-DNA adduct levels, and 3 genetic polymorphisms in different repair genes (XRCC1, XRCC3, and XPD) was studied. The XRCC3 variant was strongly associated with higher DNA adduct levels,30 a factor previously associated with low birth weight.
With this background information, it makes sense to study these genetic polymorphisms as risk factors on the one hand, but mainly in the framework of their ability to modify the effect of maternal smoking on SGA.
We carried out a case–control study and a family-based study on SGA in which the first analyses examined thrombophilic31,32 and atherosclerotic genes.33 We now consider polymorphisms in xenobiotic-metabolizing genes as risk factors for SGA, but mostly as potential modifiers of the effects of maternal smoking. One advantage of having a family design in addition to a case–control design is that the former (which includes father, mother, and newborn) is not vulnerable to population structure bias.34
Selection of Cases and Controls
The case–control study as well as the family-based study are described in detail elsewhere.31–33 Briefly, all cases (birth weight below the 10th percentile according to gestational age and sex, based on national standards)35 seen at our university center between May 1998 and June 2000 were eligible if they were born singleton, alive after the 24th week of gestation, and without severe congenital anomalies. During that period, 505 cases were seen and 493 participated in the study (98%). The same criteria applied to the selection of controls whose birth weights were at or above the 10th percentile. They were matched to cases for gestational week, sex, and race. The mothers of 480 controls were invited to participate and 472 accepted (98%). The project was approved by the Institutional Review Board of the hospital. An informed consent form was signed by the mother to collect cord and maternal blood.
A face-to-face interview, in French or English, was carried out at the hospital with all mothers of cases and controls, generally within 2 days of delivery. The interview included questions about potential confounding factors such as demographic characteristics, anthropometric measures before and after pregnancy, complications of pregnancy, maternal chronic diseases, obstetric history, and smoking habits. We asked smokers about number of cigarettes smoked per day in the month before pregnancy and for each trimester separately.
Approximately midway through the study, we began collecting buccal swabs from fathers of cases and of controls. An informed consent form was signed by all participant fathers for this collection. The goal was to analyze parental trios (mother, father, newborn) to test for association and linkage.34 The response rate among the targeted fathers of cases and controls was 86%. We collected genetic material on 258 case fathers and 248 control fathers.
Genomic DNA was extracted from cells derived either from peripheral blood or mouth epithelium, as previously described.31 Polymerase chain reaction (PCR) allele-specific oligonucleotide hybridization assays were used to genotype 3 polymorphisms in CYP1A136: T6235C (m1) in the 3′ flanking region, A4889G (m2), and C4887A (m4) in exon 7 of the gene. These were then used to define distinct allelic variants: *2A (presence of m1 mutation only), *2B (presence of both m1 and m2), and *4 (m4 only). A comparison of nomenclatures for CYP1A1 can be found in the report by Bartsch et al37; the system we are using has been applied by many others.38–40 The coding-region variant in XRCC3 resulting in the amino acid change Thr241Met (or T241M) was investigated using an allele-specific PCR with subsequent detection of amplicon using a dsDNA-specific dye.41 For GSTT1 and GSTM1, we used PCRs specific for the homozygous deletion (−/−), and for GSTT1, an additional PCR to determine the heterozygous status (+/−) as well as the presence (+/−) of the gene42; PCR products were detected and sized by agarose gel electrophoresis. The primers used are described elsewhere42,43; additional primers were used for GSTM1 as internal controls. For GSTT1, internal control was achieved by the 2 PCRs.
Of 493 cases and 472 controls, 451 were matched for gestational week, sex, and race. Because the matching involved only categorical factors, odds ratios (ORs) and 95% confidence intervals (CIs) were calculated using unconditional logistic regression analysis with all study subjects included. The unit of analysis was the mother–newborn pair, allowing simultaneous adjustment of maternal effects for effects of the newborn genotype and also adjustment of newborn genotype effects for the maternal genotype.
Maternal cigarette smoking was analyzed according to 4 periods: 1 month before pregnancy and the 3 pregnancy trimesters. We categorized the quantity of cigarettes smoked per day in 2 ways. One method categorized smoking according to none, one to 20, and more than 20. The other method collapsed the 2 higher categories into one. The second categorization was used as defense against small numbers in the analyses of interactions. Two models were used. In one model, we adjusted for race (gestational age and sex having been accounted for in the definition of SGA) using either the 3 categories for smoking (model 1) or the 2 categories (model 1a). Another model used the 2 categories for smoking and included race, mother's prepregnancy body mass index, and primiparity as potential confounders.
To analyze the role of genotypes on SGA, we used the case–control data to fit 3 models. The first model included race and the genotype for either mother or newborn. The second model included race and both maternal and newborn genotypes. The third model included race, maternal and newborn genotypes, and the following covariates: prepregnancy body mass index, parity, and smoking during the third trimester. The genotypes were analyzed as variables with 3 categories (one or 2 copies of the deleterious allele vs the wild type), except for the GSTM1 in which the null genotype was compared with the wild type, and for maternal CYP1A1*4, as well as the newborn CYP1A1*2B and *4, in which the numbers in the 2-copies category were too small and thus regrouped with the one-copy category. Hardy-Weinberg equilibrium was tested among controls within racial groups.44 Using the case–control data, we analyzed the interactions between maternal smoking and maternal genotype and between maternal smoking and fetal genotype. We based our assessment on a multiplicative formulation by testing whether the odds ratio for maternal smoking differed across genotype categories. Maternal smoking was used as a 2-category variable (0 vs >0 cigarettes per day) and the genotypes were categorized as baseline (wild type) and one or 2 copies of the variant allele (except for GSTM1 in which the null genotype is compared with the wild type, and for CYP1A1*2B and *4, in which one or 2 copies were regrouped because of small numbers). A test of heterogeneity based on a chi-squared statistic with one or 2 degrees of freedom was used to test for interaction between maternal smoking and the genotypes.
We used a log-linear model to analyze the family trio data from the case group and used the family trio data from controls to confirm the required assumption of Mendelian transmissions.45 In this model, likelihood-ratio tests are used to compare larger models to reduced submodels. Each model achieves robustness against genetic population structure through stratification on the 6 possible parental “mating types” as defined by Schaid and Sommer.46 We used 4 models, in which the risks associated with one and 2 copies (in the newborn or in the mother) are estimated separately when included. The first model included maternal and newborn variant alleles; another (model 2) included maternal but not newborn alleles; another included newborn but not maternal alleles (model 3); and a null model (model 4) included only the mating types: comparing the first with the second model tests for newborn genotype while adjusting for maternally mediated genotype effects; comparing the first with the third tests for maternal effects while adjusting for newborn effects; comparing model 3 with the null model tests for the contribution of the fetal alleles to risk of SGA (without adjustment for the maternal genotype). The log-linear likelihood-ratio test is a powerful and more flexible alternative to the transmission disequilibrium test (TDT) and tests the same null hypothesis of no linkage disequilibrium.45 Weinberg47 also proposed an expectation–maximization algorithm set in a likelihood framework to allow the inclusion of incomplete trios in the analysis. This approach is fully implemented in the LEM software48 that was used here. For newborns who had only one available parent, Sun et al49 proposed another approach to study transmission of alleles, a method that has lower power than the TDT. One cannot test for maternal effects using only cases and their mothers, and must assume there are no maternal effects. Finally, a set of control trios, in numbers similar to the case trios, was used to check for transmission distortion (the preferential transmission of an allele to surviving unaffected infants, suggesting effects of the allele on survival).34
Table 1 shows the characteristics of cases and controls. Case and control newborns were similar in sex distribution, gestational age, and race by study design. Among risk factors for SGA, case mothers had lower body mass index and were more often primiparous.
In Table 2, cigarette smoking before pregnancy is seen to be common (30% among cases and 27% among controls). In all trimesters, more case than control mothers smoked. Odds ratios were highest in the category defined as 20 or more cigarettes daily. The odds ratios for smoking any quantity during pregnancy were elevated in all trimesters and were all close to 2.0.
The contribution of genotype to the risk of SGA was analyzed with the case–control approach (Tables 3 and 4) and with the family-based approach (Table 5). Hardy-Weinberg equilibrium was found among controls in the respective racial groups. In Table 3, the case–control analysis gives evidence that mothers carrying the partial or complete deletion genotypes (−/+ and −/−) for GSTT1 have a substantial reduction in risk of SGA. The newborns (Table 4) showed some patterns not seen in the mothers. Although the numbers with CYP1A1*2A homozygous variant genotype were small, this genotype was associated with an increased risk in the first 2 models. The partial deletion for GSTT1 was associated with an increase in risk of approximately 40% in the adjusted model 2. Finally, the complete deletion for GSTM1 in the newborn was associated with a reduction in risk.
We also computed the crude results for maternal and newborn genotypes for whites and blacks separately (table available with the electronic version of this article); the numbers are small for the group of black mothers and newborns. There are no marked differences between the 2 subgroups for maternal genotypes. With respect to the newborn results, the increase in risk with GSTT1 and the reduction in risk with GSTM1 apparently come from the white group.
Another way to study the contribution of maternal and newborn genes to the risk of SGA is with case–parent trios (Table 5). Whereas a maximum of 258 case fathers were genotyped for CYP1A1, none was genotyped for GSTT1, GSTM1, or XRCC3. For CYP1A1, no maternally mediated effects were found. The maternally mediated effects could not be tested with the other genes because we did not have the required paternal genotypes. On the other hand, there were indications for the reduced transmission of the variant allele for CYP1A1*2A (P = 0.02) and excess transmission of the variant CYP1A1*4 (P = 0.07). The conservative 1-TDT test developed by Sun et al29 showed evidence of an excess in transmission for GSTT1 (TDT = 2.08; P = 0.03) and a deficit in transmission for GSTM1 (TDT = 1.3; P = 0.19) (data not shown). Using the 1-TDT test in our control trios, we found no evidence suggesting an excess or deficit in transmission for any of the studied variants.
Finally, in Table 6, the results showed an interaction between maternal smoking before pregnancy and the maternal CYP1A1*2A genotype in a pattern consistent with a smoking effect restricted to mothers who are homozygous for the variant. However, the data are presented mainly as descriptive because they are too sparse to lead to any strong conclusion. Maternal smoking during pregnancy was not associated with SGA in mothers with the XRCC3 homozygous variant genotype, whereas it was associated with increased risks in the other genotype categories, resulting in estimated effects that varied across strata. The estimated effect of maternal smoking also differed across the newborn genotypes; newborn carriers of one or 2 copies of the CYP1A1*4 variant were at markedly increased risk of SGA from maternal smoking. In addition, newborns carrying one or 2 deletions of GSTT1 showed evidence for increased susceptibility to effects of maternal smoking. Finally, the results for newborn XRCC3 were in the same direction as those observed for mothers.
Maternal smoking during pregnancy is a well-known risk factor for SGA. Our results suggest a protective effect for the partial and complete deletion in maternal GSTT1 and, for the newborn, complete deletion in GSTM1. On the other hand, there was a suggestive decrease in risk for the newborn carrying one copy of CYP1A1*2A variant and an increase for those with the partial deletion for GSTT1. In the trio data, the same variant alleles (CYP1A1*2A and GSTT1) are transmitted to the SGA cases in excess or deficit of expectation. Transmission of the GSTM1 deletion did not show a statistically significant departure from the expected either in cases or controls when we used a conservative test; nonetheless, there was a deficit among cases and an excess among controls, the ratio of which leads to an overall protective effect, as seen in the case–control analysis. The analysis of maternally mediated effects, limited to CYP1A1, showed no indication of this type of effect. Finally, the CYP1A1*2A and XRCC3 maternal variants and the newborn GSTT1 deletion show statistically significant interactions with maternal smoking, although the CYP1A1*2A data are too sparse to produce more than tentative results. An additional caution with our results is the fact that we tested a number of genes.
In other studies, Yamada et al6 report no association between maternal CYP1A1 (m1) (or *2A in our study), GSTM1, or GSTT1 and SGA in Japanese women. However, the study was small (a total of 134 mothers, 17 of whom were SGA). The homozygous variant CYP1A1 (m1) was associated with a risk of 0.35 (95% CI = 0.06–1.83), whereas the null GSTM1 genotype was associated with an odds ratio of 1.14 (0.36–3.66) and the null GSTT1 genotype with an OR of 0.84 (0.25–2.66). Hong et al,8 in a study from Korea, included 266 pregnant women. They found no relationship of the maternal null deletion genotypes for GSTT1 and GSTM1 with birth weight. Carriers of the null GSTM1 genotype had a mean birth weight of 2,970 g (95% CI = 2,788–3,153) as compared with 3,007 g (2,814–3,200) for those without the deletion. For GSTT1, those results were 2,946 g (2,760–3,131) as compared with 3,025 g (2,838–3,231). However, they found indications of interactions between GSTT1 and maternal smoking; birth weight was lowest among women with the null genotype who were smokers. Results were similar in our study because the effect of smoking among those with the deletion increased the risk of SGA.
Finally, Wang et al7 studied the interaction between maternal smoking and maternal CYP1A1 (m1) and GSTT1 null deletion polymorphisms in a group of 741 mothers in Boston. For intrauterine growth restriction, they found a marginally significant interaction between CYP1A1 (m1) and maternal smoking (the OR was highest (4.1; 95% CI = 2.0–8.6) among carriers of the variant who smoked throughout pregnancy), but no interaction for GSTT1. However, for low birth weight, the interaction between maternal smoking and CYP1A1 (m1) was again marginally significant, with the greatest birth weight reduction observed among the carriers of the variant allele who smoked throughout pregnancy (−520 g; 95% CI = −276.9 to −763.0). For GSTT1, the odds ratios for maternal smoking were heterogeneous between the genotype categories showing the highest birth weight reduction (−642 g; −340.1 to −943.8) among continuous smokers with the deletion. Overall, the latter case–control results are similar to ours, although the studies considered only 2 polymorphisms, had no newborn genetic information, and did not analyze the transmission in families of cases or of controls.
It is not surprising that genetic polymorphisms known to be involved in the metabolism of xenobiotics might modify the effects of exposure to such contaminants. However, very few studies of gene–environment interactions have been conducted regarding association with SGA, and the present results confirm what is theoretically suspected from toxicologic studies. Of particular interest is the fact that not only maternal genotypes are involved, but also newborn genotypes. In addition, although maternally mediated effects could be suspected, at least for the 3 CYP1A1 variants, the data suggest otherwise. These findings suggest that it is not through the influence of maternal CYP1A1 variants on uterine environment that the effect of cigarette smoking occurs, but rather through transmission of the genes to the fetus. More research, particularly from family-based studies, is needed to examine the role of these genes. Finally, the 2 other studies that considered the genotypes as risk factors were too small to detect effects.6,8 In the present study, the case–control approach and the case–parental design show similar results. These results suggest possible mechanisms for SGA; some polymorphic genes involved in many mechanisms of defense against xenobiotics are found to have been transmitted more frequently to conceptuses who became small babies, presumably because they had become more vulnerable to a large number of contaminants to which their mothers had been exposed from diverse sources.
Our study's strengths include its size and the low probability for selection bias given the high response rate. Furthermore, it includes not only a case–control study, but a family-based study among cases, as well as among controls. Another advantage of this genetic study is that the selected candidate genes are known to play a role in the studied metabolism.
Among the limitations of this study is the fact that not all polymorphisms were determined among fathers and that not all fathers were included. In addition, the metabolism of cigarette smoke is complex, and most certainly other genes are involved. Our analysis provides evidence that there is genetic modification of the effect of maternal smoking on risk of SGA through CYP1A1*2A (maternal genotype) and GSTT1 (fetal genotype), but our maternal model was not adjusted for the (correlated) fetal genotype or vice versa. An even larger study would be required to distinguish with confidence between modifications acting through the maternal genotype or through the fetal genotype.
In conclusion, the study provides new information on the role of polymorphic genes involved in the metabolism of tobacco smoke with respect to SGA. It suggests again that environmental factors and genes influence SGA, both independently and through their interactions.
We thank Emily Kistner for her comments on a previous version of this paper.
1. Intrauterine growth restriction. Clinical management guidelines for obstetrician–gynecologists. ACOG Practice Bulletin. 2000;12:1–11.
2. Higgins S. Smoking in pregnancy. Curr Opin Obstet Gynecol. 2002;14:145–151.
3. Andres RL, Day MC. Perinatal complications associated with maternal tobacco use. Semin Neonatol. 2000;5:231–241.
4. Shiverick KT, Salafia C. Cigarette smoking and pregnancy. I: ovarian, uterine and placental effects. Placenta. 1999;20:265–272.
5. Rama Sastry BV. Placental toxicology: tobacco smoke, abused drugs, multiple chemical interactions, and placental function. Reprod Fertil Dev. 1991;3:355–372.
6. Yamada H, Sata F, Kato EH, et al. A polymorphism in the CYP17 gene and intrauterine fetal growth restriction. Mol Hum Reprod. 2004;10:49–53.
7. Wang X, Zuckerman B, Pearson C, et al. Maternal cigarette smoking, metabolic gene polymorphism, and infant birth weight. JAMA. 2002;287:195–202.
8. Hong YC, Lee KH, Son BK, et al. Effects of the GSTM1 and GSTT1 polymorphisms on the relationship between maternal exposure to environmental tobacco smoke and neonatal birth weight. J Occup Environ Med. 2003;45:492–498.
9. Wu D, Zhang XQ, Yang F, et al. Analysis of association between gene polymorphisms of microsomal epoxide hydrolase (EPHX1) and infant birthweight (Chinese). Acta Genetica Sinica. 2001;28:595–600.
10. Besarati NA, Kleinjans JCS, Van Schooten FJ. Biomonitoring of tobacco smoke carcinogenicity by dosimetry of DNA adducts and genotyping and phenotyping of biotransformational enzymes: a review of polycyclic aromatic hydrocarbons. Biomarkers. 2002;7:209–229.
11. Hirnoven A. Polymorphisms of xenobiotic-metabolizing enzymes and susceptibility to cancer. Environ Health Perspect. 1999;107(suppl 1):37–47.
12. Smith G, Stanley LA, Sim E, et al. Metabolic polymorphisms and cancer susceptibility. Cancer Surveys. 1995;25:27–65.
13. Everson RB, Randerath E, Santella RM, et al. Quantitative associations between DNA damage in human placenta and maternal smoking and birth weight. J Natl Cancer Inst. 1988;80:567–576.
14. Perera FP, Jedrychowski W, Rauh V, et al. Molecular epidemiologic research on the effects of environmental pollutants on the fetus. Environ Health Perspect. 1999;107(suppl 3):451–460.
15. Barouki R, Morel Y. Repression of cytochrome P450 1A1 gene expression by oxidative stress: mechanisms and biological implications. Biochem Pharmacol. 2001;511–516.
16. Li Y, Wang H. In utero exposure to tobacco and alcohol modifies neurobehavioral development in mice offspring: consideration a role of oxidative stress. Pharmacol Res. 2004;49:467–473.
17. Kim YJ, Hong YC, Leeb KH, et al. Oxidative stress in pregnant women and birth weight reduction. Reprod Toxicol. 2005;19:487–492.
18. Myatt L, Cui X. Oxidative stress in the placenta. Histochem Cell Biol. 2004;122:369–382.
19. Loscalzo J. Ozone—from environmental pollutant to atherosclerotic determinant. N Engl J Med. 2004;350:834–835.
20. Shimada T, Yun CH, Yamazaki H, et al. Characterization of human lung microsomal cytochrome P450 1A1 and its role in oxidation of chemical carcinogens. Mol Pharmacol. 1992;41:856–864.
21. Guengerich FP, Shimada T. Activation of procarcinogens by human cytochrome P450 enzymes. Mutat Res. 1998;400:201–213.
22. Hakkola J, Pelkonen O, Pasanen M, et al. Xenobiotic-metabolizing cytochrome P450 enzymes in the human feto-placental unit: role in intrauterine toxicity. Crit Rev Toxicol. 1998;28:35–72.
23. Yang H-YL, Namkung MJ, Juchau MR. Expression of functional cytochrome P4501A1 in human embryonic tissues during organogenesis. Biochem Pharmacol. 1995;49:717–726.
24. Landi S. Mammalian class theta GST and differential susceptibility to carcinogens: a review. Mutat Res. 2000;463:247–283.
25. Norppa H. Cytogenetic biomarkers and genetic polymorphisms. Toxicol Lett. 2004;149:309–334.
26. Pasanen M. The expression and regulation of drug metabolism in human placenta. Adv Drug Delivery Rev. 1999;38:81–97.
27. van Lieshout EM, Knapen MF, Lange WP, et al. Localization of glutathione S-transferases alpha and pi in human embryonic tissues at 8 weeks gestational age. Hum Reprod. 1998;13:1380–1386.
28. Ronen A, Glickman BW. Human DNA repair genes. Environ Mol Mutagen. 2001;37:241–283.
29. Indahl T, Karran P, Wood RD. DNA excision repair pathways. Curr Opin Genet Dev. 1997;7:158–169.
30. Matullo G, Palli D, Peluso M, et al. XRCC1, XRCC3, XPD gene polymorphisms, smoking and 32P-DNA adducts in a sample of healthy subjects. Carcinogenesis. 2001;22:1437–1445.
31. Infante-Rivard C, Rivard GE, Yotov WV, et al. Absence of association of thrombophilia polymorphisms with intrauterine growth restriction. N Engl J Med. 2002;347:19–25.
32. Infante-Rivard C, Rivard GE, Guiguet M, et al. Thrombophilic polymorphisms and intrauterine growth restriction. Epidemiology. 2005;16:281–287.
33. Infante-Rivard C, Lévy E, Rivard GE, et al. Small babies receive the cardiovascular protective apolipoprotein ε2 allele less frequently than expected. J Med Genet. 2003;40:626–629.
34. Spielman RS, McGinnis RE, Ewens WJ. Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet. 1993;52:506–516.
35. Arbuckle TE, Wilkins R, Sherman GJ. Birth weight percentiles by gestational age in Canada. Obstet Gynecol. 1993;81:39–48.
36. Labuda D, Krajinovic M, Richer C, et al. Rapid detection of CYP1A1, CYP2D6, and NAT variants by multiplex polymerase chain reaction and allele-specific oligonucleotide assay. Anal Biochem. 1999;275:84–92.
37. Bartsch H, Nair U, Risch A, et al. Genetic polymorphism of CYP genes, alone or in combination, as a risk modifier of tobacco-related cancers. Cancer Epidemiol Biomarkers Prev. 2000;9:3–28.
38. Krajinovic M, Labuda D, Mathonnet G, et al. Polymorphisms in genes encoding drugs and xenobiotic metabolizing enzymes, DNA repair enzymes, and response to treatment of childhood acute lymphoblastic leukemia. Clin Cancer Res. 2002;8:802–810.
39. Krajinovic M, Ghadirian P, Richer C, et al. Genetic susceptibility to breast cancer in French-Canadians: role of carcinogen-metabolizing enzymes and gene–environment interactions. Int J Cancer. 2001;92:220–225.
40. D'Alo F, Voso MT, Guidi F, et al. Polymorphisms of CYP1A1 and glutathione S-transferase and susceptibility to adult acute myeloid leukemia. Haematologica. 2004;89:664–670.
41. Moran P, Dightman DA, Park LK. Nonelectrophoretic genotyping using allele-specific PCR and a dsDNA-specific dye. Biotechniques. 1998;24:206–212.
42. Sprenger R, Schlagenhaufer R, Kerb R, et al. Characterization of the glutathione S-transferase GSTT1 deletion: discrimination of all genotypes by polymerase chain reaction indicates a trimodular genotype–phenotype correlation. Pharmacogenetics. 2000;10:557–565.
43. Stücker I, Jacquet M, de Waziers I, et al. Relation between inducibility of CYP1A1, GSTM1 and lung cancer in a French population. Pharmacogenetics. 2000;10:617–627.
44. Schaid DJ, Jacobsen SJ. Biased tests of association: comparisons of allele frequencies when departing from Hardy-Weinberg proportions. Am J Epidemiol. 1999;149:706–711.
45. Wilcox AJ, Weinberg CR, Lie RT. Distinguishing the effects of maternal and offspring genes through studies of ‘case–parent triads’. Am J Epidemiol. 1998;148:893–901.
46. Schaid DJ, Sommer SS. Genotype relative risks; methods for design and analysis of candidate-gene association studies. Am J Hum Genet. 1993;53:1114–1126.
47. Weinberg CR. Allowing for missing parents in genetic studies of case-parent triads. Am J Hum Genet. 1999;64:1186–1193.
48. van den Oord EJCG, Vermunt JK. Testing for linkage disequilibrium, maternal effects, and imprinting with (in)complete case-parent triads, by use of the program LEM. Am J Hum Genet. 2000;66:335–338.
49. Sun F, Flanders WD, Yang Q, Khoury MJ. Transmission disequilibrium test (TDT) when only one parent is available: the 1-TDT. Am J Epidemiol. 1999;150:97–104.
© 2006 Lippincott Williams & Wilkins, Inc.