Molecular Human Reproduction Vol.11, No.5 pp.

357360, 2005 Advance Access publication April 22, 2005

doi:10.1093/molehr/gah175

Caffeine intake, CYP1A2 polymorphism and the risk of recurrent pregnancy loss
Fumihiro Sata1,3, Hideto Yamada2, Kana Suzuki1, Yasuaki Saijo1, Emi H Kato2, Mamoru Morikawa2, Hisanori Minakami2 and Reiko Kishi1
Department of Public Heath and 2Department of Obstetrics and Gynecology, Hokkaido University Graduate School of Medicine, Kita 15, Nishi 7, Kita-ku, Sapporo 060-8638, Japan
3 1

To whom correspondence should be addressed. E-mail: fsata@med.hokudai.ac.jp

Some case control studies have demonstrated that caffeine intake and high CYP1A2 activity increase risks of recurrent pregnancy loss (RPL) but the multifactorial effect is obscure. To investigate whether susceptible women who have more caffeine intake are at high risk of RPL, a case control study of 58 cases with two or more RPL and fertile 147 controls was performed. The association between daily caffeine intake together with CYP1A2*1F (AA versus CA and CC) genotype and RPL was assessed. Without consideration of the genotype, there were no signicant differences of the RPL risk in proportion to daily caffeine intake [less than 100 mg (reference); 100299 mg: odds ratio (OR), 1.29; 95% condence interval (CI), 0.66 2.50; 300 mg or more: OR, 1.82; 95% CI, 0.72 4.58; P for trend, 0.20]. However, the RPL risk signicantly increased only among women who had homozygous CYP1A2*1F alleles with a dosage effect of daily caffeine intake [less than 100 mg (reference); 100299 mg: OR, 1.94; 95% CI, 0.57 6.66; 300 mg or more: OR, 5.23; 95% CI, 1.05 25.9; P for trend, 0.03]. It was demonstrated for the rst time that an increase in caffeine intake deteriorates the fecundity among susceptible women. Key words: caffeine/CYP1A2/genetic polymorphism/molecular epidemiology/recurrent pregnancy loss

Introduction
About 10 14% of clinically recognized pregnancies end in pregnancy loss in the Japanese population as well as in Caucasians. The etiology of recurrent pregnancy loss (RPL) remains largely unclear (Stirrat, 1990; Parazzini et al., 1991; Cramer and Wise, 2000; Yamada et al., 2001). Epidemiological studies have suggested that the condition might be multifactorial with a possible genetic predisposition and involvement of environmental factors in its pathogenesis (Parazzini et al., 1991; Cramer and Wise, 2000; Yamada et al., 2005). Some epidemiological studies have suggested that maternal caffeine consumption increases a risk of sporadic pregnancy loss, but the results are inconsistent (Fenster et al., 1991; Infante-Rivard et al., 1993; Mills et al., 1993; Cnattingius et al., 2000; Signorello and McLaughlin, 2004). CYP1A2 is an enzyme primarily responsible for caffeine metabolism, and caffeine clearance rates differ among individuals through genetic propensity and in response to environmental factors (Landi et al., 1999). An increase in caffeine intake has been found to be a risk factor for pregnancy loss among women with high but not low CYP1A2 activity (Signorello et al., 2001). It has been suggested that the homozygous CYP1A2*1F (A/A) genotype, thought to represent a CYP1A2 high inducibility genotype, may either be a direct cause of an increased CYP1A2 activity, or be genetically linked to polymorphisms conferring high inducibility (Sachse et al., 1999). The aim of this study was to investigate whether the susceptible women who have more caffeine intake are at high risk of RPL.

Materials and methods

This casecontrol study was performed in the city of Sapporo, Japan, during the years 20032004. We conducted the self-administered questionnaire survey to 113 women with a history of RPL and 298 women whose pregnancies ended in live births, who were obstetrically managed in the Hokkaido University Hospital. The questionnaire provided information on medical and obstetric history, maternal age, smoking habits before and during previous pregnancy, and daily caffeine intake during previous pregnancy (i.e. before miscarriages of RPL women and in early pregnancies of fertile women). We estimated caffeine intake according to the frequency questions for coffee (caffeinated and decaffeinated), green tea, black tea, oolong tea and cola as previously described (Nagata et al., 1998). The estimation was based on the assumption that caffeinated coffee contained 40 mg of caffeine per 100 ml, green tea 20 mg/100 ml, black tea 50 mg/100 ml, oolong tea 20 ml/100 ml and cola 30 ml/100 ml. Sixty-four RPL women and 173 fertile women answered the questionnaire and the response rates were 56.6 and 58.1%, respectively. Six RPL women with a uterine conformational abnormality or with couples balanced type chromosomal translocation and 26 fertile women who had a history of endometriosis, infertility or who delivered small-for-gestationalage babies were excluded from this study. Thereafter, we studied 58 patients aged 25 43 years with a history of RPL and 147 controls aged 1944 years. All the patients with RPL and controls were resident in Sapporo and the surrounding areas in Japan; all the patients and the controls were native Japanese. In recent times, the geographical region has had little population immigration by different ethnic groups. The characteristics of the study groups are shown in Table I. RPL was dened as having a history of two or more consecutive pregnancy losses. The primary RPL group compromised 46 women with a history of two or more pregnancy losses but no live birth. The 12 secondary RPL women experienced three or more pregnancy losses after at least one live birth. A total of 53 RPL women experienced all their

Molecular Human Reproduction q The Author 2005. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oupjournals.org 357

F.Sata et al.
Table I. Characteristics of 58 cases with recurrent pregnancy loss (RPL) and 147 controls in a Japanese population Cases Number Age (years) # 29 12 3039 36 $ 40 10 Pregnancy loss 2 26 3 17 $4 15 Primary RPL 46 Secondary RPL 12 , 9 weeks 27 913 weeks 26 $ 14 weeks 5 CYP1A2 genotype C/C 5 C/A 33 A/A 20 Caffeine intake (mg/day) 099 21 100299 27 $ 300 10 Smoking during pregnancy Never 45 Quitter 8 Continuous 5 Passive smoking during pregnancy* No 27 Yes 31 *One control was unknown. miscarriages in the rst trimester. Twenty-seven of 53 women with RPL experienced all their miscarriages before 9 weeks of gestation, but another 26 women experienced at least one miscarriage between 9 and 13 weeks of gestation, and the other ve women experienced at least one miscarriage after 14 weeks of gestation. All women with RPL were subjected to examination by ultrasound and hysterosalpingography for detection of anatomical abnormalities of the genital tract, to chromosome karyotypic analyses of peripheral blood, and to the other RPL screening. The women who had RPL were signicantly older than the control subjects (mean 33.8 versus 31.2 years, P , 0.01). This study was conducted with all the subjects informed consent and approved by the institutional ethical board for human gene and genome studies of Hokkaido University Graduate School of Medicine. % Controls Number % Their blood samples had been obtained at the department of Obstetrics in the Hokkaido University Hospital. Genomic DNA was extracted from the peripheral blood using the QIAamp DNA Blood Kit (QIAGEN GmbH, Hilden, Germany) according to the manufacturers instructions. Genotyping of CYP1A2*1F polymorphism was determined by either PCR RFLP (Christiansen et al., 2000; Saijo et al., 2004b) or allelic discrimination using uorogenic probes and the 50 nuclease (TaqMan) assay as described (Ranade et al., 2001). Odds ratios (OR) and 95% condence intervals (CI) were calculated by unconditional logistic regression analysis. HardyWeinberg equilibrium analyses were performed to compare observed and expected genotype frequencies using a chi-square test. Furthermore, a caffeine dosage effect (P for trend) was assessed by modelling a linear effect on the log odds scale for each category of caffeine intake (less than 100 mg/day, 100 299 mg/day and 300 mg/day or more) in a logistic regression model. All analyses were conducted using SPSS software for Windows version 11.0 (SPSS Inc., Chicago, USA).

20.7 62.1 17.2 44.8 29.3 25.9 79.3 20.7 46.6 44.8 8.6 8.6 56.9 34.5 36.2 46.6 17.2 77.6 13.8 8.6 46.6 53.4

52 87 8 16 59 72 65 65 17 100 25 22 64 82

35.4 59.2 5.4 10.9 40.1 49.0 44.2 44.2 11.6 68.0 17.0 15.0 43.8 56.2

Results
Without consideration of the genotype, there was a slight tendency to increase the RPL risk in proportion to daily caffeine intake [less than 100 mg (reference); 100 299 mg: OR, 1.29; 95% CI, 0.66 2.50; 300 mg or more: OR, 1.82; 95% CI, 0.72 4.58; P for trend, 0.20], but with no signicance (Table II). The distribution of the CYP1A2 genotype in each group was in Hardy Weinberg equilibrium. Thirty-three (56.9%) cases were heterozygous, and 20 (34.5%) were homozygous for CYP1A2*1F, compared with 59 (40.1%) and 72 (49.0%) controls. The RPL risk signicantly increased only among women who had homozygous CYP1A2*1F alleles with a dosage effect of daily caffeine intake [less than 100 mg (reference); 100 299 mg: OR, 1.94; 95% CI, 0.57 6.66; 300 mg or more: OR, 5.23; 95% CI, 1.05 25.9; P for trend, 0.03]. However, caffeine intake had no effect on the RPL risk among women who had CYP1A2 CC (*1A*1F) or CA (*1A*1F) genotype (Table II).

Discussion
Recently, many investigations have demonstrated that the maternal genetic polymorphisms related to RPL risk without consideration of the effects of environmental factors (Yamada et al., 2005); these genes included factor V Leiden and prothrombin mutations (Rey

Table II. Associations of maternal caffeine intake during pregnancy with RPL by maternal CYP1A2 genotypes Genotype Caffeine intake (mg/day) Cases Number Total sample 099 100299 $ 300 CYP1A2 CC CA 099 100299 $ 300 099 100299 $ 300 21 27 10 16 17 5 5 10 5 36.2 46.6 17.2 42.1 44.7 13.2 25.0 50.0 25.0 65 65 17 31 33 11 34 32 6 44.2 44.2 11.6 41.3 44.0 14.7 47.2 44.4 8.3 1.00 (reference) 1.29 (0.662.50) 1.82 (0.724.58) P for trend, 0.20 1.00 (reference) 1.03 (0.422.52) 1.03 (0.293.70) P for trend, 0.87 1.00 (reference) 1.94 (0.576.66) 5.23 (1.0525.9) P for trend, 0.03 % Controls Number % OR* (95% CI)

CYP1A2 AA

*Adjusted to age and smoking status during pregnancy.

358

Caffeine, CYP1A2 and the risk of recurrent pregnancy loss et al., 2003), plasminogen activation inhibitor I and factor XIII (Dossenbach-Glaninger et al., 2003), coagulation factors (Wramsby et al., 2000; Pihusch et al., 2001), methylenetetrahydrofolate reductase (Lissak et al., 1999), HLA-G (Aldrich et al., 2001; Pfeiffer et al., 2001), glutathione S-transferase M1 (GSTM1) (Sata et al., 2003a), glutathione S-transferase P1 (GSTP1) (Zusterzeel et al., 2000), interleukin (IL)-1 (Unfried et al., 2001; Wang et al., 2002; Karhukorpi et al., 2003), IL-6 (Saijo et al., 2004a), cytochrome P450 17alpha-hydroxylase/17,20-lyase (CYP17) (Sata et al., 2003b) and endothelial nitric oxide synthase (NOS3) (Tempfer et al., 2001). However, other genes of xenobiotic-metabolizing enzymes including the CYP1A2 gene by themselves did not signicantly affect RPL risk (Saijo et al., 2004b). In the present study, our ndings suggest that possession of homozygous CYP1A2*1F alleles may have an increased risk of RPL with a dosage effect of caffeine intake, and provide additional evidence of RPL being a multifactorial disease affected by not only genetic factors but also environmental ones. Previous studies have assessed the effect of CYP1A2 activity on the risk of sporadic pregnancy loss (Fenster et al., 1998; Signorello et al., 2001). Fenster et al. (1998) performed a case control study enrolling 73 cases and 141 controls, whereas they found no interaction between CYP1A2 phenotype and caffeine intake in relation to pregnancy loss. On the other hand, Signorello et al. (2001) performed a case control study enrolling 101 cases and 953 controls and demonstrated that caffeine intake was a risk factor for pregnancy loss among women with high but not low CYP1A2 activity. Both studies were based on the caffeine test using urinary metabolic ratio as an indicator of CYP1A2 phenotype and activity. It was known that the homozygous CYP1A2*1F genotype carried a higher CYP1A2 inducibility than other genotypes (C/C and C/A) (Sachse et al., 1999). A trend towards enhanced activity was observed in pregnant women who had the CYP1A2*1F allele (Nordmark et al., 2002). However, there have been no studies assessing the CYP1A2 genotype and caffeine intake in relation to the RPL risk. Serum caffeine and paraxanthine, the primary metabolite of caffeine, concentrations were positively related with the caffeine consumption reported by pregnant women, but the serum paraxanthine concentration was more closely related with caffeine consumption than was the serum caffeine concentration, particularly among smokers (Klebanoff et al., 1998). It was reported that serum concentration of paraxanthine was higher in women with RPL than in women who give birth to live infants (Klebanoff et al., 1999). The novel methods for the determination of CYP1A2 activity using paraxanthine/caffeine ratio were proposed to show good predictive performance compared with the conventional method (Doude van Troostwijk et al., 2003). It was reported that homozyous CYP1A2*1F alleles inuenced the induction of CYP1A2 activity in vivo (Han et al., 2002). It is possible that paraxanthine may be a key metabolite of caffeine to predispose to RPL because women with CYP1A2*1F alleles could generate paraxanthine more abundantly. The present study discovered that the caffeine intake during early pregnancy was associated with an increased risk of RPL only among susceptible women. Women who had this high inducible CYP1A2 genotype, homozygous CYP1A2*1F alleles and ingested daily 300 mg or more possessed the highest risk of RPL (OR 5.23). However, caffeine intake had no effects among women who had other genotypes. Thus, we for the rst time demonstrated the susceptible genotype with environmental factors such as caffeine intake affecting the human fecundity. We believe that results of this study contain useful information in clinical practice to prevent RPL. It is of interest whether reduction in caffeine consumption early during pregnancy alter pregnancy outcome among women who had a history of pregnancy loss and the homozygous CYP1A2*1F alleles. In the present study, the response rate of the questionnaire was relatively low, just over 50%. It was possible that the women who took an interest in caffeine consumption or experienced pregnancy loss would answer the questionnaire more frequently. However, in our questionnaire survey we did not inform the women that the frequency of soft drinks questions contained caffeine. There was no signicant difference between the response rate of patients with RPL and that of the controls. There was also no signicant difference between age or CYP1A2 allelic frequency of the responders and that of the non-responders. Although these two groups, the responders and the non-responders, seemed to be similar according to our limited information, a response bias could not generally be avoidable. Further prospective studies should be performed to conrm our preliminary ndings. Recent investigations outlined that RPL is a multifactorial-polygenetic disease (Yamada et al., 2001; Yamada et al., 2005). Molecular epidemiological studies are further needed to unequivocally elucidate the multifactorial effects of both genetic and environmental factors in human fecundity.

Acknowledgements
This work was supported in part by Grants-in-aid for Scientic Research from the Japan Society for the Promotion of Science and the Japan Ministry of Health, Labour and Welfare. We thank Dr T. Kondo, Ms M. Sakuramachi and Ms T. Kunita for their technical assistance.

References
Aldrich CL, Stephenson MD, Karrison T, Odem RR, Branch DW, Scott JR, Schreiber JR and Ober C (2001) HLA-G genotypes and pregnancy outcome in couples with unexplained recurrent miscarriage. Mol Hum Reprod 7,11671172. Christiansen L, Bygum A, Jensen A, Thomsen K, Brandrup F, Horder M and Petersen NE (2000) Association between CYP1A2 polymorphism and susceptibility to porphyria cutanea tarda. Hum Genet 107,612614. Cnattingius S, Signorello LB, Anneren G, Clausson B, Ekbom A, Ljunger E, Blot WJ, McLaughlin JK, Petersson G, Rane A et al. (2000) Caffeine intake and the risk of rst-trimester spontaneous abortion. N Engl J Med 343,18391845. Cramer DW and Wise LA (2000) The epidemiology of recurrent pregnancy loss. Semin Reprod Med 18,331339. Dossenbach-Glaninger A, van Trotsenburg M, Dossenbach M, Oberkanins C, Moritz A, Krugluger W, Huber J and Hopmeier P (2003) Plasminogen activator inhibitor 1 4G/5G polymorphism and coagulation factor XIII Val34Leu polymorphism: impaired brinolysis and early pregnancy loss. Clin Chem 49,1081 1086. Doude van Troostwijk LJ, Koopmans RP and Guchelaar HJ (2003) Two novel methods for the determination of CYP1A2 activity using the paraxanthine/caffeine ratio. Fundam Clin Pharmacol 17,355 362. Fenster L, Eskenazi B, Windham GC and Swan SH (1991) Caffeine consumption during pregnancy and spontaneous abortion. Epidemiology 2,168174. Fenster L, Quale C, Hiatt RA, Wilson M, Windham GC and Benowitz NL (1998) Rate of caffeine metabolism and risk of spontaneous abortion. Am J Epidemiol 147,503510. Han XM, Ouyang DS, Chen XP, Shu Y, Jiang CH, Tan ZR and Zhou HH (2002) Inducibility of CYP1A2 by omeprazole in vivo related to the genetic polymorphism of CYP1A2. Br J Clin Pharmacol 54,540543. Infante-Rivard C, Fernandez A, Gauthier R, David M and Rivard GE (1993) Fetal loss associated with caffeine intake before and during pregnancy. J Am Med Assoc 270,29402943. Karhukorpi J, Laitinen T, Kivela H, Tiilikainen A and Hurme M (2003) IL-1 receptor antagonist gene polymorphism in recurrent spontaneous abortion. J Reprod Immunol 58,6167.

359

F.Sata et al.
Klebanoff MA, Levine RJ, Dersimonian R, Clemens JD and Wilkins DG (1998) Serum caffeine and paraxanthine as markers for reported caffeine intake in pregnancy. Ann Epidemiol 8,107111. Klebanoff MA, Levine RJ, DerSimonian R, Clemens JD and Wilkins DG (1999) Maternal serum paraxanthine, a caffeine metabolite, and the risk of spontaneous abortion. N Engl J Med 341,16391644. Landi MT, Sinha R, Lang NP and Kadlubar FF (1999) Human cytochrome P4501A2. IARC Sci Publ 148,173195. Lissak A, Sharon A, Fruchter O, Kassel A, Sanderovitz J and Abramovici H (1999) Polymorphism for mutation of cytosine to thymine at location 677 in the methylenetetrahydrofolate reductase gene is associated with recurrent early fetal loss. Am J Obstet Gynecol 181,126 130. Mills JL, Holmes LB, Aarons JH, Simpson JL, Brown ZA, JovanovicPeterson LG, Conley MR, Graubard BI, Knopp RH and Metzger BE (1993) Moderate caffeine use and the risk of spontaneous abortion and intrauterine growth retardation. J Am Med Assoc 269,593597. Nagata C, Kabuto M and Shimizu H (1998) Association of coffee, green tea, and caffeine intakes with serum concentrations of estradiol and sex hormone-binding globulin in premenopausal Japanese women. Nutr Cancer 30,21 24. Nordmark A, Lundgren S, Ask B, Granath F and Rane A (2002) The effect of the CYP1A2*1F mutation on CYP1A2 inducibility in pregnant women. Br J Clin Pharmacol 54,504 510. Parazzini F, Bocciolone L, Fedele L, Negri E, La Vecchia C and Acaia B (1991) Risk factors for spontaneous abortion. Int J Epidemiol 20,157 161. Pfeiffer KA, Fimmers R, Engels G, van der Ven H and van der Ven K (2001) The HLA-G genotype is potentially associated with idiopathic recurrent spontaneous abortion. Mol Hum Reprod 7,373378. Pihusch R, Buchholz T, Lohse P, Rubsamen H, Rogenhofer N, Hasbargen V, Hiller E and Thaler CJ (2001) Thrombophilic gene mutations and recurrent spontaneous abortion: prothrombin mutation increases the risk in the rst trimester. Am J Reprod Immunol 46,124131. Ranade K, Chang MS, Ting CT, Pei D, Hsiao CF, Olivier M, Pesich R, Hebert J, Chen YD, Dzau VJ et al. (2001) High-throughput genotyping with single nucleotide polymorphisms. Genome Res 11,12621268. Rey E, Kahn SR, David M and Shrier I (2003) Thrombophilic disorders and fetal loss: a meta-analysis. Lancet 361,901908. Sachse C, Brockmoller J, Bauer S and Roots I (1999) Functional signicance of a C ! A polymorphism in intron 1 of the cytochrome P450 CYP1A2 gene tested with caffeine. Br J Clin Pharmacol 47,445449. Saijo Y, Sata F, Yamada H, Kondo T, Kato EH and Kishi R (2004a) Single nucleotide polymorphisms in the promoter region of the interleukin-6 gene and the risk of recurrent pregnancy loss in Japanese women. Fertil Steril 81,374 378. Saijo Y, Sata F, Yamada H, Suzuki K, Sasaki S, Kondo T, Gong YY, Kato EH, Shimada S, Morikawa M et al. (2004b) Ah receptor, CYP1A1, CYP1A2 and CYP1B1 gene polymorphisms are not involved in the risk of recurrent pregnancy loss. Mol Hum Reprod 10,729733. Sata F, Yamada H, Kondo T, Gong Y, Tozaki S, Kobashi G, Kato EH, Fujimoto S and Kishi R (2003a) Glutathione S-transferase M1 and T1 polymorphisms and the risk of recurrent pregnancy loss. Mol Hum Reprod 9,165169. Sata F, Yamada H, Yamada A, Kato EH, Kataoka S, Saijo Y, Kondo T, Tamaki J, Minakami H and Kishi R (2003b) A polymorphism in the CYP17 gene relates to the risk of recurrent pregnancy loss. Mol Hum Reprod 9,725 758. Signorello LB and McLaughlin JK (2004) Maternal caffeine consumption and spontaneous abortion: a review of the epidemiologic evidence. Epidemiology 15,229239. Signorello LB, Nordmark A, Granath F, Blotm WJ, McLaughlin JK, Anneren G, Lundgren S, Ekbom A, Rane A and Cnattingius S (2001) Caffeine metabolism and the risk of spontaneous abortion of normal karyotype fetuses. Obstet Gynecol 98,10591066. Stirrat GM (1990) Recurrent abortion. II. Clinical associations, causes, and management. Lancet 336,728733. Tempfer C, Unfried G, Zeillinger R, Heer L, Nagele F and Huber JC (2001) Endothelial nitric oxide synthase gene polymorphism in women with idiopathic recurrent miscarriage. Hum Reprod 16,644 647. Unfried G, Tempfer C, Schneeberger C, Widmar B, Nagele F and Huber JC (2001) Interleukin 1 receptor antagonist polymorphism in women with idiopathic recurrent miscarriage. Fertil Steril 75,683687. Wang ZC, Yunis EJ, De los Santos MJ, Xiao L, Anderson DJ and Hill JA (2002) T helper 1-type immunity to trophoblast antigens in women with a history of recurrent pregnancy loss is associated with polymorphism of the IL1b promoter region. Genes Immun 3,3842. Wramsby ML, Sten-Linder M and Bremme K (2000) Primary habitual abortions are associated with high frequency of factor V Leiden mutation. Fertil Steril 74,987991. Yamada H, Kato EH, Kobashi G, Ebina Y, Shimada S, Morikawa M, Yamada T, Sakuragi S and Fujimoto S (2001) Recurrent pregnancy loss: etiology of thrombophilia. Semin Thromb Hemost 27,121129. Yamada H, Sata F, Saijo Y, Kishi R and Minakami H (2005) Genetic factors of fetal growth restriction and miscarriage. Semin Thromb Hemost (in press). Zusterzeel PL, Nelen WL, Roelofs HM, Peters WH, Blom HJ and Steegers EA (2000) Polymorphisms in biotransformation enzymes and the risk for recurrent early pregnancy loss. Mol Hum Reprod 6,474478. Submitted on February 28, 2005; accepted on April 4, 2005

360