2.4 (95% CI 0.9-6.1) for women smoking s=10 cigarettes/ day, but the trend in risk with number of cigarettes smoked per day and duration of the habit was not ...
Human Reproduction vol 11 no 10 pp 2306-2309, 1996
Coffee and alcohol intake, smoking and risk of multiple pregnancy
Fabio Parazzini1*2-4, Liliane Chatenoud1, Guido Benzi 2 , Elisabetta Di Cintio1, Daniela Dal Pino 2 , Luca Tozzi1 and Luigi Fedele3 'istituto di Ricerche Farmacologiche 'Mario Negri', via Eritrea, 62, 20157 Milano, ^ e n t r o Medicina della Riproduzione, Universita di Milano, via Commenda, 12, 20122 Milano and 3Clinica Ostetrico Ginecologica, Universita di Verona, 37100 Verona, Italy 4
To whom correspondence should be addressed
Introduction The frequency of spontaneous dizygotic multiple pregnancies has been declining in several countries during the last decades, while the number of monozygotic pregnancies remains constant (Botting et al., 1987; Parazzini et al, 1991). The determinants of these trends are however largely unknown (Parazzini et al, 1994b). Recently, it has been suggested from a case-control 2306
Materials and methods The general design of this study has been previously described (Parazzini et al, 1993). Briefly, between January 1988 and 1990 we conducted a case-control study on risk factors for multiple pregnancies Trained interviewers identified and questioned cases and controls using a standard questionnaire. Multiple births in patients who received treatments for infertility (assisted reproduction techniques or treatment for ovulation induction) were not included. An effort was made to define the zygosity. Between January 1988 and February 1989 the cases included in the study were women who delivered multiple births regardless of sex at the 'Clinica Luigi Mangiagalh' (the largest maternity hospital in Milan) Between March 1989 and September 1990 zygosity was defined on the basis of chromosome polymorphism in leukocytes (Paris Conference, 1972). A total of 133 cases (median age 31) were interviewed: 33 women delivered monozygouc and 100 dizygotic multiple births. Of these, 132deuvered twin births and one three babies. The control group comprised women who gave birth at term (>37 weeks gestation) to healthy infants on randomly selected days at the same clinic. Women were specifically excluded if they reported a history of multiple pregnancy or they had received treatment for infertility for the index pregnancy. A total of 395 control women (median age 30 years) were interviewed. No eligible case or control refused to be interviewed, but seven mothers refused to have their children's zygosity investigated using blood samples and were not included in this analysis. Information was obtained on general socio-demographic habits, personal characteristics and habits, gynaecological and obstetric history. Cases and controls were also asked about their consumption of coffee at conception in terms of cups per day and duration of the habits in years. Questions © European Society for Human Reproduction and Embryology
Downloaded from humrep.oxfordjournals.org by guest on July 21, 2011
We analysed the relationship between coffee and alcohol intake, smoking and risk of multiple pregnancies using data from a case-control study on risk factors for multiple births conducted in Italy. Cases were 133 women who delivered multiple births not related to treatment for infertility (33 monozygotic and 100 dizygotic twins). Controls were 395 women admitted for normal delivery at the same clinic where cases had been identified. The odds ratios (OR) of multiple pregnancy were 1.5 [95% confidence interval (CI) 0.8-2.8] and 2.0 (95% CI 1.0-3.7) for women drinking respectively one to two or three or more cups of coffee per day in comparison with non-coffee drinkers. Considering separately dizygotic and monozygotic pregnancies, the estimated OR were respectively for women drinking three or more cups of coffee, 1.7 and 3.1 for dizygotic and monozygotic pregnancies. The risk of multiple pregnancy tended to be higher in women drinking 2*15 alcohol drinks per week: in comparison with tea-totallers the estimated OR for drink >15 glasses per week were 2 3 and 2.6 respectively for dizygotic and monozygotic pregnancies. Heavy smokers (5*10 cigarettes per day) were at increased risk of multiple pregnancy: in comparison with never smokers, the estimated OR for multiple pregnancy was 1.6 (95% CI 0.9-2.7). Considering separately the two groups of multiple pregnancy, the OR of dizygotic and monozygotic pregnancy were 1.4 (95% CI 0.8-2.5) and 2.4 (95% CI 0.9-6.1) for women smoking s=10 cigarettes/ day, but the trend in risk with number of cigarettes smoked per day and duration of the habit was not significant Key words: epidemiology/multiple pregnancies/risk factors
study conducted in Greece that coffee intake may be associated with an increased risk of dizygotic twinning (Kapidaki et al., 1995). The association persisted after allowance for smoking and alcohol consumption. With regard to smoking habits, an increased risk of dizygotic multiple pregnancies has been observed in smokers in a large Danish case-control study (Olsen et al., 1988) but no association emerged between smoking and risk of multiple birth in the Greek study (Kapidaki et al., 1995). In biological terms it has been suggested that the association between coffee intake and smoking and risk of dizygotic multiple pregnancies may be linked to the effect of coffee and smoking in lowering oestrogen concentrations, and consequently in increasing pituitary gonadotrophins and ovarian stimulation (Olsen et al., 1988), which, in turn, may increase the frequency of dizygotic twins through increased ovulation. To provide further data, we have analysed the relationship between coffee and alcohol intake, smoking and risk of multiple pregnancies using data from a case-control study on risk factors for multiple births conducted m Italy.
Coffee and alcohol intake, smoking and multiple pregnancies on alcohol included the number of days per week each type of alcoholic beverage (wine, beer and spirits) was consumed before pregnancy and the average number of dnnks per day and the duration of the habit in years. From these data, an estimate of the total daily average alcohol intake was derived assuming a comparable pure alcohol content in each type of dnnk (125 ml wine = 333 ml beer = 30 ml spirits = -12 g pure alcohol). Wine alone accounted for -90% of the alcohol consumed by women in this population. Cases and controls were further asked about smoking habits (current smoking, former smoking, never smoking; number of cigarettes smoked per day and duration of the habit in current and former smokers). Exsmokers were defined as women reporting to have quit smoking 5* 1 year before conception. Specific attention was paid to obtaining from cases and controls information on coffee drinking, alcohol consumption and smoking at conception.
Results Table I shows the distribution of cases and controls according to age, education, parity, family history of multiple pregnancies and body mass index Dizygotic cases tended to be less educated and nulliparate. The nsk of dizygotic and monozygotic pregnancy was increased in women reporting a history of multiple pregnancies in first degree relatives. No association emerged between multiple birth and body mass index. Coffee and alcohol intake and smoking habits are considered in Table n. The risk of multiple pregnancy was 1.5 (95% CI 0.8-2.8) and 2.0 (95% CI 1.0-3 7) for women drinking respectively one, two or three or more cups of coffee per day in comparison with non drinkers. Considenng separately dizygotic and monozygotic pregnancies, the estimated OR were respectively for women drinking three or more cups of coffee, 1.7 and 3.1 for dizygotic and monozygotic pregnancies. No relationship emerged between multiple pregnancy risk and duration of coffee drinking. The nsk of multiple pregnancy tended to be higher in women drinking 5=15 alcohol drinks per week in companson with non-drinkers the estimated OR for women drinking 5=15 drinks per week were 2.3 and 2 6 respectively for dizygotic and monozygotic pregnancies. Heavy smokers (5= 10 cigarettes per day) tended to be at increased risk of multiple pregnancy: in comparison with never smokers, the estimated OR of multiple pregnancy was 1 6 (95% CI 0.9-2 7). Considering
Discussion One drawback of this study is the small sample size In particular, absolute numbers are too limited for formal statistical evaluation or inference on the difference between the two genetic entities. We have presented the results for dizygotic and monozygotic twins separately because there are very few data on the issue. There are some suggestions that alcohol drinking and smoking may increase the nsk of abortions and preterm births (Mills et al, 1993; Parazzini et al, 1994a) of low birth weight. Thus the choice of including in the control group only women who delivered at term may tend to overestimate the ORs. However, the role of potential selective mechanisms are, however, difficult to quantify. The assessment of coffee and alcohol consumption and smoking before pregnancy was based on self-reporting, so some underestimates could have occurred. However, in Italy, coffee and alcohol consumption is socially accepted and recommendations to avoid coffee and alcohol in pregnancy have not received widespread attention and are not routinely advocated by gynaecologists This kind of bias may have a major impact on information regarding smoking. However, there is no clear reason why misclassification of smoking habits may differ between cases and controls. We did not collect information on the size of cups of coffee, but this should not be a problem in Italy, since the majonty of women drink 'espresso' or 'mocha' (i.e. small cup unfiltered coffee). Other sources of bias, including selection or confounding factors, are also unlikely to have produced marked effects, especially considering that cases and controls were interviewed in the same institution and that participation was practically complete. The selected logistic regression model may have caused some overadjustment, for example by including level of education: the higher rate of twinning m women with a low educational level might be due to their higher level of smoking or alcohol consumption. However, the exclusion from the model of terms for education did not markedly change the estimated OR (date not shown). Available data on the relationship between coffee, alcohol and smoking and multiple pregnancies are very scanty. An association between coffee intake and twin pregnancy risk was reported from a study conducted in Greece (Kapidaki et al., 1995) A Danish study suggested a relationship between smoking and multiple pregnancies (Olsen et al, 1988). However, the study failed to take into account several potential covariates No association between alcohol intake and twin pregnancies was observed in the previously quoted Greek case-control study (Kapidaki et al, 1995). The potential associations between coffee, alcohol consumption, smoking and nsk of multiple pregnancies have, however, some speculative interest. High doses of alcohol intake may alter gonadotrophin and prolactin concentrations and cause 2307
Downloaded from humrep.oxfordjournals.org by guest on July 21, 2011
Data analysis We computed the odds ratios (OR), as estimators of the relative risks, of multiple pregnancy with their 95% approximate confidence intervals (CI), from data stratified for age using the Mantel—Haenszel procedure (Mantel and Haenszel, 1959) Further, to account simultaneously for the effects of several potential confounding factors, we used unconditional multiple logistic regression, with maximum likelihood fitting, to obtain OR and their corresponding 95% CI (Breslow and Day, 1980). Included in the regression equations were terms for age, education, previous live births plus terms for family history of multiple birth [a factor that was significantly associated in this data set with the nsk of twin pregnancies (Parazzini et al, 1993)] and indicators of alcohol and coffee intake and smoking habits The inclusion in the model of terms for calendar period at the interview (1988-1989 and 1989-90) did not change the estimated OR (data not shown)
separately the two groups of multiple pregnancy, the OR of dizygotic and monozygotic pregnancy were 1.4 (95% CI 0.82.5) and 2.4 (95% CI 0.9-6.1) for women smoking s=10 cigarettes/day, but the trend in risk with number of cigarettes smoked per day was not significant
F.Parazzini et at
Table L Distribuuon, and corresponding relative risks, of twin pregnancies and controls according to selected charactensucs Milan, Italy, 1988-1990 Cases
Controls
Dizygouc
Monozygouc
Total
45 144 122 84
_ _ _ _
— _ _
_ _ _ _
22 137 236
lb 0 4 (0 2-0.9) 0.4 (0 2-0.9) 1 41 (P - 0.23)
lb 16(0 2-13.6) 2 1 (0.3-16 6) 0.75 (P = 0.38)
lb 0 5 (0 2-1 0) 0 6(03-1.2) 0.37 (P = 0 54)
209 136 50
lb 0.9 (0 5-1 4) 0 5 (0 2-1.1) 2 15 (P = 0.14)
lb 0 8 (0.3-1 9) 0 8 (0.2-2 5) 0 22 (P = 0 63)
lb 0 9 (0 6-1 4) 0 6 (0 3-1.2) 2 03 (P = 0 15)
265 113
lb 2.4 (15-3 8)
lb 2.7(1.3-5 6)
lb 2 4 (16-3 7)
74 169 152
lb 1 3 (0 7-2.5) 1 3 (0 6-2.4) 0 32 (P = 0 57)
lb 1 5 (0.5-4 6) 1 5 (0 5-4 7) 0.32 (P = 0.57)
lb 1 3 (0 7-2.4) 1 3 (0 7-2 4) 0 54 (P = 0 46)
'Adjusted for age OR = odds ratios. CI = confidence interval ''Reference category. T h e sum does not add up to the total because of missing values
Table II. Distribution, and corresponding relative risks. of twin pregnancies and controls according to coffee and alcohol intake and smoking Milan, Italy, 1988-1990 Cases
Coffee consumption (cups/day) 0 1-2 *3 X2i trend Duration of coffee consumption (years) *10 >10 Alcohol consumption (glasses/week) 0 1-14 3=15 X2i trend Duration of alcohol consumption (years) *10 >10 Smoking habits Never smokers Ex smokers Current smokers < 1 0 cig^day 5»10 cig7day X2i trend Duration of smoking (years)
