Cigarette, Alcohol, and Coffee

Consumption and Congenital Defects

Alison D. McDonald, MD, Ben G. Amnstrong, PhD, and Maigaret Sloan, BA
Inltdwton Congenital defects were ascertained
(blind) in current pregnancies from pedi-
A report that smoking in pregnancy atric records.8 For previous pregnancies,
was associated with congenital heart de- the mother was asked whether any abnor-
fects, was not confirmed,2- even in two mality had been noted in the child. If a
large series of births.5 Heavy alcohol con- defect was suspected, further inquiries
sumption has been reported to cause re- were made. The defects were classified
tarded physical and mental development into 90 groups. Minor abnormalities such
and certain facial and neurological abnor- as hydrocele, glandular hypospadias, un-
malities (fetal alcohol syndrome). Moder- descended testes, skin tags, nevi, and her-
ate alcohol consumption has also been re- nias were omitted, as were known inher-
ported to affect fetal growth and ited defects, those that were poorly
morphogenesis.6 Some concern has been described (less than 1%), and those that
expressed that heavy coffee consumption became apparent only at a later stage of
may be teratogenic,7 although a case-con- development, such as mental retardation,
trol study in Finland was negative.8 We epilepsy, cerebral palsy, and deafness. In-
investigated the relationships between terauricular and interventricular septal de-
cigarette, alcohol, and coffee consump- fects were retained if accompanied by an-
tion and congenital defects, using data other heart abnormality. Cases with
from a survey of occupational factors in multiple defects were classified under the
pregnancy conducted in Montreal from one that appeared to be most serious.
1982 to 1984 (see the first paper in this Eight types of congenital defect (shown in
series). In this survey no association was Tables 1 and 2) were entered into the pre-
found between congenital defects and sent analyses.
chemical exposures estimated from job ti- Possible confounding by maternal
tle,9 but in a case-referent study an asso- age, educational level, and ethnic (color
ciation with aromatic solvents was and language) group was controlled by in-
found.10 The excess defects were mainly cluding these factors together with ciga-
renal and urinary. The association of cig- rette, alcohol, and coffee consumption in
arette, alcohol, and coffee consumption logistic regression analyses, which were
with congenital defects has now been an- performed separately for current and pre-
alyzed for all current and previous preg- vious pregnancies. Regression coeffi-
nancies of the 56 000 women inter- cients for cigarette, alcohol, and coffee
viewed-104 000 pregnancies in all. consumption from the two series were
combined as a weighted mean to obtain a
single odds ratio, as they did not signifi-
cantly differ for any factor at a 5% prob-
Medhds ability level.
The study design and methods are
briefly descnbed in the first of these re- Results
ports, and have been described in detail There was no consistent evidence of
elsewhere.-1 All current and previous an association between cigarette con-
pregnancies of more than 20 weeks' du- sumption and any defect group, after
ration, whether or not the woman was
employed during the pregnancy, were in-
cluded in the present analysis. Therapeu-
tic abortions for fetal defect were included The authors are with the School of Occupa-
tional Health, McGill University, Montreal,
irrespective of gestation length. Informa- Quebec.
tion on cigarette and alcohol consump- Requests for reprints should be sent to
tion, and on potential confounding factors Ben G. Armstrong, PhD, School of Occupa-
was available for 89 317 of these pregnan- tional Health, McGill University, 1130 Pine Av-
enue West, Montreal, Quebec, Canada H3A
cies. As questions on coffee consumption 1A3.
in previous pregnancies were introduced This paper was submitted to the journal
after the survey began, only 80 319 re- July 19, 1990, and accepted with revisions May
mained for study of this factor. 6, 1991.

American Journal of Public Health 91

Public Health Brefs

accounting for possible confounding (Ta- took seven or more drinks per week was abnormality, hemivertebrae, limb reduc-
ble 1). Odds ratios (relative to nonsmok- significant at P = .05. Among the 13 de- tion deformity (2), and fibular abnormal-
ers) for mothers who smoked one to nine fects in this group were 6 cases of con- ity. Facial deformities were not ob-
cigarettes per day were 1.84 for neural genital dislocation of the hip, whereas 2.3 served.
tube, 1.53 for renal-urinary, and 1.14 for were expected based on the proportion of Of the odds ratios for defects in
all defects. These values were on the bor- musculoskeletal defects that were of this babies born to coffee drinkers (Table
derline of statistical significance (at about type overall. The specific risk for con- 3), only that for heart defects among
P = .05); however, odds ratios were genital dislocation of the hip (6/ the children of women who drank three
lower in heavier smokers. 2.3 = 2.61) was thus somewhat higher or more cups a day (1.52) was signifi-
There was some evidence of an in- than for musculoskeletal defects overall, cantly elevated (P = .02). No specific
creased risk of musculoskeletal defects in but because of small numbers it is not type of heart defect was overrepresented
women who took alcohol, with risk in- statistically significant (P > .05). The in these 58 cases compared with defects
creasing with the amount drunk (Table other 7 musculoskeletal defects were in babies born to women who did not
2). The odds ratio of 1.82 in women who craniosynostosis, scaphocephaly, cranial drink coffee.

92 American Joumal of Public Health January 1992, Vol. 82, No. 1

 Pubfic Health Biiefs

E. X, X,

..............
M M M ............................................................. .....
............
........
:-,
...................
...........

Y ..........
I.....................

%......W
..........................................
...............................................................................
..........

..................
...............
................. ..................

.......................................%%--.---- -.-,- -

...............................................................

............

...

..............

.................

W.. I'M
..........
X
..........
...........

.............................

............................

............
..................

......................
C............................
.............................................

..........................
X

day. The power of our study to detect in- tality rate related to maternal smoking and
Discussion creased risk at these levels was thus very socioeconomic status. Am J Obstet Gyne-
col. 1971;111:53-59.
Given that over 70 odds ratios were limited, as is evident from the wide con- 3. Kullander S, Kallen, B. A prospective
tested, the three that were significantly fidence limits for the odds ratios shown in study of smoking and pregnancy. Acta
(P < .05) raised could have occurred by Table 2. No case ofsuspected fetal alcohol Obstet Gynecol Scand. 1971;50:83-94.
4. Andrews J, McGarry JM. A community
chance. The one association with smok- syndrome was noted, although we were study of smoking in pregnancy. J Obstet
ing, namely an increase of neural tube not specifically looking for the syndrome. Gynecol BC 1972;79:1057-1073.
defects in babies born to women who The association of alcohol with muscu- 5. Shiono PH, Klebanoff MA, Berendes HW.
smoked 1 to 9 cigarettes per day, would Congenital malformations and maternal
loskeletal defects has not been reported smoking during pregnancy. Teratology.
have been more convincing if an excess elsewhere, and was of low statistical sig- 1986;34:65-71.
had been seen in the heavier-smoking nificance. 6. Hanson J, Streissguth A, Smith D. The ef-
groups. Even quite heavy cigarette Coffee consumption was associated fects of moderate alcohol consumption
smoking was common in this survey-in during pregnancy on fetal growth and mor-
only with heart defects, and the evidence phogenesis. JPediatr. 1978;92:457-460.
30% of the pregnancies the mother was not strong. In Finland, where coffee 7. Soyka L. Caffeine ingestion during preg-
smoked during the first trimester of preg- consumption is high (29% of those sur- nancy: in utero exposure and possible ef-
nancy, and in 15% the mother smoked 20 fects. Sernin Pennatol. 1981;5:305-309.
veyed reported taking three or more cups 8. Kurppa K, Holmberg P, Kuosmer E,
or more cigarettes per day. Thus our of coffee a day, compared with 12% in Saxen L. Coffee consumption during preg-
study had the power to detect important Montreal), no association was found be- nancy and selected congenital malforma-
risks due to smoking, if such risks were tween coffee consumption and clefts of tions: a nationwide case-control study.Am
present. The risks we can exclude are J Public Health. 1983;73:1397-1399.
the lip and palate or defects of the central 9. McDonald AD, McDonald JC, Armstrong
specified by the upper confidence limits nervous, cardiovascular, or musculoskel- B, et al. Congenital defects and work in
of the odds ratios, which are quite low for etal systems. LI pregnancy. BrJIndMed. 1988;45:581-588.
smoking. 10. McDonald JC, Lavoie J, Cote R, McDon-
Alcohol consumption, on the other ald AD. Chemical exposures at work in
hand, although perhaps underreported,
References early pregnancy and congenital defect: a
1. Fredrick J, Alberman E, Goldstein H. Pos- case-referent study. Br J Ind Med.
was generally low. Only 3% of the women sible teratogenic effect of cigarette smok- 1987;44:527-533.
surveyed reported taking one or more ing. Nature. 1971;231:529-530. 11. McDonald AD, McDonald JC, Armstrong
drinks per day (seven perweek), and 0.3% 2. Comstock GW, Shah FK, Meyer MB, Ab- B, et al. Occupation and pregnancy out-
reported taking three or more drinks per bey H. Low birth weight and neonatal mor- come. Br J Ind Med 1987;44:521-526.

January 1992, Vol. 82, No. 1 American Journal of Public Health 93