Caffeine and Spontaneous Abortion:
Cause or Consequence?
The omnipresence of caffeine consumption
has given rise to concern among the medical and scientific communities
regarding adverse reproductive outcomes. Literature is replete with human and
animal research on intake of caffeine and its effect on the occurrence of
spontaneous abortion, delayed conception/fertility issues, low birth weight,
birth defects, sudden infant death syndrome (SIDS) and other unfavorable
pregnancy outcomes. According to the National Coffee Association, 54% of the
adult population in the
Caffeine
pharmacology and physiology.
Caffeine (C8H10N4O2,
1,3,7-trimethylxanthine) is an alkaloid compound naturally occurring in
tea leaves, coffee beans, kola nuts and cocoa beans[2]. It is a white, crystalline powder with a
distinctly bitter taste that is light sensitive and soluble in boiling water2. Related
compounds in the methylxanthine family are theophylline and theobromine, which
are components of tea and chocolate.
The chemical composition of caffeine closely resembles that of uric acid
and xanthine. It has been labeled as a possible mutagenic compound because of
its structural similarity to the purines adenine and guanine, two of the major
components of DNA. Evidence from some
animal studies support the claim of mutagenicity which prompted the Food and
Drug Administration to issue a warning advising pregnant women to avoid or
restrict caffeine intake[3].
Findings from these animal studies do not necessarily equate to mutagenicity in
humans due to differences in human and animal caffeine metabolism.
Caffeine, once ingested, is rapidly
absorbed from the digestive tract and distributed among all bodily tissues in
proportion to their water content[4]. It is a lipolytic compound and readily
crosses the blood-brain barrier and the placenta. The plasma half-life is about 2.5-4.5 hours
in a healthy adult[5]. The half-life of caffeine in a pregnant woman
is similar to the non-pregnant state in the first trimester, however it doubles
during the second trimester and triples during the third20. The plasma
half-life of caffeine for the neonate ranges from 32-149 hours4 because infants lack the enzyme to metabolize
caffeine until several days after birth[6]. The main breakdown products of caffeine are
paraxanthine and monomethylxanthine, which are mainly excreted through urine.
Physiologically, caffeine affects many of
the organ systems. It acts as a chemical
stimulus to the central nervous system, relaxes smooth muscle and stimulates
gastric secretions in the digestive system.
Caffeine affects the renal system by acting as a diuretic and
stimulating the release of renin. Cardiovascular effects include stimulation of
the cardiac muscle, bradycardia, tachycardia, dilation of coronary, renal and
peripheral vasculature, and constriction of central vasculature. Other
physiological effects include stimulation of lipolysis in adipose tissue and
inhibition of platelet aggregation7.
Caffeine may also alter calcium balance due to its ability to mobilize calcium
from storage sites within skeletal and cardiac muscle7.
Caffeine influences the reproductive
system by altering levels of circulating catecholamines, specifically
epinephrine6. Elevated
levels of catecholamines may restrict uteroplacental circulation via vasoconstriction,
possibly resulting in fetal hypoxia, which can potentially result in fetal
death, decrease in fetal weight, fetal malformations and shortened gestational
age4.
Developmental modulations may also result from altered fetal
catecholamine balance6. Data from animal studies resulted in the
hypothesis that some birth defects normally seen with caffeine administration
were attributable to increased fetal catecholamines released in response to the
presence of caffeine6.
Additionally, caffeine is known to inhibit the enzyme phosphodiesterase, which
is responsible for the hydrolysis of cyclic adenosine monophosphate (cAMP),
resulting in increased cellular levels of cAMP.
This could change the hormonal profile within the fetal and maternal
blood supply and may interfere with fetal cell growth6,4.
Caffeine Content in Foods and Beverages
The primary food sources of caffeine are
coffee, tea, chocolate and cola/soft drinks.
Analgesics and caffeine pills also contribute to caffeine consumption. Table 1 delineates the amount of caffeine in
various food substances.
Table 1: Caffeine Content
of Various Beverages and Foods9
|
Food Type |
Serving Size |
Caffeine (mg) |
|||
|
Coffee: |
|
|
|||
|
Automatic Drip/Brewed |
5.0 oz |
115 (80-175) |
|||
|
Percolated |
5.0 oz |
80 (40-170) |
|||
|
Instant |
5.0 oz |
60 (46-71) |
|||
|
Espresso |
1.5-2 oz. |
100 |
|||
|
Cappuccino |
6.0 oz |
60-120 |
|||
|
Decaffeinated |
6.0 oz |
<1 |
|||
|
Tea: |
6.0 oz |
30-80 |
|||
|
Iced |
12.0 oz |
70 |
|||
|
Brewed, imported |
7.0 oz |
60 |
|||
|
Brewed, U.S |
7.0 oz |
40 |
|||
|
Chocolate: |
|
|
|||
|
Chocolate milk
|
8.0 oz |
2-8 |
|||
|
Hot chocolate
|
8.0 oz |
10 |
|||
|
Milk chocolate candy |
1.0 oz |
6 |
|||
|
Dark chocolate |
1.0 oz |
20 |
|||
|
Chocolate cake |
1 slice |
20-30 |
|||
|
Soft
Drinks: |
|
|
|||
|
Jolt |
12 oz. |
100.0 |
|||
|
Mountain Dew |
12 oz. |
54.0 |
|||
|
Mello Yello |
12 oz. |
52.8 |
|||
|
Coca-Cola |
12 oz. |
45.6 |
|||
|
Dr. Pepper |
12 oz. |
39.6 |
|||
|
Pepsi |
12 oz. |
38.4 |
|||
|
RC Cola |
12 oz. |
36.0 |
|||
|
Diet Rite |
12 oz. |
36.0 |
|||
The
information base in Table 1 is merely an estimate. Numerous other sources portray similar data;
however, these numbers are very difficult to quantify accurately.
The two main plant species from which coffee is manufactured are arabica and
robusta. Robusta beans have roughly
twice the amount of caffeine as arabica beans; therefore, the blend of coffee
will impact the amount of caffeine contained in the final product. The roasting process may also affect caffeine
content. Many coffee companies consider
blends and roasting methods proprietary information. This hinders the attempt to accurately
ascertain the amount of caffeine in a cup of coffee. Other variations in the caffeine content of
coffee are attributable to plant variety, country of origin and cultural
harvesting practices2. Caffeine variation in tea is affected by
parameters that lead to differences in plant composition. Fertilizers containing nitrogen can increase
caffeine content by as much as 40%23. Seasonal variations, plant variety and leaf
position on the cutting are other examples of plant-related caffeine diversity23. Black tea,
which is fermented (oxidized) generally has more caffeine than green tea. In brewing tea, caffeine is extracted at
about the same rate as the tea solids23, so it can
be deduced that increased brewing time will increase caffeine content. Extrinsic caffeine, sometimes purchased from
coffee roasters as a by-product of the decaffeination process, is added to many
soft drinks during manufacturing. Some
medications, including decongestants, stimulants, weight loss aids and pain
relief medicines also contain caffeine9.
Review of Literature
Various
studies will be outlined in this section.
Please refer to Appendix 1 for
statistics related to each study.
The earliest published human study on
caffeine and its relationship to spontaneous abortion was done by Weathersbee,
et al in 1977. A retrospective survey was
conducted on a population of women who had received obstetric care at selected
hospitals in
Table 2: Outcome of
pregnancy in relation to caffeine consumption10
|
Daily caffeine consumption (estimate) |
Number of households (pregnancies) |
Outcome of pregnancy |
||||
|
|
|
Spontaneous abortion |
Stillbirth |
Premature birth |
Uncomplicated delivery |
|
|
>= 600mg by woman |
16 |
8 |
5 |
2 |
1 |
6.3% |
|
> 600mg by man
< 400mg by woman |
13 |
4 |
2 |
2 |
5 |
38.5% |
|
300-450mg by man, woman or both |
23 |
0 |
0 |
0 |
23 |
100.0% |
|
<300 mg by man, woman or both |
81 |
7 |
17 |
5 |
52 |
64.2% |
|
None |
356 |
33 |
38 |
6 |
279 |
78.4% |
Weathersbee,
et al concluded from the above data that daily caffeine intake of 600mg or more
may predispose a woman to complications in pregnancy. An additional hypothesis that resulted from
the data was that outcome of pregnancy may be male mediated, as evidenced by
the 13 households of group 2, where the man consumed more than 600mg of
caffeine daily. The authors concluded
that comparatively high intake of caffeine by men in households where the level
of intake by women was high may contribute to altered reproductive outcomes10. It is assumed
that the results of the study reflect caffeine intake during pregnancy rather
than before or after. The validity of
the study can be questioned due to the low response rate and the predominantly
Mormon population which may indicate selection bias. Those who had unfavorable outcomes of pregnancy
may not have responded and those who did respond may have overestimated their
intake of caffeine, possibly leading to recall bias. Also, the study did not control for the
potential confounding variables of educational levels, ethnicity, maternal age,
smoking and alcohol consumption.
Srisuphan and Bracken4 (1986) investigated the relationship between caffeine
and late-first and second trimester spontaneous abortion in a prospective
cohort study. Information was gathered
from a sample of 3,135 women to determine their consumption of coffee,
non-herbal tea, caffeinated soft drinks and caffeine containing drugs during
pregnancy. Smoking and drinking habits
during pregnancy were also assessed.
Caffeine use before pregnancy and changes in caffeine consumption during
pregnancy was not ascertained. All
spontaneous abortions in the study occurred between 8 and 26 weeks of gestation
and were defined as a nondeliberate interruption of an intrauterine pregnancy
of less than 28 weeks gestation in which
the fetus was dead when expelled4. Total caffeine use was broken into three
categories; none, light (1-150 mg per day) and moderate to heavy (ł151 mg per day). Results indicate that moderate to heavy use
of caffeine almost doubled the risk of spontaneous abortion (relative risk =
1.95, p=0.07). Levels of caffeine intake
greater than 150 mg daily indicated an increase in the adjusted relative risk
for spontaneous abortion, however, consumption of greater than 200 mg did not
increase the risk further4. The authors point out some possible
limitations. They did not include the
occurrence of maternal illness that may have predisposed the subjects to avoid
or limit caffeine intake nor did they evaluate the association of caffeine with
early first trimester spontaneous abortions.
Caffeine intake recorded at the time of interview was assumed to be
constant during the pregnancy. Factors
were controlled for at the level of significance p < 0.10, which included
age (maternal and gestational), prior gynecologic surgery, Jewish faith and
last pregnancy terminating in spontaneous abortion, however, smoking and
alcohol consumption were not controlled for because they did not fit this
criterion. Research indicates that there
is a positive correlation between smoking and high levels of caffeine
consumption.
Fenster et al introduced a new variable in
their 1991 study: nausea. Several
studies have found nausea to be more frequent in pregnancies carried to term
than in those that resulted in miscarriages, especially those that miscarried
early in the pregnancy12. The authors’ goal was to evaluate the
potential association between caffeine consumption and spontaneous abortion
during the first trimester of pregnancy (first 20 weeks). The study population was comprised of 607
cases and 1291 controls. Spontaneous abortions, more than half of which
happened before 13 weeks, were ascertained by pathology reports. Participants were asked about consumption of
coffee, tea and cola during the month before pregnancy, changes in consumption
habits during pregnancy and the degree of change. Controls were requested to
recount information about the first 20 weeks of their pregnancies in order to
make their exposure periods comparable with cases. Caffeine usage was grouped by 1-150mg/day
(light use), 151-300mg/day (moderate use) and >300mg/day (heavy use). It is postulated that since nausea is
associated with reduced risk of spontaneous abortion, the presence of nausea
may reduce caffeine consumption13,
therefore reports of nausea were stratified to determine whether it modified
the caffeine-spontaneous abortion association. The results indicate a slightly
increased risk of spontaneous abortion associated with heavy caffeine
consumption. The risk doubled for heavy caffeine users who reported nausea,
while there was no such association in those who did not report nausea. Heavy
caffeine users who reduced their caffeine intake to less than 300 mg daily
within 6 weeks of their last menstrual period had a risk of spontaneous
abortion equivalent to non-users13. Both cases and controls changed their
caffeine intake during the weeks they were pregnant, however cases had a
reduced opportunity to do so. Cases were substantially less likely to have
reported nausea and the presence of nausea was associated with a decrease in
caffeine consumption. This tendency
varied by gestational age, therefore women with earlier spontaneous abortions
were less likely to report nausea13. Stein and
Susser12 question whether the change
in caffeine consumption was the cause or the effect of miscarriage. They postulate that if nausea begins to fade
when the pregnancy fails, then normal caffeine intake may be resumed after the
failure of pregnancy and before the extrusion of the conceptus, therefore, a
false appearance of raised caffeine intake before spontaneous abortion will
take place12. Fenster and co-authors may have controlled
inadequately for nausea leading to some confounding in the results. Another limitation is the possibility of
selection bias due to measurement of spontaneous abortions using pathology
records. Women may abort before
beginning obstetrical care and therefore would not be included in the study.
Karyotyping of the aborted fetuses may have given the results more
validity. Maternal caffeine consumption
would most likely not have affected the fetal survival if genetic damage was
present14. Caffeine intake from
chocolate, cocoa or drugs was not ascertained and therefore subjects in the
light use group may have had exposure from these sources. Serving sizes and method of preparation,
which can also modify the levels of caffeine exposure, were not defined.
Kline et al (1991)14 examined whether caffeine ingestion
shortly before or during pregnancy led to an association with specific chromosomal
aberrations in spontaneous abortions or with loss of chromosomally normal
conceptions14. The authors postulate that a chromosomally
aberrant fetus would not have survived regardless of maternal caffeine
intake. Caffeine exposure in
chromosomally normal losses, however, may be relevant before conception, soon
after, or later in gestation. Unlike
some earlier studies, caffeine intake was the primary focus of this study. In a case-control, retrospective design,
subjects were questioned on caffeine consumption during pregnancy and during
the perifertilization period. Pregnancy
was defined as one month after the last menstrual period until the time of loss
for cases or time of interview for controls, and the perifertilization period
was characterized as two months prior to and one month after the last menstrual
period. Average frequency of consumption was calculated for women who offered
information on altered intake of caffeine during either of these periods. Spontaneous abortion was defined as involuntary
termination of intrauterine pregnancy before 28 weeks gestation. Caffeine exposure was measured based on
serving size of beverages, usual frequency of consumption, brand names of soda,
coffee brands and brewing methods and intake of chocolate and cocoa products as
well as decaffeinated coffee and tea. Caffeine consumption was estimated to be
higher in the perifertilization period than during pregnancy. Results were compiled using two comparisons. The first contrasted caffeine use in cases,
grouped by karyotype, with controls. The
second contrasted chromosomally normal and chromosomally aberrant losses
combined and by type. The findings
demonstrated that caffeine intake in the perifertilization period at levels of
225 mg or more per day was not related to chromosomally normal or trisomic
spontaneous abortions and that caffeine consumption does not influence
chromosomally normal loss during pregnancy14. Gestational
ages were later in controls at time of interview than in cases at time of
loss. Since it has been documented that
caffeine intake is generally decreased during pregnancy, this difference may
result in artificially lower estimates of caffeine intake during pregnancy in
controls as compared with cases. Case-control differences were noted due to the
increased magnitude of caffeine reduction during pregnancy for controls. The authors did not inquire specifically
about change in caffeine intake during pregnancy, therefore reported caffeine
intake levels may have been reflective of more recent consumption rather than
average over the entire period of pregnancy.
Also, population bias may have been introduced because all cases did not
receive pre-natal care, while all controls did.
Adjusting for nausea and vomiting did not alter results. There is no evidence that smoking and alcohol
consumption was controlled for. A
retrospective study by Armstrong et al. examined the relationship of smoking,
alcohol and coffee intake to the occurrence of spontaneous abortion. 56,000 women were interviewed, all of whom
had either a delivery or spontaneous abortion in one of eleven
consumed
0-2 cups of coffee daily while only 596 consumed 5 or more cups per day, which
may have resulted in a skewed outcome.
The authors concluded that the weak association of coffee intake and
spontaneous abortion might possibly be due to confounding factors. The researchers did not include questions on
caffeine intake from other sources besides coffee, nor did they define serving
size. They acknowledge that recall bias
may be a factor due to the fact that the abortions occurred closer in time to
the interview than did live births11.
Two studies published in the Journal of the American Medical Association in 1993 netted two very different outcomes. The first was published in February by Mills, et al and the second in December by Infante-Rivard et al.
Mills and colleagues sought to determine
the relationship between caffeine consumption during pregnancy and occurrence
of spontaneous abortion, among other factors15. They conducted a prospective cohort study
of 431 women who enrolled in a
multi-center study either before conception or within 21 days of conception. Caffeine use was measured at the time of
pregnancy diagnosis and repeated at 6, 8, 10, 12, 20, 28 and 36 weeks
thereafter. Participants were asked
about daily caffeine consumption, based on eight-ounce servings of regular and
decaffeinated coffee, tea (hot and iced), cocoa, cola and other soft drinks. Women were also asked to report on (caffeine
containing) drug use for the same time intervals. Results indicated that caffeine was not a
significant risk factor for pregnancy loss.
After adjusting for risk factors such as maternal age, smoking, parity,
prior spontaneous abortion, income, alcohol use and maternal education, the
odds ratio for spontaneous abortion for users of caffeine was 1.15 (95% CI
.89-1.49). In addition, there was no
significant increase in spontaneous abortion in women who consumed over 300
mg/day and caffeine did not affect the time at which miscarriage occurred15. Table 3
delineates the results of the study based on first trimester caffeine
consumption.
Table 3: Pregnancy Outcomes by Mean daily first
trimester consumption (partial)15
|
|
Caffeine Consumption, mg
(%) |
|
|||
|
|
|
|
|
|
|
|
|
None |
1-99 |
100-199 |
200-299 |
300+ |
|
|
(n=39) |
(n=187) |
(n=110) |
(n=63) |
(n=24) |
|
|
|
|
|
|
|
|
Deliveries |
32 (82.1) |
162 (86.6) |
97 (88.2) |
51 (81.0) |
19 (79.2) |
|
|
|
|
|
|
|
|
Losses |
|
|
|
|
|
|
Spontaneous Abortion |
6 (15.4) |
24 (12.8) |
13 (11.8) |
11 (17.5) |
5 (20.8) |
|
Late Loss (>20 wk) |
0 |
1 (0.5) |
0 |
0 |
0 |
Most of the losses in the study occurred in the first trimester. The strength of this study lies in the prospective design and the close tracking of caffeine consumption during the pregnancy to reduce the possibility of recall bias. All spontaneous abortions after 21 days were recorded due to early identification of pregnancy. A major limitation appears to be population bias. Those participating in this study had carefully planned their pregnancies and took actions to ensure the pregnancy turned out favorably. The sample size was also relatively small, especially the high-consumption (300+) group. In addition, as in most other studies except Kline et al14, the method of preparation of caffeinated beverages was not established. The authors discussed nausea as a factor as well. Women were most likely to experience nausea between weeks 8 and 10. Those women who experienced nausea were more likely to decrease their caffeine consumption than those who did not, however it does not appear that the authors controlled for nausea as a modifying factor in caffeine consumption.
Infante-Rivard, et al16 analyzed the relationship between fetal loss and caffeine consumption before and during pregnancy using a case-control design. Cases consisted of women hospitalized with a diagnosis of spontaneous abortion or fetal death. Controls were matched to be at the same period of pregnancy as cases but had not had fetal loss. They were matched by periods of pregnancy (Ł16 weeks, 17-20 weeks, 21-27 weeks, ł 28 weeks) on a 3:1 control to case ratio. Both cases and controls with previous spontaneous abortions were excluded from the study. The sample population was asked about their intake of beverages containing caffeine such as coffee, tea and cola one month before pregnancy and during pregnancy. Intake during pregnancy was averaged over the time of pregnancy up until the time of the enrollment in the study. Results indicated a pattern of increased risk for fetal loss with each category of caffeine intake (see table 4 for groupings). After controlling for confounding factors, there was a moderate association of pre-pregnancy caffeine intake and fetal loss. A strong correlation of caffeine consumption with spontaneous abortion was recognized during pregnancy, with each 100 mg increase in daily consumption resulting in an odds ratio increase by a factor of 1.22. Table 4 depicts the results of the study.
Table 4:
Adjusted Odds Ratios and 95% Confidence intervals for caffeine intake before
and during pregnancy16
|
|
Controls |
Cases |
Odds Ratio |
p |
95% CI |
|
Caffeine
intake before pregnancy, mg/day |
|
|
|
|
|
|
< 48 |
279 |
79 |
1.00 |
|
|
|
48-162 |
230 |
74 |
1.29 |
0.22 |
0.85-1.95 |
|
163-321 |
288 |
96 |
1.37 |
0.16 |
0.92-2.04 |
|
> 321 |
195 |
82 |
1.85 |
0.01 |
1.18-2.89 |
|
Caffeine intake during pregnancy, mg/day |
|
|
|
|
|
|
< 48 |
530 |
148 |
1.00 |
|
|
|
48-162 |
277 |
91 |
1.15 |
0.43 |
0.82-1.63 |
|
163-321 |
145 |
71 |
1.95 |
0.001 |
1.29-2.93 |
|
> 321 |
41 |
21 |
2.62 |
0.003 |
1.38-5.01 |
55% of the subjects decreased caffeine consumption during pregnancy, 42% did not change and 3% increased consumption. Reduction in consumption was slightly greater among controls than cases, however after the first 16 weeks, cases reduced their consumption to a greater degree than controls. The results indicate that caffeine intake during pregnancy is strongly associated with spontaneous abortion and intake before pregnancy is moderately associated with spontaneous abortion. Limitations of this study include the plausibility of recall bias and the lack of questioning on all caffeine containing foods, beverages and drugs, as well as serving sizes of caffeinated beverages consumed. The authors believe recall bias was minimal. This study did not measure nausea as a variable, but it may partly account for the increased early (<16 week) reduction in caffeine consumption in controls versus cases. Lower caffeine intake in the control group may also be due to higher level of education, inferring that they were more aware of the possible harmful effects of caffeine.
In 1996, Dlugosz and co-authors17 conducted a prospective cohort study of 2,967 pregnant women to examine the relation of caffeine intake to spontaneous abortion. The population consisted of women seeking pre-natal care before the sixteenth week of gestation. Questions pertaining to lifestyle, pregnancy history and caffeine consumption were asked. The question regarding caffeine was phrased: Since you became pregnant, have you been drinking one or more cups of coffee, tea or soda with caffeine per week?17. Subjects were also questioned on consumption frequency of caffeinated soda, coffee and non-herbal tea during the first month (defined as 3-6 weeks from the last menstrual period) of pregnancy. Information on pregnancy outcome was obtained from medical records. 4.5% of the population had spontaneous abortions, defined by the authors as a nondeliberate interruption of an intrauterine pregnancy of less than 28 weeks gestation in which the fetus was dead when expelled17. The mean daily caffeine consumption during the first month was 89.0 mg/day for the women who spontaneously aborted and 72.4 mg/day for those whose pregnancy resulted in a live-born infant. Based on their analysis, the authors concluded that consumption of more than 300 mg of caffeine per day was associated with almost double the risk of spontaneous abortion (odds ratio 1.75, CI 0.88-3.47), however, the linear dose-response trend was weak. In addition, caffeine consumption was analyzed by source; coffee, tea and soda. The risk of spontaneous abortion increased when three or more cups of coffee and tea were consumed, however the trend did not extend to soda (see Table 5). This raises the question of whether caffeine is the causal factor in the occurrence of spontaneous abortion, due to the complex chemical nature of coffee and tea.
Table 5:
Risk estimates for spontaneous abortion according to coffee, tea and soda
intake during the first month of pregnancy 17
|
Beverage |
n |
% with Spontaneous Abortion |
Adjusted Odds Ratio |
95% CI |
|
Coffee (cups/day) |
|
|
|
|
|
0 |
1,814 |
4.6 |
1.00 |
Referent |
|
<1 |
353 |
4.0 |
0.93 |
0.51-1.69 |
|
1 |
356 |
5.3 |
1.08 |
0.64-1.84 |
|
2 |
219 |
3.2 |
0.71 |
0.32-1.60 |
|
>3 |
107 |
11.2 |
2.63 |
1.29-5.34 |
|
Tea (cups/day) |
|
|
|
|
|
0 |
2,168 |
4.9 |
1.00 |
Referent |
|
<1 |
360 |
3.6 |
0.76 |
0.41-1.43 |
|
1 |
186 |
4.3 |
0.88 |
0.42-1.87 |
|
2 |
86 |
2.3 |
0.45 |
0.11-1.87 |
|
>3 |
49 |
12.2 |
2.33 |
0.92-5.85 |
|
Soda (cans/day) |
|
|
|
|
|
0 |
2,034 |
5.2 |
1.00 |
Referent |
|
<1 |
522 |
3.8 |
0.70 |
0.42-1.17 |
|
1 |
142 |
0.7 |
0.13 |
0.02-0.96 |
|
2 |
93 |
5.4 |
1.14 |
0.44-2.98 |
|
>3 |
58 |
5.2 |
1.10 |
0.32-3.75 |
The number of spontaneous abortions were relatively low (4.5%), which may be due in part to population bias, since the population of this study was actively seeking pre-natal care. Nausea was not measured as a possible confounding variable. The authors utilized a previously published caffeine content for coffee, tea and soda rather than calculating an estimate based on their particular subjects habits. Caffeine consumption from all food and drug sources was not assessed.
Fenster and colleagues (1997)18 , as part of a prospective study, assessed the impact of caffeinated beverages and decaffeinated coffee on spontaneous abortion. 5,144 pregnant women were interviewed between 4 and 13 weeks of pregnancy, and information was collected regarding consumption of caffeinated coffee, tea and soda as well as decaffeinated coffee. Decaffeinated coffee was investigated as a comparison for the relationship between nausea, caffeinated coffee intake and spontaneous abortion. Subjects were asked to quantify their consumption of caffeinated beverages and decaffeinated coffee in cups or cans per day in the week before the interview, which took place at about 8 weeks gestation. The same questions were asked regarding the week prior to their last menstrual period to ascertain pre-pregnancy consumption. If there was a change between the two periods, the subjects were asked approximately what week the change occurred. Women were also asked about the presence of nausea between their last menstrual period and time of interview. Spontaneous abortion was defined as termination of pregnancy at 20 weeks or gestation or less. Results indicate that there was no increased risk for spontaneous abortion related to caffeine consumed either before or during the first trimester of pregnancy. There was, however, a risk associated with heavy consumption of decaffeinated coffee. The authors could find no causal explanation and attribute the result to bias. They speculate that the result obtained from intake of decaffeinated coffee may be due to residual confounding by smoking and alcohol. Those subjects who consumed heavy amounts of decaffeinated coffee were more likely to be older, drink alcohol, smoke cigarettes and were interviewed at an earlier gestational age. They further postulate that the phenomenon may be associated with nausea. Absence of nausea in the first trimester was related to an increased risk of spontaneous abortion (adjusted odds ratio 2.6, 95% CI 2.1-3.2)18. Heavy consumption of caffeinated and decaffeinated coffee would be more predictive of spontaneous abortion due to the absence of nausea, however, in their data, heavy decaffeinated coffee consumption was not related to absence of nausea in the first trimester. As gestational age increases, the probability of nausea increases and the probability of heavy decaffeinated coffee intake decreases. If the fetus was no longer viable at the time of interview, but not yet expelled, the subject may not have been experiencing symptoms of pregnancy and would be more likely to be a heavy decaffeinated coffee consumer. As in some previous studies, caffeine intake was adapted from published standards and was not calculated based on information from the population, leading to potential inaccuracies in the estimated amount of caffeine per serving.
In 1998 Parazzini et al19 explored factors related to spontaneous
abortion including paternal coffee consumption, history of maternal coffee
drinking in addition to maternal consumption of caffeine. Because this study was conducted in
In 1999, Klebanoff et al attempted to circumvent possible inaccurate recall of caffeine intake by directly measuring serum paraxanthine to relate the possible link of caffeine intake and spontaneous abortion. The researchers hoped to either prove or disprove that mean serum paraxanthine concentration was higher in women who had spontaneous abortion and also to determine if there was a threshold above which serum paraxanthine concentration was associated with spontaneous abortion. Serum samples collected from women who participated in the Collaborative Perinatal Project between 1959 and 1966 were utilized for analysis. During the project, serum was obtained from the participants approximately every 2 months during pregnancy, at delivery and six weeks after delivery21. Serum of 487 cases; women with early fetal losses, and 2,087 controls; women who gave birth after at least 28 weeks gestation, was analyzed for paraxanthine concentration. The serum for both cases and controls was drawn on the same day of gestation. The mean serum paraxanthine level was higher in cases (752 vs. 583 ng/ml). Increased risk of spontaneous abortion, however, was almost entirely restricted to women with serum paraxanthine concentrations of 1845 ng/ml, which is equivalent to approximately 5-6 cups of coffee daily for a non-smoker. Moderate caffeine consumption was not strongly correlated with spontaneous abortion. The methods of this study are strong in that they eliminate recall bias. Serving size and caffeine content variability are also not applicable. Serum paraxanthine is a marker only of short term caffeine intake, however, the researchers assumed that caffeine ingestion was relatively constant from day to day. Women enrolled in this study relatively late in pregnancy, with a majority of the abortions occurring in the second trimester. Karyotypes were not completed on the aborted fetuses, therefore, the possibility of abortion due to chromosomal aberrations can not be ruled out, though most chromosomally aberrant fetuses are generally aborted in the first trimester. The blood samples had been stored for over thirty years and the researchers acknowledge that some of the paraxanthine may have deteriorated over time. Coffee consumption in the 1960s was much higher than it is today and women were most likely not advised to curtail caffeine consumption during pregnancy, hence Klebanoff and colleagues had a substantial sample of high caffeine consumers. In recent studies, investigators may not have been able to recruit enough women to gauge the effects of high caffeine consumption.
A population based case-control study
conducted in
Appendix 1 summarizes key data from the research reviewed above.
Discussion
It remains to be seen whether caffeine consumption is an indicator or a cause of negative pregnancy outcome. The research completed to date has produced equivocal findings on the relationship between caffeine and spontaneous abortion. As studies have evolved over the years since Weathersbee, et al in 1977, research methodology has become more fine-tuned however limitations remain. Much of the research was confined by retrospective designs, leaving the distinct likelihood of recall bias. Assessment of exposure to caffeine and the measure of moderate to high caffeine levels was very inconsistent. The highest echelon of consumption ranged from 151 mg per day to 600 or more mg per day. There were also incongruities as to when in the pregnancy caffeine exposure was measured or what point the spontaneous abortion occurred. Subjects were not always questioned on caffeine intake from all sources, such as chocolate or caffeine containing medications. Serving sizes were not consistently ascertained. One cup or serving of coffee could range anywhere from 6 to 20 ounces and the amount of caffeine contained within it can vary greatly depending on the blend and brewing method, among other factors. None of the research accounted for individual differences in caffeine metabolism, with the possible exception of Klebanoff, et al who used the biomarker of paraxanthine to assess caffeine intake. Smoking and alcohol consumption must be adequately controlled for, which was not the case in all of the research. When a spontaneous abortion did take place, a karyotype of the fetus should determine whether the pregnancy was viable and results should be adjusted accordingly. Various other factors must be considered when assessing the association of caffeine with spontaneous abortion.
Approximately 30% of women of child bearing age smoke cigarettes. Although about 20% of smokers quit during pregnancy8, many women continue to smoke despite documented effects on negative pregnancy outcome. Smoking during pregnancy has been linked to increased second trimester spontaneous abortions, dose-related fetal growth restriction, low birth weight, preterm delivery and incidence of perinatal mortality. It is assumed that cigarette smoking acts by replacing oxygen-binding sites on hemoglobin with carbon monoxide, resulting in increased levels of carboxyhemoglobin24 and possibly fetal hypoxia. The relationship between smoking and caffeine is twofold. First, cigarette smoking decreases the half-life of caffeine. Studies show that urinary levels of caffeine were lower in smokers than in non-smokers regardless of caffeine intake8. Second, there was a strong positive correlation between smoking and high coffee consumption among women in the literature reviewed. Cnattingius, et al measured serum cotinine, a breakdown product of nicotine, and observed that risk of caffeine-related spontaneous abortion was lower in women who smoked cigarettes. Smoking may indeed mask the effects and/or accelerate the metabolism of caffeine. Some studies may have had distorted results if daily cigarette smoking was not adequately quantified. In most of the research cited, smoking was a confounding variable that was controlled for, however, none of the studies included the effect of passive smoking, with the possible exception of Cnattingius, et al. Research performed in the early to mid 1990s connects environmental tobacco smoke to modest adverse effects on fetal birth weight25. A study published in 2000 links umbilical cord serum cotinine to non-smoking mothers exposed to second hand smoke26. There have been no studies to date on the relationship of environmental tobacco smoke and spontaneous abortion. The possibility that passive tobacco smoke exposure may be a confounding factor when assessing the association between caffeine and spontaneous abortion should be addressed.
Perhaps caffeine is not the culprit. Coffee and tea are complex organic substances containing a multitude of chemicals that interact naturally as well as in the manufacturing process. The fresh tea leaf contains a high level of polyphenols23. Catechins are the main flavanols (polyphenols) in the tea leaf and may form complexes with caffeine and other methylxanthines23. Many of the reaction products in the manufacturing process of tea interact with caffeine23. Physiological effects due to this chemical interaction in relation to spontaneous abortion are generally unknown. Coffee beans both in their raw (green) and roasted form are comprised of a multitude of substances including phenols such as chlorogenic acid and heterocyclic compounds such as benzo-a-pyrene. It has been postulated that these compounds may be associated with spontaneous abortion4. Extrinsic substances not inherent to green or roasted coffee are also frequently introduced. Caffeine has antifungal properties, however, green coffee is susceptible to the growth of mold if not stored in proper temperature and humidity conditions. Generally, contaminated coffee does not reach consumers. Decaffeinated beans may be more susceptible to mycotoxins27 because their protective wax covering has been removed. Roasting and brewing reduces the level of some mycotoxins, however the extent to which they are removed has not been consistently determined. Pesticides and their breakdown products may also contaminate green coffee. Again, roasting and brewing significantly decreases levels of pesticides. Little information has been published on pesticide degradation products produced in coffee that may have significant physiological effects27. Some studies implicate coffee and tea but not cola in the incidence of spontaneous abortion17,19. Others have questioned whether tap water used in brewing coffee or tea may contain substances that may contribute to unfavorable pregnancy outcome. The possibility remains that it is not caffeine, but some other component in coffee or tea, which may be associated with spontaneous abortion. Further investigation needs to be completed to support this hypothesis.
Some issues that may warrant additional investigation include the nutritional status of the mother and paternal caffeine intake. Maternal nutritional status may modify the susceptibility of the effects of caffeine during and before pregnancy28. Increases in caffeine intake may result in increased urinary excretion of calcium and thiamin as well as decreased zinc and iron absorption24. Assessment of educational level and socioeconomic status partially control for this factor, however there is no concrete evidence on the role nutritional status may play in the occurrence of spontaneous abortion.
The relation of paternal caffeine consumption to spontaneous abortion was first investigated by Weathersbee et al6,10. Few other studies since then have accounted for this variable (Parazzini et al19). Weathersbee found that paternal caffeine use resulted in adverse pregnancy outcome, while Parazzini did not. Beach et al examined the concentration of caffeine in men after a 200 or 400mg dosage. The caffeine was rapidly distributed into semen where it reached a concentration almost identical to that which was concurrently observed in blood. The caffeine half-life in both fluids was almost the same. It was concluded that caffeine partitions rapidly into secretions that contribute to the formation of the ejaculate, hence paternal caffeine intake at the time of conception could have reproductive implications29.
Another factor that was difficult to control for in these studies was the occurrence of nausea and vomiting. Nausea usually appears about six weeks into pregnancy and disappears six to eight weeks later. It is thought that increased estrogen and human chorionic gonadotropin (hCG) is the cause of this phenomenon. Nausea is generally a positive prognostic indicator for favorable pregnancy outcome. The occurrence of nausea may cause a change in normal food consumption patterns. This change extends to consumption of caffeine. Some studies that included nausea as a variable found that women who did not experience nausea were at a higher risk for spontaneous abortion. As nausea disappears, caffeine intake increases possibly leading to spurious results. This in itself does not prove that caffeine is a cause of spontaneous abortion, but raises the question of whether there is a positive correlation. Stein and Susser12 point out that if nausea begins to fade when pregnancy fails, normal caffeine intake is likely to be resumed at some point after pregnancy failure but before the expulsion of the conceptus. This may result in an erroneous appearance of raised caffeine intake before miscarriage12. They also point out that pregnancies doomed to fail from the outset, do not face additional risk from environmental factors.
Summary and Conclusion
Two of the main explanations for the relationship between caffeine and spontaneous abortion are one, that caffeine causes chromosomally normal spontaneous abortions and two, that higher caffeine levels in pregnancies resulting in spontaneous abortion are a correlate or consequence, not a cause, of spontaneous abortion17. A large prospective longitudinal study, following caffeine consumption and pregnancy outcomes, taking into account the variables listed above may prove to be the ideal way to approach this topic. As it stands today, the relationship of caffeine and spontaneous abortion remains uncertain. Based on equivocal research, it would be wise for pregnant women to follow a conservative route by limiting or excluding caffeine from their diet.