Version May 2024
INTRODUCTION — Caffeine is the most popular pharmacologically active substance consumed in the world. It is a stimulant that is often used to enhance mental alertness.
Although there is no high-quality evidence that a modest level of caffeine consumption has adverse effects on fertility or pregnancy outcome, putative beliefs about a relationship between caffeine intake and adverse reproductive outcomes are common, and caffeine consumption is often perceived to be an unhealthy habit. The consumption of caffeinated beverages has been implicated in fertility problems, while consumption during pregnancy has been associated with an increased risk of miscarriage, congenital malformations, fetal growth restriction, stillbirth, and long-term behavioral effects in offspring. The biologic rationale proposed for these associations is caffeine's ability to accumulate in fetal tissues and produce diverse pharmacologic effects that could interfere with fetal growth and development [1,2].
This topic will review the relationship between caffeine intake and reproductive issues in females. The physiologic effects of caffeine and caffeinated beverages and the effects of caffeine on specific disease processes, including insulin resistance, cancer, cardiovascular disease, and mortality are discussed separately.
●(See "Cardiovascular effects of caffeine and caffeinated beverages".)
●(See "Benefits and risks of caffeine and caffeinated beverages".)
In this topic, when discussing study results, we will use the terms "woman/en" or "patient(s)" as they are used in the studies presented. We encourage the reader to consider the specific counseling and treatment needs of transgender and gender-expansive individuals.
SOURCES OF CAFFEINE — Caffeine originates in more than 60 plants, occurs naturally in various food products and beverages (including coffee, tea, chocolate, cocoa products, colas), and is added to some soft drinks and most "energy" drinks. Caffeine is also present in some prescription and over-the-counter medications, such as cold and flu remedies, allergy and headache treatments, diet pills, diuretics, and stimulants. Increasingly, caffeine is now present as an additive in snack foods, sports performance supplements, and dietary supplements. Examples of the caffeine content of various products can be found online (eg, http://www.cspinet.org/new/cafchart.htm) and in the table (table 1).
Coffee, tea, cocoa and carbonated soft drinks are the main sources of caffeine intake. Consumption varies throughout the world, and is influenced by culture, individual preferences, availability, and advertising. Coffee contains 50 to 70 percent more caffeine than tea and other products, accounting for the main source of caffeine in many populations.
CAFFEINE CONSUMPTION IN WOMEN OF REPRODUCTIVE AGE — Estimates of daily caffeine intake vary widely, both worldwide and within individual countries, due to the many significant difficulties in ascertaining caffeine consumption (table 1) (see 'Quality of existing evidence and methodologic considerations' below). In a study of data from a United States Department of Agriculture (USDA) Survey of Food Intakes by Individuals, caffeine was consumed by 89 percent of women aged 18 to 34 years and their average intake was 164 mg per day [3]. In a study from the United Kingdom, the average daily caffeine intake of women of childbearing age was estimated to be 174 mg per day, with up to 18 percent consuming more than 300 mg per day [4]. A study from Japan reported the mean caffeine intake of women aged 30 to 39 years was 213 mg/day, and 7 percent consumed more than 400 mg/day [5].
Although most individuals reduce their caffeine intake during pregnancy, likely in part because of nausea, pregnant persons have been reported to consume an average of 60 to 125 mg of caffeine per day [3,6], and 10 to 15 percent consumed more than 160 to 200 mg daily [6,7].
PHARMACOKINETICS, METABOLISM, AND PHYSIOLOGIC EFFECTS
General population — Caffeine (methylated xanthine 1,3,7-trimethylxantine) is a highly soluble compound that readily crosses cell membranes throughout the body. It is rapidly absorbed in the stomach and small intestine and can be detected in human tissues 30 to 45 minutes after ingestion, with peak blood concentration reached within two hours. Clearance values are 1 to 3 mL/kg/min in both men and women after low caffeine intake; however, clearance is diminished with higher doses, largely because of saturable metabolism of caffeine metabolites that accumulate in plasma and reduce caffeine clearance [2]. The volume of distribution of caffeine is 0.7 to 1.3 L/kg and elimination half-life is 4.1 to 6.4 hours. Caffeine is metabolized in the liver by the cytochrome P450 family. Cytochrome P450 1A2 (CYP 1A2) is the predominant isoform, accounting for 95 percent of primary caffeine metabolism [8]. CYP 2E1 and CYP 1B1 may also make a minor contribution to the primary metabolism of caffeine [2]. Caffeine metabolism depends on numerous genetic and environmental factors. Thus, studies attempting to assess the impact of caffeine exposure may be further confounded by genotype in addition to caffeine dose [9]. (See "Cardiovascular effects of caffeine and caffeinated beverages", section on 'Pharmacokinetics'.)
The primary metabolites of caffeine, paraxanthine (1,7 dimethylxanthine), theophylline (1,3 dimethylxanthine) and theobromine (3,7 dimethylxanthine), share many of the biologic activities of caffeine and also occur independently in natural products [10]. Paraxanthine is the main metabolite in humans and is present in coffee beans; theophylline is a significant xanthine in tea leaves; and theobromine, the weakest of xanthines, is present in cacao beans. Further metabolism leads to the formation of monomethylxanthines and methyluracils. Caffeine is excreted by the kidneys, mainly in the form of different metabolites; only 0.5 to 2.0 percent is excreted unchanged in urine [11].
Individual variations in caffeine metabolism — The clearance rate of caffeine varies among individuals and depends on factors affecting cytochrome P450 enzymes, particularly genetic polymorphisms; smoking; and exposure to other compounds that are metabolized by the same enzyme. As a result, despite equivalent caffeine intake, slow metabolizers have higher caffeine concentrations and may be at higher caffeine-related risk for loss of pregnancy than fast metabolizers [12,13].
Maternal and fetal kinetics — Caffeine and its metabolites readily cross the placenta and can be found in substantial quantities in the amniotic fluid and fetal blood [14]. Maternal caffeine metabolism declines significantly during pregnancy; the half-life increases threefold in the third trimester, reaching a t1/2 of 11.5 to 18 hours [15,16]. The fetus metabolizes caffeine very slowly, mainly due to immaturity of caffeine-metabolizing hepatic microsome enzymes and lack of CYP 1A2 activity in the placenta (the placenta does not metabolize caffeine) [1]. Therefore, even low maternal caffeine consumption can be expected to lead to prolonged fetal caffeine exposure, particularly when the mother is a genetically slow caffeine metabolizer. Infants of smokers have lower umbilical cord blood caffeine concentrations and higher concentrations of caffeine metabolites than infants of nonsmokers, reflecting faster caffeine metabolism in smokers [1].
Caffeine consumption may increase release of circulating catecholamines [17,18], which could cause uteroplacental vasoconstriction leading to fetal hypoxia, but an increase in catecholamines has not been reported consistently [19]. Increases in maternal homocysteine, cholesterol, and cellular cyclic adenosine monophosphate (cAMP) concentrations [20-24] and changes in maternal reproductive hormone levels [25] have also been reported, and potential adverse effects on pregnancy outcome have been hypothesized but have not been proven.
QUALITY OF EXISTING EVIDENCE AND METHODOLOGIC CONSIDERATIONS — Examining the association between caffeine consumption and adverse reproductive outcomes presents a considerable challenge. The consistency and quality of available data have generally been poor. Most studies have methodologic shortcomings and were based on epidemiologic research designs usually classified as level III or low quality evidence (ie, comparative studies with concurrent controls but allocation not randomized, cohort studies, case-control studies, or interrupted time series with a control group) [26].
There are many potential reasons for inconsistent outcomes among studies, including but not limited to [1]:
●Inaccurate estimation of caffeine consumption
●Recall bias due to retrospective assessment of caffeine intake
●Failure to allow for individual variations in caffeine metabolism
●Inadequate control of confounding factors
●Selection bias
●Chemicals other than caffeine in dietary products, which may contribute to the observed outcome
Comprehensive reviews of these methodologic issues are available elsewhere [1,27,28], and will be reviewed briefly here. In particular, caffeine consumption is likely to be lower in early than late pregnancy because of early pregnancy nausea [29]. The majority of existent studies primarily capture the short-term effects of caffeine by assessing consumption in the period shortly before the studied reproductive event or in individual trimesters rather than throughout pregnancy, although many health-related conditions are likely to have multifactorial causes and a long natural history. It has been difficult to document the long-term effects of caffeine because many new caffeinated products appear in the marketplace and consumption patterns change as a result of dietary advice from a variety of experts.
Assessment of caffeine intake — Assessment of caffeine consumption is methodologically challenging due, in part, to the wide variation of caffeine content in beverages and large differences in serving sizes. The amount of caffeine varies according to source (coffee, tea, chocolate, soft drink), portion size (eg, cup size), brewing method (for coffee and tea), and brand [30]. A standard cup of coffee is often assumed to provide 100 mg of caffeine; however, an analysis of caffeine content of different coffees from various coffee shops revealed variations of 72 to 130 mg/cup [31]. For espresso, an analysis of caffeine content reported caffeine content of 200 to 322 mg/cup in 20 commercially available espresso products [32].
Coffee is reported as the only source of caffeine in 19 percent of pregnant caffeine consumers, with 26 percent of caffeine intake solely from other non-coffee dietary sources [33,34]. Therefore, using coffee consumption as a proxy for caffeine intake leads to underestimation of caffeine consumption. The majority of studies on caffeine and reproductive outcomes neither made a comprehensive analysis of intake of all food, drinks, and drugs containing caffeine, nor carefully accounted for the variation of caffeine content from individual portion size, brand, or preparation method of these foods and drinks.
Furthermore, most epidemiologic studies have relied on indirect data, such as self-reported intake of caffeine, instead of laboratory assessment of biomarkers. The biomarkers of caffeine exposure (caffeine and its metabolites in blood, urine, and saliva) are of limited use in large population samples, are highly dependent on the timing of the last caffeine intake, and primarily indicate recent consumption rather than long-term exposure [7,35].
Although caffeine is detectable in amniotic fluid and in the nails of exposed newborns, studies have not used these types of assays to monitor in utero caffeine exposure.
Recall bias — Recall bias refers to differential recall between cases and controls and is of concern because it distorts the internal validity of studies using self-reported data. It affects studies (both retrospective and prospective) in which assessment of caffeine consumption is determined after, rather than before, the occurrence of the health outcome of interest. Patients who have had an adverse event may be more likely to remember perceived unhealthy habits and associate their consumption of coffee with the adverse event, and thus may overestimate their past caffeine consumption compared with patients who have not experienced an adverse event.
The importance of recall bias was highlighted by a study that found the correlation between modest caffeine consumption and pregnancy loss disappeared when pregnant women were interviewed about caffeine consumption before, rather than after, they had miscarried [36].
Potential confounders in studying caffeine consumption — Heavy coffee consumption tends to be strongly associated with increased age, cigarette smoking, alcohol consumption, a less health conscious lifestyle, and lower educational level and household income [2]. Other potential confounders include ethnicity, stress, and sleeping pattern. Trying to establish the independent health effects of coffee consumption by adjusting for these potential confounders is difficult and can be inaccurate since their assessment is also affected by measurement error.
Nausea and other symptoms of a viable early pregnancy are sometimes called the "pregnancy signal." A particular concern in studies of the effects of caffeine intake during pregnancy is that pregnancy related nausea is common and an important confounding factor [37]. Nausea in early pregnancy tends to cause dietary changes, such as an aversion to coffee, but is also strongly predictive of fetal survival, presumably because nausea reflects hormonal changes that support continued pregnancy [38]. For this reason, there has been a concern that the lower caffeine consumption and more favorable pregnancy prognosis among women with nausea compared with those who do not have nausea result in an artificial positive association between low caffeine consumption and better pregnancy outcome. Well-designed studies of caffeine intake and pregnancy outcome attempt to control for the presence or absence of pregnancy signal symptoms in their subjects.
Selection bias — Many of the studies on the effect of caffeine on reproductive adversity are based on data obtained from preconception counseling and prenatal health services. The patients attending these services are likely to be more health conscious, which may be associated with lower coffee consumption and thus lead to selection bias, but this would also tend to reduce the associated odds ratios.
Other components of caffeinated beverages that may contribute to health effects — Coffee, tea, and cocoa products contain hundreds of compounds other than caffeine that may contribute to the observed health effects of caffeine consumption [39]. In addition, coffee and black tea are often consumed with milk or a creamer. Non-dairy creamers contain partially hydrogenated oils that, when used in large amounts, could be a significant dietary source of trans fatty acids, which may have independent adverse effects on reproduction. Trans fatty acids may interfere with the desaturation and elongation of n-3 (omega-3) fatty acids. These are important for the prevention of heart disease and possibly some complications of pregnancy.
EFFECTS OF CAFFEINE ON REPRODUCTIVE OUTCOMES
Animal and in vitro studies — Studies in animals can overcome many of the methodologic limitations of studies in humans, but pose additional methodologic issues that prevent extrapolation to humans. First, humans are highly unlikely to consume the high doses of caffeine used in animal studies. Second, although many metabolic and pharmacokinetic factors related to caffeine are similar in humans and animals, there is not a direct correlation between teratogenic effects in animals and in humans, and the dose/anomaly relationship may vary between the species. Finally, in some animal studies, caffeine intake is by intravenous or intraperitoneal infusion or gavage feeding, rather than by oral consumption in food/drink over the course of a day, as in humans. This difference may result in different effects in laboratory animals versus humans. Indeed, when animals receive caffeine doses typically consumed by humans, caffeine is not mutagenic, oncogenic, or cytotoxic [27]. However, it is possible that subteratogenic doses of caffeine may have synergistic detrimental effects with other therapeutic agents consumed by pregnant women.
Some examples of findings from animal studies are summarized below:
●In rodents, a teratogenic effect was demonstrated at extremely high caffeine doses (100 to 330 mg/kg); limb deficiencies, cleft palate, and neural tube defects (NTDs) were among the most frequently observed caffeine-induced congenital malformations [40-43]. The caffeine-induced NTD phenotype has been demonstrated by others in various animal models [44-47].
●Caffeine injected into chicken eggs affected monoamine neurotransmitter synthesis in chick embryos and was not metabolized in embryonic brains, which led to caffeine accumulation and magnification of the adverse effects [44]. These changes have been implicated in subsequent neurobehavioral or neurophysiologic deficiencies [48-50].
●Rodents given a large bolus or parenteral dose of caffeine are at increased risk of growth restriction in their pups; however, no effect on birth weight has been observed when large doses of caffeine were consumed over the course of the day in drinking water [27]. Caffeine boluses also delayed conception and decreased placental weight and gain in maternal bodyweight; these effects were dose dependent [51]. Using a dose-conversion correlation between humans and rats, the lowest dose of caffeine given to pregnant rats and associated with the above effects appeared to be equivalent to 3.2 mg/kg per day in humans, which is equivalent to consumption of 1.5 to 2.2 cups of coffee by a 70 kg woman.
Several basic research studies have assessed the effect of caffeine on trophoblast biology in vitro and suggested molecular mechanisms for deleterious caffeine effect on pregnancy outcomes [52-54]. A detailed discussion of these mechanisms is beyond the scope of this topic.
Human studies
Summary — Although there is ongoing controversy regarding the safety of caffeine consumption in females planning to conceive or in those who are pregnant, most studies on the topic, and resultant recommendations, provide reassurance that small to moderate amounts of caffeine are unlikely to harm the pregnancy (table 2) [55-61]. Higher levels of maternal caffeine consumption (above 300 mg per day) have generally been associated with adverse reproductive outcomes, while the impact of lower levels of consumption is less clear. Although earlier studies had suggested lower levels of caffeine intake had little to no impact [27,28], subsequent reviews reported conflicting results, including potential for harm. Examples include:
●A 2015 systematic review that included one trial comparing the effects of caffeinated coffee intake with decaffeinated coffee use on pregnancy outcomes demonstrated that reducing the caffeine intake in high coffee consumers by an average of 182 mg per day during the second and third trimesters did not affect birth weight or length of gestation [62]. The authors concluded there was insufficient evidence to confirm or refute the effect of caffeine avoidance on pregnancy outcomes and neonatal health.
●A 2017 comprehensive and structured systematic review of 381 studies published from 2001 to June 2015 found that, for healthy pregnant women, consumption of up to 300 mg of caffeine per day was generally not associated with adverse reproductive or developmental effects [63]. It emphasized the need to support a shift in caffeine research to a better quantitative characterization of interindividual variability.
●A 2020 narrative review of 42 findings from 37 observational studies, all published since 2000 and many used in the 2017 review, concluded that caffeine consumption was significantly associated with negative outcomes, including pregnancy loss (miscarriage), stillbirth, low birth weight, and small for gestational age [64]. Negative outcomes were identified at lower levels of consumption and increased in a dose-dependent manner. Even moderate caffeine consumption (200 mg per day) was alleged to be unsafe. As discussed above, the effect of extraneous variables and possible misclassification of exposure and outcome variables remain limitations of observational studies. Evaluation of the explanation for the difference in conclusions between the 2017 and 2020 studies is pending but highlights the need for caution in advising women who are planning to become pregnant or are in the early stages of pregnancy. (See 'Quality of existing evidence and methodologic considerations' above.)
A list of the caffeine content of beverages and foods is available in the table (table 1).
Fertility
Fecundability — When adequately adjusted for other lifestyle factors related to fertility, particularly maternal age, cigarette smoking, and alcohol intake, the bulk of evidence does not support a clear detrimental effect of caffeine on fecundability (ie, probability of conception within a non-contracepting menstrual cycle) [28,65,66].
In 1988, a study of couples who failed to conceive within the first three months of trying to become pregnant observed that consumption of at least one cup of coffee, two cups of tea, or 2.5 glasses of cola daily was associated with a marked reduction in fecundability in subsequent cycles compared with women who consumed less caffeine, and a dose response effect was noted [67]. This study prompted further investigation of a possible relationship between caffeine and subfertility. Although some retrospective studies have also observed reduced fecundability, particularly with high caffeine intake [68,69], most prospective epidemiologic studies have not found a statistically significant association between caffeine consumption and fecundability [70-75]. One of these studies was particularly strong because it was the first to use daily, prospective information on caffeine intake and prospectively ascertained pregnancies, including subclinical pregnancies, while controlling for frequency of intercourse, smoking, age, body mass index, frequency of unprotected intercourse, and alcohol intake [71]. Another observed decreased fecundability among women who consumed sodas and increased fecundability among women who drank tea [75]. This indicates that a caffeine exposure variable that does not take into account the specific source may not be appropriate and suggests that these associations may reflect unmeasured confounding by diet or other lifestyle factors.
The relationship between caffeine consumption and success of in vitro fertilization (IVF) has not been studied as extensively. In one study, caffeine intake by women undergoing IVF did not affect rates of oocyte retrieval, fertilization, embryo development, or pregnancy rate when caffeine consumption was assessed just before or during the treatment cycle [76]. However, lifetime caffeine intake in the same group of women seemed to have a negative effect on achieving and maintaining pregnancy. This was a small study and the confidence intervals around the point estimates were wide. Another study concluded caffeine consumption did not affect overall IVF pregnancy rates; however, higher coffee/tea consumption was negatively associated with the number of oocytes retrieved and the number of good quality embryos [77]. Caffeine was detected in follicular fluid; its actions there are unknown.
The available data are not definitive; therefore, reproductive endocrinologists suggest that it may be prudent for women who are having difficulty conceiving to limit caffeine consumption to less than 200 to 300 mg per day, in addition to eliminating tobacco use, decreasing alcohol consumption, and other sensible lifestyle actions [58,65,66,78].
A 2017 systematic review demonstrated lack of consistency among studies reporting on effect of caffeine on male fertility. Most observational studies on the topic did not observe significant alterations in semen parameters or consistent changes in sperm DNA integrity, although it is widely acknowledged that semen quality has only limited predictive value for pregnancy [79]. Another systematic review that synthesized the available evidence on effect of nutrients on male fertility demonstrated that, while the data on association of caffeine and low semen quality are inconclusive, high caffeine intake in male partners has a negative influence on the chance of pregnancy or fertilization rates [80]. Likewise, a prospective cohort study on effects of environmental exposures on reproductive health in subfertile couples (EARTH Study), demonstrated that male caffeine intake was negatively associated with live birth following assisted reproductive technology treatments, while having little effect on semen quality [81]. In this study, adjusted live birth rate was 19 percent in couples with a male partner consuming above 272 mg caffeine/day versus 55 percent in couples with a male partner consuming below 99 mg caffeine/day [81].
Pregnancy loss (miscarriage) — As discussed in detail below, studies assessing impact of maternal caffeine consumption on risk of pregnancy conflict, which makes patient counseling challenging. Additional prospective high-quality research is required to confirm or refute the direct association.
A 2011 systematic review of both human and animal studies of the risk of spontaneous abortion from caffeine exposure concluded there was fair to good evidence that consumption of caffeinated beverages during pregnancy at a level ≤5 to 6 mg/kg body weight/day does not increase the risk of spontaneous abortion [27]. Although high caffeine intake in some animal studies increased the risk minimally, even a woman who consumed 10 cups of coffee over a period of 8 to 10 hours (approximately 1000 mg of caffeine) would not reach this harmful level of intake.
Findings among the 17 epidemiologic studies in this review were inconsistent and many studies did not observe an association at any level of reported exposure [27]. Studies that noted a link between caffeine intake and spontaneous abortion generally reported a dose-response relationship, but differed in their conclusions about a threshold for a safe amount of caffeine consumption. An increased risk of spontaneous abortion generally was not observed until self-reported intake levels were ≥3 cups or ≥300 mg caffeine per day. A study using an objective assessment of caffeine exposure, serum paraxanthine concentration, observed an increased risk beginning at caffeine doses ≥600 mg [82]. In addition, an analysis within the Danish national birth cohort, which included 565 women who consumed a prescription diet drug (Letigen) containing 200 mg of caffeine per pill (ie, approximately 600 mg/treatment day) periconceptionally, found no association between miscarriage and high-dose caffeine [83]. In contrast, a prospective study that recruited 344 couples before conception and followed them through the first seven weeks of gestation reported an association between early miscarriage and consumption of >2 daily caffeinated beverages by either parent [84]. Also, a prospective cohort study of 5132 Danish women with spontaneous pregnancy demonstrated a small increased risk of spontaneous miscarriage with caffeine consumption above 100 mg/day in early pregnancy, although preconceptional consumption did not seem to increase the risk [85].
The methodologic issues and limitations previously discussed apply to the association reported in these studies, and include confounding by the inverse relationship between maternal nausea and both coffee consumption and miscarriage rate; recall bias; and imprecise ascertainment of caffeine intake [36,38]. Other potentially distorting factors include using controls that were not well-matched for gestational age to the study population, lack of ascertainment of the multiple known causes of abortion, a tendency to confuse induced terminations with spontaneous abortions, and inadequate adjustment for multiple known confounders [27].
Fetal outcomes
General impact — The physiologic effects on the fetus of caffeine exposure have not been studied extensively in human pregnancy. Higher maternal plasma caffeine levels appear to be associated with a longer duration of fetal arousal, but fetal tolerance to caffeine exposure also appears to occur [86-88]. Other fetal effects possibly related to maternal caffeine consumption include increased fetal heart rate variability, lower basal heart rate [86,89], and increased fetal breathing activity [88,89]. An increase in beat to beat variability and accelerations of the fetal heart rate following maternal intake of even one cup of espresso coffee (60 mg caffeine) or 30 g of "dark" chocolate (70 percent cocoa) in the third trimester has been attributed to the pharmacologic action of theobromine, present in coffee and in chocolate [90].
Congenital anomalies — The overall body of evidence remains reassuring that low levels of maternal caffeine consumption are unlikely to increase the risk of congenital anomalies. Although some epidemiologic studies have reported an association between caffeine intake and certain congenital anomalies, the results should be interpreted with caution due to the methodologic flaws of these studies (eg, lack of correction for multiple comparisons, poor ascertainment of caffeine intake, etiologic heterogeneity for some malformations, biased exposure data, biased ascertainment of congenital anomalies) or because the findings were inconsistent with basic teratologic principles (eg, lack of dose effect, lack of increasing risk with increasing exposure, lack of consistent association with a specific defect or constellation of defects).
●Unlikely to increase risk – Several narrative reviews and a systematic review concluded that caffeine is unlikely to cause congenital anomalies at doses consumed by humans (eg, fractional doses totaling <5 mg/kg body weight/day) [27,28,91,92].
●Possible increased risk but no evidence of dose-response – The US National Birth Defects Study (NBDPS), 1997-2011, compared prepregnancy and first-trimester caffeine exposure (by patient self-report) for children with congenital anomalies (n = 30,285) with an unaffected control group (n = 11,502) and reported increased risk of some anomalies for both low (10 to <100 mg/day) to moderate 100 to <200 mg/day) and high (≥300 mg/day) levels of prepregnancy caffeine consumption [93]. For the anomalies with increased adjusted odds ratios, similar elevations were seen across all levels of caffeine consumption, which indicates that a dose-response relationship was not observed. We and the study authors believe this analysis further confirms previous analyses demonstrating a lack of evidence for a meaningful association between maternal caffeine consumption and the studied congenital anomalies.
Low birth weight — Although observational and cohort studies reported conflicting data on the impact of caffeine on birth weight [27,94-102], two meta-analyses found that maternal caffeine intake during pregnancy was associated with a higher risk of delivering low birth weight infants (defined as birth weight less than 2500 gm, regardless of gestational age) as compared with the reference group (no or very low caffeine intake) [103,104]. The risk of low birth weight increased as the maternal caffeine intake increased. In the dose-response analysis, each 100 mg per day increment in maternal caffeine intake was associated with a 3 to 13 percent higher risk of low birth weight, which further supports the current recommendation to minimize caffeine intake during pregnancy [103,104]. (See 'Recommendations of societies' below.)
However, a limitation of observational studies is that they cannot or may inadequately control for important risk factors for low birth weight, particularly smoking [28,78,91]. The only randomized double blind trial that analyzed the effect of reducing caffeine intake on birth weight and length of gestation found that reducing caffeine intake by greater than 50 percent had no significant effect on these outcomes [105]. In this trial, 1207 pregnant women who drank at least three cups of coffee per day were randomly assigned to receive unmarked jars of caffeinated or decaffeinated instant coffee for consumption after 20 weeks of gestation. Mean caffeine intake in the decaffeinated and caffeinated coffee groups was 117 and 317 mg/day, respectively. Mean birth weight and mean length of gestation were similar for women in the decaffeinated and caffeinated coffee groups (mean birth weight 3519 versus 3539 grams; mean length of gestation 279.3 versus 280.2 days) [105].
In light of these conflicting data, we advise women of the possible detrimental impact of caffeine consumption on fetal growth, although restricted growth is unlikely to be associated with moderate (200 to 300 mg/day) caffeine consumption; this approach is consistent with expert opinion [2]. However, lower amounts of caffeine consumption may be associated with small but statistically significant decreases in birth weight that do not meet the definition of low birth weight [106]. The clinical significance of small reductions in birth weight at term is not known.
Preterm delivery — A 2010 systematic review and meta-analysis that combined 15 cohort and 7 case-control studies did not find a significant association between maternal caffeine intake anytime in pregnancy and preterm birth [107]. The lack of association was observed when comparing "any" versus "no" caffeine consumption and for "high" versus "low/no" caffeine consumption. For the highest versus the lowest level of caffeine intake (or no intake) during the first, second, and third trimesters, the odds ratios for preterm birth in cohort studies were 1.11 (95% CI 0.96-1.28), 1.10 (95% CI 1.01-1.19), and 1.08 (95% CI 0.93-1.27), respectively. For case-control studies, the odds ratios were 1.07 (95% CI 0.84-1.37), 1.17 (95% CI 0.94-1.45), and 0.94 (95% CI 0.79-1.12), respectively. Subsequent studies have reported mixed outcomes regarding risk of preterm birth, defined as delivery less than 37 weeks gestation, and caffeine exposure, but some studies were limited by small sample size [62,102].
Fetal death — A systematic review of observational studies that examined the association between any source of exposure to caffeine from food in pregnancy and fetal mortality concluded that "the small number of publications, methodologic limitations, inaccurate exposure assessment in all the studies, overall risks only marginally significant in most cases, and the possibility of publication bias preclude stating with certainty that caffeine consumption is actually associated with fetal death" [108].
Maternal outcomes
Gestational diabetes mellitus — Consistent with epidemiologic studies in nonpregnant individuals reporting reduced risk of type 2 diabetes mellitus with long-term coffee consumption, low-dose coffee consumption during pregnancy appears to reduce the risk of gestational diabetes [109-111]. (See "Benefits and risks of caffeine and caffeinated beverages", section on 'Type 2 diabetes mellitus'.)
●In a prospective cohort study, periconceptional moderate caffeine intake from coffee appeared to protect against development of gestational diabetes mellitus (GDM; adjusted RR 0.50, 95% CI 0.29-0.85); there was no risk reduction associated with consumption of decaffeinated coffee or caffeinated tea or soda [109].
●Similarly, a population-based cohort study including over 71,000 singleton pregnancies in nondiabetic Danish women suggested that moderate first-trimester coffee and tea intake was not associated with increased risk of GDM and possibly had a protective effect [110].
Paradoxically, several studies have shown that exposure to caffeine in experimental settings mildly reduced insulin sensitivity in humans [112-114]. This may be due to combination of an indirect effect of caffeine via increasing levels of circulating epinephrine or to direct antagonism of adenosine receptors. Overall, the effect of caffeine on insulin resistance appears to be mild and has been mainly related to acute exposure. Moreover, the researchers in these studies predominantly implemented alkaloid caffeine, which may cause a different physiologic response than caffeine from dietary sources. The significance of caffeine's acute effect on the physiologic release of insulin in the natural setting, as well as the degree of tolerance to its metabolic effects during chronic consumption, remains unclear [115]. Pregnant women in whom there are physiologic changes in glucose metabolism and a pregnancy-induced state of peripheral insulin resistance may be more likely to show the acute effects of caffeine observed in experimental settings. A retrospective study on banked blood samples of 251 women in the second trimester (mean gestational age of 20.3 weeks) showed a positive association between the risk of insulin resistance (estimated by homeostasis model assessment) and higher levels of caffeine and its metabolite paraxanthine in the blood [116]. Caffeine concentrations >266 ng/mL and paraxanthine concentrations >392 ng/mL were associated with threefold higher odds of having higher insulin resistance [116]. The clinical significance of these findings was not studied and requires further investigation.
Gestational hypertension — A group from the Netherlands assessed the effect of caffeine consumption on maternal cardiovascular changes in a cohort of 7890 pregnant women [117]. Although higher caffeine intake was associated with elevated systolic blood pressure in the first and third trimesters, there was no effect on diastolic blood pressure levels and caffeine consumption appeared to be protective against preeclampsia.
Postpartum depression — A study of 662 women recruited within 24 hours of giving birth did not observe an association between prenatal caffeine intake and depressive symptoms postpartum [118].
HEALTH IMPACT ON CHILD — The impact of caffeine exposure on neurodevelopment and growth appears to vary by outcome of study and duration of follow-up. Examples include the following studies:
●Neonatal withdrawal – Several case reports describe symptoms of neonatal caffeine withdrawal including irritability, jitteriness, and vomiting. Symptoms began in immediate postnatal period and resolved without intervention within 84 hours. All the cases were associated with heavy prolonged maternal consumption of caffeine [119,120]. There are no data on neonatal withdrawal in mothers consuming fewer than six caffeinated drinks per day.
●Intelligence Quotient (IQ) – Maternal caffeine consumption does not appear to impact offspring IQ. A study of 2197 mother-child pairs reported no association between caffeine exposure in the second and third trimester with childhood IQ or behavior problems at ages four and seven years [121]. The caffeine exposure in this study was assessed by measuring serum paraxanthine levels in women before 20 weeks and after 26 weeks [3]. Similar lack of association was also reported by a secondary analysis in a different prospective study with seven years of follow-up [122].
●Overactive behavior – The data on the impact of caffeine on childhood overactive behavior are conflicting and may be dose-related. In one prospective study of over 25,000 mother-child pairs that reported an association between caffeine consumption and overactive behavior in children aged 18 months, the association only held for intrauterine exposure to soft drinks and not for coffee [123]. A prospective cohort study that assessed maternal caffeine consumption by self-report and measured childhood behavior at five to six years (over 3400 children) reported no association between maternal caffeine and behavior disorders. In contrast, in the largest study from the Danish National Birth Cohort of over 47,000 children born between 1996 and 2002, maternal consumption of 8 or more cups of coffee or tea per day at 15 weeks of gestation was associated with increased risks of behavioral disorders in children at 11 years of age [124]. Excessive coffee consumption was associated with a 50 percent increased risk of hyperactivity-inattention disorder, while excessive tea consumption was associated with a 28 percent increased risk of anxiety-depressive disorders and 24 percent increased risk of any psychiatric disorder. Strengths of the study include the older age of the assessed children and large sample size, while the study is limited by use of maternal self-report of caffeine intake. Of note, anxious mothers tend to consume more caffeine, which in itself creates an inherent selection bias or reverse causation. More work with adjustment to maternal emotional status is required to shed light on effect of caffeine consumption on behavioral disorders in offspring.
●Childhood growth and risk of overweight or obesity – Although there are conflicting data, large prospective observational studies have reported an increased risk of high weight gain and elevated body mass index (BMI), as well as shorter height, in children born to mothers who consumed large amounts of caffeine while pregnant [125-130]. Although potential mechanisms are not yet understood, these data provide further rationale to limit caffeine intake while pregnant.
•Increased childhood weight gain – In a prospective Norwegian cohort study of nearly 51,000 mothers with singleton pregnancies, children born to women with average (50 to 199 mg/day), high (≥200 to 299 mg/day), and very high (≥300 mg/day) caffeine intake during pregnancy had an increased risk of excess weight gain (defined as gender-adjusted World Health Organization weight-for-ages Z-score of >0.67) during infancy and early childhood when compared with children born to women with low caffeine intake (<50 mg/day) in adjusted analysis [131]. The odds of being overweight at three and five years rose with increasing maternal caffeine consumption (odds ratios 1.15 (95% CI 1.09-1.22), 1.30, (95% CI 1.16-1.45), and 1.66 (95% CI 1.42-1.93), respectively. There was also a linear association between maternal caffeine consumption and the risk of having an overweight or obese child (as determined by body mass index) with similar findings for caffeine from different caffeine sources (black coffee, black tea, and soda drinks). By eight years of age, only children of very high caffeine exposure mothers continued to have an increased risk of being overweight or obese. In subgroup analysis that excluded the highest caffeine consumption group, caffeine intake of <300 mg/day was still associated with increased infant growth and overweight when compared with women who did not consume caffeine and there was no threshold effect for safety. Adjusted variables included maternal age, maternal education, parity, prepregnancy BMI, paternal BMI, maternal and paternal smoking during pregnancy, maternal energy intake and nausea/vomiting during pregnancy, as well as gestational age and child's gender. Although prenatal caffeine exposure is generally associated with an increased risk of lower birth weight [104,132], in utero caffeine exposure is thought to increase postnatal growth through alterations in fetal programming [133,134]. However, a subsequent smaller prospective observational study that followed children up to age eight years did not observe an association between maternal caffeine consumption and childhood BMI, fat mass, or fat percentage [130].
•Shorter childhood height – A prospective observational study including two different cohorts of children (n = 788 and 1622) reported that higher maternal caffeine consumption was associated with shorter childhood stature [130]. The finding was present across all quartiles of caffeine exposure, dose-related, and persisted up to eight years of life. At four to eight years, height differences were between 0.68 to 2.2 cm, similar to the results from women who smoked in pregnancy [130]. As in previous studies, there was insufficient detail on potential confounding factors such as maternal diet, nausea and vomiting in early pregnancy, or paternal factors.
●Other – In addition to childhood growth and general development, associations between in utero caffeine exposure and leukemia, testicular cancer, sudden infant death syndrome, and brain development in offspring have been evaluated, but the available data are either inconclusive or of unclear clinical significance [135-139].
LACTATION — Caffeine is detectable in breast milk within 15 minutes of consumption and levels peak after approximately one hour [140]. Maternal caffeine consumption of 500 mg/day results in daily infant caffeine intake ranging from 0.3 to 1.0 mg/kg body weight; after maternal consumption of up to 335 mg of caffeine, the amount available through breast milk has been estimated at less than 2 mg per 24 hour period [141]. Adverse effects have not been reported in term infants whose mothers had moderate intake of three to five cups of coffee daily, but some case reports have described an increase in infant irritability, jitteriness, or sleep disturbance beyond this level of maternal caffeine consumption [78,142,143]. In one of the larger series (n = 885 infants), heavy (≥300 mg/day) maternal caffeine consumption during pregnancy and three months postpartum by breastfeeding mothers was associated with an increase in the percentage of infants with frequent (>3 times per night) nocturnal wakening (19.5 versus 13.3 percent with non-heavy or no maternal caffeine consumption; RR 1.65, 95% CI 0.86-3.17), but the difference was not statistically significant [144].
RECOMMENDATIONS OF SOCIETIES — No uniform counseling guidelines or evidence-based recommendations for caffeine intake are available. Many medical societies and government bodies recommend limiting caffeine intake in pregnancy to 200 to 300 mg per day (table 2) [55-60,63,145].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: General prenatal care".)
SUMMARY AND RECOMMENDATIONS
●Common sources of caffeine – In many countries, the main source of caffeine is coffee, followed by tea and caffeinated beverages (table 1). Examples of the caffeine content of various beverages are shown in the table (table 1) and available at http://www.cspinet.org/new/cafchart.htm. (See 'Sources of caffeine' above and 'Recommendations of societies' above.)
●Maternal caffeine metabolism and fetal impact – Maternal caffeine metabolism declines significantly during pregnancy. Caffeine and its metabolites readily cross the placenta and can be found in substantial quantities in the amniotic fluid and fetal blood. Higher maternal plasma caffeine levels appear to be associated with a longer duration of fetal arousal, increased fetal heart rate variability, lower basal fetal heart rate, increased fetal breathing activity, and inconsistent effects on fetal and placental blood flow. (See 'Maternal and fetal kinetics' above.)
●Data limitations and patient guidance – Given the limitations and inconsistencies of available data, we suggest individuals who are attempting to conceive or who are pregnant or breastfeeding limit caffeine consumption to less than 200 to 300 mg per day until more conclusive data are available (table 2) (Grade 2C). Although there is ongoing controversy regarding the safety of caffeine, most studies on the topic, and resultant recommendations, provide reassurance that small to moderate amounts of caffeine are unlikely to harm the pregnancy. (See 'Quality of existing evidence and methodologic considerations' above and 'Recommendations of societies' above.)
•Fertility and pregnancy outcomes – At typical levels of caffeine intake (up to 300 mg per day from all sources), there does not appear to be a convincing association between caffeine consumption and fecundability or adverse reproductive events, including congenital anomalies, spontaneous abortion, fetal growth restriction, and preterm birth. (See 'Fertility' above and 'Fetal outcomes' above.)
•Maternal outcomes – Periconceptional moderate coffee intake appears to protect against development of gestational diabetes mellitus. (See 'Gestational diabetes mellitus' above.)
•Childhood outcomes – Maternal caffeine consumption can result in neonatal withdrawal. Higher levels of maternal consumption have been associated with childhood behavioral and growth changes, although available data conflict. (See 'Health impact on child' above.)
آیا می خواهید مدیلیب را به صفحه اصلی خود اضافه کنید؟
