Original Article

Caffeine’s Implications for Women’s Health and Surveyof Obstetrician-Gynecologists’ Caffeine Knowledge

and Assessment Practices

Britta L. Anderson, B.A.,1,2 Laura M. Juliano, Ph.D.,2 and Jay Schulkin, Ph.D.1

Abstract

Objective: Caffeine has relevance for women’s health and pregnancy, including significant associations withspontaneous abortion and low birth weight. According to scientific data, pregnant women and women ofreproductive age should be advised to limit their caffeine consumption. This article reviews the implications ofcaffeine for women’s psychological and physical health, and presents data on obstetrician-gynecologists’ (ob-gyns) knowledge and practices pertaining to caffeine.Methods: Ob-gyns (N¼ 386) who are members of the American College of Obstetricians and Gynecologists’Collaborative Ambulatory Research Network responded to a 21-item survey about caffeine.Results: Although most knew that caffeine is passed through breast milk, only 24.8% were aware that caffeinemetabolism significantly slows as pregnancy progresses. Many respondents were not aware of the caffeinecontent of commonly used products, such as espresso and Diet Coke,� with 14.3% and 57.8% indicating amountswithin an accurate range, respectively. Furthermore, ob-gyns did not take into account large differences incaffeine content across different caffeinated beverages with most recommending one to two servings of coffee ortea or soft drinks per day. There was substantial inconsistency in what was considered to be ‘‘high levels’’ ofmaternal caffeine consumption, with only 31.6% providing a response. When asked to indicate the risk that highlevels of caffeine have on various pregnancy outcomes, responses were not consistent with scientific data. Forexample, respondents overestimated the relative risk of stillbirths and underestimated the relative risk ofspontaneous abortion. There was great variability in assessment and advice practices pertaining to caffeine.More than half advise their pregnant patients to consume caffeine under certain circumstances, most commonlyto alleviate headache and caffeine withdrawal.Conclusions: The data suggest that ob-gyns could benefit from information about caffeine and its relevance totheir clinical practice. The development of clinical practice guidelines for caffeine may prove to be useful.

Introduction

Caffeine (1,3,7-trimethylxanthine), a mild central ner-vous system (CNS) stimulant, is the most widely used

behaviorally active drug in the world. Caffeine is ingestedfrom various sources, including coffee, tea, soft drinks, energydrinks, foods, dietary supplements, and medications (Table 1).In the United States, more than 85% of adults and children,and 68% of pregnant women consume caffeine on a regularbasis.1 Mean daily caffeine intake has been estimated to beabout 280 mg among adults2 and 125 mg among pregnantwomen in the United States.1

Caffeine has important clinical implications in the contextof women’s emotional and physical health as well as preg-

nancy outcomes. Physicians and particularly obstetrician-gynecologists (ob-gyns), due to their expanding role as pri-mary care providers,3 may benefit by having backgroundknowledge about caffeine and its clinical implications. How-ever, little is known about ob-gyn’s general knowledge aboutcaffeine, assessment of caffeine use, and recommendationsgiven to patients.

Clinical implications of caffeine

Caffeine may affect individuals seeking medical care, andwomen in particular, in a variety of clinically significant ways.Caffeine produces various CNS and peripheral nervous sys-tem effects, primarily via antagonism of A2A and A1 adenosine

1Research Department, American College of Obstetricians and Gynecologists, Washington, D.C.2Department of Psychology, American University, Washington, D.C.

JOURNAL OF WOMEN’S HEALTHVolume 18, Number 9, 2009ª Mary Ann Liebert, Inc.DOI: 10.1089=jwh.2008.1186

1457

receptors.4 It produces positive subjective effects (e.g., hap-piness, alertness) at low to moderate doses (e.g., <200 mg)and negative subjective effects (e.g., jitteriness) at higher doses(e.g., >200 mg).5 The potential for caffeine to produce in-somnia6 and anxiety7 is well documented, and both of theseproblems occur at much higher rates in women.8,9 There arealso a number of physiological effects of caffeine that maypertain to patient care, including increased blood pressure,gastric acid secretions, colonic activity, urine volume, calciumexcretion, and increased levels of adrenocorticotropic hor-mone (ACTH), insulin, and cortisol.5 Heavy caffeine use(>300 mg per day) also has been shown to be associated withshorter menses and shorter menstrual cycles.10

The average half-life of caffeine is 4–6 h; however, there isas much as a 10-fold difference in metabolism across indi-

viduals,11 with metabolism slowed by oral contraceptives12

and hastened by cigarette smoking.13 Furthermore, caffeinemetabolism is slower during the luteal phase relative tothe follicular phase of the menstrual cycle.10 Due to hor-monal influences, caffeine metabolism slows during the sec-ond and third trimesters of pregnancy,14 which can lead toelevated caffeine levels in women who maintain their usualpre-pregnancy intake of caffeine. Elevated levels of caffeine(as low as 250 mg, according to the DSM-IV-TR) can pro-duce caffeine intoxication symptoms (e.g., heart arrhythmias,nervousness) and other physical and psychological conse-quences,15 as well as allow for greater fetal exposure to caffeine.Caffeine metabolism is very slow among fetuses and neonates(i.e., 80–100 h half-life) due to immature liver systems, whichdon’t fully develop until around 6–8 months of age.16

Table 1. Caffeine Content of Common Caffeinated Products

Servingsize

(volume orweight)

Typicalcaffeinecontent

(mg)Range(mg)

Servingsize

(volume orweight)

Typicalcaffeinecontent

(mg)

Beverages FoodsCoffee Chocolate

Brewed=drip 8 oz 133 71–280 Hershey’s Chocolate Bar 1.55 oz 9Instant 8 oz 93 27–173 Hershey’s Special Dark 1.45 oz 18Espresso 1 oz 70 60–95 Miscellaneous foodDecaffeinated 8 oz 5 0–13 Dannon Coffee Yogurt 6 oz 30SBKs drip 12 oz 260 SBKs Classic Coffee Ice Cream 4 oz 30SBKs cappuccino 12 oz 75 Powerbar Tangerine Powerge 41 g 50SBKs espresso 1 oz 75 Jolt Caffeinated Gum 1 stick 33SBKs bottled frappuccino 9.5 oz 85 Stay-Alert Caffeinated Gum 1 stick 100SBKs decaffeinated 12 oz 20 Penguin Peppermints 1 mint 7

Tea Prescription medicationsBrewed 8 oz 53 40–120 Headache=migraine=painInstant 8 oz 40 13–47 Fiorinal 2 capsules 80Canned or bottled 12 oz 20 8–32 Fioricet=Esgic=many others 2 tablets 80

Soft drinks Cafergot 2 tablets 200Typical caffeinated 12 oz 40 22–69 Norgesic 2 tablets 60Mountain Dew=Diet Mt. Dew 12 oz 55 Over-the-counter medicationsDiet-coke 12 oz 47 StimulantsDr. Pepper=Diet Dr. Pepper 12 oz 41 Vivarin 1 tablet 200Pepsi-Cola 12 oz 38 No-Doz=No-Doz Max. Strength 1 tablet 100 or 200Diet Pepsi 12 oz 36 AnalgesicsCoke Classic 12 oz 35 Anacin Advanced Headache 2 tablets 130Sunkist 12 oz 41 Excedrin Extra Strength 2 tablets 130Dr. Pepper=Diet Dr. Pepper 12 oz 41 Menstrual pain relief=diureticsBarq’s Root Beer 12 oz 23 Diurex Water Pills 2 tablets 100A & W Root Beer 12 oz 0 Midol Menstrual Complete 2 caplets 1207UP=Diet 7UP 12 oz 0 Pamprin Max 2 caplets 130Sprite=Diet Sprite 12 oz 0 Dietary supplements/weight loss productsCanada Dry Ginger Ale 12 oz 0 Dexatrim Max 1 caplet 50

Energy drinks Hydroxycut Weight Loss Formula 2 caplets 200Typical amount Varies Varies 50–505 Leptopril 2 capsules 220Red Bull 8.3 oz 80 Metabolife Ultra 2 caplets 150Rockstar 16 oz 160 Metabolife Weight Management 2 tablets 101Tab 10.5 oz 95 Stacker 2 1 capsule 253

Cocoa=hot chocolate 6 oz 7 2–10 Twinlab Ripped Fuel 2 capsules 220Chocolate milk 6 oz 4 2–7 Swarm Extreme Energizer 1 capsule 300Caffeinated water 16.9 oz 60 60–200 Xenadrine Efx 2 capsules 200

Sources: Juliano LM and Griffiths RR: Caffeine. In Substance Abuse: A Comprehensive Textbook, Fourth Edition. Lowinson JH, Ruiz P, MillmanRB, Langrod JG Baltimore: Lippincott Williams & Wilkins; 2005. McCusker RR, Fuehrlein B, Goldberger BA, Gold MS, Cone EJ: (2006a)Caffeine content of decaffeinated coffee. J Anal Toxicol. 2006;30(8):611. McCusker RR, Goldberger BA, and Cone EJ: Caffeine content ofspecialty coffees. J Anal Toxicol. 2003;27(7):520. Caffeine values for all brand name products were obtained directly from product labels, or themanufacturer’s website or customer service department. SBK, Starbucks.

1458 ANDERSON ET AL.

The potential for caffeine to produce physical dependenceis well-documented even at relatively low daily doses (e.g.,100 mg per day).17 This could have important clinical impli-cations, especially for pregnant women who may abruptlystop using caffeine due to medical advice, health concerns,nausea, or requirements of medical tests (e.g., 3-h glucosechallenge test). In fact, in one study of pregnant women, 54%of those who ceased caffeine use reported withdrawal symp-toms, and 26% of those with withdrawal symptoms reportedthat these symptoms significantly interfered with their dailyfunctioning.18 Symptoms of caffeine withdrawal (e.g., head-ache, fatigue, difficulty concentrating, mood disturbances,and flu-like symptoms17) could be misattributed to othercauses, including pregnancy symptoms. Furthermore, caf-feine abstinence has been identified as a significant causeof post-operative headaches19 and has been documented inneonates after exposure to high doses of caffeine in utero.20,21

The International Classification of Diseases (ICD-10) recog-nizes a caffeine substance dependence syndrome, which com-prises a cluster of symptoms indicative of problematic use ofa drug.22 Individuals meeting criteria for substance depen-dence on caffeine, including pregnant women, have beenidentified.18,23

It is also important for health care providers to recognizethat caffeine can interact with commonly prescribed medica-tions (e.g., benzodiazepines, cimetidine)12 and exacerbatecertain medical conditions (e.g., urinary incontinence, anxi-ety).24 It is used therapeutically as a respiratory stimulant inneonates25 and as an analgesic adjuvant (e.g., Fioricet, Ex-cedrin).26,27 Interestingly, epidemiological studies have iden-tified a potential protective effect of caffeine and=or coffeeconsumption in the risk of developing Parkinson’s disease,liver disease, and Type II diabetes.28–30 There is no conclusiveevidence that caffeine negatively affects bone density inpostmenopausal women31,32 or fibrocystic breast disease,33–35

or increases breast cancer risk.36

Caffeine and pregnancy

There has been a substantial amount of research on caffeineconsumption and pregnancy outcomes. Caffeine readilycrosses the placental barrier and is distributed to all fetal tis-sues, including the CNS,14 allowing for the fetus to be exposedto caffeine at levels similar to the mother’s. A recent large scalestudy and a meta-analysis of previous studies suggest thatmaternal caffeine use increases the rate of spontaneous abor-tion in a roughly dose-dependent fashion.37,38 Associationsbetween high caffeine use and decreased fecundity and re-duced fetal growth have also been observed,39,40 including arecent study that showed that reduced fetal growth was asso-ciated with consumption of as little as one to two cups of coffeeper day.41 It has been suggested that individual differences inthe CYP1A2 gene, which is involved in caffeine metabolism,may interact with caffeine exposure in influencing the risk ofpregnancy loss and other negative pregnancy outcomes.42,43

The most consistent generality to emerge is that, when caf-feine is associated with negative pregnancy outcomes, it is typ-ically among women who consume high doses of caffeine.40

Caffeine consumption recommendations

Comprehensive scientific reviews of research on caffeineand pregnancy have concluded that reproductive aged

women should consume no more than 300 mg caffeine perday.39,40 In line with such recommendations, governmentalagencies in North America and Europe have made statementsregarding safe consumption levels during pregnancy. HealthCanada44 and the American Dietetic Association45 advisesthat pregnant women consume no more than 300 mg caffeineper day, while the Food Standards Agency of the UnitedKingdom46 recently lowered their recommended upper limitto 200 mg per day for pregnant women. Presently, neitherthe American Medical Association nor the American Collegeof Obstetricians and Gynecologists (ACOG) has put forthguidelines for caffeine consumption in pregnant women.

Physicians’ knowledge and advice about caffeine

No previous studies have examined ob-gyns’ knowledgeand assessment practices pertaining to caffeine despite thefact that most women consume caffeine, caffeine can haveimportant clinical implications for patients, and scientific re-views have concluded that reproductive-aged women shouldlimit consumption to 300 mg of caffeine per day or less.39 Infact there has been only one published report pertaining tophysician’s recommendations pertaining to caffeine, whichwas conducted over two decades ago and consisted of onequestion.47 In this survey, medical specialists from two geo-graphical regions were asked to check medical conditions forwhich they think patients should be advised to reduce oreliminate caffeine use. The most common conditions forwhich reduction or cessation of caffeine was advised werepalpitations, arrhythmias, anxiety and insomnia. Further-more among a sub-sample of 38 ob-gyns who were surveyed,79% and 68% recommended caffeine reduction or cessationfor fibrocystic disease and pregnancy, respectively.

This report describes the first comprehensive survey toexamine knowledge and beliefs about caffeine among ob-gyns as well as their assessment and advice practices per-taining to caffeine use in pregnant and non-pregnant women.

Methods

Participants

A total of 785 Fellows of the ACOG were invited to par-ticipate in a cross-sectional survey study that asked abouttheir beliefs and knowledge of caffeine’s effects and their as-sessment and advice to pregnant and non-pregnant patientsconcerning caffeine use. Participants were all members ofACOG’s Collaborative Ambulatory Research Network(CARN), a group of ACOG Fellows who agree to participatein four to six surveys every 12 months. CARN members are arepresentative sample (by age, gender, and geographic loca-tion) of the ACOG membership, of which over 90% of ob-gynsin the United States are members. Half of CARN was ran-domly selected for this survey sample. The first mailing wassent in June 2007, and second and third mailings were sent tonon-responders between July and August 2007.

Survey questions

All questions were developed for this study due to the lackof previous survey studies on the topic.

Demographics and personal caffeine consumption.Demographic questions included gender, age, practice status

CAFFEINE AND WOMEN’S HEALTH 1459

(practicing=retired), practice focus (e.g., obstetrics, gynecol-ogy, maternal fetal medicine), and geographical location ofpractice. To assess whether personal consumption was relatedto practice patterns as has been shown in previous research,48

ob-gyns were asked to report their typical consumption ofcaffeinated beverages (average weekly number of servings ofcoffee, tea, soft drinks, or other, and the typical serving size).

Caffeine knowledge. Ob-gyns were asked to estimate thenumber of milligrams of caffeine in the following caffeinatedbeverages: an 8-oz cup of coffee, a shot of espresso, a 12-oz canof Diet Coke�, and an 8-oz cup of black tea, and were askedthe number of servings of caffeinated beverages that are safefor pregnant women to consume. They were asked whethercaffeine is passed through breast milk (true=false). They wereasked whether the rate of caffeine metabolism stays the sameor changes during the menstrual cycle and whether caf-feine metabolism becomes faster, slower, or stays the samethroughout pregnancy. Ob-gyns were asked what they con-sidered to be ‘‘high levels of maternal caffeine consumption(in mg)’’ and to rate the degree to which high levels of caffeineconsumption increases the risk of various birth complicationson a scale from 0 (no increased risk) to 10 (substantial in-creased risk) and to indicate what they considered to be ‘‘highlevels of maternal caffeine consumption’’ (in mg=day). Phy-sicians gave separate ratings for each of the following preg-nancy outcomes: overall health of fetus, low birth weight,shorter gestational age, spontaneous abortion, congenitalabnormalities, nausea, stillbirths, or long-term health of thechild. Ob-gyns were asked whether they consider caffeine tobe a drug of clinical dependence and to indicate the minimumamount of caffeine consumption per day that can lead to with-drawal symptoms if someone abruptly stops using caffeine.

Caffeine assessment and advice. Ob-gyns were askedabout the following: if and when they discuss caffeine use,methods they use to assess caffeine consumption, caffeinatedproducts they routinely ask about, how frequently they ini-tiate a conversation about caffeine use, what advice theyusually give their pregnant patients, and if they ever ad-vise their pregnant patients to consume caffeine. Finally,they indicated which of their non-pregnant patients (adoles-cent patients=patients of childbearing age=perimenopausalpatients=postmenopausal patients) they ask about caffeineconsumption.

Statistical analysis

Analyses were conducted using SPSS 15.0 (SPSS Inc.,Chicago, IL) with alpha set at p< 0.05. For the analysis ofassessment practices relating to pregnancy, only physicianswho were currently practicing (n¼ 379) and treating (n¼ 332)pregnant patients were included. Due to large variability inresponse rates across questions, raw numbers are presentedwith percentages for added clarity.

Results

Demographics

A total of 386 ob-gyns returned the survey, a response rateof 49.1%. A comparison of responders and non-respondersindicated no differences in age or gender. The sample was

predominantly white (62.7%) with an average age of 48.6(SDþ 10.1) years, which is representative of the fellows ofACOG. The majority of respondents practice general obstet-rics and gynecology (78.9%). Just under half (40.5%) alsoprovide primary care to their adolescent, women of child-bearing age, or post menopausal patients.

No meaningful differences were observed based on age,gender, or other demographic variables, and thus all data isreported for the sample as a whole.

Personal use of caffeine

Of the 83.9% (n¼ 324) who provided information abouttheir caffeine consumption, 90.1% (n¼ 292) reported con-suming caffeine on a weekly basis with an estimated meanof 1188.2 mg=week (SD¼ 1046.3) ranging from 68 to6016 mg=week.

Knowledge about caffeine

When asked to estimate the amount of caffeine in milli-grams contained in common caffeinated beverages, responseswere considered correct if they were within a wide rangeof possible values for each type of beverage. As shown inTable 2, respondents largely over estimated the amount ofcaffeine in a serving of espresso with only 14.3% (18=126)providing a value in the accepted range24 (i.e., 60–95 mg).Accurate estimates for a serving of coffee (i.e., 71–280 mg),tea (i.e., 40–120 mg), and Diet Coke� (i.e., 22–69 mg) wereprovided by 66.70% (92=138), 63.90% (85=133), and 57.80%(78=135) of respondents, respectively.

When asked to report what was considered to be ‘‘highlevels of maternal caffeine consumption,’’ 15.8% (61=386)wrote in that they did not know and 52.6% (203=386) left thequestion blank. Of the 31.6% (122=386) that did provide ananswer, the mean response was 242.2 mg (SD¼ 215.7), with amedian of 200 mg. Only 13.2% (51=386) of all participants and41.8% (51=122) of those who provided an answer indicated300 mg or above.

When asked to rate the effect of high maternal caffeineconsumption on various pregnancy outcomes, low birthweight had the highest mean response (greatest increase risk;M¼ 4.5, SD¼ 2.5) and congenital abnormalities had thelowest mean response (least increased risk; M¼ 1.9, SD¼ 2.1;Fig. 1). The mean response for spontaneous abortion was3.62 (SD¼ 2.88). Low birth weight was rated significantlyhigher than spontaneous abortion (t¼ 5.28, p< 0.001).

As shown in Table 2, 96.0% (313=326) of those who pro-vided an answer correctly indicated that caffeine can be pas-sed through breast milk. However, only 24.8% (77=310)indicated that the metabolism of caffeine slows as womenprogress through pregnancy, with the majority indicating thatthe rate of metabolism becomes faster (39.7%, 123=310) orstays the same (35.5%, 110=310). Furthermore, only 46.3%(143=309) indicated that caffeine metabolism changes duringdifferent stages of the menstrual cycle.

Only 34.2% (132=386) provided an estimate for the mini-mum daily amount of caffeine consumption that will lead towithdrawal symptoms when someone abruptly stops usingcaffeine, the mean response was 197.01 (SD¼ 209.49), thoughthere was great variability in responses. Nearly half of allrespondents (49.2%, 65=132) incorrectly overestimated theamount that can cause physical dependence.

1460 ANDERSON ET AL.

Almost three-fourths of respondents (73%, 265=365) con-sider caffeine to be a drug of clinical dependence. Only 5.4%(21=386) did not respond to this question.

Assessment of caffeine use

Notably, 58.2% (188=323) reported that they always discusscaffeine consumption with their pregnant patients and 31.3%(101=323) never do. Ten and a half percent (34=323) reported

discussing caffeine under certain patient circumstances, withthe following conditions indicated by respondents: osteopo-rosis, irregular heartbeat, insomnia, diabetes, hypertension,poor weight gain, fetal arrhythmia, vascular disease, head-aches, breast pain, and palpitations.

When asked when ob-gyns discuss caffeine consumptionwith their pregnant patients, most participants indicateddoing so at patients’ first visit (64%, 218=386) or when thepregnant patient brings up the topic of caffeine consumption

Table 2. Empirical Data Compared with Ob-Gyns’ Responses to Questions

about Caffeine and Women’s Health

Empirical dataResponses of ob-gynsin the present study

What is the minimumdaily amount that willcause withdrawal?

As little as 100 mg=daycan cause withdrawal59

Mean¼ 197 (SD¼ 210)

Is caffeine a drug ofclinical dependence?

ICD-10 recognizescaffeine as a drugof dependence; theDSM-IV does not.15,22

73% said yes

Is caffeine passedthrough breast milk?

Yes, caffeine is ingestedby an infant througha mother’s breast milk.39

96.0% said yes

Does caffeine metabolismchange throughout themenstrual cycle?

Yes, metabolism slowsduring the luteal phase.60

46.3% said yes

Does caffeine metabolismchange throughoutpregnancy?

Yes, the half-life of caffeineslows from an averageof 5 h during month4 to 18 h by month 9.14

24.8% said caffeine metabolism slows39.7% said caffeine metabolism increases

35.5% said caffeine metabolism stays the same

How many mg of caffeineare in the following:

Estimate rangefor size and type

of product.2 Mean (SD)

% withinestimated

range% who

underestimated% who

overestimated

8-oz cup of coffee 71–280 mg 110.5 mg (84.3) 66.70% 27.5% 2.2%One shot of espresso 60–95 mg 142.8 mg (156.9) 14.30% 25.4% 40.5%12-oz can of Diet Coke� 22–69 mg 70.5 mg (70.7) 57.80% 8.9% 33.3%8-oz cup of tea 40–120 mg 70.0 mg (80.8) 63.90% 26.3% 9%

FIG. 1. Ob-gyns’ rating of the impact that high caffeine use has on birth complications.

CAFFEINE AND WOMEN’S HEALTH 1461

(45.9%, 177=386). Twenty-five percent (97=386) discussedcaffeine consumption when discussing other drug use (e.g.,alcohol, nicotine), and 18.4% (71=386) indicated that theydiscuss caffeine consumption with their pregnant patientswhen a ‘‘pregnant patient is complaining of what seem to becaffeine-related effects’’ with the following types of effectsreported: headache, palpitations, breast tenderness, and dif-ficulty sleeping.

When asked how often ob-gyns initiate a conversationabout caffeine consumption with their pregnant patients,24.0% (78=325) always do, 20.0% (65=325) often do, 26.2%(85=325) sometimes do, and 23.1% (75=325) rarely do. Caffeineconsumption is most often assessed via a routine questionasked by the ob-gyn or a member of his=her staff (40.4%,156=386), followed by assessment using questions on a form(20.2%, 78=386). One-third (30.8%, 119=386) indicated thatthey had no formal routine method. Most ob-gyns ask aboutcaffeine consumed from coffee (73.7%, 241=327), soft drinks(71.8%, 234=326), and tea (62.0%, 202=326) but fewer askabout energy drinks (38%, 124=326), foods (29.4%, 96=326),medications (21.2%, 69=326), and energy aids (19.3%, 63=326).However, 13.2% (43=326) indicated that they do not ask aboutspecific sources of caffeine.

The majority of respondents (64.5%, 238=369) reported thatthey do not assess caffeine use in their non-pregnant patients.Around one quarter assess caffeine use in adolescents, womenof child bearing age, perimenopausal, and postmenopausalwomen (20.3%, 75=369; 28.7%, 106=369; 26%, 96=369; and25.2%, 93=369, respectively).

Caffeine consumption advice

Most respondents reported that it is safe for pregnantwomen to consume either one or two servings of coffee orsoft drinks or tea per day. To evaluate how consistent theirrecommendations were across different products in terms ofcaffeine exposure, all responses were converted to milligramsamounts, and the differences among the various products(coffee–soft drinks, coffee–tea, tea–soft drinks) were com-puted for each respondent. We chose a very liberal cutoff of a100-mg or less difference to be considered to be a consistentrecommendation. Only 10.5% (29=277) of ob-gyns were con-sistent across coffee and tea, and 11.0% (31=282) across coffeeand soft drinks. However, 98.5% (271=275) were consistentacross tea and soft drinks.

Most ob-gyns indicated that they did not know what theyconsidered to be a ‘‘healthy daily intake’’ of caffeine for non-pregnant women (64.8%, 250=386) or left it blank (10.6%,41=386). Of those who responded (24.6%, 95=386), the meanamount reported was 199.3mg (SD¼ 153.1) and responsesranged from 0 to 1000 mg, with 87.4% (83=95) of responders or21.5% (83=386) of all participants reporting an amount 300 mgor below.

Participants were asked what advice they usually give theirpatients about caffeine, one quarter (25.4%, 82=323) of ob-gyns tell their patients that research about caffeine’s effect inpregnant women is largely inconclusive. Still, 31% (82=266)wrote in a response that indicated that they tell their pregnantpatients to limit caffeine consumption to one drink=glass=serving, 20% (53=266) recommend one to two servings perday, and 18% (48=266) recommend two servings per day.

Only 2% (5=266) wrote in trimester specific advice for caf-feine use. Only 11.3% (37=326) of respondents advise theirpatients that ‘‘it is best to stop all caffeine consumption duringpregnancy.’’

Most participants indicated if they advise their patients toconsume caffeine (84.5%, 326=386). More than half of the re-spondents (61.7%, 201=326) advise some patients to consumecaffeine during pregnancy for the following reasons: alleviateheadaches (51.8%, 169=326), relieve caffeine withdrawalsymptoms (32.8%, 107=326), improve mood (2.8%, 9=326),other (4.0%, 13=326; e.g., fatigue, constipation, low bloodpressure). Thirty-eight percent of respondents (125=326) in-dicated that they never advise their pregnant patients toconsume caffeine.

Discussion

Caffeine is a widely used drug that has many clinicallyrelevant physiological and psychological effects, some ofwhich are specifically related to women’s health and preg-nancy outcomes. This is the first study to examine ob-gyn’sknowledge and practices pertaining to patients’ use of caf-feine.

Ob-gyn’s knowledge about caffeine

It appears that many ob-gyns are not aware of the caffeinecontent of commonly used beverages. As shown in Table 2,more than a third of respondents did not provide valueswithin a wide range of possible values for coffee, tea, and softdrinks; with a tendency to overestimate the caffeine content ofsoft drinks and underestimate the caffeine content of coffeeand tea. More than 85% of respondents either significantlyoverestimated or underestimated the caffeine content ofespresso.

When ob-gyns were asked to define ‘‘high levels’’ of caf-feine consumption there was great variability in their re-sponses and 13.2% of all respondents indicated 300mg orabove. However, it is interesting to note that more than two-thirds did not answer the question or wrote that they did notknow. Together, the wide variability, inaccuracy, and lowresponse rate suggest that physicians need guidance and in-formation about acceptable levels of caffeine use for pregnantwomen. Scientific reviews have concluded that pregnantwomen should consume no more than 300 mg of caffeine perday,45,46 while some recent studies suggest that no more than200 mg may be a more appropriate cut off.

While the actual risk of caffeine causing adverse pregnancyoutcomes is a complicated and often discrepant body of lit-erature, our data suggests that perhaps ob-gyns are over-estimating the relative risk of stillbirths and underestimatingthe relative risk of spontaneous abortion. Empirical evidenceis inconclusive about caffeine’s effect on many adverse preg-nancy outcomes, however, research supports that spontane-ous abortion and low birth weight are associated with caffeineconsumption.37,38,41,49–51 For example, a recent study found aslittle as two cups of coffee per day is associated with an in-creased risk of spontaneous abortion,37 and one to two cups ofcoffee per day is associated with low birth weight.41 TheACOG has produced a Practice Bulletin on the Managementof Recurrent Early Pregnancy Loss for ob-gyns that statesthat caffeine and other drugs ‘‘may act in a dose-dependent

1462 ANDERSON ET AL.

fashion or synergistically to increase the rate of sporadicpregnancy loss,’’ but is not associated with recurrent preg-nancy loss.52

Although most respondents knew that caffeine is passedthrough breast milk, fewer than one-fourth knew that caffeinemetabolism slows during pregnancy, and nearly 40% incor-rectly answered that caffeine metabolism increases duringpregnancy. Because caffeine remains in the body longer dur-ing this period of slower metabolism, the physiological andpsychological effects of caffeine are likely to be exacerbated,particularly for women in their final months of pregnancy.

A substantial amount of research supports that caffeine is adrug of physical dependence,17 meaning that its cessation willproduce withdrawal symptoms. About three-fourths of ob-gyns indicated that they consider caffeine to be a drug ofclinical dependence. However, ‘‘clinical dependence’’ was notoperationalized, and some may have taken this to meanphysical dependence rather than a cluster of symptoms in-dicative of problematic drug use. Nevertheless, features ofclinical dependence (e.g., use despite harm, unsuccessful ef-forts to reduce consumption, tolerance) have been notedamong caffeine consumers across various studies.18,23,53 Inone recent study, a sample of pregnant women meeting cri-teria for substance dependence on caffeine was identified,18

and a caffeine dependence diagnosis predicted greater use ofcaffeine during pregnancy despite being advised by their ob-gyn to eliminate caffeine.

Ob-gyns appeared to overestimate the amount of caffeineexposure that is necessary to become physically dependent,and thus experience withdrawal symptoms upon acute ab-stinence, with the mean response being about twice as muchcaffeine as research has shown is necessary to producephysical dependence. This along with the fact that over two-thirds of the respondents failed to respond, suggests thatcaffeine withdrawal is poorly understood by ob-gyns. It isimportant that health care providers be aware that if a patientabruptly stops consuming even relatively small amounts ofcaffeine (i.e., 100 mg per day), as some pregnant women arelikely to do,18 they may experience caffeine withdrawalsymptoms such as headache, fatigue, difficulty concentrating,mood disturbances and flu-like symptoms.17 Thus, it is re-commended that when patients complain of such symptomsphysicians assess caffeine use and recent consumption pat-terns. Patients who would like to cease their caffeine useshould be advised to gradually reduce their caffeine con-sumption to reduce the severity of withdrawal symptoms.There is no empirical data available to suggest a specific time-frame for reduction, but some have suggested that reducingconsumption by 10–25% every few days or so may be effec-tive. Additional guidelines for reducing or eliminating caf-feine can be found elsewhere.5

Assessment of caffeine use

Compared with other commonly used recreational drugs,ob-gyns discuss consumption of caffeine, with their pregnantpatients less often. In our sample, 58% of ob-gyns discusscaffeine, where other studies have reported that 97% discussalcohol54 and 98% discuss tobacco.55 There was little consis-tency in the methods, frequency, and timing of the assessmentof caffeine among ob-gyns. It is possible that ob-gyns do not

believe that caffeine is an important topic to discuss and=or itcould be a result of the lack of practice guidelines pertaining tocaffeine. It is also possible that it is due to difficulty inter-preting often conflicting research findings pertaining to caf-feine associated health risks. Lastly, the amount of timenecessary to accurately assess caffeine exposure may presentanother barrier.

When assessing caffeine use, most ob-gyns reported thatthey ask about coffee, soft drinks, and tea; however, only 30%ask about caffeine intake from food and 21% ask aboutmedications. A recent study estimated that 80.6% of pregnantwomen consume caffeine from multiple sources with 36.7%obtaining caffeine from all non-coffee sources.37 It is impor-tant to note that food and medications can be a significantsource of caffeine as shown in Table 1.

Some, though a minority, of our respondents discusscaffeine use under specific patient circumstances includinginsomnia, diabetes, hypertension, poor weight gain, fetal ar-rhythmia, vascular disease, headaches, breast pain, and pal-pitations and=or assess for caffeine use when pregnant patientscomplain of what seem to be caffeine-related effects. Due tocaffeine’s well-established effects on sleep and anxiety,6,7 ob-gyns should assess caffeine consumption whenever patientscomplain of sleep disruption or anxiety-related symptoms.Women presenting with urinary incontinence may benefitfrom advice to reduce or cease caffeine use.56 Caffeine useshould also be assessed in patients presenting with hyperten-sion as it has been shown that caffeine can raise blood pressureby 5–15 Hg systolic and 5–10 Hg diastolic in healthy adults.39

Symptoms of caffeine excess=intoxication and caffeine with-drawal have symptoms that overlap with many other healthconditions, including some pregnancy symptoms and ob-gynsshould keep these in mind as potential differential diagnoses.For example, caffeine excess=intoxication should be consideredin differential diagnoses of conditions such as medication=drug induced side effects (e.g., akathisia), substance with-drawal (e.g., benzodiazepines), hyperthyroidism, anxiety,mania, insomnia, and pheochromocytoma. Caffeine with-drawal can mimic medication-induced side effects, migraineand other headache disorders, viral illness, dehydration,withdrawal from other drugs, and pregnancy symptoms.

Advice

Many ob-gyns in our study advise patients to limit caffeineconsumption to one or two servings of coffee or tea or softdrinks per day. Thus, it does not appear that physicians aretaking into account the wide differences in caffeine contentacross different products when providing advice to their pa-tients. For example, the amount of caffeine in two servings ofcoffee (roughly 266 mg) is about three times as much as theamount of caffeine in two servings of soft drinks (roughly80 mg). Knowledge of the various sources of caffeine andvariability across products should help physicians provideconsistent advice to their patients.

Thirty two percent of respondents reported that theywould never advise their pregnant patients to consume caf-feine, while 50.8% of ob-gyns reported that they sometimesadvise their patients to consume caffeine. Reasons providedincluded consumption to alleviate headache, withdrawalsymptoms, or mood. Indeed caffeine typically alleviates

CAFFEINE AND WOMEN’S HEALTH 1463

caffeine withdrawal symptoms including headache and poormood in about 30–60 min.17 There is evidence that caffeineincreases the effectiveness of analgesic medications in thetreatment of headache with more modest therapeutic effectswhen given.26

Study limitations

As this was the first study of its kind, replication andpsychometric evaluation of the questionnaire is warranted.Furthermore, the data collected is limited by self-report andpotential retrospective biases. This sample of ob-gyns re-ported using less caffeine than the typical American adultcaffeine consumer (i.e., 170 mg vs. 280 mg per day).2 It ispossible that individuals who chose to complete this surveywere more sensitive to issues pertaining to caffeine and thusmay use less caffeine than the general population. It is alsopossible that our crude measurement of caffeine consumptionunderestimated use, or respondents underreported their use.At least one study has found that self-reported caffeine con-sumption has questionable validity.57 Furthermore, some ofthe open ended questions had very low response rates.However, the low response rates appeared to be questionspecific and not a function of the location in the survey, whichmay suggest a lack of knowledge rather than lack of desire tocomplete the survey. The response rate for the survey was49.1%, which is consistent with typical response rates for theCARN group. There were no differences on any demographicvariable between those that returned the survey and theCARN group at large. Furthermore, previous analyses havebeen done to assess whether responses from CARN membersare different from responses from ACOG members in gen-eral,58 and have concluded that there are few, if any, differ-ences between samples.

Conclusions

Although it is acknowledged that additional research isneeded to clarify limits of caffeine consumption duringpregnancy, major health reviews have suggested that preg-nant women or women trying to conceive should limit theircaffeine consumption to no more than 300 mg per day. It isacknowledged that physicians have limited time with pa-tients and that other issues may take priority over routineassessment of caffeine consumption. At the very least, it isimportant for health care providers to be aware of the clini-cally relevant effects that caffeine has on physical and psy-chological health and to consider these effects when treatingpatients. Furthermore, as obstetricians and gynecologistscontinue to provide an expanding range of care to women,their knowledge and awareness of caffeine’s general effects onhealth and psychological well-being is becoming even moresignificant to their practice.

Our study confirms that physicians frequently counselpatients to reduce or eliminate caffeine use during pregnancy.Among those who reported providing recommendations,their advice is in line with current research recommendations,that pregnant women limit their caffeine consumption to�300 mg per day (i.e., many said they advise limiting use toone or two servings of coffee or tea or soft drinks per day).However, there was great variability in what was consideredto be ‘‘high levels’’ or ‘‘safe levels’’ of caffeine use for pregnant

women and the data suggests that many simply do not knowan appropriate limit of caffeine for pregnant women. Ob-gynsin our survey were unaware of the amount of caffeine invarious products, and they tended to limit their assessment ofcaffeine exposure to caffeinated beverages. They were alsolargely unaware that caffeine metabolism significantly slowsas pregnancy progresses as well as changes across the men-strual cycle, or that as little as 100 mg per day of caffeine canresult in withdrawal symptoms upon acute abstinence.

Future research should be aimed at education and assess-ment of physicians’ knowledge about caffeine and its clinicalimplications for patients. We hope that this article serves toincrease knowledge of caffeine and its clinically relevantpharmacological effects and recommend the development ofpractice guidelines for ob-gyns and other health care pro-viders.

Acknowledgments

This study was supported by the Maternal and ChildHealth Bureau, Health Resources and Services Administra-tion, Department of Health and Human Services (grant R60-MC-05674).

Disclosure Statement

The authors have no conflicts of interest to report.

References

1. Frary CD, Johnson RK, Wang MQ. Food sources and intakesof caffeine in the diets of persons in the United States. J AmDiet Assoc 2005;105:110–113.

2. Barone JJ, Roberts HR. Caffeine consumption. Food ChemToxicol 1996;34:119–129.

3. Jacoby I, Meyer GS, Haffner W, et al. Modeling the futureworkforce of obstetrics and gynecology. Obstet Gynecol1998;92:450–456.

4. Ferre S. An update on the mechanisms of the psychosti-mulant effects of caffeine. J Neurochem 17 Jan 2008.

5. Juliano LM, Ferre S, Griffiths RR. Caffeine: pharmacologyand Clinical Effects. In: Graham AW, Schultz TK, Mayo-Smith M, et al., eds. Principles of Addiction Medicine 4th ed.Chevy Chase, MD: American Society of Addiciton Medicine;2009 (in press).

6. Snel J. Coffee and caffeine: sleep and wakefulness. In: Gar-attini S, ed. Caffeine, Coffee, and Health. New York: RavenPress; 1993:255–290.

7. Greden JF. Anxiety or caffeinism: a diagnostic dilemma. AmJ Psychiatry 1974;131:1089–1092.

8. Krystal AD. Insomnia in women. Clin Cornerstone 2003;5:41–50.

9. Pigott TA. Gender differences in the epidemiology andtreatment of anxiety disorders. J Clin Psychiatry 1999;60(Suppl 18):4–15.

10. Vo HT, Smith BD, Elmi S. Menstrual endocrinology andpathology: caffeine, physiology, and PMS. In: Smith BD,Gupta U, Gupta BS, eds. Caffeine Activation Theory: Effectson Health and Behavior. Boca Raton, FL: Taylor & FrancisGroup; 2007:181–197.

11. Denaro CP, Benowitz NL. Caffeine metabolism: dispositionin liver disease and hepatic-function testing. In: Watson RR,ed. Drug and Alcohol Abuse Reviews, Vol. 2: Liver Pathol-

1464 ANDERSON ET AL.

ogy and Alcohol. Totowa, NJ: The Human Press; 1991:513–539.

12. Carrillo JA, Benitez J. Clinically significant pharmacokineticinteractions between dietary caffeine and medications. ClinPharmacokinet 2000;39:127–153.

13. Parsons WD, Neims AH. Effect of smoking on caffeineclearance. Clin Pharmacol Ther 1978;24:40–45.

14. Aldridge A, Bailey J, Neims AH. The disposition of caffeineduring and after pregnancy. Semin Perinatol 1981;5:310–314.

15. APA. Diagnostic and Statistical Manual of Mental Disorders,4th ed [text revision]. Washington, DC: American Psychia-tric Press, 2000.

16. Parsons WD, Neims AH. Prolonged half-life of caffeine inhealthy term newborn infants. J Pediatr 1981;98:640–641.

17. Juliano LM, Griffiths RR. A critical review of caffeinewithdrawal: empirical validation of symptoms and signs,incidence, severity, and associated features. Psychophar-macology (Berl) 2004;176:1–29.

18. Svikis DS, Berger N, Haug NA, et al. Caffeine dependence incombination with a family history of alcoholism as a pre-dictor of continued use of caffeine during pregnancy. Am JPsychiatry 2005;162:2344–2351.

19. Fennelly M, Galletly DC, Purdie GI. Is caffeine withdrawalthe mechanism of postoperative headache? Anesth Analg1991;72:449–453.

20. McGowan JD, Altman RE, Kanto WP, Jr. Neonatal with-drawal symptoms after chronic maternal ingestion of caf-feine. South Med J 1988;81:1092–1094.

21. Martin I, Lopez-Vilchez MA, Mur A, et al. Neonatal with-drawal syndrome after chronic maternal drinking of mate.Ther Drug Monit 2007;29:127–129.

22. World Health Organization. The ICD-10 Classification ofMental and Behavioural Disorders: Clinical Descriptionsand Diagnostic Guidelines. Geneva, Switzerland: WHO,1992.

23. Strain EC, Mumford G, Silverman K, et al. Caffeine depen-dence syndrome: evidence from case histories and experi-mental evaluations. JAMA 1994;272:1043–1048.

24. Juliano LM, Griffiths RR. Caffeine. In: Lowinson JH, Ruiz P,Millman RB, et al., eds. Substance Abuse: A ComprehensiveTextbook, 4th ed. Baltimore, MD: Lippincott, Williams, &Wilkins; 2005:403–421.

25. Schmidt B, Roberts RS, Davis P, et al. Long-term effects ofcaffeine therapy for apnea of prematurity. N Engl J Med2007;357:1893–1902.

26. Shapiro RE. Caffeine and headaches. Neurol Sci 2007;28(Suppl 2):S179–S183.

27. Sawynok J, Yaksh TL. Caffeine as an analgesic adjuvant: areview of pharmacology and mechanisms of action. Phar-macol Rev 1993;45:43–85.

28. Ross GW, Abbott RD, Petrovitch H, et al. Association ofcoffee and caffeine intake with the risk of Parkinson disease.JAMA 2000;283:2674–2679.

29. Ruhl CE, Everhart JE. Coffee and tea consumption are as-sociated with a lower incidence of chronic liver disease inthe United States. Gastroenterology 2005;129:1928–1936.

30. van Dam RM, Hu FB. Coffee consumption and risk of type 2diabetes: a systematic review. JAMA 2005;294:97–104.

31. Lloyd T, Rollings N, Eggli DF, et al. Dietary caffeine intakeand bone status of postmenopausal women. Am J Clin Nutr1997;65:1826–1830.

32. Rapuri PB, Gallagher JC, Kinyamu HK, et al. Caffeine intakeincreases the rate of bone loss in elderly women and inter-

acts with vitamin D receptor genotypes. Am J Clin Nutr2001;74:694–700.

33. Heyden S. Coffee and fibrocystic breast disease. Surgery1980;88:741–742.

34. Heyden S, Fodor JG. Coffee consumption and fibrocysticbreasts: an unlikely association. Can J Surg 1986;29:208–211.

35. Heyden S, Muhlbaier LH. Prospective study of ‘‘fibrocysticbreast disease’’ and caffeine consumption. Surgery 1984;96:479–484.

36. Ishitani K, Lin J, Manson JE, et al. Caffeine consumption andthe risk of breast cancer in a large prospective cohort ofwomen. Arch Intern Med 2008;168:2022–2031.

37. Weng X, Odouli R, Li DK. Maternal caffeine consump-tion during pregnancy and the risk of miscarriage: a pro-spective cohort study. Am J Obstet Gynecol 2008;198:e271–e278.

38. Fernandes O, Sabharwal M, Smiley T, et al. Moderate toheavy caffeine consumption during pregnancy and rela-tionship to spontaneous abortion and abnormal fetalgrowth: a meta-analysis. Reprod Toxicol 1998;12:435–444.

39. Nawrot P, Jordan S, Eastwood J, et al. Effects of caffeine onhuman health. Food Addit Contam 2003;20:1–30.

40. Higdon JV, Frei B. Coffee and health: a review of recenthuman research. Crit Rev Food Sci Nutr 2006;46:101–123.

41. The Care Study Group. Maternal caffeine intake duringpregnancy and risk of fetal growth restriction: a large pro-spective observational study. BMJ 2008;337:a2332.

42. Grosso LM, Bracken MB. Caffeine metabolism, genetics, andperinatal outcomes: a review of exposure assessment con-siderations during pregnancy. Ann Epidemiol 2005;15:460–466.

43. Sata F, Yamada H, Suzuki K, et al. Caffeine intake, CYP1A2polymorphism and the risk of recurrent pregnancy loss. MolHum Reprod 2005;11:357–360.

44. Health Canada. Caffeine and your health. Available at:www.hc-sc.gc.ca=fn-an=securit=facts-faits=caffeine-eng.php.Accessed August 30, 2008.

45. Kaiser L, Allen LH. Position of the American Dietetic As-sociation: Nutrition and Lifestyle for a Pregnancy Outcome.J Am Diet Assoc 2008;108:553–561.

46. Food Standards Agency. COT statement on the reproduc-tive effects of caffeine. Available at: http:==cot.food.gov.uk=cotstatements=cotstatementsyrs=cotstatements2001=caffeine.Accessed August 30, 2008.

47. Hughes JR, Amori G, Hatsukami DK. A survey of physicianadvice about caffeine. J Subst Abuse 1988;1:67–70.

48. Aalto M, Hyvonen S, Seppa K. Do primary care physicians’own AUDIT scores predict their use of brief alcohol inter-vention? A cross-sectional survey. Drug Alcohol Depend2006;83:169–173.

49. Bech BH, Nohr EA, Vaeth M, et al. Coffee and fetal death: acohort study with prospective data. Am J Epidemiol 2005;162:983–990.

50. Tolstrup JS, Kjaer SK, Munk C, et al. Does caffeine and al-cohol intake before pregnancy predict the occurrence ofspontaneous abortion? Hum Reprod 2003;18:2704–2710.

51. Rasch V. Cigarette, alcohol, and caffeine consumption: riskfactors for spontaneous abortion. Acta Obstet Gynecol Scand2003;82:182–188.

52. Management of Recurrent Early Pregnancy Loss. ACOGPractice Bulletin. Washington, DC: American College ofObstetricians and Gynecologists, 2001.

CAFFEINE AND WOMEN’S HEALTH 1465

53. Bernstein GA, Carroll ME, Thuras PD, et al. Caffeinedependence in teenagers. Drug Alcohol Depend 2002;66:1–6.

54. Diekman ST, Floyd RL, Decoufle P, et al. A survey ofobstetrician-gynecologists on their patients’ alcohol useduring pregnancy. Obstet Gynecol 2000;95:756–763.

55. Price JH, Jordan TR, Dake JA. Obstetricians and gynecolo-gists’ perceptions and use of nicotine replacement therapy.J Commun Health 2006;31:160–175.

56. Bryant CM, Dowell CJ, Fairbrother G. Caffeine reductioneducation to improve urinary symptoms. Br J Nurs 2002;11:560–565.

57. Bracken MB, Triche E, Grosso L, et al. Heterogeneity in as-sessing self-reports of caffeine exposure: implications forstudies of health effects. Epidemiology 2002;13:165–171.

58. Hill LD, Erickson K, Holzman GB, et al. Practice trends inoutpatient obstetrics and gynecology: findings of the Colla-borative Ambulatory Research Network, 1995–2000. ObstetGynecol Surv 2001;56:505–516.

59. Evans SM, Griffiths RR. Caffeine withdrawal: a parametricanalysis of caffeine dosing conditions. J Pharmacol Exp Ther1999;289:285–294.

60. Lane JD, Steege JF, Rupp SL, et al. Menstrual cycle effects oncaffeine elimination in the human female. Eur J Clin Phar-macol 1992;43:543–546.

Address correspondence to:Britta L. Anderson, B.A.

Research DepartmentAmerican College of Obstetricians and Gynecologists

409 12th Street, SWWashington, DC 20024

E-mail: banderson@acog.org

or

Laura M. Juliano, Ph.D.Associate Professor of Psychology

American University4400 Massachusetts Avenue NW

Washington, DC 20016

E-mail: juliano@american.edu

1466 ANDERSON ET AL.