Recent research suggests that chocolate, particularly dark chocolate, may benefit cardiovascular health. Chocolate contains over 600 chemicals including flavanoids, magnesium, and theobromine. Flavanoids (including flavanols, flavones, flavanones, and others) are potent antioxidants capable of inducing nitric oxide-dependent vasodilation, as well as having antiplatelet and anti-inflammatory effects.1,2 Magnesium deficits have been linked to hypertension, and other cardiovascular disease.3,4 The methylxanthine theobromine is present in very high quantities, with dark chocolate containing the most.5 The primary pharmacologic effects of theobromine include diuresis, myocardial stimulation, vasodilation, and smooth muscle relaxation,6 and it has been used to treat hypertension, angina, and atherosclerosis.7 Theobromine is widely consumed in the form of chocolate and cocoa products, and although theobromine is one of the 3 primary metabolites of caffeine, it accounts for only about 12% of total metabolized caffeine, compared with 70% to 80% for paraxanthine.6 Thus, theobromine is a useful, specific biomarker for chocolate consumption. In addition, theobromine, along with the other methylxanthines, freely crosses the placental barrier in pregnancy.
Preeclampsia is a serious maternal complication of pregnancy that affects 3% to 8% of pregnancies.8 Preeclampsia shares many characteristics and risk factors of cardiovascular disease, including endothelial dysfunction, oxidative stress, hypertension, insulin resistance, and hypertriglyceridemia.9 Cardiovascular manifestations of preeclampsia include changes in vascular reactivity, hypertriglyceridemia, endothelial dysfunction, and hypertension.8,10,11 Women with preeclampsia may also be at increased risk of cardiovascular disease and metabolic disturbances in the years following pregnancy.12–16
We investigate whether chocolate consumption, measured by self-reported maternal intake and fetal cord serum concentrations of theobromine, is associated with preeclampsia.
Pregnant women were recruited September 1996 to January 2000 from 56 obstetric practices and 15 clinics associated with 6 hospitals in Connecticut and Massachusetts.17 Women were excluded if they were more than 24 weeks’ gestational age at enrollment, had insulin-dependent diabetes mellitus, did not speak English or Spanish, or intended to terminate their pregnancy.
Of 11,267 women screened for study, 9576 met eligibility criteria. To ensure an adequate number of higher caffeine consumers for a larger study, all eligible women who reported drinking ≥150 mg of caffeine per day in the prior week were invited to participate (n = 715; 11% of final sample). The remaining population consisted of random samples of women who drank <150 mg caffeine per day (n = 839; 45% of final sample) and nonconsumers (n = 2077; 44% of final sample). A total of 3631 women were invited; 2478 (68%) enrolled, 639 (18%) declined, 424 (12%) were lost to follow-up, 72 (2%) miscarried prior to enrollment, and 20 (<1%) were not eligible at enrollment interview. Among the 2478 enrolled women, 2291 (92%) delivered a singleton infant.
Cord blood biomarker data were available for 1611 infants. A total of 1995 women provided data on both first-trimester and third-trimester chocolate consumption. Preeclampsia status was determined for 1943 women; the remaining 348 were excluded because they had preexisting hypertension, indication of gestational hypertension but no proteinuria, or incomplete information to definitively classify preeclampsia. After these exclusions, the biomarker exposure analyses included 1346 women; analyses of reported chocolate consumption included 1681 women.
Most women were interviewed at home by 14 weeks (mean = 14.9 weeks, interquartile range = 12–17 weeks, min = 6.1 and max = 24.3 weeks). The structured interview collected detailed information on dietary intake of caffeinated beverages and chocolate products since becoming pregnant. Mothers reported on potential confounders including race/ethnicity, education, smoking, age, prepregnancy weight, height, and prior pregnancy history. Respondents were reinterviewed postnatally, usually during the delivery hospitalization, to obtain third-trimester exposure information.17
Reported Chocolate Consumption
Women were asked if they drank hot chocolate, cocoa, or chocolate milk since becoming pregnant, and how many cups they had on a daily or weekly basis; if they ate milk or dark chocolate candy, cake, cookies, or ice cream, and how many servings of milk chocolate and dark chocolate they had on a daily or weekly basis. From this information, we calculated variables for the reported number of chocolate servings per week (<1, 1–4, and ≥5) for first and third trimesters.
Cord Serum Theobromine Concentration
At delivery, obstetricians cut the umbilical cord and collected venous and arterial cord blood, which was immediately refrigerated. The hospital laboratory separated and froze the serum within 24 hours of collection. Frozen samples were transported to the study laboratory on ice and stored at −80°C. Chemists at the Clinical Pharmacology Laboratory at the University of California, San Francisco, who were blind to exposure and pregnancy information, analyzed samples for theobromine (the major metabolite of chocolate), caffeine, paraxanthine (the major metabolite of caffeine), and theophylline.18 Concentrations of these methylxanthines were determined using liquid chromatography coupled with tandem mass spectrometry. Stable isotope-labeled analogs were used as internal standards. The limit of quantitation was 10 ng/mL. The precision of the assay (within-run coefficient of variation) ranged from 1.7% to 10.3%, and accuracy (percent of expected values) ranged from 88% to 118% for plasma concentrations from 10 to 5000 ng/mL, respectively. All assays below the detection limit (n = 63, 5% of all samples) were assigned a value of zero. Such assays are not considered biologically significant, and assigning an alternative value (eg, halfway between 0 and the detection limit) would not affect results since methylxanthine concentrations were evaluated in quartiles.
Obstetric records were abstracted to identify pregnancy outcomes. Abstractors were blind to exposure status. Maternal blood pressure throughout pregnancy, International Classification of Diseases (9th Revision) diagnoses, and maternal and infant medical conditions were recorded on structured abstraction forms. Blood pressure and urinary protein values from ante-, intra-, and postpartum periods were recorded if the abstractor noted 2 or more blood pressure readings ≥140 mm Hg systolic or ≥90 mm Hg diastolic during the delivery hospitalization, or an ICD-9 diagnosis indicating pregnancy-induced hypertension, preeclampsia, or HELLP syndrome for that subject.
Preeclampsia was defined according to National Heart, Lung and Blood Institute (NHLBI) guidelines.19 The criteria required (1) de novo hypertension (≥140 mm Hg systolic or ≥90 mm Hg diastolic) on 2 or more occasions at least 6 hours apart beginning after the 20th week of gestation; (2) accompanying proteinuria, defined as urinary protein concentrations of 30 mg/dL or greater, equivalent to dipstick value of 1+ from 2 or more specimens collected at least 4 hours apart, or one or more urinary dipstick values of 2+ near the end of pregnancy, or one or more catheterized dipstick values of 1+ during delivery hospitalization, or 24-hour urine collection with protein of ≥300 mg. We excluded women for whom pre-existing hypertension could not be ruled out (eg, no readings available prior to 20 weeks’ gestation; physician notes indicating chronic hypertension in the patient) or who met partial criteria for preeclampsia (eg, pregnancy-induced hypertension; proteinuria with no hypertension).
Because of the high correlation between concentrations of theophylline and caffeine (r = 0.96; theophylline is a minor metabolite of caffeine), all reported analyses included caffeine rather than theophylline. Analyses replacing caffeine with theophylline did not materially change any findings.
Separate regression models were run for reported chocolate consumption and cord blood theobromine concentrations. We calculated unadjusted and adjusted odds ratios logistic regression using PC-Statistical Analysis System v. 9.1 (SAS Institute, Inc., Cary, NC). Adjusted models controlled for race/ethnicity, age, education, parity, maternal smoking, prepregnancy body mass index (BMI), and prenatal care provider (private/clinic). Models of the association between theobromine and preeclampsia also adjusted for cord blood caffeine and paraxanthine concentrations.
Table 1 describes the study population's characteristics and the distribution of exposure measures. Reported chocolate consumption was high, particularly in the third trimester. Consumption was higher among younger women, less well educated women, Hispanic women, women who smoked in pregnancy, and women receiving prenatal care in clinics. Obese women were less likely to report chocolate consumption than normal or overweight women in the third trimester, but not the first trimester. Cord theobromine levels were similarly higher with younger age, less education, and clinic prenatal care provider. In addition, white and parous women had higher levels of theobromine. BMI was not associated with theobromine levels.
Reported chocolate consumption was only modestly correlated with cord blood theobromine levels (quartiles): first trimester, rspearman = 0.15 and third trimester, rspearman = 0.29. Median theobromine concentration in women who reported consuming less than 1 serving weekly in the third trimester was low (211–237 ng/mL), regardless of how much chocolate was consumed in the first trimester. Theobromine concentrations increased with increasing third trimester consumption, with the highest median levels among women consuming 5 or more servings in both first and third trimesters (674 ng/mL).
Table 2 shows unadjusted associations between potential confounders and preeclampsia, which developed in 3.7% of 1681 women (NHLBI criteria n = 63). Higher BMI, education, and nulliparity were most strongly associated with increased risk of preeclampsia.
In unadjusted logistic regression models, reported chocolate consumption in the third trimester and cord serum theobromine concentrations were inversely (and significantly) associated with risk of preeclampsia. Point estimates for reported chocolate consumption in the first trimester were protective although with wide confidence intervals (Table 3).
In adjusted analyses, serum theobromine remained inversely associated with risk of preeclampsia (P for trend = 0.008). Point estimates were strikingly similar to those in the unadjusted analyses. Women with cord serum theobromine in the highest quartile had a 69% reduction (95% confidence interval [CI] = 0.11–0.87) in risk compared with women whose concentrations were in the lowest quartile. In adjusted analyses of reported chocolate consumption in the third trimester, estimates remained protective (adjusted odds ratio 0.60 [95% CI= 0.30–1.24] for women consuming 5+ versus <1 weekly serving of chocolate). Adjusted estimates of consumption in the first trimester were less strongly associated with risk of preeclampsia (0.81 [0.37–1.79] for women consuming 5+ versus <1 weekly serving).
In this prospective cohort of pregnant women, we observed that chocolate consumption, as measured by cord serum levels of the biomarker theobromine, was associated with lower risk of preeclampsia. As measured by self-reported maternal intake, increased chocolate consumption in both first and third trimesters was suggestive of reduced preeclampsia risk. Our findings are consistent with other studies that have investigated vascular and metabolic effects of chocolate. Grassi et al20 found that consumption of dark (vs. white) chocolate reduced blood pressure and insulin resistance, and improved nitric oxide-dependent vasorelaxation in men and women with untreated essential hypertension. In healthy men and women dark chocolate consumption lowered blood pressure and insulin sensitivity.21 Fisher and Hollenberg22 reported that consumption of flavanol-rich cocoa improved measures of endothelial function. A recent meta-analysis23 of 5 trials showed significant and clinically important drops in systolic and diastolic blood pressure after cocoa administration.
A major strength of this study is use of umbilical cord blood theobromine as a biomarker for cocoa and chocolate consumption. Flavanoids and magnesium are found in numerous other substances, but theobromine is primarily found in cocoa and tea leaves. Quantifying self-reported chocolate and cocoa consumption is extremely difficult due to considerable variation in the cocoa content of chocolate products. In addition, it is difficult to standardize self-reported chocolate consumption for serving size, or in any other way. Theobromine concentrations in chocolate also vary widely from 0.15% to 0.46%.6 Such sources of misclassification most likely drive effect estimates toward the null. These measurement issues may account for some of the differences in the magnitude of effects between reported consumption and cord serum theobromine. Umbilical cord blood levels of theobromine provide an objective indicator of recent maternal cocoa and chocolate intake since theobromine is rapidly absorbed from the gastrointestinal tract24 and freely crosses the placental barrier25 and are not hampered by possible recall bias of self-reported measurements.
One limitation of our study is the possibility of reverse causality. If women diagnosed with preeclampsia reduced their calorie intake (including chocolate) subsequent to their diagnosis, and if the reported third trimester consumption or cord theobromine concentration represented exposure after the time of diagnosis, reverse causality could explain some of our findings. (Reverse causality could not explain the first trimester findings.) We conducted several analyses to help elucidate the possible role of reverse causality in our data.
Examination of correlations between reported consumption in the first and third trimesters by preeclampsia status (rspearman = 0.34 for women who developed preeclampsia; rspearman = 0.35 for women who did not), suggested that women did not change consumption differentially based on preeclampsia diagnosis. Similarly, women diagnosed with preeclampsia were no more likely to change consumption than unaffected women. Restricting adjusted analyses to the 785 women whose category of chocolate consumption did not change from first to third trimester of pregnancy (Table 3, last column) produced estimates of associations of cord theobromine levels strikingly similar to the adjusted estimates in all women. Considering the possibility that women with preeclampsia consumed less chocolate because they were admitted to hospitals earlier than healthier women, we also analyzed times from hospitalization to delivery. Such times were essentially identical in mothers who had and had not developed preeclampsia (96% and 97% of women with or without preeclampsia, respectively, were admitted on the same or previous day as date of delivery). These analyses failed to support a role of reverse causation, although they cannot rule out this possibility.
Another potential limitation of our study is residual confounding by smoking or BMI. To address such confounding, we repeated analyses (1) restricting the sample to nonsmoking women and (2) excluding obese women (but still controlling for BMI). In both analyses, we found no change in results. Results were similar when we further restricted the sample to women with normal BMI only. Finally, the small number of women with preeclampsia and the potential misclassification of exposure may have reduced the precision of these estimates.
Our findings of an inverse relationship between cord serum theobromine concentrations and risk of preeclampsia may be due to a direct role of theobromine. During pregnancy, theobromine (or the other methylxanthines in chocolate) may improve placental circulation and inhibit xanthine oxidase, which, in the setting of hypoxia, increases production of reactive oxygen species and free radicals.26 Alternatively, theobromine concentrations could play an indirect role by (1) acting as a proxy for others chemicals (such as flavanols or magnesium) found in cocoa, (2) their correlation with other unmeasured dietary factors that influence risk of preeclampsia or (3) acting as a proxy for maternal metabolism of theobromine whereby enzymatic activity associated with metabolism, rather than actual theobromine concentrations, is responsible for influencing the risk of maternal outcomes.27
We repeated analyses (not shown) using a physician diagnosis of preeclampsia in the medical chart instead of our own designation based strictly on NHLBI preeclampsia criteria. Such analyses (n = 1907) consistently suggested an inverse relationship between all measures of chocolate consumption and preeclampsia risk. Interestingly, all the point estimates were practically unchanged, except that adjusted estimates of reported first trimester consumption were more strongly inversely associated with risk of preeclampsia (adjusted odds ratio 0.37 [95% CI= 0.13–1.08] for women consuming 5+ versus <1 weekly serving).
Our results raise the possibility that chocolate consumption by pregnant women may reduce the occurrence of preeclampsia. Because of the importance of preeclampsia as a major complication of pregnancy, replication of these results in other large prospective studies with a detailed assessment of chocolate consumption is warranted. Measurements of chocolate exposure should be designed to permit careful examination of the temporal relationship between chocolate consumption in pregnancy and subsequent risk of preeclampsia.
We thank Peyton Jacob III and Lisa Yu for developing the methylxanthine assay and Masae Ahmann for conducting the chemical analyses. We also thank the following for their assistance with data collection. Baystate Health System (MA): R. Burkman, K. Troczynski, P. O'Grady; Bridgeport Hospital (CT): E. Luchansky, I. San Pietro, J. Collins, R. Torres, C. Presnick; Danbury Hospital (CT): L. Silberman; Hartford Hospital (CT): S. Curry, C. Mellon; Hospital of St. Raphael (CT): W. Reguero, B. McDowell; Yale-New Haven Hospital (CT): J. Coppel, A. Somsel, and S. Updegrove.
1.Fisher ND, Hughes M, Gerhard-Herman M, et al. Flavanol-rich cocoa induces nitric-oxide-dependent vasodilation in healthy humans. J Hypertens
2.Heiss C, Kleinbongard P, Dejam A, et al. Acute consumption of flavanol-rich cocoa and the reversal of endothelial dysfunction in smokers. J Am Coll Cardiol
3.Song Y, Ridker PM, Manson JE, et al. Magnesium intake, C-reactive protein, and the prevalence of metabolic syndrome in middle-aged and older US women. Diabet Care
4.Weglicki W, Quamme G, Tucker K, et al. Potassium, magnesium, and electrolyte imbalance and complications in disease management. Clin Exp Hypertens
5.Smit HJ, Gaffan EA, Rogers PJ. Methylxanthines are the psycho-pharmacologically active constituents of chocolate. Psychopharmacology (Berl)
6.Stavric B. Methylxanthines: Toxicity to humans. 3. Theobromine, paraxanthine and the combined effects of methylxanthines. Chem Toxic
7.Kelly CJ, Effects of theobromine should be considered in future studies. Am J Clin Nutr
. 2005;82:486–7; [author reply 487–488].
8.ACOG Practice Bulletin No. 33: Diagnosis and management of preeclampsia and eclampsia. Obstet Gynecol
9.Roberts JM, Gammill HS. Preeclampsia: recent insights. Hypertension
10.Cunningham FG, Gant NF, Leveno KJ, et al. Hypertensive disorders in pregnancy. In: Williams Obststrics
, 21st ed. New York: McGraw-Hill; 2001, pp. 567–618.
11.Hauth JC, Ewell MG, Levine RJ, et al. Pregnancy outcomes in healthy nulliparas who developed hypertension. Calcium for Preeclampsia Prevention Study Group. Obstet Gynecol
12.Brown DW, Dueker N, Jamieson DJ, et al. Preeclampsia and the risk of ischemic stroke among young women: results from the stroke prevention in young women study. Stroke
13.Forest J-C, Girouard J, Masse J, et al. Early occurrence of metabolic syndrome after hypertension in pregnancy. Obstet Gynecol
14.Pouta A, Hartikainen A-L, Sovio U, et al. Manifestations of metabolic syndrome after hypertensive pregnancy. Hypertension
15.Seely EW, Solomon CG, Insulin resistance and its potential role in pregnancy-induced hypertension. J Clin Endocrinol Metab
16.Wilson BJ, Watson MS, Prescott GJ, et al. Hypertensive diseases of pregnancy and risk of hypertension and stroke in later life: results from cohort study. BMJ
. 2003;326:7394–7845. Doi: 10.1136/bmj.326.
17.Bracken MB, Triche EW, Belanger K, et al. Association of maternal caffeine consumption with decrements in fetal growth. Am J Epidemiol
18.Grosso LM, Triche EW, Belanger K, et al. Caffeine metabolites in umbilical cord blood, Cytochrome P-450 1A2 activity, and intrauterine growth restriction. Am J Epidemiol
19.Gifford RW, August PA, Cunningham G, et al. Report on the National High Bood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am J Obstet Gynecol
20.Grassi D, Necozione S, Lippi C, et al. Cocoa reduces blood pressure and insulin resistance and improves endothelium-dependent vasodilation in hypertensives. Hypertension
21.Grassi D, Lippi C, Necozione S, et al. Short-term administration of dark chocolate is followed by a significant increase in insulin sensitivity and a decrease in blood pressure in healthy persons. Am J Clin Nutr
22.Fisher ND, Hollenberg NK, Aging and vascular responses to flavanol-rich cocoa. J Hypertens
23.Taubert D, Roesen R, Schomig E. Effect of cocoa and tea intake on blood pressure: a meta-analysis. Arch Intern Med
24.Rodopoulos N, Hojvall L, Norman A. Elimination of theobromine metabolites in healthy adults. Scand J Clin Lab Invest
25.Grosso LM, Bracken MB, Caffeine metabolism, genetics, and perinatal outcomes: A review of exposure assessment considerations during pregnancy. Ann Epidemiol
26.Many A, Westerhausen-Larson A, Kanbour-Shakir A, et al. Xanthine oxidase/dehydrogenase is present in human placenta. Placenta
27.Bracken MB. Cotinine and spontaneous abortion: might variations in metabolism play a role? Epidemiology