CAFFEINE AND CARDIOVASCULAR DISEASE RISK

For more than 30 years caffeine has been of interest in the etiology of heart disease particularly because it may be associated with alterations in blood lipids and blood pressure, arrhythmias, and other adverse cardiac functions. While earlier studies suggested an effect of caffeine on blood lipids, Sedor et al. (1991) found no influence of coffee on serum lipoproteins in women with normal cholesterol levels. These different results may be accounted for by the finding that the method of preparing coffee could influence the relationship between caffeine and blood lipids. Only one fraction of boiled coffee was found to significantly increase blood cholesterol and low-density cholesterol in a dose-dependent manner (Pirich et al., 1993). In 1,074 adults studied in the United Kingdom, coffee consumption was not found to have a significant effect on total or high-density lipoprotein cholesterol (Lancaster et al., 1994). Similarly, Lewis et al. (1993) found no consistent associations between caffeine-containing beverages and serum lipoproteins in 5,115 healthy black and white, men and women aged 18–30 years. In contrast, in a 17-month follow-up of 2,109 healthy nonsmokers, Wei et al. (1995) found that total serum cholesterol increased by about 2 mg/dL for each subsequent increase in cups of regular coffee per day. Furthermore, a dose-response in serum lipoproteins was found among those who increased consumption, continued the same dose, or decreased consumption of regular coffee. This association was not observed with the consumption of decaffeinated coffee, regular or decaffeinated tea, or caffeine-containing colas. However, in a double-blind, randomized trial of 69 young healthy subjects whose habitual coffee consumption was 5.9 cups (140 mL) of filtered regular coffee per day, abstinence from caffeine resulted in no effect on serum lipids (Bak and Grobbee, 1991). Urgert and Katan (1997), in an extensive review, found effects of coffee brewing techniques on serum concentrations of total and low-density lipoprotein cholesterol. The compounds responsible for this effect are the diterpene lipids cafestol and kahweol, which make up about 1 percent (wt:wt) of coffee beans. These diterpenes are extracted by hot water but are retained by paper filters, thus explaining why filtered coffee has no effect on serum cholesterol, while boiled coffees such as Scandinavian cafetiere and Turkish coffees do.

Several approaches have been utilized to investigate the possible relationship between caffeine intake and blood pressure. Results summarized in recent reviews by Myers (in press) and Green and Suls (1996) suggested that caffeine-naive individuals may experience a small increase in blood pressure after acute dosing with caffeine. During chronic administration of caffeine, tolerance appears to develop, and chronic long-lasting changes in blood pressure are usually not seen in individuals who routinely consume caffeine. Coffee consumption was shown to have no significant effects on blood pressure in the 1,074 adults studied by Lancaster et al. (1994). Similarly, in the Coronary Artery Risk Development in Young Adults study of 5,115 black and white men and women aged 18–30 years, no consistent association was found between consumption of caffeine-containing beverages and blood pressure (Lewis et al., 1993). The 6-year data from the Multiple Risk Factor Intervention Trial showed a significant, independent, inverse relation between caffeine intake and both systolic and diastolic blood pressure (Stamler et al., 1997). In contrast, a recent report of a meta-analysis of 11 controlled trials showed an independent, positive relationship between cups of coffee consumed (median dose=5 cups/day) and subsequent change in systolic blood pressure (Jee et al., 1999). Another recent review critically examined 30 years of controlled clinical and epidemiological studies on the blood pressure effects of coffee and caffeine (Nurminen et al., 1999). The authors concluded that the acute pressor effects of caffeine are well documented, but that at present there is no clear epidemiological evidence that caffeine consumption is causally related to hypertension. They also concluded, however, that high caffeine intake may be an additional risk factor for hypertension at the individual level due to long-lasting stress or a genetic susceptibility to hypertension.

In general, controlled clinical attempts to demonstrate the effects of caffeine on increasing heart rate or inducing arrhythmias have been unsuccessful (Myers, in press). Chelsky et al. (1990) reported that in patients with clinical ventricular arrhythmias, ingestion of 275 mg of caffeine did not significantly alter the inducibility or severity of arrhythmias. Newby et al. (1996) conducted a randomized, double-blind, 6-week intervention trial using dietary caffeine restriction, caffeinated coffee, and decaffeinated coffee in 13 patients with symptomatic frequent idiopathic ventricular premature beats. Results showed no significant changes in palpitation scores of premature beat frequencies during the intervention weeks and no significant correlation between these variables and serum caffeine concentrations.

A possible association between coffee and risk of coronary heart disease has been examined in case-controlled as well as longitudinal cohort studies. Case-control studies have produced variable results (Myers, in press). However, a meta-analysis of 11 prospective, longitudinal cohort studies showed no increased risk of coronary heart disease associated with consumption of up to 6 cups of coffee per day (Myers and Basinski, 1992). Based on a meta-analysis of 8 case-control studies and 15 cohort studies, Kawachi et al. (1994) reported a pooled case-control odds ratio of 1.63 (95 percent confidence interval [CI], 1.50–1.78) for the effect on coronary heart disease of drinking 5 cups of coffee per day versus none. However, the odds ratio from the 15 cohort studies was not statistically significant. A study of 10,359 men and women in the Scottish Heart Health Study showed a significantly higher prevalence of coronary heart disease in subjects who were nonusers of coffee than in those who drank varying amounts of coffee (Brown et al., 1993). A more recent follow-up of subjects in the Scottish Heart Health Study showed that for many conventional risk factors, coffee had a weak but beneficial gradient with increasing consumption (Woodward and Tunstall-Pedoe, 1999). A 10-year follow-up of North American women participating in a large prospective cohort study showed no evidence for any positive association between coffee consumption and risk of subsequent coronary heart disease (Willett et al., 1996). For women initially drinking 6 cups or more of caffeine-containing coffee per day, the relative risk was 0.95 (95 percent CI, 0.73–1.26) compared to women who did not consume regular coffee.

Despite numerous studies attempting to show a relationship between caffeine and serum lipoproteins, blood pressure, cardiac arrhythmias, and risk of coronary heart disease, results have failed to show a consistent adverse effect of ingestion of moderate amounts of caffeine. Thus, increased risk of cardiovascular problems resulting from the use of caffeine supplements by the military would, in most cases, not appear to be a major concern.

One potential risk should be noted, however. A number of studies have demonstrated that caffeine consumption produces a transient elevation in blood pressure and that this occurs regardless of whether the individual is or is not a habitual user of caffeine (James, 1990; Lane et al., 1990, 1998). Caffeine consumption has also been demonstrated to potentiate the effects of acute exercise and mental stress in increasing blood pressure (Höfer and Bättig, 1993; Lane et al., 1990; Myers et al., 1989). This effect of caffeine is more pronounced in those with high stress reactivity (i.e., high levels of anxiety) and those who are borderline hypertensive or are hypertensive (James, 1990; Lane et al., 1998; Lovallo et al., 1991; Sung et al., 1995). Lovallo et al. (1996) demonstrated that in borderline hypertensive men, the use of caffeine in situations of behavioral stress may elevate blood pressure to a clinically meaningful degree and that these types of blood pressure rises in hypertensives would be large enough to transiently reduce the therapeutic effects of antihypertensive medication. However, earlier work by Greenberg and Shapiro (1987) compared two levels of caffeine in males with or without a family history of hypertension and found systolic blood pressure levels were significantly greater in individuals with a family history of hypertension across all conditions, but not specifically in response to caffeine. Wise et al. (1996) examined the effects of placebo or 6 mg of caffeine per kg lean body mass on calcium metabolism in normotensive and hypertensive individuals. Urinary excretion of calcium over 72 hours following caffeine/placebo dosing was not different with respect to caffeine treatment, or between hypertensive and normotensive subjects. Both Eggertsen et al. (1993) and MacDonald et al. (1991) reported 24-hour ambulatory blood pressures were not different between decaffeinated and caffeinated coffee in treated hypertensives.

Since military scenarios in which the use of caffeine supplements might be desirable would frequently occur when personnel are also under acute mental and/or physical stress, this could be a concern to those personnel with family histories of hypertension.

CAFFEINE EFFECTS ON REPRODUCTION

Caffeine consumption has been suggested as the cause of numerous negative reproductive outcomes, from shortened menstrual cycles to reduced conception, delayed implantation, spontaneous abortions, premature birth, low infant birthweight, and congenital malformations. As with most other aspects of caffeine consumption, there is a paucity of reliable data concerning the metabolic effects of caffeine on reproductive processes. As a general conclusion, no adverse effect on reproduction (e.g., conception, pregnancy, lactation) has been linked consistently to caffeine consumption (Christian and Brent, 2001; Leviton, 1998). Similarly, the effects of small amounts of caffeine on infants and children seem to be modest and typically innocuous (Castellanos and Rapoport, in press). Nevertheless, physicians conventionally recommend that caffeine intake be limited in pregnant women and nursing mothers. This position is also taken by the FDA (Williams, 1999). Such recommendations are in keeping with pharmacological data showing that caffeine is distributed throughout body water, crosses the placenta to enter the fetus, and is secreted in milk.

A number of reviews have examined the association between caffeine consumption and fertility, as well as its effects during pregnancy on risk of premature births, spontaneous abortions, and fetal problems including low birthweight and congenital malformations. In an epidemiological study of 403 healthy premenopausal women, heavy caffeine consumption (more than 300 mg of caffeine per day) was associated with a shortened menstrual cycle, but not with anovulation or short luteal or long follicular phase (Fenster et al., 1999). The conflicting results and methodological inadequacies of some studies surveying the association in humans between caffeine intake and effects on fertility, birthweight, premature births, or congenital malformation (when malformations of all organs is used as the outcome measure) suggest that caffeine has no consistent effect on these outcomes in humans (Leviton, 1993, 1998) despite the findings in animal studies.

Extremely high doses of caffeine in pregnant rats (well outside the range of normal human consumption) are associated with teratogenicity and fetal and maternal loss (Christian and Brent, 2001; Leviton 1993, 1998; Purves and Sullivan, 1993). Similar teratogenic effects have not been confirmed in humans, and the relevance of the route of administration (intraperitoneally) in animal studies is dubious. More recent animal studies, using lower doses of caffeine, have indicated that preconceptual exposure of rats to caffeine reduced fertility due to effects on implantation rather than fertilization rate and was associated with lower birthweight and lower neonatal and prepubertal growth rates (Pollard et al., 1999). Other studies by these same authors have indicated that in rats, caffeine administration during pregnancy is also associated with increased fetal mortality, impaired sexual differentiation, and reduced maturation of neuronal mechanisms controlling respiration and parturition.

Recent reviews of human studies suggest that some of the initial reported associations between caffeine and reduced fertility, teratogenicity, and other fetal and maternal effects in humans may be explained by confounders such as associated cigarette smoking, reporting inaccuracies, and other methodological errors (Christian and Brent, 2001; Leviton, 1998). A prospective study of 210 women consuming moderate and high levels of caffeine showed no association between birthweight of offspring and caffeine consumption (Caan et al., 1998). In contrast, a population-based study of 7,855 live births showed a small but significant increase in the odds ratio for low birthweight and preterm delivery in mothers consuming both caffeinated and decaffeinated coffee, compared to those consuming neither (Eskenazi et al., 1999). Cigarette smoking was controlled in this study; however the authors could not rule out reporting confounders for caffeine consumption. This would seem to imply that some compound in coffee other than caffeine is the potential cause. A recent meta-analysis of studies encompassing 42,988 pregnancies indicated that there was a small but statistically significant increase in risk of spontaneous abortion and low-birthweight babies in pregnant women consuming more than 150 mg of caffeine per day; however, contributing factors such as maternal age, smoking, ethanol use, or other confounders could not be excluded (Fernandes et al., 1998). A recent population-based, case-control study that controlled for confounders showed no effect of caffeine on low birthweight, preterm births, or intrauterine growth retardation (Santos et al., 1998).

Early spontaneous abortions in caffeine-consuming women have been reported in some studies but not others (Dlugosz and Bracken, 1992). Theoretically, early spontaneous abortions could be related to a caffeine-induced depressed production of placental hormones and a vulnerable implantation. On the other hand, pregnancy slows caffeine metabolism and clearance, especially in the last trimester.

One approach to obtaining an objective assessment of caffeine intake and exposure is to use biomarkers such as serum paraxanthine levels. A recent, well-controlled study of 487 women with spontaneous abortions and 2,087 normal controls, in which caffeine exposure was quantitated objectively by serum paraxanthine levels, showed that the mean serum paraxanthine concentration was significantly higher in women who had spontaneous abortions than in controls (752 versus 583 ng/mL). However, the odds ratio for spontaneous abortion was not significantly increased except in subjects with extremely high paraxanthine levels (>1,845 ng/mL). The authors concluded that moderate consumption of caffeine was not likely to increase the risk of spontaneous abortion (Klebanoff et al., 1999).

Taken together, these studies suggest that the effects of caffeine on pregnancy and fetal health vary according to the route of exposure and dosing schedule in animals, and according to caffeine dose and levels of exposure in both animals and humans. In humans, confounders that may account for many of the observed effects of caffeine in pregnancy include concurrent cigarette smoking, maternal age, ethanol consumption, and inaccuracies in reporting of caffeine consumption.

Early reports of delayed conception in women who chronically consume as little as 100 mg of caffeine per day have been confirmed in some but not all subsequent studies. However, this relationship is often confounded by coexisting cigarette smoking, which does lead to subfecundity. Based on the available evidence, some physicians recommend that caffeine consumption be avoided entirely by women who wish to become pregnant (Jensen et al., 1998; Stanton and Gray, 1995).

In the 1970s caffeine consumption was linked to benign lumps and fibrocystic disease of the breasts. However, extensive subsequent research has failed to establish a causal relationship between caffeine use and either benign or malignant diseases of the breasts. Wolfrom and Welsh (1990) concluded that the scientific literature to that point demonstrated no consistent role of methylxanthines in the etiology of fibrocystic breast disease and no consistent beneficial effect on the disease of reducing or eliminating methylxanthine consumption.

Possible effects on caffeine metabolism that are caused by hormonal changes during the menstrual cycle are unclear and poorly studied. Elimination of caffeine from the diet has been recommended to lessen premenstrual symptomology (Rossignol, 1985), but the evidence for such an effect remains inconclusive.

CAFFEINE EFFECTS ON BONE MINERAL DENSITY

Caffeine consumption has been proposed as a risk factor for osteoporosis. One of the first papers indicating a deleterious effect of caffeine came from the laboratory of Heaney and Recker (1982). Metabolic studies conducted in a large number of middle-aged women showed that caffeine intake contributed to a negative calcium balance. However, the overall loss amounted to less than 5 mg of calcium per cup of coffee. This original observation stimulated several observational studies that examined the possible relationship between caffeine consumption, calcium intake, and various indices of bone health. In the large number of studies that have since been conducted, there appears to be no consistent trend linking caffeine consumption and negative effects on bone mineral density or incidence of fracture. A moderate increase in hip fracture risk was seen in subjects in the Framingham Study who consumed more than 2 cups of coffee or 4 cups of tea per day (Kiel et al., 1990). In a prospective study of a large number of women aged 34–59 years, a positive relation was observed between caffeine intake and risk of hip but not forearm fracture (Hernandez-Avila et al., 1991). In contrast to these findings, more recent studies have failed to show a detrimental effect of caffeine on total bone mineral gain in three groups of teenage women with mean daily intakes of caffeine ranging from 14 to 77 mg (Lloyd et al., 1998), or in college-age women with a mean caffeine intake of 103 mg/day (Packard and Recker, 1996). No effect of caffeine on hip fracture rates was found in women with coffee intakes of 5 cups or more per day (Tavani et al., 1995). No effect was observed on bone loss in postmenopausal women whose habitual dietary caffeine intake ranged from 0 to 1,400 mg/day (Lloyd et al., 1997) or on bone mineral density in older men (Glynn et al., 1995). Although early experimental studies also indicated a significant effect on acute calcium diuresis (Massey and Hollingbery, 1988; Massey et al., 1989), subsequent work indicated that this acute phase of excretion was compensated by a later decrease in excretion of calcium in the urine (Kynast-Gales and Massey, 1994). Moreover, in contrast to initial studies, later studies found either no significant effect of caffeine on calcium balance (Barger-Lux et al., 1990) or negative balance only in subjects consuming less than about 660 mg of calcium per day, or half of the currently recommended intake of calcium. After a comprehensive evaluation of currently available data, Heaney (in press) concluded that any deleterious effect of caffeine on calcium balance could be offset by only 1 or 2 tablespoons of milk added to coffee and that the real issue of concern is low calcium intake rather than high caffeine intake.

CAFFEINE EFFECTS ON FLUID HOMEOSTASIS

Wemple et al. (1997) found that the consumption of approximately 2,500 mL of a carbohydrate-electrolyte beverage containing caffeine led to a greater mean 3-hour urine output than the carbohydrate-electrolyte drink alone in a resting condition (1,843 mL with caffeine versus 1,411 mL without caffeine). During exercise, however, the difference between treatments was not significant (398 mL in 5.75 hours with caffeine and 490 mL in 5.75 hours without caffeine with 2,200 mL of fluid consumed during exercise). It should be noted that the caffeine dose in this experiment was extremely low (approximately 1 mg/kg) and was not sufficient to produce a positive effect on cycling performance. The fact that urine volume was affected by this dose could be of significance in military situations where significantly higher caffeine doses may be used. Caffeine ingestion is of particular concern in situations where water balance is already in jeopardy, such as in hot or high-altitude environments.

Although moderate- to high-dose caffeine consumption (e.g., 600–900 mg) may increase fluid and electrolyte losses in urine, a normal diet will replace these losses in most military scenarios (Maughan and Leiper, 1994).

Nussberger et al. (1990) administered an oral dose of 250 mg of caffeine to eight healthy subjects and found an increase in diuresis, and increased sodium, potassium, and osmol excretion within 1 hour post-treatment. However, aldosterone and vasopressin concentrations remained unchanged. Neuhauser-Berthold et al. (1997), in a controlled experiment with 12 healthy volunteers, administered enough coffee to provide 642 mg of caffeine in a single day, and monitored fluid homeostasis in comparison with a group with an equal amount of fluid consumption from mineral water only. Subjects given the caffeine had a highly significant increase in 24-hour urine output of 753±532 ml, a corresponding negative fluid balance, and a corresponding decrease in body weight of 0.7 kg. Total body water as measured by bioelectrical impedance decreased by 2.7 percent, and sodium and potassium excretion increased by 66 and 28 percent, respectively. Caffeine use during prolonged operations in hot environments increases the risk of dehydration because such operations involve large sweat losses and/or inadequate fluid and electrolyte intake. The scientific literature indicates that a total body water deficit may (Gonzalez-Alonso et al., 1992; Maughan and Leiper, 1994; Neuhauser-Berthold et al., 1997) or may not (Brouns et al., 1998; Massey and Wise, 1984) occur. The deficit depends on the amount of caffeine consumed, the individual's history of acute and chronic caffeine use, and the total solute load of the beverage plus accompanying meals (Brouns et al., 1998; Wemple et al., 1997).

Finally, a recent study by Kiyohara et al. (1999), in an attempt to determine if serum uric acid concentration could be used as an indicator of increased urination, examined 2,240 Japanese men. They found that men consuming less than 1 cup of coffee per day had a mean serum uric acid concentration of 60 mg/L, while those consuming 5 or more cups of coffee per day had a mean concentration of 56 mg/L.

DETRIMENTAL EFFECTS OF HIGH DOSES OF CAFFEINE

High doses of caffeine can be acutely toxic. Ingestion of caffeine in doses up to 10 g has caused convulsions and vomiting with complete recovery in 6 hours (Dreisbach, 1974). The fatal acute oral dose of caffeine in humans is estimated to be 10–14 g (150–200 mg/kg) (Hodgman, 1998), but numerous factors can alter an individual's sensitivity to caffeine (e.g., smoking, age, prior caffeine status, pregnancy status, concurrent drug use) and thus alter the toxic dose. Doses of 1,000 mg (approximately 15 mg/kg body weight) have generated detrimental side effects, with early symptoms being insomnia, restlessness, and agitation. These symptoms may progress to mild delirium, emesis, and convulsions. Other symptoms can include tachycardia, asystole, and rapid respiration (Kamimori et al., 1999).

One potential risk of high doses of caffeine that needs further substantiation is dose-related decrements in mental functioning (Kaplan et al., 1997; Kuznicki and Turner, 1986; Lieberman, 1999). A number of researchers have found that high doses of caffeine can adversely affect mental performance. Kaplan and colleagues (1997) reported that although a relatively low dose of caffeine (250 mg) produced favorable subjective effects (e.g., elation, pleasantness) and enhanced performance on cognitive tasks in healthy volunteers, higher doses (500 mg) led to less favorable subjective reports (e.g., tension, nervousness, anxiety, restlessness) and less improvement in cognitive performance than placebo. Negative effects may be more pronounced in nonusers than in regular users of caffeine (Kuznicki and Turner, 1986). Excessive intake of caffeine (caffeinism) may be mistaken for anxiety disorder (Benowitz, 1990). Caffeine has been shown to produce anxiety or panic attacks in individuals with agoraphobia or panic disorders, but not in healthy controls (Boulenger et al., 1984; Charney et al., 1985).

Foreman et al. (1989) examined the effects of 0, 125, and 250 mg of caffeine on performance of a numerical version of the Stroop test, which requires sustained vigilance and intense cognitive effort as well as fast responses. Subjects receiving 250 mg of caffeine had significantly slower response times. Streufert et al. (1997) investigated the impact of 400 mg of caffeine in excess of normal consumption by persons who were already moderate to heavy caffeine consumers (400–1,000 mg/day) on performance of complex managerial tasks. Increased caffeine consumption in these individuals had mixed results. Speed of response to incoming information was faster with added caffeine; however, the managers' capacity to utilize opportunity decreased. The authors postulated that increased response speed in association with decreased effectiveness in immediate recall (Warburton, 1995) may have had unfavorable effects on a performance that requires bringing together events and actions that occur across a time dimension.

SUMMARY

Caffeine is approved as a food additive with provisional status by the FDA, thus indicating that the agency concludes there is no evidence of a human health hazard arising from consumption of caffeine added to foods and cola beverages. However, controversy continues with respect to caffeine's role in cardiovascular disease, negative reproductive outcomes, physical dependency and withdrawal, and excessive intake. The preponderance of evidence indicates that the use of caffeine by the military would not place personnel at increased risk of cardiovascular disease. Evidence on the risk of large doses of caffeine for individuals who are hypertensive or borderline hypertensive is inconclusive. For women there may be a small increase in risk of spontaneous abortion in the first trimester of pregnancy. The effects of caffeine on calcium metabolism may be of some concern only for those with very low calcium intakes (less than 50 percent of the current recommended intake). Caffeine can significantly increase 24-hour urine output, and may or may not alter total body water. Therefore, if caffeine supplements are used, emphasis should be placed on adequate fluid consumption, particularly in hot or high-altitude environments.

High doses of caffeine can have a negative effect on mood and cognitive performance, and thus the maximum content of caffeine in the delivery form of choice should not exceed 600 mg. In addition, caffeine potentiates the effects of physical, physiological, and psychological stress. Military personnel who are habitual caffeine consumers should not be denied access to caffeine in order to maximize effects of a caffeine supplement.