Caffeine is usually ingested. Caffeine is soluble in water and lipids, easily crosses the blood–brain barrier, and can be found in all body fluids, including saliva and cerebrospinal fluid. Importantly, caffeine ingested by women perinatally will be present in the umbilical cord and breast milk. Hence, it will also be present in the fetus and in breastfed infants. Caffeine is absorbed rapidly and totally in the small intestine in less than 1 h (44) and diffuses rapidly in other tissues (45). Absorption by the small intestine does not seem to vary by sex, genetic background, environmental factors, or other variables (46), although specific studies are still needed to confirm this premise. Caffeine concentrations peak in saliva 45 min after ingestion (47) and in serum after about 2 h (48). Caffeine has a relatively long half-life of 3–7 h in adults. In neonates, the half-life is even longer—between 65 and 130 h—because of their immature kidneys and liver. Peak concentrations are important because the effects of caffeine depend in part on the length of time it remains in tissues. Clearly, the effects are age dependent and depend on complex genetic and environmental interactions.

Caffeine is primarily metabolized in the liver by the cytochrome P450 oxidase enzyme system; in particular, by the CYP1A2 enzyme. However, this oxidase enzyme system is also present in other tissues, including the brain (49). Caffeine metabolism is affected by several factors, described in detail below.

The CYP1A2 gene, which encodes for a cytochrome P450 enzyme, has a large genetic variability. At least 150 single-nucleotide polymorphisms can accelerate caffeine clearance (50). The metabolic consequences of this polymorphism on caffeine downstream effects should be studied in humans. Genetic variation (i.e., increased or decreased activity of the cytochrome P450 oxidase enzyme) may increase or decrease the possible harmful effects of caffeine (e.g., during pregnancy) and any beneficial effects (e.g., on memory and learning during aging or in pathologies, such as Alzheimer’s disease). The half-life of caffeine may also be increased in liver diseases, which decreases P450 activity (50).

The molecular targets of caffeine, namely the adenosine receptors, also have great genetic variability. For example, common variants of the gene encoding for the A2a receptor can disrupt sleep (51) or cause anxiety in some individuals (52) after ingesting caffeine. More studies are needed to determine the effects of genetic variants on the consequences of caffeine consumption (53), not only in the central nervous system but also in other organs, such as the heart (40).

The expression of the cytochrome P450 epoxygenases is regulated in a circadian manner (54). Although this effect was discovered in cultured rodent cells, it may apply to many species, including humans (55). The implications are particularly important because the effects of caffeine (at least the duration of its activity) will differ during the circadian cycle. Because caffeine can alter sleep, it may also change the circadian rhythm, leading to a change in expression patterns for the cytochrome P450. One interesting hypothesis is whether caffeine consumption in adolescents and adults disrupts the expression of P450 in relation to its circadian rhythm. If the expression is downregulated, the effects of caffeine could be prolonged and produce a negative feedback loop.

The cytochrome P450 oxidase enzyme system is the same enzyme that metabolizes steroid hormones (56). Thus, steroid hormones slow caffeine metabolism. In women, this effect slows the metabolism of caffeine during pregnancy and when taking oral contraceptives (57). However, studies have not found marked differences in caffeine metabolism between the luteal and follicular phases of the menstrual cycle (57, 58). Oral contraceptives tend to double the half-life of caffeine (59).

The half-life of caffeine is on average 8.3 h longer during pregnancy and may be as much as 16 h longer than usual (60, 61). This longer half-life means that the effects of caffeine will be longer lasting in women and in the fetus. Given the effects that caffeine may have on brain development, this increased half-life in pregnant women should be taken into account when considering safety issues.

Caffeine is eliminated more slowly during early infancy, requiring perhaps 80 h in preterm and healthy-term neonates, because of the reduced efficiency of cytochrome P450 (62, 63). Elimination is likely to be at least as slow in the fetus. Fetal exposure to caffeine during pregnancy may potentially have long-lasting effects, especially in the brain. By age 6 months, infants eliminate caffeine at the same rate as that of adults (62).

Cigarette smoking doubles the rate of caffeine clearance by increasing liver enzyme activity, which may explain the higher rate of caffeine consumption among smokers (64). Substantial alcohol intake increases the half-life of caffeine and decreases its clearance (65).

Caffeine can influence objective and perceived cognitive performance by increasing alertness and wakefulness (66–68). Acute caffeine can also improve performance on memory tasks (69, 70). Finally, caffeine improves psychomotor vigilance, such as reaction time (71–73). The impact of caffeine appears to be greater under conditions that would negatively impact performance, such as acute caffeine withdrawal (74–76) or sleep deprivation (71, 77). In fact, studies that have employed long-term caffeine withdrawal methodology have consistently failed to find cognitive enhancing effects of acute caffeine (78–82). Nevertheless, in 2001, the Institute of Medicine’s Food and Nutrition Board Committee on Military Nutrition Research reported that ingesting 150 mg of caffeine enhances cognitive performance for at least 10 h (83), and this recommendation has not been updated in light of more recent empirical findings.

Numerous preclinical studies have found that antagonizing adenosine receptors, including with caffeine, has neuroprotective effects during aging and in neurological disorders by slowing cognitive decline and the progression of the disorders [reviewed in Ref. (84, 85)]. Based on these animal studies, several large longitudinal clinical studies in different countries have established an inverse relationship between coffee consumption and memory decline during normal aging (86–88). However, a study of 4,200 women and 1,800 men reported that caffeine consumption reduced cognitive decline only in women (69). In addition, a more recent study in a small group of women (89) failed to replicate the findings of the Ritchie study, demonstrating that more work is needed to understand the relationship between habitual caffeine consumption and cognitive performance. Finally, large cohort studies of men and women have also found an inverse relationship between caffeine consumption and the risk of Parkinson’s disease (90–92) and Alzheimer’s disease (93–95).

Caffeine has long been used to treat pain. However, its pain-reduction effects were not properly studied until 1984, when Lachance (96) documented that additive caffeine reduced the dose of acetaminophen necessary to achieve the target of a 40% reduction in pain scores (96). Since then, the vasoconstricting action of caffeine, secondary to adenosine receptor antagonism, has been associated with pain relief (97). Several studies have reported that acute dietary caffeine consumption can reduce pain (98, 99). In addition, caffeine in doses of between 300 and 500 mg can soothe post-dural puncture headaches, which is the most common complication of lumbar puncture procedures (100).

In general, acute intake of caffeine stimulates a modest increase in blood pressure (both systolic and diastolic), effects on heart rate (bradycardia or tachycardia, depending on dose), and neuroendocrine effects (release of epinephrine, norepinephrine, and renin) (101). These effects suggest that the mechanism of action is an increase in intracellular calcium concentrations, the release of norepinephrine, and the sensitization of dopamine receptors. These events may lead to supraventricular and ventricular tachyarrhythmias, especially at high doses. One proposed mechanism for caffeine-related cardiac arrhythmias is, again, the blockade of adenosine receptors (102, 103).

Patients with cardiac disease are often warned about the potential harmful effects of caffeine. For example, 94% of several hundred physicians from Minnesota and Vermont recommended reducing or stopping caffeine for patients reporting heart palpitations (104). However, this advice has been based primarily on anecdote and folklore (105, 106). Many of caffeine’s health effects occur after sympathetic excitation. Today, however, data suggest that caffeine does have cardiac effects, and arrhythmia is among them (107). Moreover, effects that do exist differ by dose and between habitual and non-habitual users. This severity of these threats often depends on such factors as preexisting medical conditions as well as the quantity of the ingredients taken and the length of time a person has been exposed to these substances. Many of the ingredients that include caffeine alone or in combination with other active substances have the potential to interact with prescription and over-the-counter medications. At typical caffeine doses, however, studies have documented mild changes in heart rate and blood pressure, a slight increase in sympathetic activity, and small changes in cardiac electrophysiological properties (105, 108–110).

Caffeine is believed to improve endothelial cell function at rest by increasing intracellular calcium concentrations, which stimulates the expression of endothelial nitric oxide synthase, which in turn stimulates the endothelial cells to produce nitric oxide. The nitric oxide then diffuses into vascular smooth muscle, which lies just underneath the endothelial cells, causing vasodilation (111). Caffeine can also bind directly to the vascular smooth muscle cell receptors and, through similar mechanisms, cause vasoconstriction (112).

The above information not withstanding, consuming caffeine immediately before or during exercise can be harmful and may increase the risk for myocardial ischemia (113). Indirect laboratory measures indicate that caffeine consumed immediately before exercising substantially reduces myocardial blood flow in healthy individuals (114). Several mechanisms may explain this reduction (114), including the ability of caffeine to block adenosine receptors that modulate coronary vasomotor tone. This vasoconstrictive effect might be more pronounced among caffeine-naïve individuals or in those who quickly ingest higher amounts of caffeine: for example, by consuming energy drinks. When caffeine blocks adenosine receptors, it reduces the ability of the coronary arteries to improve their flow commensurate with the increased myocardial demand of exercise, which could result in supply demand ischemia (114).

Death from caffeine ingestion appears to be rare. This rarity may be related, in part, to the marked gastric irritation from caffeine that results in spontaneous emesis. Nevertheless, several hospitalizations and some deaths from caffeine toxicity have been reported (121). For example, between 2005 and 2011, there were 79,438 emergency room visits attributable to overconsumption of energy products containing high levels of caffeine in patients aged 12 years and older (121). In most of these cases, the mechanism seems to be tachyarrhythmia and involves unusually high doses of caffeine (>3 mg/kg) (121). Most deaths after caffeine intoxication were caused by overdoses of diet pills and stimulants, and most have occurred in young patients without known underlying heart disease or any variant of normal, such as mitral valve prolapse. In one non-fatal adverse event report, no predisposing factors or structural cardiac abnormality were associated with atrial fibrillation (122). In this case, caffeine-induced atrial fibrillation spontaneously reverted to normal sinus rhythm.

Caffeine consumption is associated with fertility indices in some studies but not in others. An extensive literature review by the Oak Ridge National Laboratory concluded that chronic caffeine intake in humans is related to adverse effects on conception and reproduction, such as delayed conception and decreased fecundity. These effects appeared at caffeine doses above 200 mg/day (121). A separate review concluded that for healthy adults, intakes below 400 mg/day were not associated with adverse reproductive effects; however, the authors recommended consumption below 300 mg/day for women of reproductive age (123). In addition, some researchers argue that any association between caffeine intake and reproductive outcomes may be explained by other variables, such as maternal smoking or substance use and that research should address confounding, as well as errors in measuring exposure (124).

Reports regarding caffeine consumption and spontaneous abortions have also been conflicting. Weng et al. (125) reported a hazard ratio of 2.23 for spontaneous abortion among 164 women who consumed 200 mg/day or more of caffeine and of 1.34 for 899 women who consumed less than 200 mg/day (125). After adjusting for pregnancy symptoms, such as nausea and vomiting, other researchers found that consuming doses of 200 mg/day or more still almost doubled the risk of spontaneous abortion. A meta-analysis by Chen et al. (126) reported that, compared to a no or very low caffeine intake reference group (0–50 mg/day during pregnancy), every additional 100 mg/day of caffeine (about the amount contained in a typical cup of coffee) increased the risk of pregnancy loss (both miscarriage and stillbirth) by 7% (126). In addition, among women consuming more than 700 mg/day, the risk of pregnancy loss was increased by 72%. Similar findings were reported by Li et al. (127), who found in a separate meta-analysis of 26 studies that the risk of pregnancy loss increased by 19% for every additional 150 mg of caffeine consumed per day and by 8% for every additional 2 cups of coffee (about 200 mg) per day (127). However, Savitz et al. (128) reported no association among 2407 women who were interviewed regarding caffeine intake before experiencing spontaneous abortion (128). This finding suggests that recall bias may explain the increased hazards of spontaneous abortion reported by Weng et al. (125) and potentially other researchers (125). Other comprehensive reviews have reported some evidence that caffeine intakes of more than 300 mg/day have been associated with spontaneous abortion and low birth weight, but all have stressed the need for further research before a causal relationship can be established (129, 130). A recent study from the Nurses Health Study shows pre-pregnancy coffee consumption at levels ≥4 serving/day is associated with an increased risk of spontaneous abortions, particularly at 8–19 weeks gestational age (131).

No clear association has been found in humans between moderate doses of caffeine ingestion during pregnancy and birth defects, including congenital heart disease (132). For example, the National Birth Defects Prevention Study found variable results for this possible association (133). In another study of 2,030 malformed infants, the risk of congenital anomalies was not related to the total maternal daily caffeine ingestion below 400 mg/day (or up to 4 cups of coffee) during pregnancy (134). Other studies have found that the frequency of all congenital malformations, including congenital heart defects, was no higher than expected among women who drank between four and eight cups of coffee daily during their pregnancy (135, 136). The Institute of Medicine’s Workshop on Potential Health Hazards Associated with Consumption of Caffeine in Food and Dietary Supplements found that risk of congenital defects from caffeine was not increased in the range of amounts women typically consumed during pregnancy (121).

The consequences of caffeine consumption during pregnancy on offspring have recently been studied in mice (137). Caffeine consumption by the dam (the human equivalent of two to three cups of coffee per day) was associated with caffeine concentrations in the offspring brain that were similar to those in the umbilical cords of women drinking two to three cups of coffee per day (138). At early stages of development, specific types of neurons arise in particular brain regions and then migrate to their target areas. Caffeine slowed the migration of these neurons by 50% by antagonizing adenosine type 2A receptors. As a result, these neurons were late at being incorporated into the circuitry, with negative consequences: pups were more susceptible to seizures, and in adulthood, in utero exposed mice had mild cognitive deficits. This study was the first to document that caffeine exposure during pregnancy could harm the offspring. Generalizing the results of animal studies to humans is always speculative, but these results strongly justify conducting prospective studies in humans. Interestingly, in keeping with animal data, greater exposure to caffeine during pregnancy is associated with a lower IQ in children at age 5.5 years (139). This finding again supports the need for additional studies in humans.

Several studies have reported a significant negative association between maternal caffeine consumption and birth weight (84, 85, 140–142). However, two other large prospective cohort studies reported a dose-dependent positive association between caffeine intake during pregnancy and the risk of adverse birth weight-related outcomes, such as fetal growth restriction and small for gestational age babies (143, 144). In these studies, caffeine intake and adverse birth weight-related outcomes were found at all amounts of maternal caffeine intake. In both studies, the risk for adverse birth-related outcomes increased notably at a caffeine dose of 200 mg/day from all nutritional sources. In addition, one study of 1,207 pregnant women reported that, although they tended to reduce consumption of caffeine during pregnancy, a moderate decrease in caffeine intake to 100 mg/day in the third trimester of pregnancy did not decrease the risk of adverse birth weight-related outcomes (145).

Two separate meta-analyses of different sets of studies by Rhee et al. (142) and Chen et al. (146) reported odds ratios of having a newborn classified as low-birth weight (less than 2,500 g) for maternal caffeine consumption above 50 mg/day when compared to intakes below 50 mg/day. Furthermore, both meta-analyses found an increased risk of low-birth weight offspring for every 100 mg/day increase in maternal caffeine consumption (OR, 1.03–1.62). Another study by Hoyt et al. (140) found the odds ratios of having a low-birth weight baby increased to a range of 1.3–2.1 in women consuming more than 300 mg/day of caffeine during pregnancy (140).

Taken together, these studies provide substantial evidence of a negative association between maternal caffeine consumption and infant birth weight. Even so, the studies all relied on maternal self-report about caffeine intake; thus, the data may not be accurate. Furthermore, it is possible that additional variables, not controlled for in the analyses, could explain these relationships. For example, chronic sleep loss during pregnancy is also associated with poor birth outcomes, including low birth weight (147). Thus, pregnant women with disrupted sleep might use more caffeine to increase alertness, so the impact on birth weight could be related to short sleep duration and not to caffeine. Although this conclusion is speculative, it highlights the importance of considering additional variables when interpreting correlational data.

Caffeine may cause irritability and sleep disruption in nursing infants whose mothers consume caffeine (148), but the findings are equivocal (149). In addition, some evidence indicates that caffeine intake can reduce production of breast milk (148). Mothers are often advised by their doctors to reduce or eliminate caffeine intake if they feel that their infant shows signs of caffeine sensitivity, but there is no evidence in the literature of detrimental effects of caffeine ingestion during lactation in the general population. Behavioral issues, such as fussiness, jitteriness, and poor sleep patterns, have been reported among infants breastfed by mothers who consumed 10 or more cups of coffee (~1 g of caffeine) per day (121). The effects of caffeine in breast milk can be amplified in preterm infants or infants less than 5 months old because they metabolize caffeine so slowly (121). In addition, an intake of more than 450 mL (about two cups) of coffee per day may decrease breast milk iron concentrations, which could contribute to infant anemia (150). However, the European Food Safety Authority concluded that a single dose of 200 mg or less of caffeine (about two cups) consumed by lactating women, as well as chronic intakes at or below 200 mg, pose no safety concerns for breastfed infants (151).

Few studies have examined the impact of maternal caffeine intake on outcomes after infancy. One study by Klebanoff and Keim (152, 153) using 2,197 mother–child dyads measured child IQ and problem behaviors and examined correlations with maternal paraxanthine concentrations (a metabolite of caffeine) taken between 20 and 26 weeks of gestation (152, 153). This study found no meaningful relationship between maternal caffeine intake during pregnancy and a range of behavioral and cognitive measures in children 4–7 years old. However, another study of 1,083 mother–child pairs revealed that children who were born to mothers who estimated caffeine intake >200 mg/day during pregnancy had an odds ratio of 2.3 (95% confidence interval of 1.13–4.69) of having a child with a lower IQ at age of 5.5 years compared to the reference population of mothers reporting <100 mg/day of caffeine consumption (139). A study by Li et al. (154) reported that maternal caffeine intake was associated with increased odds of childhood obesity, with each 100-mg increase in daily maternal caffeine intake being associated with a 23% higher odds of obesity at age 15 years (127), although a study by Klebanoff and Keim found no relationships between maternal caffeine consumption and childhood obesity (152, 153).

The above studies are correlational; thus, causation cannot be determined. In addition, the maternal caffeine intake in these studies was estimated based on self-reports. One potential explanation for the discrepancies described above is the method used to determine caffeine use. In the study by Klebanoff and Keim (152, 153), which found no significant relationship between maternal caffeine intake and outcomes after infancy, measured serum caffeine concentrations and did not use self-report (152, 153). By contrast, the studies that found significant relationships between maternal intake and measures in the offspring after infancy relied exclusively on retrospective self-reports, several years after the fact, about prenatal caffeine consumption by mothers after they gave birth and during the first two trimesters of pregnancy, respectively. Caffeine intake was estimated from food-frequency questionnaires or interviews in which women reported how often and how much they consumed coffee, tea, and soda. Other variables affecting self-reported caffeine consumption and offspring behavioral outcomes might explain these relationships, but in the study that relied entirely on serum concentrations, such variables were not identified. These studies also measured different outcomes in the offspring. Klebanoff and Keim (152, 153) had the most comprehensive battery of cognitive and behavioral outcomes, but Galera et al. (139) only measured IQ (The Wechsler Preschool and Primary Scale of Intelligence Third Edition), and Li et al. (127) only measured weight and weight gain in the offspring (139, 152–154). Meaningful comparisons of studies are difficult when the methods for assessing caffeine intake and the outcomes are different. Research with objective measures of caffeine intake and standard outcomes is needed.

Most of the research examining linkages between caffeine and cancer has been conducted on coffee and tea and not on caffeine specifically, which makes it difficult to determine the mechanism. The International Agency for Research on Cancer has concluded that the evidence is insufficient to conclude that caffeine, as consumed by a typical coffee drinker, is carcinogenic (121). Several large prospective trials have reached the same conclusion (123, 155, 156). Furthermore, Nawrot et al. (123) concluded in their review of the research that caffeine is unlikely to be a human carcinogen at levels less than 500 mg/day, to the equivalent of five cups of coffee (123).

Excessive caffeine intake (more than 400 mg/day) may increase the risk of detrusor instability (unstable bladder) in women (157). For women with preexisting bladder symptoms, even moderate caffeine intake (200–400 mg/day) may increase the risk for unstable bladder (157). This finding was confirmed in another case–control study of women who were given 200 mg of caffeine citrate (158). In addition, caffeine intake of 4.5 mg/kg/day caused early urgency and frequency of urination in men and women with overactive bladder (159). However, these studies did not examine whether a decrease in caffeine intake was associated with improvements in overactive bladder symptoms. Studies should address this issue.

According to www.drugs.com (a site owned by The Drugsite Trust, a privately held Trust administered by two New Zealand Pharmacists), 85 drugs (430 brand and generic names) are known to interact with caffeine, of which 11 can lead to major interactions.² Because caffeine consumption is at an all-time high and prescription drug use is more prevalent than ever, the risk of negative caffeine and prescription drug interactions is increasing (160, 161). Because of the popularity of caffeine, clinicians should be conscious of the pharmacokinetic interactions between dietary caffeine and over-the-counter and prescription medications, and they should provide the necessary guidance to the patient including dietary restrictions. We also recommend that the potential interaction with these drugs be appropriately addressed on the labeling.

Caffeine has a diuretic effect (123, 162, 163). However, in one clinical trial, different doses of caffeine (up to 6 mg/kg body weight) consumed by 59 habitual caffeine consumers after a 6-day run-in period of 3 mg/kg of caffeine did not markedly change hydration-related biomarkers, suggesting that increasing doses of caffeine did not induce hypohydration in these participants (164). These findings are supported by two similar studies, one in which 5 mg/kg body weight of caffeine was consumed daily for 4 consecutive days by 30 men who normally consumed less than 100 mg/day (42) and one in which 4 mg/kg body weight/day of caffeine from coffee was consumed for 3 consecutive days by 50 adult male habitual coffee consumers who usually consumed 3–6 cups of coffee/day (165). These findings suggest that the diuretic effects from consuming between 4 and 6 mg/kg body weight/day of caffeine are not likely to have adverse consequences for healthy adults who are habitual consumers of caffeine. Similar studies should be conducted in populations that vary by health status, age, and sex.

Pregnant women and fetuses may be particularly vulnerable to the effects of caffeine. Caffeine is a biologically active molecule that can act on multiple targets and affect numerous functions positively or negatively. At early stages of fetal development, caffeine may have deleterious effects (137). A recent prospective study suggests that preconception caffeine consumption may also pose a risk to pregnancy, with pre-pregnancy consumption of >400 mg of caffeine/day increasing the risk of spontaneous abortion by 11% compared with women who consumed <50 mg of caffeine/day (131). Many psychoactive compounds can cross the placental barrier and alter the development of the fetal brain. Once caffeine enters the fetal circulation, it is metabolized slowly because neither the placenta nor the fetus itself has cytochrome P450, the enzyme that metabolizes caffeine (166). This reduced caffeine metabolism results in a longer half-life and increased caffeine exposure to the fetus (141, 167). The American College of Obstetricians and Gynecologists recommends limiting caffeine consumption during pregnancy to less than 200 mg/day (168). In the late 1970s, most women maintained their intake during pregnancy at an average of about 190 mg/day³ (5). In the 1980s and 1990s, the average maternal caffeine consumption declined to about 125 mg/day (5). Consumption was reported to be about 123 mg/day between 1997 and 2007 (84, 85) and was even lower (58 mg/day) in a 1999 survey (169). This decline has been attributed to FDA warnings that excess caffeine consumption during pregnancy may adversely affect neonates (170). Interestingly, however, in a small cohort of 105 women who drank coffee before pregnancy, 65% reported an aversion to coffee during the first trimester, and 95% voluntarily reduced their consumption during this trimester (171), so perhaps women might be naturally averse to caffeinated products during pregnancy.

Data on caffeine consumption during lactation are limited. One small study from Poland reported that average caffeine intake in a sample of lactating women ranged from 127 to 163 mg/day (172).

Young children may be vulnerable to the effects of caffeine because they weigh less. For example, a typical can of soda contains about 45 mg of caffeine on average. In an adult weighing 70 kg, the effective dose is 0.6 mg/kg, but in a child weighing 20 kg, the effective dose of the same soda would be 2.25 mg/kg. In comparison, the average caffeine intake in adults is 180 mg/day, resulting in an average effective dose of 2.5 mg/kg. Thus, the physiological impact of a single soda in a child may be equivalent to the impact of two cups of coffee in an average-sized adult. Adolescents may also be particularly vulnerable to the sleep-disrupting effects of caffeine because they may also use caffeinated beverages to stay awake (173, 174).

Data have been collected in children and adolescents using dose–response and placebo-controlled research methods. Outcomes, such as cardiovascular function (175–178), mood (179–181), and cognitive performance (82, 182), have all been measured at caffeine doses ranging from 50 to 300 mg. None of the results suggest that caffeine at these doses is acutely harmful to children and adolescents (183).

Some studies suggest an association between caffeine consumption and longer term behavioral problems in youth, such as anger, violence, sleep disturbances, and alcohol and drug use (180, 184). Researchers in Iceland surveyed 7,400 adolescents (aged 14 and 15 years) and found that most reported consuming caffeine on a typical day and that caffeine intake (primarily from soda and energy drinks) was related to daytime sleepiness and anger for both sexes (185). In a 2013 study of 3,747 15- to 16-year olds, self-reported caffeine intake was strongly associated with self-reported violent behavior and conduct disorders (186). In this study, 21% of participants consumed at least one energy drink per day.

Other studies have found that anxiety can be produced at a wide range of doses (200–2,000 mg of caffeine/day), but many of these studies have used psychiatric patients or patients with a preexisting anxiety disorder (123). Other effects in these studies included nervousness, fidgeting, jitteriness, restlessness, hyperactivity, and sleeplessness (123, 187, 188). When children were stratified by prestudy caffeine intake, emotions and behaviors differed between low- and high-dose consumers (187, 188). Children consuming high doses were more easily frustrated and were more nervous during baseline tests than were the children consuming lower doses. Other studies have found that children with attention-deficit/hyperactivity disorder have higher rates of caffeine abuse, perhaps due to the additive effects of caffeine on dopamine action at the dopamine D2 dopamine receptor, similar to the way guanfacine works for children with this disorder (189, 190).

The safety of high-dose caffeine and energy drinks in younger individuals and caffeine-naïve individuals has not yet been determined. The consumption of highly caffeinated energy drinks has been associated with elevated blood pressure, altered heart rates, and severe cardiac events in children, adolescents, and young adults, especially those with underlying cardiovascular diseases (115, 177, 191, 192). For example, a study of 50 young adults found that consuming a sugar-free energy drink containing 80 mg of caffeine (slightly less than the caffeine contained in one cup of coffee) was associated with changes in platelet and endothelial function great enough to increase the risk for severe cardiac events in susceptible individuals (193). These findings show how the acute effects of caffeine on heart rate might result in cardiovascular events requiring hospitalization, especially in at-risk young adults. In addition, caffeine’s effects on blood pressure are more pronounced among African-American children than among Caucasian children (mean difference in blood pressure averaging 6.5 mm Hg) (175, 194). High doses of caffeine may exacerbate cardiac conditions for which stimulants are contraindicated (195–198). In particular, ion channelopathies and hypertrophic cardiomyopathy, which is the most prevalent genetic cardiomyopathy in children and young adults (0.2% of the population), are of concern because of the risk of hypertension, syncope, arrhythmias, and sudden death (197, 199).

Another population that may be at risk for adverse effects of caffeine are patients with mental illness. Caffeine antagonism of adenosine receptors can result in enhanced dopaminergic signaling, thought to be due to a combination of increased dopamine release (200, 201), upregulation of dopamine receptors, and increased affinity of dopamine receptors for dopamine in the striatum and nucleus accumbens (202). Furthermore, adenosine receptors can form heterodimers with dopamine receptors (203), which can modulate dopamine signaling. For some psychiatric illness, such as Parkinson’s disease, Alzheimer’s disease, and depression, caffeine antagonism of adenosine receptors may improve symptoms (204, 205) and slow the progression of neurodegeneration (206, 207), although these findings are equivocal with some studies reporting caffeine increases depressive symptoms (208). For other mental illness, such as schizophrenia, caffeine may exacerbate psychotic symptoms (209), although the majority of this literature is informed by case studies, with very few double-blind placebo-controlled studies (210). There is also good evidence that higher caffeine use is associated with greater reporting of anxiety symptoms (211, 212) and may increase risk of symptom relapse (213) and suicide among bipolar disorder patients (214). Finally, there is strong empirical evidence that caffeine potentiates the rewarding effects of drugs of abuse (215–217), which suggests that caffeine use can increase vulnerability to substance use disorder (218). The lack of randomized control trials on the impact of caffeine in patients with mental illness makes it difficult to determine safe doses, effects of acute and chronic caffeine, and potential interactions between caffeine and medications. Currently, there are no specific recommendations for caffeine consumption for individuals with mental or psychiatric illness, but it may be worth consideration by physicians and psychologists treating patients with mental illness.