Caffeine in Food and Dietary Supplements: Examining Safety: Workshop Summary (2014)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4 Exploring Safe Caffeine Exposure Levels for Vulnerable Populations Although the challenge of gaining a better scientific understanding of potentially vulnerable populations was addressed throughout the workshop in different contexts, the Day 2, Session 3, panel was de- signed to focus specifically on vulnerable populations. This chapter summarizes that panel. Moderator Mark Feeley, M.Sc., Health Canada, Ottawa, Ontario, asked the panelists to consider three key questions: (1) Are there specific vulnerable subpopulations that can be defined? (2) Are there additional variables beyond individual caffeine sensitivity that should be considered in these subpopulations as a means to help identify caffeine sensitivity? (3) What are the most relevant end points for defining a sensitive population? Many of the panelists’ remarks and subsequent discussion revolved around currently established safe levels and the evidence underlying those levels. Box 4-1 describes key mes- sages from the speakers. BOX 4-1 Key Points Made by Individual Speakers As noted by Mark Feeley, on the basis of an extensive scientific eval- uation conducted about 10 years ago, Health Canada recommended that individuals consume no more than 400 mg of caffeine daily. Since then, Health Canada has identified two potentially vulnerable subpopulations, women of reproductive age and children, and has further recommended that children consume no more than 2.5 mg of caffeine per kilogram body weight per day and that women of reproductive age consume no more than 300 mg a day. Feeley discussed these standards, noting that the challenge with caffeine-containing energy drinks is similar to what the U.S. Food and Drug Administration is facing. 51 

52 CAFFEINE IN FOOD AND DIETARY SUPPLEMENTS Christina Chambers said that several studies have indicated an in- creased risk for spontaneous abortion with caffeine consumption in preg- nant women, prompting the American College of Obstetrics and Gynecol- ogy to recommend that pregnant women restrict their intake to less than 200 mg per day. Chambers described these studies and other evidence of health effects in pregnant women and made a call for better and more con- tinuous measures of caffeine exposure during pregnancy. Steven Lipshultz described the efforts of a working group in South Flori- da that was formed in 2007–2008 to see whether the safety signals being observed in children who had consumed caffeinated energy drinks were of concern. He identified children and children with underlying cardiac condi- tions as two potentially vulnerable populations. He encouraged banning the sale of caffeinated energy drinks to children and teenagers until and unless their safety can be demonstrated through scientific research. HEALTH CANADA’S APPROACH Introductory Remarks by Mark Feeley, M.Sc., Health Canada Health Canada’s position on caffeine is not much different from the Food and Drug Administration (FDA) position, according to Mark Feeley. When added directly to a food, caffeine is regulated as a food additive, requiring an application to Health Canada and a standard safety assess- ment. Approvals are granted on a case-by-case basis. Initially, the use of caffeine as a food additive in Canada was restricted to cola-type beverages; in the mid-2000s it was expanded to include all carbonated soft drinks. So today, theoretically, adding caffeine to a food for sale in Canada is restrict- ed to carbonated soft drinks. Regulations regarding energy drinks are cur- rently in flux, with most energy drinks on the market in the United States also having market access in Canada with some slight modifications to their composition. About 10 years ago, Health Canada undertook a relatively extensive scientific evaluation of caffeine, from which it determined that a level of up to 400 mg of caffeine on a daily basis would likely not be associated with any adverse health effects in the general population. Subsequent to that review, Health Canada identified two potentially vulnerable subpop- ulations: women of reproductive age and children. Women of reproductive age were identified as a potentially vulnerable group on the basis of the literature. Most studies examined by Health Cana- da focused on either reproductive or developmental outcomes and were based on coffee consumption as a surrogate for caffeine exposure. Accord- 

EXPLORING SAFE CAFFEINE EXPOSURE LEVELS 53 ing to Feeley, most of the studies examined by Health Canada are described in an Oak Ridge National Laboratory (ORNL) report (2011), a very exten- sive evaluation that was conducted at the request of the FDA and whose conclusions were consistent with those of Health Canada. The studies did not show a clear cause-effect relationship between caffeine consumption and adverse health effects, with about 20 percent showing no effect, but they did show associations between caffeine consumption and some fertility indices. Health Canada decided that it would be prudent to suggest limiting caffeine intake among women of reproductive age to not more than 300 mg per day. For children, the majority of evidence reviewed by Health Canada focused on children between 7 and 12 years of age. Most studies in- volved children being brought into a clinical setting and provided with defined caffeine doses through beverage exposure and with doses rang- ing from 2.5 to 10 mg per kg of body weight per day. Again, Feeley re- ferred workshop participants to the ORNL (2011) review for what he described as an “elegant compilation” of the evidence. Beneficial effects were observed in terms of task, motor activity, attention, and other out- comes. But negative effects were observed as well. The one negative ef- fect that was observed both consistently and at the lowest dose level was anxiety (i.e., both subjective and objective measurements of anxiety). On the basis of that research, although limited, Health Canada decided that it would be prudent to limit caffeine intake among children up to the age of 12 to no more than 2.5 mg per kg of body weight per day. Feeley observed that median intakes in these subpopulations do not come close to these recommended maximum values. Only above the 90th percentile in either subpopulation have researchers observed indi- viduals exceeding the recommended maximum levels. Health Canada is currently exploring options for product labeling and consumer education as ways to control the intakes of caffeine below the recommended daily intakes. SAFE CAFFEINE EXPOSURE LEVELS IN VULNERABLE POPULATIONS: PREGNANT WOMEN AND INFANTS Presented by Christina Chambers, Ph.D., M.P.H., University of California, San Diego Christina Chambers listed several end points of interest with any ex- posure during pregnancy (i.e., not just caffeine): increased risk of spon- 

54 CAFFEINE IN FOOD AND DIETARY SUPPLEMENTS taneous abortion or spontaneous pregnancy loss, typically defined as loss prior to 20 weeks’ gestation; increased risk for major congenital anoma- lies; intrauterine growth restriction; and preterm birth. End points of in- terest with exposure to caffeine in infants, which occurs primarily through lactation, include increased risk for central nervous system ef- fects, such as irritability and sleeplessness, and infant growth problems. Chambers said that researchers who study caffeine exposure in preg- nant women are challenged by several measurement and design issues, including limited capability to ethically conduct randomized clinical tri- als. Observational studies themselves are limited by their frequent reli- ance on maternal report of major sources of caffeine, some of which the mother may not be aware of, and misclassification of exposure in terms of timing and dose. Mothers are often being asked months or years after pregnancy to recall specific times that they consumed caffeine- containing products and doses of caffeine consumed. Observational stud- ies that incorporate biomarker assessment (i.e., biomarkers of exposure) typically do so only at select time points and do not necessarily reflect exposure over time. Observational studies can also be challenged by ei- ther bias issues or confounding comorbidities such as depression or other maternal psychiatric disorders, autoimmune diseases, maternal diet (i.e., it may be different in high versus low caffeine consumers), and maternal body mass index. Also, there may be coexposures associated with higher caffeine intake, such as alcohol and tobacco. Finally, an important problem, especially with studying spontaneous abortion, is the change in usual caffeine consumption because of the symptoms of pregnancy. More than half of pregnant women experience nausea and vomiting, which, in turn, may lead to reduced caffeine con- sumption. But nausea and vomiting are also highly protective against spontaneous abortion, potentially confounding any found association between caffeine consumption and spontaneous abortion. For this and other reasons, according to Chambers, spontaneous abortion is one of the most difficult pregnancy outcomes to study appropriately. The challenge to studying spontaneous abortion is compounded by the fact that although spontaneous abortion is an extremely common event, occurring in 60 percent or more of pregnancies, is usually not clin- ically recognized. Also, because it may be more socially acceptable or preferable for a mother to report that she had a spontaneous abortion than a medically induced abortion, spontaneous abortion is sometimes mis- classified. Finally, although most spontaneous abortions are believed to be caused by chromosome aberrations, not environmental exposure, 

EXPLORING SAFE CAFFEINE EXPOSURE LEVELS 55 many known environmental risk factors that might be associated with high caffeine exposure, such as the quantity and frequency of tobacco use, are poorly measured. Despite these challenges, said Chambers, several studies have sug- gested increased risk for spontaneous abortion with caffeine consumption, particularly with higher doses. She described one study that drew atten- tion in the past several years by Weng et al. (2008). The study focused on 1,063 women enrolled in the Kaiser health care plan in northern Califor- nia. The women, all of whom tested positive for pregnancy, were inter- viewed subsequent to their pregnancy tests, sometime in the first 15 to 16 weeks of gestation, at a median age of 10 weeks of gestation. They were asked about previous caffeine consumption up to that time. Chambers noted that some women had already experienced a spontaneous abortion at the time of the interview and that those women were being interviewed retrospectively. The researchers reported a hazard ratio of 1.42 for a spontaneous abortion among women who consumed, on average, less than 200 mg of caffeine per day. That value, she said, was not statistical- ly significant. For women who consumed 200 mg or more per day, the hazard ratio rose to 2.23, with a lower bound to the confidence interval of 1.34. Accounting for nausea and vomiting, the researchers concluded that doses of 200 mg or more per day were associated with an increased risk, about a doubling of risk, for spontaneous abortion. A conflicting study was published the same year. Savitz et al. (2008) collected data on caffeine consumption by interviewing 2,407 women enrolled in a cohort either prior to pregnancy or during early pregnancy. The researchers reported that caffeine consumption was unrelated to an increased risk of spontaneous abortion if the event occurred after data collection on caffeine, that is, among women who were interviewed prior to knowing they were going to have spontaneous abortions. Among women who were interviewed after already experiencing spontaneous abortions, the researchers found an association. They concluded that there was no evidence of an increased risk for spontaneous abortion with caffeine consumption at any level within the range of the study and that the association found among women who had already experienced spon- taneous abortions was a spurious association resulting from recall bias. Despite these conflicting results, concern for risk of spontaneous abortion led the American College of Obstetrics and Gynecology to issue guidelines recommending that pregnant women restrict their caffeine intake to less than 200 mg per day. 

56 CAFFEINE IN FOOD AND DIETARY SUPPLEMENTS Evidence for congenital anomalies as a potential adverse health effect associated with caffeine consumption during pregnancy has been some- what unremarkable, according to Chambers. A number of recent studies have examined increased risk relative to a variety of specific defects that might be expected if caffeine were a teratogen. Most recently, the National Birth Defects Prevention Study,1 a Centers for Disease Control and Pre- vention (CDC) multicenter study, which now has sufficient numbers of specific birth defects to provide adequate statistical power, has yielded variable and not very compelling results. There has been little evidence of a dose–response relationship. In addition, animal models at doses relevant to human exposure have not been concerning. In sum, in Chambers’ opinion, the evidence is not compelling enough to suggest that there is an increased risk for any specific pattern of congenital anomalies in humans with caf- feine consumption in the range that women would typically consume. Nor have researchers found a consistent association between caffeine consumption during pregnancy and various measures of fetal growth, in- cluding increased risk for growth, small for gestational age, and low birth weight. Chambers described a 2008 study on small for gestational age and low birth weight among 2,635 women that showed little evidence of a dose relation in various dosing ranges (all greater than 100 mg per day) com- pared to women who consumed less than 100 mg per day. Odds ratios ranged from 1.2 to 1.5, with two showing borderline statistical signifi- cance, but again, with no evidence of a dose–response relationship with increasing levels of consumption (CARE Study Group, 2008). Likewise, even studies showing positive associations have suggested relatively small effects in terms of magnitude, too small to be clinically significant. Finally, there have been a couple of noteworthy studies on preterm birth. Bech et al. (2007) conducted a randomized trial of caffeine reduc- tion with two groups of pregnant women: one group received caffeinated coffee, the other decaffeinated coffee. Among the 1,207 women, the re- searchers found no effect of caffeine consumption on length of gestation at an average intake of 182 mg per day. Clausson et al. (2002) reported no association among 873 pregnant women between caffeine consump- tion at any level and preterm birth. Among all these various studies of pregnant women, few have incor- porated urine blood or core blood markers of exposure. In those studies that have included biomarkers, the biomarkers appear to be correlated with maternal report. According to Chambers, caffeine does not seem to carry the same stigma that alcohol and perhaps tobacco consumption do. 1 Available at http://www.nbdps.org/index.html (accessed December 25, 2013). 

EXPLORING SAFE CAFFEINE EXPOSURE LEVELS 57 Still, Chambers cautioned that most studies involve spot measures, which may not reflect individual genetic variability or changes over the course of pregnancy. A few studies have been conducted on the relation be- tween polymorphisms for metabolizing enzymes and adverse health out- comes to see whether rapid metabolizers are at different risks, but the results have been conflicting. With respect to infant exposure through caffeine in breast milk, there have been anecdotal individual case reports or small case series of fussi- ness, jitteriness, and poor sleep patterns among infants born to mothers who consume the equivalent of 10 or more cups of coffee daily. As with many agents in infants being breastfed, the effects can be amplified in preterm or very young infants (until the age of about 4 to 5 months) be- cause they metabolize caffeine more slowly and may attain similar levels to their mothers. One study from Costa Rica suggested that coffee intake of more than 450 ml caffeine daily may decrease breast milk iron con- centrations (Munoz et al., 1988). Data Gaps To conclude, Chambers identified several data gaps:  Better and more continuous measures of exposure at specific time points and at repeated time points over pregnancy. Such data would help researchers to tease apart real risks and determine whether there is a peak exposure effect or critical time of exposure.  Data that would help determine whether there are any neurobe- havioral effects of high dose exposure that are independent of the effects of coexposure, such as alcohol and tobacco.  Data on the dose of exposure and potential effects during the 5- to 6-week period prior to a woman realizing that she is pregnant. This is an important piece of information, in Chambers’s opinion, because more than half of all pregnant women do not plan their pregnancies. So, for example, among women who are not aware that they are pregnant, what is the pattern of energy drink con- sumption? What is the pattern of alcohol consumption? Are wom- en who consume energy drinks binge drinking? Is high caffeine consumption in an unrecognized pregnancy associated with poor dietary habits? Is high caffeine consumption in an unrecognized pregnancy associated with lack of use of folic acid supplements, which in turn may lead to an increased risk of birth defects? 

58 CAFFEINE IN FOOD AND DIETARY SUPPLEMENTS  Data in humans on the effects of other ingredients in caffeinated energy drinks and other products. RISK OF ADVERSE EFFECTS OF CAFFEINE AND CAFFEINATED PRODUCTS IN CHILDREN AND OTHER VULNERABLE GROUPS Presented by Steven E. Lipshultz, M.D., University of Miami In 2007, physicians in south Florida were seeing health effects in children that seemed to be temporally related to the use of caffeinated energy drinks. Steven Lipshultz described how he and his colleagues formed a working group to determine whether those safety signals were of concern. The group interviewed the south Florida poison control center data- base. In 2007, across the state of Florida, 39 persons, aged 2 to 20 years, were being tracked for health concerns associated with caffeine con- sumption. In 2008, 125 persons aged 2 to 20 years were being tracked. Those cases were only for general caffeine toxicity because at the time there was no reporting mechanism in the United States for caffeine tox- icity related to energy drink consumption. Next, the group interviewed the National Poison Control Center da- tabase, again focusing on caffeine. They reported their results in Pediat- rics (Seifert et al., 2011). They found that children from infants to age 19 years accounted for 46 percent of all calls reporting caffeine toxicity in 2007. About 10 percent of these cases had moderately severe symptoms that often required treatment. There were deaths in all years examined (2006–2008). The working group, Lipshultz noted, eventually became aware that the Substance Abuse and Mental Health Services Administration was collecting data on U.S. emergency department visits involving caffeinat- ed energy drink consumption in patients 12 years old and older. Those data, which were published in a 2011 DAWN report (SAMHSA, 2013), suggested that such visits increased exponentially between 2005 and 2011 (see Figure 4-1). He also noted that the working group sought data from outside the United States as well. Caffeinated energy drink toxicity has been tracked elsewhere, notably in Australia, Germany, Ireland, and New Zealand. The German Federal Institute for Risk Assessment assessed caffeinated 

EXPLORING SAFE CAFFEINE EXPOSURE LEVELS 59 energy drink toxicity from 2002 to 2008 and reported several serious outcomes: liver damage, kidney failure, respiratory disorders, agitation, seizures, psychotic conditions, rhabdomyolysis, tachycardia, cardiac ar- rhythmias, hypertension, heart failure, and death. From 1999 through 2005, Ireland’s poison center reported seventeen adverse events associat- ed with energy drinks, including confusion, tachycardia, and seizures and two deaths. Furthermore, Lipshultz said, from 2005 to 2009, the New Zealand poison center reported 20 energy drink– or energy shot–related adverse events, including vomiting, nausea, abdominal pain, jitteriness, racing heart, agitation, and myocardial infarction. Between 2004 and 2010, the New South Wales Australian Poison Information Center re- ported increases in both recreational ingestions and accidental pediatric ingestions of caffeinated energy drinks (Gunja and Brown, 2012; see Figure 4-2). This information led the group to discuss the need to track caffeinat- ed energy drink toxicity with the U.S. National Poison Data System (NPDS). Lipshultz and his colleagues were pleased to learn that the NPDS had recognized the same need. Data from the first year of report- ing (from October 1, 2010, through September 30, 2011) were published in Clinical Toxicology (Seifert et al., 2013). During that period, there were 1,480 cases of nonalcoholic caffeinated energy drink toxicity, half of which involved children younger than 6 years old. Specifically, 51 percent involved children 0 to 5 years old; 11 percent, children 6 to 12 years old; 18 percent, teenagers 13 to 19 years old; and 21 percent, adults 20 years and older. Children under the age of 6 years also had the highest proportion of unintentional exposures to nonalcoholic energy drinks (76 percent), and teenagers between 13 and 19 years had the highest propor- tion of intentional exposures (49 percent). In 2012, the New Zealand Food Safety Authority compared the ef- fects of exposure to caffeine from energy drinks or energy shots to back- ground dietary exposure from naturally occurring caffeine in foods and beverages and in cola-type soft drinks (Thompson and Schiess, 2010). They found that 68 percent of children and 42 percent of teenagers ex- ceeded the adverse-effect level of 3 mg caffeine per kg body weight after consuming one retail unit of energy drink or energy shot beyond their baseline dietary exposure (Thompson and Schiess, 2010; Seifert et al., 2011). 

60 CAFFEINE IN FOOD AND DIETARY SUPPLEMENTS 15,000 Adverse Reactions 14,042 Misuse or Abuse 12,000 10,770 Number of Emergency Department Visits 9,618 9,000 8,798 6,996 6,000 6,090 5,368 5,284 4,312 3,000 1,973 3,060 840 1,704 583 0 2005 2006 2007 2008 2009 2010 2011 Year FIGURE 4-1 U.S. emergency department visits involving caffeinated energy drinks in patients age 12 years and older, 2005–2011. NOTE: An adverse reaction is defined as an adverse reaction or side effect to the use of energy drinks as documented in the chart; misuse or abuse is broadly defined to include all visits associated with inappropriate use of energy drinks. SOURCE: SAMHSA, 2013. Presented to the Planning Committee for a Work- shop on Potential Health Hazards Associated with Consumption of Caffeine in Food and Dietary Supplements on August 5, 2013. 60 Recreational: intentional ingestion for the purpose of gaining euphoria or other psychotropic effect 50 Accidential paediatric: non-deliberate ingestion by a person under 18 years of age Number of Exposures 40 30 Recreational Accidental paediatric 20 10 0 2004 2005 2006 2007 2008 2009 2010 Year FIGURE 4-2 Calls regarding the intentional or accidental consumption of ener- gy drinks to individual New South Wales, Australia, poison control centers, 2004–2010. SOURCE: Gunja and Brown, 2012. 

EXPLORING SAFE CAFFEINE EXPOSURE LEVELS 61 Children with Underlying Cardiac Disease as a Potentially Vulnerable Population Lipshultz elaborated on the lack of screening in U.S. children for un- derlying heart disease and other susceptibilities. In his experience, many children are not aware of their underlying cardiac conditions. Sometimes the first event is terminal, and the underlying cardiac disease is identified only at autopsy. The American College of Cardiology recommends restricting activi- ties that increase adrenergic stimulation for groups at risk of sudden car- diac death. According to Lipshultz, some of the risks are the same as those associated with sudden death from exposure to caffeinated energy drinks, specifically hypertrophic cardiomyopathy and long QT syn- drome. Dufendach et al. (2012) reported on a 13-year-old girl with no known history of cardiac disease who was taken to the emergency de- partment. She had palpitations, chest pain, shakiness, and dizziness and had recently consumed a 16-ounce can of an energy drink containing 160 mg of caffeine (amounting to 4.1 mg per kg of body weight). After ex- tensive testing, doctors diagnosed long QT syndrome. Lipshultz noted that other regulatory groups already recognize risk groups for stimulants in addition to caffeine, such as those at risk from amphetamines used to treat attention deficit hyperactivity disorder (ADHD). In February 2005, Health Canada suspended the use of the am- phetamine Adderall XR as a result of concerns about an increased risk of sudden cardiac death. In August of the same year, they reinstated the product but with revised labeling that identified rare heart-related side effects. In May 2006, the FDA directed manufacturers to strengthen the warning section of their labeling for Adderall XR by listing potentially serious cardiovascular adverse events. For Lipshultz, it is not clear how best to balance the need to provide necessary medications to children with the need to protect children who have underlying heart disease. The south Florida working group asked itself whether any of its own specialists would recommend that children on hypertensive therapy for high blood pressure or children on anticon- vulsant therapy for seizures not consume caffeine. These specialists be- lieved that children on either type of therapy should probably be advised not to consume caffeine from any source. High blood pressure, Lipshultz said, occurs in 3 to 5 percent of chil- dren in the United States, with 2.5 percent of NPDS calls for caffeinated energy drink consumption toxicity in the 2000–2013 period related to 

62 CAFFEINE IN FOOD AND DIETARY SUPPLEMENTS hypertension. Over the same period, 24 percent of reported caffeinated energy drink poisonings involved seizures. Caffeinated energy drinks may interfere with anticonvulsant therapy and lower the threshold for seizures. The situation is the same for children with syncopal disorders. Most physicians would advise children with syncopal disorders to not consume caffeine. Similarly, about 10 percent of U.S. children have a diagnosis of ADHD. About 70 percent of these children are treated with prescription therapy, most commonly with stimulants. Most physicians would advise patients on stimulant therapy not to consume an additional stimulant. Lipshultz mentioned caloric intake and diabetes as another area of concern, with caffeine potentially exacerbating the adverse health outcomes associ- ated with these conditions. In addition, among adolescents with eating dis- orders, caffeine can potentially lead to adverse health outcomes. For children with underlying heart disease, any stimulant is of con- cern, whether it be a prescription medication or a caffeinated energy drink. Unfortunately, noted Lipshultz, routine physical exams for high school athletes do not identify everyone at risk for sudden cardiac death, nor are children in the United States routinely examined with electrocar- diography or echocardiography. Various national organizations and individuals have put forth rec- ommendations for the use of energy drinks among children, including the American Academy of Pediatrics, perhaps the largest pediatric organiza- tion in the world. In 2013, the American Medical Association voted that marketing energy drinks to children and adolescents less than 18 years old should be suspended. According to Lipshultz, these various educa- tional campaigns, as well as the banning of the sale of alcohol-containing energy drinks by the FDA, appear to be associated with decreasing calls to poison centers for energy drink consumption (see Figure 4-3). In his conclusion, Lipshultz reiterated that not only do caffeine- containing energy drinks have no therapeutic or nutritional benefit for children less than 18 years of age, but also all available databases suggest that certain subgroups may experience serious adverse events after con- suming these drinks. Are there safety signals? “There are absolute safety signals,” Lipshultz said. “It’s consistent in every study that we can get our hands on.” He identified two probable vulnerable populations: chil- dren and children with underlying disease. Lipshultz said that until re- search establishes the children’s safety, “as a pediatrician, as somebody who’s run one of the largest children’s hospitals for a decade, it is part of 

EXPLORING SAFE CAFFEINE EXPOSURE LEVELS 63 my responsibility to protect children when I see safety signals until I real- ly know there’s a therapeutic benefit or at least no increased risk.” PANELIST DISCUSSION WITH THE AUDIENCE Lipshultz clarified that no data indicate any specific safety threshold in suspected highly vulnerable populations, such as children with underlying cardiac or other conditions. Until such data become available and without any data to suggest that there is a therapeutic advantage to consuming caffeine-containing energy drinks, he reiterated his recommendation that children with underlying cardiac conditions not consume such products. The National Poison Data System: An Imperfect System An audience member commented on Dr. Lipshultz’s observation that 51 percent of all reported energy drink poisonings involved children Slope before vs. 180 Non-alcoholic energy drinks after p = 0.001 Alcoholic-related 160 140 Number of Cases 120 Educational initiatives 100 80 60 Banning of alcohol-containing 40 energy drinks 20 0 Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Months (2010–2011) FIGURE 4-3 Poison center energy drink calls over time, 2010–2011. NOTE: Calls made before and after the FDA ban on the sale of alcohol- containing energy drinks and before and after the initiation of public education campaigns about the risks of caffeine-containing energy drinks. SOURCE: Seifert et al., 2013. 

64 CAFFEINE IN FOOD AND DIETARY SUPPLEMENTS younger than 6 years of age. The commenter observed that most of the 717 cases involving children under age 6 are trivial cases in which a par- ent calls and the poison center specialist learns, for example, that the child came into contact with an energy drink by touching a can to their mouths. On the basis of his own calculations, the commenter concluded that 82 percent of the 717 cases were “no effect” cases. The commenter remarked on the difference between a “poisoning” and an “exposure” as reported in the NPDS database. Lipshultz replied that such a calculation is possible because those da- ta are now available and that the NPDS is operating under full disclosure. He noted that these data were initial first-year data and acknowledged that such limited data can be overinterpreted. Thus, he and his colleagues are in the process of analyzing the first 3 years of data. As part of that analysis, they are comparing caffeine exposures associated with energy drinks to caffeine exposures associated with other products. One of the frustrations for him concerning NPDS data is the low level of follow-up to verify facts and to ascertain outcomes. He reiterated, “It’s a very im- perfect system.” With respect to the word “poisoning,” he noted that it is a definition- al term. Because the calls are made to poison centers, they are logged as “poisonings.” For Lipshultz, the key point is that, regardless of which database one examines, whether in Australia, Germany, Ireland, New Zealand, the Unit- ed States, or elsewhere, one observes very similar findings. He reiterated, “There’s not one perfect way to ascertain adverse effects of energy-drink consumption when cases are not being tracked in a systematic way.” He cautioned against relying on consumption data only, without examining NPDS health-consequence data. Consumption data suggest no consump- tion among young children, yet real toxicities are being observed in such children worldwide. In fact, that apparent discrepancy may indicate an even more vulnerable population than children, that is, children with un- derlying medical conditions. Lipshultz referred to earlier presentations on consumption data analyzed by two separate groups and the challenges those groups are having with respect to quantifying consumption. In Lipshultz’s opinion, safety signals are being observed. Calls are being received by a variety of people that relate to temporal associations with these products. The next step is to verify those signals at “the next level of higher-quality science.” When asked about the need for a national registry to track adverse events associated with caffeinated energy drink consumption or the need 

EXPLORING SAFE CAFFEINE EXPOSURE LEVELS 65 for mandated reporting of such events, Lipshultz replied that the NPDS is an imperfect system but that it is the best means currently available for tracking such events in any “semi-systematic” way. That said, he sup- ports efforts to more carefully examine safety issues related to caffeinat- ed energy drink consumption. Following Lipschultz’s presentation, audience members were invited to ask questions of the panelists. This section summarizes the discussion that took place. Most of the questions revolved around how some of the data presented are being interpreted and the gaps in data. Questions About Data on Pregnant Women as a Vulnerable Population Chambers was asked whether there is any evidence for the mutagen- icity of caffeine. She replied that there is none. In response to a question about monotonic linear dose responses, Chambers clarified that the lack of such a response for small for gestation- al age does not necessarily suggest that there is no effect. It could be a non- linear relationship. Smoking, for example, has an effect on craniosynosto- sis at low and moderate doses but not at high doses. Or it could be that there is an effect but that the effect is impacted by other phenomena. Finally, Chambers was asked about compliancy among pregnant women with respect to recommendations for reduced caffeine consump- tion. The commenter referred to a study showing that women who met diagnostic criteria for caffeine dependence were less likely to reduce caf- feine consumption below the Health Canada recommended 300 mg daily. Although not familiar with any data on compliance, Chambers suspected that most pregnant women are not compliant. She wonders whether pregnant women who are addicted to caffeine and to other substances are “particularly recalcitrant” to reducing their caffeine consumption. “That’s an important question,” she said. She also pointed to the need to examine caffeine withdrawal during pregnancy. Some of the symptoms of withdrawal are similar to those of pregnancy. As far as she knows, no one has examined that yet. 

66 CAFFEINE IN FOOD AND DIETARY SUPPLEMENTS REFERENCES Bech, B. H., C. Obel, T. Brink Henrickson, and J. Olsen. 2007. Effect of reducing caffeine intake on birth weight and length of gestation: Randomised con- trolled trial. British Medical Journal. doi:10.1136/bmj.39062.520648.BE. CARE Study Group. 2008. Maternal caffeine intake during pregnancy and risk of fetal growth restriction: A large prospective observational study. British Medical Journal 337:a2332. doi:10.1136/bmj.a2332. Clausson, B., F. Granath, A. Ekbom, S. Lundgren, A. Nordmark, and L. B. Signorello. 2002. American Journal of Epidemiology 155(5):429–436. Dufendach, K. A., J. M. Horner, B. C. Cannon, and M. J. Ackerman. 2012. Congenital type 1 long QT syndrome unmasked by a highly caffeinated en- ergy drink. Heart Rhythm 9:285–288. Gunja, N., and J. A. Brown. 2012. Energy drinks: Health risks and toxicity. Medical Journal of Australia 196:46–49. Munoz, L. M., B. Lonnerdale, C. L. Keen, and K. G. Dewey. 1988. Coffee con- sumption as a factor in iron deficiency anemia among pregnant women and their infants in Costa Rica. American Journal of Clinical Nutrition 48:645– 651. ORNL (Oak Ridge National Laboratory). 2011. Adverse health effects of caf- feine: Review and analysis of recent human and animal research. http://iom.edu/~/media/Files/Activity%20Files/Nutrition/PotentialEffectsof Caffeine/caffeineORNLreport.pdf (accessed February 10, 2014). SAMHSA (Substance Abuse and Mental Health Services Administration). 2013. Drug Abuse Warning Network (DAWN) report: Update on emergency de- partment visits involving energy drinks: A continuing public health concern. http://www.samhsa.gov/data/2k13/DAWN126/sr126-energy-drinks-use.pdf (accessed November 25, 2013). Savitz, D. A., R. L. Chan, A. H. Herring, P. P. Howards, and K. E. Hartmann. 2008. Caffeine and miscarriage risk. Epidemiology 19:55–62. Seifert, S. M., J. L. Schaechter, E. R. Hershorin, and S. E. Lipshultz. 2011. Health effects of energy drinks on children, adolescents, and young adults. Pediatrics 127:511–528. Seifert, S. M., S. A. Seifert, J. L. Schaechter, A. C. Bronstein, B. E. Benson, E. R. Hershorin, K. L. Arheart, V. I. Franco, and S. E. Lipshultz. 2013. An analysis of energy-drink toxicity in the National Poison Data System. Clini- cal Toxicology 51:566–574. Thompson, B., and S. Schiess. 2010. Risk profile: Caffeine in energy drinks and energy shots. Report prepared for New Zealand Food Safety Authority un- der project CFS/09/04. Christchurch, New Zealand: Institute of Environ- mental Science and Research Limited. Weng, X., R. Odouli, and D. K. Li. 2008. Maternal caffeine consumption during pregnancy and the risk of miscarriage: A prospective cohort study. Ameri- can Journal of Obstetrics and Gynecology 198:279.e1–279.e8.