In recent years, sugar-sweetened beverages (SSBs) have accounted for 80% of the rise in sugar consumption worldwide. Furthermore, SSBs are the largest single source of added sugar consumption [1
]. SSBs containing high sugar content may cause insulin secretion, drive reactive hypoglycemia and stimulate an increase in epinephrine, which activates hyperactivity disorder behaviors [3
]. Additionally, SSBs often contain other additives, such as artificial food colorings (AFCs) and preservatives that could affect children’s behavior [4
Increasing evidence from animal studies clearly indicates that sugar consumption can cause deficits in cognitive and behavioral functions [8
]. While many clinical studies on obesity and metabolic disturbances resulting from SSB consumption have been conducted, few studies have examined how SSBs affect development and mental health outcomes [8
]. Diet factors, such as sugar consumption, have been linked to an increased risk of attention deficit/hyperactivity disorder (ADHD) [20
]. However, this reported effect is controversial [20
]. One Norwegian population-based study of 15- to 16-year-old students adopted checklists and questionnaires to assess metal health problems and showed associations between high SSB consumption levels and hyperactivity [20
]. A US study of middle school students found that those who consumed SSBs and those who consumed energy drinks had 1.14 and 1.66 odds, respectively, of having ADHD [21
]. In a study of disruptive patterns in eating behaviors, Ptacek et al. revealed that male children diagnosed with ADHD exhibited increased sweetened beverage consumption [19
]. Their findings suggested that younger children might be susceptible to the adverse effects of SSBs. By contrast, Kim and Chang conducted a Korean study of 107 school-aged children, with only 8.5% categorized as having ADHD, and found no significant association between the consumption of simple sugars, including those in sweets and SSBs, and an increased risk of ADHD [25
Taiwan has the highest density of chain bubble tea shops and 24-h convenience stores in the world, and both types of stores sell SSBs [26
]. At the end of 2015, Taiwan had 16,836 bubble tea shops and 10,131 chain convenience stores for a total population of 23.5 million, indicating an average of one store for every 870 people [28
]. Researchers have measured the amount of sugar, mainly high-fructose corn syrup (HFCS), added to SSBs and found that it ranged from 22 g to 68 g for a 750 mL serving size [27
]. One cross-sectional study evaluated the association between adolescent obesity-related health outcomes and SSB consumption in Taiwan and found that 87.7% of adolescents drank SSBs daily and 25.1% consumed more than 500 mL of SSBs per day [26
]. Bubble tea shops have recently arrived in the United Kingdom, Europe, Asia, Australia, Canada and the U.S. [29
]. Attention should be paid to the intake of these drinks not only because excessive sugar consumption induces metabolic syndrome, which results in excess and imbalanced caloric intake, but also because of its potential association with mental health problems [9
], especially among school-aged children.
ADHD is a chronic disease and one of the most common behavioral disorders among children [34
]. The prevalence of ADHD in children has been reported to range from 7.5% to 9.9% in Taiwan and from 5.9% to 7.1% worldwide [36
]. ADHD impairs learning and social development; the condition typically develops during the years before school and often persists into adulthood [38
]. The importance of ADHD prevention and intervention is highlighted by the high prevalence and psychosocial impacts of the disorder. The etiology of ADHD is complex and involves genetic, dietary, and environmental factors that have strong genetic associations, with >70% heritability [40
Several studies have assessed the association between SSB consumption and childhood ADHD cross-sectionally [20
]. These studies have measured mental health problems through self-report or interviews with teachers and/or parents. To avoid discordance with the standard nosology of ADHD and prevent misclassification, we ensured that our ADHD cases were diagnosed by doctors based on the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, revised criteria (DSM-IV-TR). Only few doctor-diagnosed case-control studies have been reported; however, these studies have not controlled potential confounding variables [19
]. This is the first case-control study to analyze the association between school-aged children’s SSB consumption and ADHD diagnosis by a board-certificated doctor while controlling for covariates. This study hypothesized that children with ADHD drink more SSBs.
2. Materials and Methods
2.1. Study Participants and Recruitment
The study protocol was approved by the Taipei Veterans General Hospital and Taipei City Hospital institutional review boards. The project identification codes were 97-01-52A, TCHIRB-1010216 and TCHIRB-1030510 (approved on 4 March 2008, 1 July 2012 and 26 August 2014, respectively). Written informed consent was obtained from the participants’ parents or guardians. All participants provided their oral or written assent. This study was conducted in accordance with the revised guidelines of the Declaration of Helsinki. We recruited 4- to 15-year-old subjects from outpatient waiting rooms in both hospitals. In this study, the cases were defined as children with ADHD, as identified by board-certificated pediatricians or psychiatrists after at least a three-visit clinical assessment. Children consecutively admitted for initial and follow-up ADHD treatments were also recruited as cases during the study period. Children with neurological deficits or mental retardation were excluded. The ADHD diagnosis was made in accordance with the DSM-IV-TR criteria [43
We recruited control subjects by randomly selecting normal 4- to 15-year-old children who visited either of the hospitals for reasons unrelated to ADHD during the same study period, and the same exclusion criteria as those applied to the cases were used. We did not attempt to match cases and controls on gender; however, we matched the two groups on age. Controls were screened for the absence of ADHD symptoms by teachers and parent(s), who assessed children’s behaviors in the classroom and at home, respectively, according to the Chinese version of the Swanson, Nolan and Pelham, Fourth Revision (SNAP-IV) questionnaire [44
]. The rating results were evaluated by pediatricians to confirm the absence of ADHD symptoms. The SNAP-IV Teacher and Parent Rating Scale, which directly adopted DSM-IV symptoms and used the same format, was translated into Chinese and was found to be a reliable and valid tool for screening for ADHD in clinical and research settings in Taiwan [46
]. The 26-item SNAP-IV questionnaire was based on a four-point (0–3) rating scale and consisted of the DSM-IV ADHD criteria for inattention (items 1–9) and hyperactivity/impulsivity (items 10–18) as well as criteria for oppositional defiant disorder symptoms (items 19–26). If more than six of the nine inattention or hyperactivity/impulsivity items were given a score of a 2 (quite a bit) or 3 (very much) on either the parent’s or teacher’s form, then the children were defined as at risk of ADHD; otherwise, children were rated as non-ADHD [48
]. Children with scores indicating a risk of ADHD were referred to pediatricians or psychiatrists for further confirmation. All cases also received a SNAP-IV score at the initial visit.
We recruited a total of three hundred and thirty-two subjects, one hundred and seventy-three ADHD subjects (n = 173) and one hundred and fifty-nine normal controls (n = 159). The response rates were 63.6% and 58.9% in the case and control groups, respectively. In examining the association between the intake of SSBs and ADHD, we investigated the following covariates: family factors, maternal lifestyle, participants’ dietary habits, blood lead levels (BLLs) and gene polymorphisms. In this study, participating subjects’ demographic features, dietary habits, mother’s lifestyle during pregnancy, and family history of nervous system diseases were collected via questionnaires. For the measurement of BLLs and analysis of gene polymorphisms, subjects provided a blood (or saliva) sample during the clinic visit.
2.2. Dietary Habits, Consumption of SSBs and Intake of Sugar and Calories
We collected information on each child’s dietary habits and SSB consumption. A trained interviewer administered a 30-min questionnaire to the mother or other caretaker. The average weekly consumption of meat (including poultry and livestock), milk, eggs, fish, shellfish and other types of seafood was gathered for the previous month. The questionnaire also included questions that estimated the average weekly consumption of vegetable (light-colored vegetables, such as lettuce, cabbage, and bean sprouts, and dark-colored vegetables, such as spinach and broccoli) and fruit servings. The weekly serving was calculated by multiplying the daily servings with frequency of consumption per week over the previous month. One serving of meat, fish, shellfish or other types of seafood was approximately 35 g (for meat and fish, this was approximately the size of one’s palm). One cup of milk (approximately 250 mL) and one egg or six quail eggs represented a serving of milk and eggs, respectively. One cup of raw leafy vegetables (approximately the size of the participant’s fist) or a half cup of other vegetables represented a serving of vegetables. One serving of fruit was a medium-sized fruit (medium was defined as the size of a baseball) or a half cup of chopped fruit. One serving of SSBs contained 600 mL. Parents or caregivers were asked separate questions about how many servings of milk tea, juice or fruit-flavored drinks, Yakult drinks, and other sugar-sweetened drinks the participants normally consumed. The milk tea referenced in this study is also known as boba milk tea, bubble or pearl tea [49
] and is a Taiwanese tea-based drink. It contains a tea base that is mixed/shaken with milk or fruit, and chewy tapioca balls or jellies are often added. Milk tea drinks are part of the larger group of SSB because these beverages are typically HFCS sweetened [50
]. Yakult drink is a probiotic dairy product made by fermenting a mixture of milk with a special strain of the bacterium Lactobacillus casei
]. This beverage is very popular in Taiwan because it contains probiotics and, thus, is likely considered a healthy drink. However, it also contains high content of sugar (approximately 12 grams for every 100 g of Yakult drink) and is classified as a SSB [52
]. When questionnaire items were incomplete, we contacted the mothers or caretakers via telephone to obtain the missing information.
Sugar ingested from SSBs was calculated by multiplying the servings by the sugar content in each type of SSB based on a study of the Ministry of Health and Welfare (MOHW) in Taiwan [52
]. In addition, the Taiwan Food and Drug Administration (TFDA) Nutrients Database was used to analyze daily caloric intake [53
2.3. Measurement of BLLs and Gene Polymorphism Analysis
Peripheral blood was drawn using a syringe or venoclysis needle and then sealed in a heparin-containing vacuum tube and immediately transported at 4 °C to the laboratory. If individuals were not able to supply a blood sample, DNA was extracted from a saliva sample. Saliva was spat into the Saliva DNA Collection and Preservation Kit (Norgen Biotek Corporation, Thorold, ON, Canada) and stored at room temperature until analysis. The sample preparation and analysis of BLLs and gene polymorphism are described elsewhere [42
]. The BLLs were measured using inductively coupled plasma-mass spectrometry (Thermo Scientific, Waltham, MA, USA). Trace Elements Serum L-2 (Seronorm™, Billingstad, Norway) was used to verify the precision and accuracy of the analytical measurements. The limit of detection (LOD) for lead was 0.001 μg/dL. Regarding gene polymorphism, 10 tagged SNPs (rs7395429, rs3758653, rs11246228, rs752306, rs6347, rs2975292, rs37022, rs40358, rs10040882, and rs464049) of DRD4 and DAT1 were identified and analyzed.
Risk factors potentially associated with ADHD were examined. Predictors were chosen according to their association with ADHD in previous studies. The following variables were considered covariates: gender, body weight, child’s age, maternal age at childbirth, gestational age at birth (<37 weeks or ≥37 weeks), parity (primiparous or multiparous) [55
], birth order (1st, 2nd and 3rd and above), parent’s education (high school education and below or college or advanced training), maternal history of still or dead birth (yes or no), maternal smoking during pregnancy (yes or no) [57
], maternal alcohol consumption during pregnancy (yes or no) [59
], and maternal chronic disease during pregnancy (yes or no). In addition to examining these risk factors, we included family history of nervous system diseases in this study. The nervous system diseases listed in the questionnaire included Parkinson’s disease, Alzheimer’s disease, ADHD, mental retardation, cerebral palsy, autism, epilepsy, developmental delay, multiple sclerosis and peripheral neuromuscular disease, and the family members considered for this list were the subjects’ grandparents, parents and siblings. The variables were obtained from clinical records or questionnaires completed by the mothers or caretakers.
2.5. Statistical and Probabilistic Analysis
SPSS Version 17.0 (SPSS, Chicago, IL, USA) was used for the statistical analysis. We calculated descriptive statistics based on sample sizes and percentages. Next, we assessed the significance of differences between the case and control groups. A 2-sided nonparametric statistical Mann-Whitney U test was used for consecutive data, and a chi-squared test or Fisher’s exact test was adopted for categorical data when appropriate [61
]. Then, we performed logistic regression analysis with ADHD and the consumption of SSBs. Statistical significance was set at p
< 0.05. The covariates related to ADHD at p
< 0.05 were controlled in the multivariate analyses. A Kruskal-Wallis test was used to analyze the relationship between ADHD severity and the increased intake of SSBs.
Monte Carlo (MC) simulation, which is a probabilistic analysis, was performed to quantify the theoretical exposure dose range of sugar, AFCs and preservatives. Probability density functions for each parameter were log-normally distributed. Distributions of exposed concentrations and population exposures were estimated by a MC simulation with 5000 replications. Microsoft Excel™ 2010 (Microsoft Inc., Redmond, WA, USA) and Oracle Crystal Ball, Fusion Edition, Release 220.127.116.11.000 software (Oracle Corporation, Redwood City, CA, USA) were used for MC simulation.