Potentially addictive properties of sugar-sweetened beverages among adolescents

Accepted Manuscript Potentially addictive properties of sugar-sweetened beverages among adolescents Jennifer Falbe, Hannah R. Thompson, Anisha Patel, Kristine A. Madsen PII:

S0195-6663(18)30174-0

DOI:

https://doi.org/10.1016/j.appet.2018.10.032

Reference:

APPET 4073

To appear in:

Appetite

Received Date: 8 February 2018 Revised Date:

24 October 2018

Accepted Date: 27 October 2018

Please cite this article as: Falbe J., Thompson H.R., Patel A. & Madsen K.A., Potentially addictive properties of sugar-sweetened beverages among adolescents, Appetite (2018), doi: https:// doi.org/10.1016/j.appet.2018.10.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Abstract Sugar-sweetened beverages (SSBs) increase risk of cardiometabolic disease. Young people consume the largest amounts of SSBs and have experienced the greatest relative gains

RI PT

in obesity in the past several decades. There is evidence of addictive properties of both caffeine and sugar, the primary ingredients in SSBs, but little research into such properties of SSBs in naturally occurring consumption patterns. Thus, in this exploratory study, we sought to examine

SC

potentially addictive properties of SSBs during a 3-day SSB cessation intervention in overweight and obese adolescents who typically consume ≥3 SSBs daily. Participants (n=25) were aged 13-18 years, mostly female (72%), and African American (56%) or Hispanic (16%) with a

M AN U

BMI≥95th %tile (76%). Withdrawal symptoms and SSB craving were assessed approximately 1week apart, during both regular SSB consumption and a 3-day period of SSB cessation in which participants were instructed to drink only plain milk and water. During SSB cessation, adolescents reported increased SSB cravings and headache and decreased motivation,

TE D

contentment, ability to concentrate, and overall well-being (uncorrected Ps<0.05). After controlling the false discovery rate, changes in motivation, craving, and well-being remained significant (corrected Ps<0.05). Using 24-hr recalls and drink journals, participants reported

EP

lower total daily consumption of sugar (-80 g) and added sugar (-16 g) (Ps<0.001) during cessation. This study provides preliminary evidence of withdrawal symptoms and increased

AC C

SSB cravings during cessation in a diverse population of overweight or obese adolescents.

Keywords: obesity, sugar-sweetened beverages, addiction, withdrawal, craving, adolescence

ACCEPTED MANUSCRIPT

Potentially Addictive Properties of Sugar-Sweetened Beverages among Adolescents Jennifer Falbe,a* ScD, MPH, Hannah R. Thompson,b PhD, MPH, Anisha Patel,c MD, MSPH, MSHS, Kristine A. Madsen,b MD, MPH a

University of California, Berkeley Community Health Sciences, School of Public Health 50 University Hall, Berkeley, CA 94720 (510) 664-9070 [email protected] & [email protected] c

SC

b

RI PT

University of California, Davis Human Development and Family Studies, Department of Human Ecology One Shields Ave, Davis, CA 95616 (415) 779-5363 [email protected]

M AN U

Stanford University Division of General Pediatrics, Lucile Salter Packard Children’s Hospital 1265 Welch Rd, MSOB X240, Mailcode 5459, Stanford, CA 94305 (650) 725-8314 [email protected] *Corresponding author

EP

TE D

Clinicaltrials.gov 12-08545

AC C

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

1

ACCEPTED MANUSCRIPT

27 28 29

Introduction

30

public health priority (Lobstein, 2017). This is due to strong evidence that SSB consumption

31

increases risk of obesity and cardiometabolic disease (Hu, 2013; Malik, Popkin, Bray, Despres,

32

& Hu, 2010) and the fact that SSBs are the largest contributor to added sugar in the diet across

33

multiple countries (Block, 2004; Brisbois, Marsden, Anderson, & Sievenpiper, 2014; Reedy &

34

Krebs-Smith, 2010; Sanchez-Pimienta, Batis, Lutter, & Rivera, 2016). Excessive SSB

35

consumption is a particular concern among adolescents, who consume more SSBs than any

36

other age group in the U.S. (Kit, Fakhouri, Park, Nielsen, & Ogden, 2013). Adolescents also

37

experienced the greatest relative increase in obesity prevalence, which quadrupled over the

38

past several decades (U.S. Department of Health and Human Services, 2014).

RI PT

SC

M AN U

39

Reducing consumption of sugar-sweetened beverages (SSBs) has become a global

Compared to other foods, SSBs have unique attributes that may promote overconsumption of calories. First, compared to calories from solid food, liquid calories appear

41

to be less satiating and incompletely compensated for by eating fewer calories later in the day

42

(Almiron-Roig et al., 2013; Mattes, 1996). Second, SSBs contain primary ingredients—caffeine

43

and sugar—with evidence of addictive potential. Caffeine withdrawal is an established disorder

44

recognized by the Diagnostic and Statistical Manual of Mental Disorders (DSM-V) and the

45

International Statistical Classification of Diseases and Related Health Problems (ICD-10)

46

(American Psychiatric Association, 2013; Meredith, Juliano, Hughes, & Griffiths, 2013; World

47

Health Organization, 1992). The continued use of caffeine despite harm (e.g., insomnia,

48

hypertension) has been documented, and caffeine dependence syndrome is recognized by ICD-

49

10 (Meredith et al., 2013; World Health Organization, 1992).

50

AC C

EP

TE D

40

In contrast, research on the addictive potential of sugar is a relatively new but

51

burgeoning area. The majority of such studies have used animal models, which reveal parallels

52

between added sugars and substances of abuse in bingeing, craving, tolerance, and withdrawal

53

(DiNicolantonio, O'Keefe, & Wilson, 2017). Also, human neuroimaging studies show that high 2

ACCEPTED MANUSCRIPT

sugar intake activates similar neural circuitry and reward systems as substances of abuse

55

(Smith & Robbins, 2013). Although there is an evolutionary rationale for liking sweetness to

56

encourage consumption of nutrient- and energy-containing foods (e.g., mother’s milk or fruits),

57

added sugars are distinct in that they are extracted, concentrated, and separated from the

58

nutrients in their original forms (e.g., beets or corn). This extraction and concentration process is

59

akin to that of addictive substances like opium from poppies or ethanol from fruit (Davis, 2014).

RI PT

54

Adolescence is a particularly susceptible period for addiction, when still-developing

61

brains are highly sensitive to substances and when risk-taking is more likely due to a faster

62

developing limbic system (which supports emotion) than prefrontal cortex (site of executive

63

control) (IOM (Institute of Medicine) and NRC (National Research Council), 2011). Although still

64

minors, the transition to the teenage years is marked by greater autonomy in food purchasing

65

and more frequent exposure to advertising for SSBs (Powell, Szczypka, & Chaloupka, 2010). In

66

2009, carbonated beverages were the food category with the highest marketing expenditures for

67

teens, totaling nearly $400 million (Federal Trade Commission, 2012).

M AN U

TE D

68

SC

60

Although there is evidence of addictive potential of ingredients in SSBs, there is a dearth of research on such properties of SSBs themselves, particularly among adolescents. However,

70

the five-fold increase in SSB consumption since the 1950’s (Economic Research Service, 2017)

71

coupled with still high levels of consumption (Rosinger, Herrick, Gahche, & Park, 2017) despite

72

increased publicity of the associated health consequences, is consistent with the “use despite

73

harm” manifestation of addiction. Given the heavy marketing, consumption, and health

74

consequences of SSBs among youth, determining the extent to which SSBs exhibit addictive

75

properties could have important regulatory, communications, and obesity treatment implications.

76

Thus, in this exploratory study, we sought to examine the extent to which SSBs exhibit

77

potentially addictive properties, namely symptoms of withdrawal and craving, among overweight

78

and obese adolescents who consume high amounts of SSBs.

AC C

EP

69

79 3

ACCEPTED MANUSCRIPT

80

Methods

81

Study Design

82

We employed a single arm intervention trial in which pre-post measurements were taken approximately 1-week apart. During the first 5 days, participants were instructed to continue

84

their usual beverage consumption, followed by a 3-day period of SSB cessation (Figure 1). A 3-

85

day period was selected to overlap with peak severity for symptoms of caffeine, alcohol, and

86

nicotine withdrawal (Hughes, 2007; Juliano & Griffiths, 2004; Kosten & O'Connor, 2003;

87

Shiffman et al., 2006). This study was approved by the institutional review boards at the UC San

88

Francisco, UC Berkeley, and Children’s Hospital and Research Center Oakland.

89 90 91

Participants and Recruitment

M AN U

SC

RI PT

83

Eligible participants were aged 13-18 years who consumed ≥3 SSBs daily, had a BMI≥85th percentile for age and sex, spoke English, and had a parent/guardian who spoke

93

English or Spanish. Eligibility criteria included BMI≥85th percentile due to evidence that those

94

with high BMIs may be more susceptible than their normal BMI peers to the potentially additive

95

properties of foods/beverages (Epel et al., 2014; Schulte & Gearhardt, 2017). Teens were

96

ineligible if they were pregnant, nursing, or undergoing pharmacological treatment for mental

97

health. The study was advertised through parent listservs and by flyers at local hospitals, clinics,

98

health fairs, and high schools. Participants were screened for eligibility and recruited over the

99

phone. Figure 2 shows participant flow. Of 28 eligible participants identified, 25 enrolled in the

101 102

EP

AC C

100

TE D

92

study. For all participants, we obtained assent and informed parental consent.

Intervention

103

The purpose of the study was described to participants as exploring “how soda affects

104

teenagers’ health” and learning “how teenagers react when they stop drinking sugary drinks.”

105

Participants were instructed to drink their usual beverages over a 5-day period (Phase 1: Usual 4

ACCEPTED MANUSCRIPT

SSB intake) and then to drink only plain water or plain milk over the next 3-day period (Phase 2:

107

SSB cessation). Participants received these instructions verbally, in writing, and graphically

108

through a calendar. Researchers reminded participants of these instructions during data

109

collection visits at a local children’s hospital. On the last day of usual intake and the first day of

110

cessation (days 5 and 6), a researcher called and/or texted participants to remind them to drink

111

only plain milk and water on days 6-8. Participants were reimbursed for travel and received up

112

to $160 for participation.

SC

RI PT

106

113

115

Measures

M AN U

114

The primary outcomes were symptoms of withdrawal, measured by (1) a modified scale of mood, behavior, and physical symptoms (Griffiths et al., 1990) for identifying symptoms of

117

caffeine withdrawal, and (2) dimensions of mood, including SSB craving, assessed five times

118

per day via check-in questions. We focused on withdrawal and craving, common manifestations

119

of addiction / substance use disorders, because (1) withdrawal perpetuates continued use of a

120

substance to alleviate symptoms, and (2) craving predicts frequency of substance use and

121

relapse (American Psychiatric Association, 2013; D. H. Epstein, Marrone, Heishman,

122

Schmittner, & Preston, 2010; Oslin, Cary, Slaymaker, Colleran, & Blow, 2009). Figure 1 shows

123

the timing of all measures. Although we drew upon an assessment tool for caffeine withdrawal

124

and craving, many of these symptoms may also occur in response to cessation from high sugar

125

intake, as evidenced by animal models (DiNicolantonio et al., 2017).

EP

AC C

126

TE D

116

We used 11 items sensitive to caffeine cessation from a scale by Griffiths and

127

colleagues (1990). This scale asked, “How do you feel right now?” for each dimension: having a

128

headache, feeling energetic, awake/alert, tired, good overall, drowsy/sleepy, content/satisfied,

129

motivated to do work, outgoing, self-confident, and able to concentrate. Wording was modified

130

for some items to use terminology more familiar to adolescents (i.e., “outgoing” instead of

131

“social disposition,” “tired” instead of “lethargy/fatigue,” “feel good overall” instead of “well5

ACCEPTED MANUSCRIPT

being,” and “awake/alert” instead of “alert/attentive”). We used a 5- (instead of 4-) point

133

response scale to capture greater variability: 0 (not at all), 1 (a little), 2 (somewhat), 3

134

(moderately), and 4 (very much). Participants completed the scale on the first two days of usual

135

SSB intake and one-week later, during each cessation day. We calculated total scale score by

136

reverse coding positive items (e.g., “I feel good overall”) before summing, such that higher

137

scores indicated greater severity of withdrawal. Responses on each day and responses

138

averaged across days of the cessation phase were compared with responses averaged across

139

days of the usual intake phase.

SC

RI PT

132

To assess craving and the times at which certain withdrawal symptoms may be most

141

severe, participants completed four check-in questions five times daily, either by mobile text

142

message or a check-in diary. These questions assessed SSB cravings as well as perceptions of

143

overall well-being, tiredness, and irritability. Griffiths et al. (1990) identified craving and irritability

144

as also sensitive to caffeine withdrawal(Griffiths et al., 1990). Check-in questions were

145

completed at 10am, 12pm, 4pm, 6pm, and 9pm on days 2-3 of the usual SSB phase and on

146

each day of SSB cessation. Each of the following questions used a 10-point response scale:

147

How do you feel? 1 (really bad) – 10 (really good); How much do you want [favorite SSB] right

148

now? 1 (I don’t want at all) – (I really want); How tired are you? 1 (not tired at all) – 10 (really

149

tired); and How irritable are you? 1 (not at all irritable) – 10 (really irritable). During the usual

150

SSB phase, responses at each time point were averaged across days (e.g., response at 10am

151

on day 2 was averaged with the response at 10am on day 3) and served as the comparison for

152

each day of SSB cessation responses.

TE D

EP

AC C

153

M AN U

140

Additional outcomes assessed (see Figure 1 for timing) included dietary intake, salivary

154

caffeine, and secondary outcomes of attention, impulsivity, and the relative reinforcing value of

155

SSBs. Through the Automated Self-Administered 24-Hour (ASA24®) dietary assessment tool,

156

participants recalled the types and amounts of foods and beverages consumed the previous

157

day. We additionally assessed beverage intake using a self-administered daily beverage recall 6

ACCEPTED MANUSCRIPT

(i.e., “drink journal”), which probed for how often and in what quantities the following were

159

consumed: plain milk (whole, 2%, 1%, and skim); flavored milk; plain non-dairy milk; flavored

160

non-dairy milk; horchata; smoothies; 100% fruit juice; tamarindo, jamaica, or aguas frescas;

161

fruit-flavored drinks; sports drinks; energy drinks; tap water; bottled water; and caffeinated and

162

uncaffeinated regular soda, diet soda, unsweetened coffee or tea, and coffee or tea with added

163

sugar. Drink journal beverages that were not reported on ASA24, or that were reported in larger

164

quantities, were coded for nutrients using the USDA’s Food and Nutrient Database for Dietary

165

Studies and the Food Patterns Equivalents Database (U.S. Department of Agriculture, 2016).

166

The corresponding differences in nutrients and quantities were added to those from ASA24 and

167

averaged across days for each study phase.

SC

M AN U

168

RI PT

158

Saliva samples were used to assess caffeine intake and provided an indicator of adherence among caffeinated SSB consumers, which was communicated to participants.

170

Participants provided 5mL samples after abstaining from food and drink for 30 minutes, rinsing

171

their mouth, and removing lipstick or balm. Samples were stored at -20°C and analyzed by the

172

UCSF Clinical Pharmacology Laboratory (San Francisco, CA).

173

TE D

169

Computer tasks assessed attention via the Stroop tasks for color and word naming (MacLeod, 1991; Stroop, 1935) (Inquisit 3, Millisecond Software, LLC); impulsivity via delay and

175

probabilistic discounting (Millisecond Software, LLC) (Mazur, 1987; Rachlin, Raineri, & Cross,

176

1991); and the relative reinforcing value of SSBs (RRVSSBs) via the Behavioral Choice Computer

177

Task (L. H. Epstein, Carr, Lin, & Fletcher, 2011; Goldfield, Epstein, Davidson, & Saad, 2005).

178

Other measures taken at the start of the study included age, sex, race/ethnicity,

AC C

EP

174

179

anthropometrics, parental education, living arrangement, and fasting serum glucose, insulin,

180

hemoglobin A1c, lipids, adiponectin, and leptin. The fasting blood draw occurred during

181

scheduled morning appointments on days 2-5 of the usual intake phase.

182 183

Analytic sample 7

ACCEPTED MANUSCRIPT

184

All 25 participants provided responses to the 11-item withdrawal scale and dietary data. Three had missing values for one withdrawal scale item, which was imputed from the mean of

186

the individual’s other scale items. Twenty-four participants provided responses to check-in

187

questions during both phases, which was 84% non-missing at each time point through 12 pm on

188

day 8 (last day of cessation). Thereafter, check-in data had missingness ≥40% and thus were

189

not analyzed. Computer task data were available from 23 participants, but for delay/probability

190

discounting, only 15 participants had non-outlying data that could be fit using the hyperbolic

191

equations (Mazur, 1987; Rachlin et al., 1991). Complete case analyses were performed on

192

these secondary outcomes.

M AN U

193 194 195

SC

RI PT

185

Analysis

Internal consistency of the withdrawal scale was assessed by Cronbach’s alpha. Change in individual item scores and total scale scores from the withdrawal scale were examined using

197

paired t-tests. Daily check-in responses at each time of day during the cessation phase were

198

compared to responses at the same times during the usual intake phase using post hoc

199

contrasts after fitting linear regression models with indicators for each time point and robust

200

standard errors (Huber-White sandwich). The remaining outcomes were compared between

201

phases using paired t-tests or Wilcoxon signed-rank tests in the event of non-normality. In this

202

exploratory study, results of the above statistical tests are presented both uncorrected for

203

multiplicity and corrected after applying the Benjamini-Hochberg procedure to control the false

204

discovery rate to <5% (Benjamini & Hochberg, 1995). Although the sample size limited our

205

ability to test for separate or interactive effects of caffeine and sugar, in additional exploratory

206

analyses uncorrected for multiplicity, we restricted analyses to those who reported consuming

207

low levels of caffeine from beverages (<55mg/day) at baseline, less than the amount in a 20 fl

208

oz bottle of cola. Analyses were conducted in STATA13 (StataCorp LP, College Station, TX).

AC C

EP

TE D

196

209 8

ACCEPTED MANUSCRIPT

210

Results Table 1 shows participant baseline characteristics assessed during the usual SSB intake

211

phase. The majority of teens were female (72%), non-Hispanic African American (56%), with a

213

BMI≥95th %tile (76%), and a mother with greater than a high school education (64%). Fewer

214

than half lived with both parents (44%) and had a father with greater than a high school

215

education (45%). Mean consumption of SSBs was 32 ± 22 fl oz/day with soda accounting for

216

the largest share (42%) of SSBs, followed by fruit-flavored drinks (35%). However, the majority

217

of participants (n=19, 76%) reported consuming low amounts of daily caffeine (<55 mg). Only

218

five participants reported consuming no caffeine from beverages at baseline. Mean fasting

219

hemoglobin A1c level was within the diagnostic range for pre-diabetes.

M AN U

SC

RI PT

212

220 221

Primary Outcome

Cronbach’s alpha for the modified withdrawal scale (Griffiths et al., 1990) ranged from

222

0.82 to 0.91 across days of administration, indicating good to excellent internal consistency.

224

Reported symptoms from this scale (Table 2) included increased ratings of headache and

225

decreased feelings of contentment/satisfaction, motivation to do work, and ability to concentrate

226

during the SSB cessation phase, which were most pronounced on the first two days of

227

cessation (days 6 and 7). After applying the Benjamini–Hochberg procedure, only feeling

228

motivated to do work significantly decreased during cessation on day 2 (corrected P<0.05).

229

Supplementary Table S1 contains fully reported statistics. Before correction, there was also a

230

borderline significant decrease in reported wellbeing and increase in the total 11-item scale

231

score on the first and second day of cessation (Table S1).

AC C

EP

TE D

223

In additional exploratory analyses examining these symptoms among low-caffeine

232 233

consumers (n=19), the magnitude of change was similar for headache (+0.4 ± 0.9) and

234

motivation (-0.5 ± 1.0). Change in ability to concentrate in this subgroup was attenuated (-0.3 ±

235

1.0). 9

ACCEPTED MANUSCRIPT

236

Symptoms of withdrawal assessed by the daily check-in questions five times daily are shown in Figure 3 (full statistics reported in Supplemental Table S2). When uncorrected for

238

multiplicity, ratings of overall wellbeing and SSB craving appeared significantly lower at several

239

time points during the SSB cessation phase. After the Benjamini–Hochberg procedure, overall

240

wellbeing at 12pm the second day of cessation and SSB cravings at 9pm the first day of

241

cessation differed significantly from those times during the usual intake phase (corrected

242

Ps<0.05). There were no significant differences between phases for feeling tired or irritable.

243

Results were similar when restricted to those who reported consuming low levels of caffeine

244

from beverages at baseline (not shown).

M AN U

245 246 247

SC

RI PT

237

Other Outcomes

Table 3 shows changes in dietary outcomes reported on 24-hr recalls and drink journals. Compared to the usual SSB intake phase, during the cessation phase, participants reported

249

decreased daily consumption of SSBs and sugar, energy, and caffeine from beverages without

250

compensatory increases in sugar from food. Consumption of total daily sugar and added sugar

251

from all sources significantly decreased.

Most (n=16) reported only consuming water and plain milk during the cessation phase,

EP

252

TE D

248

but 9 participants consumed some SSBs, ranging from 1-8 fl oz among five participants, 14-15 fl

254

oz among three participants, and 42 fl oz in one participant. Flavored milk accounted for 67% of

255

non-adherent SSB consumption. Salivary caffeine, another indicator of adherence, significantly

256

decreased by 63% during the SSB cessation phase (-0.22 µg/ml; P=0.02) (Table 3) to levels

257

similar to those in other studies among participants undergoing caffeine cessation (James &

258

Gregg, 2004; James, Gregg, Kane, & Harte, 2005). Non-adherence was likely in two

259

participants, whose salivary caffeine levels increased from the usual intake phase to 0.51 and

260

0.96 µg/ml during cessation. The individuals described here were included in all analyses.

AC C

253

10

ACCEPTED MANUSCRIPT

261 262

Lastly, there were no significant changes in Stroop tasks, delay/probabilistic discounting, or RRVssb (all P-values>0.05).

263 Discussion

RI PT

264

In this exploratory study of the potentially addictive properties of SSBs , we detected an

266

increase in symptoms of craving for SSBs and withdrawal among overweight and obese teens

267

during a 3-day period of SSB cessation. The teens, who were regular consumers of SSBs

268

before cessation, reported the following specific withdrawal symptoms during SSB cessation:

269

increased headache; decreased motivation to do work, contentment, and ability to concentrate;

270

and lower ratings of overall wellbeing. After controlling the false discover rate, increased SSB

271

craving and decreased feelings of motivation and overall wellbeing remained significant.

272

Withdrawal and craving are key manifestations in response to other legal addictive substances

273

with large public impacts, like tobacco and alcohol, and like these substances, SSBs are

274

consumed excessively despite health harms (Bray & Popkin, 2014). This is the first study we

275

are aware of to examine potentially addictive properties of SSBs during a cessation intervention

276

in free-living participants.

M AN U

TE D

In this study, decreased well-being and increased cravings were most pronounced in the

EP

277

SC

265

afternoon and evening, perhaps because this is when youth typically consume SSBs or are

279

exposed to SSB cues through peer consumption or advertising. This timing could also result

280

from circadian rhythm or timing of meals. Other withdrawal symptoms that were measured only

281

once per day—headache and decreased contentment, motivation, and concentration—were

282

more pronounced on days 1 and 2 of cessation than on day 3 of cessation. Other findings

283

included clinically meaningful reductions in total sugar and sugar from beverages, the latter of

284

which was not compensated for by increased sugar from foods. This finding is consistent with

285

trials demonstrating decreased total sugar intake and/or weight loss from SSB replacement (de

286

Ruyter, Olthof, Seidell, & Katan, 2012; Ebbeling et al., 2012).

AC C

278

11

ACCEPTED MANUSCRIPT

287

The symptoms we observed in this study overlap with those observed in studies of withdrawal from caffeine (De Biasi & Salas, 2008; Juliano & Griffiths, 2004), the only

289

psychoactive substance that can be legally purchased by children (Bernstein et al., 1998).

290

However, in the current study, mean daily caffeine consumption was low (50 mg/d from

291

beverages) compared to studies of caffeine cessation (100-900 mg/d) (De Biasi & Salas, 2008;

292

Juliano & Griffiths, 2004). Yet, even low levels of caffeine may promote SSB consumption. A

293

recent randomized controlled trial blindly assigned 99 adults to caffeinated SSBs (66 mg /

294

110ml, similar to common cola brands) or non-caffeinated SSBs, ad libitum for 6-weeks (Keast,

295

Swinburn, Sayompark, Whitelock, & Riddell, 2015). The caffeinated group consumed

296

significantly more SSBs and reported greater “liking” of SSBs than the non-caffeinated group.

297

Therefore, caffeine, which is added to soda by manufacturers, may be contributing to high SSB

298

consumption though its psychoactive properties. Manufacturers claim caffeine is added as a

299

flavor enhancer, yet multiple studies have found evidence against flavor activity at

300

concentrations common to SSBs (Keast & Riddell, 2007; Keast et al., 2015).

SC

M AN U

TE D

301

RI PT

288

Although this study was not powered to compare withdrawal symptoms between caffeinated and non-caffeinated SSBs consumers, when we restricted to those usually

303

consuming no/low caffeine from beverages, results were similar for headache, motivation,

304

overall wellbeing, and craving. Therefore, we hypothesize that beyond caffeine, sugar content

305

may have contributed to the observed symptoms, consistent with human neuorimaging studies

306

of sugar consumption (Jastreboff et al., 2016; Smith & Robbins, 2013). Also, in a recent fMRI

307

study examining adolescent brain response to SSBs, overweight teens were assigned to

308

consume 1 ml of water and 1 ml of their preferred soda (Feldstein Ewing, 2016). Compared to

309

water, soda resulted in brain activation in “...the regions that have been implicated in the

310

neurobiological phenotype of substance use severity” and mapped onto “....existing addiction

311

cue exposure literature” (Feldstein Ewing, 2016). Because only 25% of participants selected a

312

caffeinated soda, it is unlikely that caffeine accounted for the results. In another fMRI study, in

AC C

EP

302

12

ACCEPTED MANUSCRIPT

which adolescents consumed 75 g of glucose and 75 g fructose dissolved in cherry-flavored

314

water (without caffeine), obese but not lean adolescents exhibited both attenuated prefrontal

315

executive control responses and heightened homeostatic and hedonic responses to glucose

316

and fructose (Jastreboff et al., 2016). Although this provides further evidence of the addictive

317

potential of liquid sugar without caffeine in obese youth, the cross-sectional design makes it

318

unclear if this neural response led to or was a consequence of obesity.

Questions have arisen as to whether certain foods/beverages like SSBs are addictive in

SC

319

RI PT

313

vulnerable individuals (substance-based “food addiction” framework), or if instead, some

321

individuals are addicted to eating (behavioral addiction framework) (Hebebrand, 2014). To

322

explore this issue, college students and diverse adult populations were surveyed to determine if

323

particular foods played a role in addictive eating (Schulte, Avena, & Gearhardt, 2015; Schulte,

324

Smeal, & Gearhardt, 2017). In these studies, soda was ranked high on indicators of addictive

325

eating, such as loss of control (Schulte et al., 2015; Schulte, Smeal, et al., 2017). Other high-

326

ranking foods, like pizza and cookies, were highly processed with refined carbohydrates and/or

327

added fat. Foods that ranked low were those that were less/not processed like fruits,

328

vegetables, salmon, and beans. These, and other studies (Schulte, Potenza, & Gearhardt,

329

2017), provide support for the food addiction framework wherein only particular foods or

330

beverages, such as SSBs, trigger addictive behavior. However, not all studies using survey-

331

based approaches have found that soda (Lemeshow et al., 2018) and/or high sugar (mainly

332

without fat or protein) foods were related to food addiction (Markus, 2017).

TE D

EP

AC C

333

M AN U

320

The current study, with its prospective design, both complements and addresses some

334

limitations of survey-based approaches for identifying potentially additive foods, which have

335

been mostly cross-sectional and reliant on participants’ cognizance of the harms caused by

336

particular foods and beverages. However, this study’s small sample size limited our ability to

337

examine differences in symptoms by participant characteristics, such as caffeine consumption,

338

and likely resulted in insufficient power to detect effects of SSB cessation on secondary and 13

ACCEPTED MANUSCRIPT

even primary outcomes. Generalizability is limited to overweight, predominantly female teens.

340

Although the employed design has been used to examine caffeine withdrawal symptoms and

341

has good ecological validity because it involves a naturally occurring pattern of consumption

342

followed by abrupt cessation (De Biasi & Salas, 2008; Juliano & Griffiths, 2004), there are

343

associated limitations. Because participants were free-living, adherence could not be

344

guaranteed, but salivary caffeine and 24-hr recalls provided a degree of reassurance. Also, the

345

use of a single group intervention design, relative to a crossover or two-group design, may be

346

biased by “testing”, or the possibility that taking a baseline survey influenced subsequent

347

responses to the same survey/task during SSB cessation. In particular, testing may have

348

influenced the cognitive and behavioral tasks, for which we did not find significant changes.

349

However, it is also possible that the tasks used were not sensitive enough. Even experimental

350

studies of caffeine cessation do not consistently demonstrate impaired cognitive or behavioral

351

performance (Juliano & Griffiths, 2004). This study was limited to three days of cessation, which

352

overlaps with timing for peak symptoms of withdrawal from caffeine of 20-51 hours (Juliano &

353

Griffiths, 2004) and several symptoms of withdrawal from tobacco and alcohol (Hughes, 2007;

354

Kosten & O'Connor, 2003; Shiffman et al., 2006). However, for other substances, symptoms can

355

peak around 3-5 days, so it is possible our cessation period was not long enough to capture

356

greatest severity for all symptoms. Relatedly, insufficient data for analysis was available from

357

daily check-in questions after 12 pm on day 8. Consequently, we may have missed detecting

358

symptoms that peaked around this time.

SC

M AN U

TE D

EP

AC C

359

RI PT

339

Future studies with larger sample sizes should examine possible withdrawal symptoms

360

from SSBs over longer periods of time using a cross-over design, enroll normal and overweight

361

participants and regular consumers and non-consumers of SSBs, examine differences between

362

responses to caffeinated and non-caffeinated SSBs as well as liquid and solid sugar, and

363

examine other potential manifestations of addiction. Neural imaging studies among younger

364

children, coupled with longitudinal assessments of SSB consumption, anthropometrics, and 14

ACCEPTED MANUSCRIPT

365

metabolic outcomes, could also help clarify the direction of the association between addictive

366

response to SSBs and weight status.

367

369

Conclusion

RI PT

368

This exploratory study provides early evidence of potentially addictive properties of

SSBs in a diverse population of overweight adolescents. These results, combined with present

371

and future corroborating evidence, could inform clinical practice around helping adolescents

372

reduce SSB intake, have important implications for messaging in public health campaigns, and

373

inform the need for efforts to reduce SSB advertising to youth and SSB availability in and

374

around schools.

M AN U

SC

370

375 Acknowledgements

377

We would like to thank the participants who made this research possible, as well as Ely

378

Niroomand and Elena Barbot-Wheaton for their assistance with this study.

379

TE D

376

Funding

381

This work was supported by a University of California Office of the President Multi-Campus

382

Research Initiative grant (PI, Elissa Epel). J. Falbe’s work was supported in part by the

383

American Heart Association (grant #14POST20140055) and National Institute of Diabetes and

384

Digestive and Kidney Diseases of the National Institutes of Health (grant #1K01DK113068-

385

01A1). The content is solely the responsibility of the authors and does not necessarily represent

386

the official views of the National Institutes of Health. These funders had no role in the study

387

design; collection, analysis or interpretation of data; writing of the report; or the decision to

388

submit the article for publication.

AC C

EP

380

389

15

ACCEPTED MANUSCRIPT

References

EP

TE D

M AN U

SC

RI PT

Almiron-Roig, E., Palla, L., Guest, K., Ricchiuti, C., Vint, N., Jebb, S. A., & Drewnowski, A. (2013). Factors that determine energy compensation: a systematic review of preload studies. Nutr Rev, 71(7), 458-473. doi:10.1111/nure.12048 American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (DSM-5). Arlington, VA: American Psychiatric Association. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the royal statistical society. Series B (Methodological), 289-300. Bernstein, G. A., Carroll, M. E., Dean, N. W., Crosby, R. D., Perwien, A. R., & Benowitz, N. L. (1998). Caffeine Withdrawal in Normal School-Age Children. Journal of the American Academy of Child & Adolescent Psychiatry, 37(8), 858-865. doi:https://doi.org/10.1097/00004583-199808000-00016 Block, G. (2004). Foods contributing to energy intake in the US: Data from NHANES III and NHANES 1999–2000. J Food Composition Anal, 14, 439-447. Bray, G. A., & Popkin, B. M. (2014). Dietary sugar and body weight: have we reached a crisis in the epidemic of obesity and diabetes?: health be damned! Pour on the sugar. Diabetes Care, 37(4), 950-956. doi:10.2337/dc13-2085 Brisbois, T. D., Marsden, S. L., Anderson, G. H., & Sievenpiper, J. L. (2014). Estimated intakes and sources of total and added sugars in the Canadian diet. Nutrients, 6(5), 1899-1912. doi:10.3390/nu6051899 Davis, C. (2014). Evolutionary and neuropsychological perspectives on addictive behaviors and addictive substances: relevance to the "food addiction" construct. Subst Abuse Rehabil, 5, 129-137. doi:10.2147/sar.s56835 De Biasi, M., & Salas, R. (2008). Influence of neuronal nicotinic receptors over nicotine addiction and withdrawal. Exp Biol Med (Maywood), 233(8), 917-929. doi:10.3181/0712mr-355 de Ruyter, J. C., Olthof, M. R., Seidell, J. C., & Katan, M. B. (2012). A trial of sugar-free or sugar-sweetened beverages and body weight in children. N Engl J Med, 367(15), 13971406. doi:10.1056/NEJMoa1203034 DiNicolantonio, J. J., O'Keefe, J. H., & Wilson, W. L. (2017). Sugar addiction: is it real? A narrative review. Br J Sports Med. doi:10.1136/bjsports-2017-097971 Ebbeling, C. B., Feldman, H. A., Chomitz, V. R., Antonelli, T. A., Gortmaker, S. L., Osganian, S. K., & Ludwig, D. S. (2012). A randomized trial of sugar-sweetened beverages and adolescent body weight. N Engl J Med, 367(15), 1407-1416. doi:10.1056/NEJMoa1203388 Economic Research Service. (2017). Food Availability (Per Capita) Data System. Retrieved from Washington, DC: https://www.ers.usda.gov/data-products/food-availability-percapita-data-system/ Epel, E. S., Tomiyama, A. J., Mason, A. E., Laraia, B. A., Hartman, W., Ready, K., . . . Kessler, D. (2014). The reward-based eating drive scale: a self-report index of reward-based eating. PLoS One, 9(6), e101350. doi:10.1371/journal.pone.0101350 Epstein, D. H., Marrone, G. F., Heishman, S. J., Schmittner, J., & Preston, K. L. (2010). Tobacco, cocaine, and heroin: Craving and use during daily life. Addict Behav, 35(4), 318-324. doi:10.1016/j.addbeh.2009.11.003 Epstein, L. H., Carr, K. A., Lin, H., & Fletcher, K. D. (2011). Food reinforcement, energy intake, and macronutrient choice. Am J Clin Nutr, 94(1), 12-18. doi:10.3945/ajcn.110.010314 Federal Trade Commission. (2012). A Review of Food Marketing to Children and Adolescents: Follow-up Report. Retrieved from Washington DC: http://www.ftc.gov/os/2012/12/121221foodmarketingreport.pdf

AC C

390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440

16

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Feldstein Ewing, S. W. C., E. D.; Hudson, K. A.; Filbey, F. M.; Yakes Jimenez, E.; Lisdahl, K. M.; Kong, A. S. (2016). Overweight adolescents' brain response to sweetened beverages mirrors addiction pathways. Brain Imaging Behav. doi:10.1007/s11682-0169564-z Goldfield, G. S., Epstein, L. H., Davidson, M., & Saad, F. (2005). Validation of a questionnaire measure of the relative reinforcing value of food. Eat Behav, 6(3), 283-292. doi:https://doi.org/10.1016/j.eatbeh.2004.11.004 Griffiths, R. R., Evans, S. M., Heishman, S. J., Preston, K. L., Sannerud, C. A., Wolf, B., & Woodson, P. P. (1990). Low-dose caffeine physical dependence in humans. J Pharmacol Exp Ther, 255(3), 1123-1132. Hebebrand, J. A., O.; Adan, R.; Antel, J.; Dieguez, C.; de Jong, J.; Leng, G.; Menzies, J.; Mercer, J. G.; Murphy, M.; van der Plasse, G.; Dickson, S. L. (2014). "Eating addiction", rather than "food addiction", better captures addictive-like eating behavior. Neurosci Biobehav Rev, 47, 295-306. doi:10.1016/j.neubiorev.2014.08.016 Hu, F. B. (2013). Resolved: there is sufficient scientific evidence that decreasing sugarsweetened beverage consumption will reduce the prevalence of obesity and obesityrelated diseases. Obes Rev, 14(8), 606-619. doi:10.1111/obr.12040 Hughes, J. R. (2007). Effects of abstinence from tobacco: valid symptoms and time course. Nicotine Tob Res, 9(3), 315-327. doi:10.1080/14622200701188919 IOM (Institute of Medicine) and NRC (National Research Council). (2011). The Science of Adolescent Risk-Taking: Workshop Report. Retrieved from Washington, DC: James, J. E., & Gregg, M. E. (2004). Effects of dietary caffeine on mood when rested and sleep restricted. Hum Psychopharmacol, 19(5), 333-341. doi:10.1002/hup.589 James, J. E., Gregg, M. E., Kane, M., & Harte, F. (2005). Dietary caffeine, performance and mood: enhancing and restorative effects after controlling for withdrawal reversal. Neuropsychobiology, 52(1), 1-10. doi:10.1159/000086172 Jastreboff, A. M., Sinha, R., Arora, J., Giannini, C., Kubat, J., Malik, S., . . . Caprio, S. (2016). Altered Brain Response to Drinking Glucose and Fructose in Obese Adolescents. Diabetes, 65(7), 1929-1939. doi:10.2337/db15-1216 Juliano, L. M., & Griffiths, R. R. (2004). A critical review of caffeine withdrawal: empirical validation of symptoms and signs, incidence, severity, and associated features. Psychopharmacology (Berl), 176(1), 1-29. doi:10.1007/s00213-004-2000-x Keast, R. S., & Riddell, L. J. (2007). Caffeine as a flavor additive in soft-drinks. Appetite, 49(1), 255-259. doi:10.1016/j.appet.2006.11.003 Keast, R. S., Swinburn, B. A., Sayompark, D., Whitelock, S., & Riddell, L. J. (2015). Caffeine increases sugar-sweetened beverage consumption in a free-living population: a randomised controlled trial. Br J Nutr, 113(2), 366-371. doi:10.1017/s000711451400378x Kit, B. K., Fakhouri, T. H., Park, S., Nielsen, S. J., & Ogden, C. L. (2013). Trends in sugarsweetened beverage consumption among youth and adults in the United States: 19992010. Am J Clin Nutr, 98(1), 180-188. doi:10.3945/ajcn.112.057943 Kosten, T. R., & O'Connor, P. G. (2003). Management of drug and alcohol withdrawal. N Engl J Med, 348(18), 1786-1795. doi:10.1056/NEJMra020617 Lemeshow, A. R., Rimm, E. B., Hasin, D. S., Gearhardt, A. N., Flint, A. J., Field, A. E., & Genkinger, J. M. (2018). Food and beverage consumption and food addiction among women in the Nurses' Health Studies. Appetite, 121, 186-197. doi:10.1016/j.appet.2017.10.038 Lobstein, T. (2017). Reducing consumption of sugar-sweetened beverages to reduce the risk of childhood overweight and obesity. Retrieved from http://www.who.int/elena/titles/ssbs_childhood_obesity/en/

AC C

441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490

17

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

MacLeod, C. M. (1991). Half a century of research on the Stroop effect: An integrative review. Psychol Bull, 109(2), 163-203. doi:10.1037/0033-2909.109.2.163 Malik, V. S., Popkin, B. M., Bray, G. A., Despres, J. P., & Hu, F. B. (2010). Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk. Circulation, 121(11), 1356-1364. doi:10.1161/circulationaha.109.876185 Markus, C. R. R., P. J.; Brouns, F.; Schepers, R. (2017). Eating dependence and weight gain; no human evidence for a 'sugar-addiction' model of overweight. Appetite, 114, 64-72. doi:10.1016/j.appet.2017.03.024 Mattes, R. D. (1996). Dietary compensation by humans for supplemental energy provided as ethanol or carbohydrate in fluids. Physiol Behav, 59(1), 179-187. Mazur, J. E. (1987). An adjusting procedure for studying delayed reinforcement. In J. E. M. M. L. Commons, J. A. Nevin, & H. Rachlin (Ed.), Quantitative analysis of behavior: Vol. 5. The effect of delay and of intervening events of reinforcement value (pp. 55-73). Hillsdale, NJ: Erlbaum. Meredith, S. E., Juliano, L. M., Hughes, J. R., & Griffiths, R. R. (2013). Caffeine Use Disorder: A Comprehensive Review and Research Agenda. J Caffeine Res, 3(3), 114-130. doi:10.1089/jcr.2013.0016 Oslin, D. W., Cary, M., Slaymaker, V., Colleran, C., & Blow, F. C. (2009). Daily ratings measures of alcohol craving during an inpatient stay define subtypes of alcohol addiction that predict subsequent risk for resumption of drinking. Drug Alcohol Depend, 103(3), 131-136. doi:10.1016/j.drugalcdep.2009.03.009 Powell, L. M., Szczypka, G., & Chaloupka, F. J. (2010). Trends in exposure to television food advertisements among children and adolescents in the United States. Arch Pediatr Adolesc Med, 164(9), 794-802. doi:2010.139 [pii] 10.1001/archpediatrics.2010.139 [doi] Rachlin, H., Raineri, A., & Cross, D. (1991). Subjective probability and delay. J Exp Anal Behav, 55(2), 233-244. doi:10.1901/jeab.1991.55-233 Reedy, J., & Krebs-Smith, S. M. (2010). Dietary sources of energy, solid fats, and added sugars among children and adolescents in the United States. J Am Diet Assoc, 110(10), 14771484. doi:10.1016/j.jada.2010.07.010 Rosinger, A., Herrick, K., Gahche, J., & Park, S. (2017). Sugar-sweetened beverage consumption among U.S. youth, 2011–2014. NCHS data brief. Retrieved from Hyattsville, MD: Sanchez-Pimienta, T. G., Batis, C., Lutter, C. K., & Rivera, J. A. (2016). Sugar-Sweetened Beverages Are the Main Sources of Added Sugar Intake in the Mexican Population. J Nutr, 146(9), 1888S-1896S. doi:10.3945/jn.115.220301 Schulte, E. M., Avena, N. M., & Gearhardt, A. N. (2015). Which foods may be addictive? The roles of processing, fat content, and glycemic load. PLoS One, 10(2), e0117959. doi:10.1371/journal.pone.0117959 Schulte, E. M., & Gearhardt, A. N. (2017). Associations of Food Addiction in a Sample Recruited to Be Nationally Representative of the United States. Eur Eat Disord Rev. doi:10.1002/erv.2575 Schulte, E. M., Potenza, M. N., & Gearhardt, A. N. (2017). A commentary on the "eating addiction" versus "food addiction" perspectives on addictive-like food consumption. Appetite, 115, 9-15. doi:10.1016/j.appet.2016.10.033 Schulte, E. M., Smeal, J. K., & Gearhardt, A. N. (2017). Foods are differentially associated with subjective effect report questions of abuse liability. PLoS One, 12(8), e0184220. doi:10.1371/journal.pone.0184220 Shiffman, S., Patten, C., Gwaltney, C., Paty, J., Gnys, M., Kassel, J., . . . Balabanis, M. (2006). Natural history of nicotine withdrawal. Addiction, 101(12), 1822-1832. doi:10.1111/j.1360-0443.2006.01635.x

AC C

491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541

18

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Smith, D. G., & Robbins, T. W. (2013). The neurobiological underpinnings of obesity and binge eating: a rationale for adopting the food addiction model. Biol Psychiatry, 73(9), 804-810. doi:10.1016/j.biopsych.2012.08.026 Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18(6), 643-662. doi:10.1037/h0054651 U.S. Department of Agriculture, Agricultural Research Service,. (2016). Food Surveys Research Group. Retrieved from https://www.ars.usda.gov/northeast-area/beltsville-md/beltsvillehuman-nutrition-research-center/food-surveys-research-group/ U.S. Department of Health and Human Services, Health Resources and Services Administration, Maternal and Child Health Bureau. (2014). Child Health USA 2014. Retrieved from https://mchb.hrsa.gov/chusa14/health-statusbehaviors/adolescents/adolescent-overweight-obesity.html#source1 World Health Organization. (1992). The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines. Geneva: World Health Organization.

AC C

542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557

19

ACCEPTED MANUSCRIPT

Table 1. Characteristics of 25 participants

RI PT

14 (56) 3 (12) 2 (8) 1 (4) 1 (4) 4 (16) 33 ± 7 96.1 ± 0.04 19 (76) 102 ± 24

AC C

EP

TE D

M AN U

Age (years) Female Race/ethnicity Non-Hispanic African American More than one race White Asian/Pacific Islander Hispanic African American Hispanic/Latino white BMI BMI%tile BMI≥95th %tile Waist circumference, cm Socioeconomic characteristics Mother’s education >high school degree Father’s education >high school degree Lived with both parents Daily self-reported beverage intake, fl oz/day Sugar-sweetened beverages Soda with added sugar Fruit drinks and smoothies Sports drinks Sweetened coffees or teas Energy drinks Flavored milks Total sugar-sweetened beverages Non-sugar-sweetened beverages, fl oz/day Water Juice, 100% Unflavored milk, dairy or non-dairy Diet soda Unsweetened coffee or tea Fasting serum measures Glucose, mg/dl Insulin, µg/ml Hemoglobin A1c, % Triglycerides, mg/dl Total cholesterol, mg/dl HDL, mg/dl LDL, mg/dl Adiponectin, µg/ml Leptin, µg/ml

Mean±SD or n (%) 15.3 ± 1.4 18 (72)

SC

558

559

16 (64) 9 (45) 11 (44)

14 ± 17 11 ± 13 4±8 2±6 1±5 0.4 ± 1.2 32 ± 22 18 ± 26 5±8 2±4 1±2 1±2

96.2 ± 7.8 23.6 ± 9.9 5.9 ± 0.6 75.5 ± 36.5 150.3 ± 30.1 42.9 ± 12.5 92.3 ± 23.1 7.4 ± 3.2 42.1 ± 29.1

Note: Observations with missing data were excluded from calculations of means, SD, and percentages.

20

ACCEPTED MANUSCRIPT

Table 2. Symptoms of Withdrawal from 11-item modified Scale of Caffeine Withdrawal Average change (n=25) M (%) ± SD

Phase 2 Day Phase 2 Day Phase 2 Day 1 of 2 of 3 of cessation cessation cessation M ± SD M ± SD M ± SD

RI PT

b

Phase 1 Phase 2 a a Average average Usual SSB SSB Intake Cessation (n=25) (n=25) M ± SD M ± SD 16.6 ± 7.6 19.2 ± 7.0

AC C

EP

TE D

M AN U

SC

Total scale score +2.6 (+16%) ± 7.6 19.8 ± 9.2 20.0 ± 8.3 17.6 ± 7.0 c Individual items 0.5 ± 0.8 0.9 ± 0.9* +0.4 (+86%) ± 0.8 1.0 ± 1.2* 1.0 ± 1.3* 0.6 ± 1.0 I have a headache d I feel energetic/have energy 2.0 ± 1.1 1.8 ± 0.9 -0.2 (-11%) ± 1.1 1.8 ± 1.1 1.6 ± 1.1 1.9 ± 1.0 d I feel awake/alert 2.2 ± 1.1 2.0 ± 0.9 -0.2 (-9%) ± 1.4 2.0 ± 1.3 2.0 ± 1.1 2.0 ± 1.1 I feel tired 1.9 ± 1.2 1.9 ± 1.0 0.0 (-1%) ± 1.7 1.8 ± 1.4 1.8 ± 1.4 1.9 ± 1.3 d I feel good overall 2.7 ± 0.9 2.4 ± 0.8 -0.4 (-13%) ± 0.9 2.5 ± 1.2 2.1 ± 1.2* 2.5 ± 1.0 I feel drowsy/sleepy 1.5 ± 1.1 1.2 ± 1.0 -0.2 (-16%) ± 1.3 1.2 ± 1.3 1.2 ± 1.4 1.2 ± 1.1 d I feel content/satisfied 2.3 ± 0.9 2.0 ± 0.9 -0.3 (-11%) ± 0.8 1.8 ± 1.2* 1.9 ± 1.0* 2.4 ± 0.9 I feel motivated to do workd 2.3 ± 0.9 1.7 ± 1.0** -0.6 (-25%) ± 0.9 1.7 ± 1.2* 1.5 ± 1.1***¥ 1.9 ± 1.0 d I feel outgoing 2.2 ± 0.9 2.1 ± 1.0 -0.1 (-6%) ± 1.0 2.0 ± 1.3 2.1 ± 1.1 2.2 ± 0.9 d I feel self-confident 2.8 ± 0.9 2.5 ± 1.0 -0.4 (-13%) ± 1.1 2.2 ± 1.4 2.4 ± 1.1 2.8 ± 1.1 d 2.7 ± 0.9 2.3 ± 1.0* -0.4 (-15%) ± 1.0 2.2 ± 1.3* 2.4 ± 1.0 2.3 ± 1.3 I am able to concentrate *P<0.05, **P<0.01, and ***P<0.01 uncorrected for multiplicity from paired t-tests comparing cessation (phase 2: average across days and values for each day 1-3) to usual intake (phase 1: average across days). ¥ Benjamini–Hochberg corrected P-value<0.05. Note: Complete data (n=25) were available on day 1 of usual intake and day 3 of cessation. Last value carried forward or next value carried backward were used to impute missing data on day 2 of phase 1 (n=3), day 1 of phase 2 (n=3), and day 2 of phase 2 (n=5). a Averaged across days of the phase without missing data. Higher scores indicate greater severity of symptoms. b Possible range: 0-44 by summing responses to 11 items. c Individual items from the Modified Griffith Withdrawal Scale. Possible range: 0 (not at all) - 4 (Very much). d In the calculation of the total scale score, items were reverse coded.

21

ACCEPTED MANUSCRIPT

a

AC C

EP

TE D

M AN U

SC

RI PT

Table S1. Symptoms of Withdrawal from 11-item Modified Scale: T-statistics and Incorrected P-values Comparing Cessation to Usual Intake t(24): difference t(24): difference between t(24): difference between t(24): difference between between phase day 1 of cessation and day 2 of cessation and day 3 of cessation and b c c c averages average of usual intake average of usual intake average of usual intake (n=25) b Total scale score 1.74, p=0.09 1.82, p=0.08 2.05, p=0.05 0.61, p=0.55 c Individual items I have a headache 2.34, p=0.03 2.52, p=0.02 2.29, p=0.03 0.48, p=0.64 d I feel energetic/have energy -0.98, p=0.34 -0.82, p=0.42 -1.51, p=0.14 -0.33, p=0.74 d I feel awake/alert -0.72, p=0.48 -0.65, p=0.52 -0.61, p=0.55 -0.71, p=0.48 I feel tired -0.08, p=0.94 -0.20, p=0.85 -0.21, p=0.83 0.00, p=1.00 I feel good overalld -1.86, p=0.08 -1.03, p=0.31 -2.11, p=0.045 -0.98, p=0.34 I feel drowsy/sleepy -0.91, p=0.37 -0.79, p=0.43 -0.85, p=0.41 -0.77, p=0.45 d I feel content/satisfied -1.58, p=0.13 -2.25, p=0.03 -2.15, p=0.04 0.60, p=0.55 d ¥ -3.12, p=0.005 -2.46, p=0.02 -3.88, p=<0.001 -1.67, p=0.11 I feel motivated to do work d I feel outgoing -0.69, p=0.50 -0.75, p=0.46 -0.67, p=0.51 0.21, p=0.83 d I feel self-confident -1.65, p=0.11 -1.98, p=0.06 -1.82, p=0.08 -0.26, p=0.80 d I am able to concentrate -2.21, p=0.04 -2.07, p=0.049 -1.53, p=0.14 -1.77, p=0.09 a Uncorrected for multiplicity b Average across days of each phase c Average across days of usual intake phase ¥ Benjamini–Hochberg corrected P-value<0.05 Note: Complete data (n=25) were available on day 1 of usual intake and day 3 of cessation. Last value carried forward or next value carried backward were used to impute missing data on day 2 of phase 1 (n=3), day 1 of phase 2 (n=3), and day 2 of phase 2 (n=5).

22

ACCEPTED MANUSCRIPT

M AN U

TE D

a

Pa

RI PT

Za

4.00 4.21 4.00 3.62 3.94

<0.001 <0.001 <0.001 <0.001 <0.001

1.06 4.21 4.27 4.12 0.38 -0.63

0.29 <0.001 <0.001 <0.001 0.70 0.53

3.40 3.89 0.18 -0.42 0.66 1.31 2.30

<0.001 <0.001 0.86 0.68 0.51 0.19 0.02

SC

Table 3. Daily dietary intake and salivary caffeine Phase 1: Phase 2: Change Usual SSB SSB (n=25) M ± intake Cessation SD (n=25) M ± (n=25) M ± SD SD Daily intakes Sugar sweetened beverages Amount, fl oz 32 ± 22 5±9 -27 ± 25 Added sugar, g 20 ± 13 2±3 -18 ± 15 Sugar, g 80 ± 56 11 ± 23 -69 ± 65 Energy, kcals 344 ± 238 69 ± 151 -275 ± 303 Caffeine, mg 40 ± 59 1±2 -39 ± 59 Total beverages -6 ± 26 Amount, fl oz 59 ± 32 53 ± 36 Sugar, g 100 ± 58 18 ± 26 -82 ± 59 Energy, kcals 458 ± 262 150 ± 197 -308 ± 272 Caffeine, mg 50 ± 60 4 ± 14 -46 ± 57 Caffeine from food, mg 1.5 ± 0.6 1.5 ± 0.5 0.0 ± 0.8 Sugar from food, g 54 ± 39 47 ± 33 -7 ± 45 Total daily foods and beverages Added sugar, g 26 ± 18 10 ± 9 -16 ± 19 Sugar, g 144 ± 83 64 ± 43 -80 ± 78 Protein, g 71 ± 67 64 ± 37 -6 ± 63 Solid fat, g 25 ± 20 25 ± 14 0 ± 19 Fiber, g 11 ± 6 10 ± 5 -1 ± 7 Energy, kcals 1945 ± 1129 1610 ± 761 -334 ± 1021 Salivary caffeine (µg/ml)b 0.35 ± 0.62 0.13 ± 0.24 -0.22 ± 0.58

AC C

EP

Z-statistics and P-values from Wilcoxon Signed Rank Tests comparing usual intake to SSB cessation phases. b Amounts below the quantifiable limit of 0.100 µg/ml were coded as 0 µg/ml.

23

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Figure 1. Study timeline

24

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Figure 2. Participant flow

25

ACCEPTED MANUSCRIPT

Figure 3. Daily check-in questions assessing symptoms of withdrawal at regular intervals during usual intake and SSB cessation phases (n=24) A. Overall wellbeing B. SSB craving

*

6 5 4

7 6 5 4

3

3

2

2

1

1

C. Tiredness

D. Irritability

10

10

9

9

7 6

3 2 1

7 6 5 4 3 2 1

10 12 16 18 21 34 36 40 42 45 58 60 Hours since midnight of phase

AC C

EP

10 12 16 18 21 34 36 40 42 45 58 60 Hours since midnight of phase

Response

8

TE D

Response

8

4

10 12 16 18 21 34 36 40 42 45 58 60 Hours since midnight of phase

M AN U

10 12 16 18 21 34 36 40 42 45 58 60 Hours since midnight of phase

5

*

8

7

*

RI PT

*

*

9

Response

Response

8

*

**

***

9

¥

10

¥

SC

10

*P<0.05, **P<0.01, and ***P<0.01 uncorrected for multiplicity for differences between the usual intake and SSB cessation phases. ¥ Benjamini–Hochberg corrected P-value<0.05. Plots show average responses at each time point as well as error bars for 95% CIs. CIs and post hoc pairwise contrasts between each time point were estimated using the margins and pwcompare commands in STATA for postestimation after fitting generalized linear regression models containing indicators for each time point with robust (Huber-White sandwich) standard errors. Usual intake phase responses at each time were averaged across days to serve as the comparisons for each time/day of cessation. Full results are reported in Table S2. Questions and 10-point response scale anchors: (A) How do you feel? 1 (really bad) – 10 (really good); (B) How much do you want your favorite sugar-sweetened beverage right now? 1 (I don’t want at all) – (I really want); (C) How tired are you? 1 (not tired at all) – 10 (really tired); (D) How irritable are you? 1 (not at all irritable) – 10 (really irritable).

26

ACCEPTED MANUSCRIPT

Table S2. Full results for daily check-in questions assessing symptoms of withdrawal at regular intervals during usual intake and SSB cessation phases Usual intake phase c

c

b, c

Post hoc pairwise contrasts

n

Mean (SE )

n

6.39 (0.43) 7.41 (0.33) 7.24 (0.39) 6.63 (0.41) 5.89 (0.48 6.39 (0.43) 7.41 (0.33) 7.24 (0.39) 6.63 (0.41) 5.89 (0.48 6.39 (0.43) 7.41 (0.33)

23 23 23 23 23 23 23 23 23 23 23 23

6.32 (0.53) 6.60 (0.53) 5.87 (0.54) 5.13 (0.58) 5.18 (0.63) 5.91 (0.55) 5.13 (0.55) 5.64 (0.65) 5.85 (0.57) 5.96 (0.60) 6.40 (0.60) 7.17 (0.67)

22 24 23 23 22 23 23 22 23 23 21 21

4.78 (0.51) 5.99 (0.53) 6.46 (0.55) 5.96 (0.59) 4.92 (0.55) 4.78 (0.51) 5.99 (0.53) 6.46 (0.55) 5.96 (0.59) 4.92 (0.55) 4.78 (0.51) 5.99 (0.53)

23 23 23 24 22 23 23 23 24 22 23 23

5.73 (0.78) 6.75 (0.64) 7.11 (0.69) 7.74 (0.60) 7.57 (0.67) 6.54 (0.73) 7.78 (0.54) 7.23 (0.69) 7.04 (0.66) 6.30 (0.75) 5.81 (0.76) 7.12 (0.54)

22 24 23 23 23 23 23 22 23 23 21 21

1.01 0.92 0.74 2.11 3.07 1.98 2.38 0.88 1.23 1.49 1.12 1.50

0.31 0.36 0.46 0.04 0.002¥ 0.05 0.02 0.38 0.22 0.14 0.26 0.13

6.02 (0.48) 4.67 (0.65) 5.74 (0.52) 6.20 (0.58) 7.79 (0.49) 6.02 (0.48) 4.67 (0.65) 5.74 (0.52) 6.20 (0.58) 7.79 (0.49) 6.02 (0.48) 4.67 (0.65)

23 23 23 23 24 23 23 23 23 24 23 23

5.41 (0.71) 4.81 (0.64) 5.00 (0.70) 6.09 (0.62) 8.09 (0.58) 5.15 (0.63) 5.74 (0.63) 5.55 (0.72) 5.78 (0.57) 7.43 (0.56) 5.14 (0.69) 4.69 (0.64)

24 23 23 23 23 23 23 22 23 23 21 21

-0.72 0.15 -0.85 -0.13 0.39 -1.1 1.17 -0.22 -0.51 -0.48 -1.05 0.02

0.47 0.88 0.40 0.90 0.70 0.27 0.24 0.83 0.61 0.63 0.30 0.99

4.15 (0.57) 3.85 (0.43) 3.89 (0.42) 4.50 (0.60) 4.39 (0.56) 4.15 (0.57)

22 22 22 22 22 22

4.43 (0.71) 4.07 (0.62) 4.91 (0.59) 5.27 (0.65) 5.59 (0.69) 4.23 (0.59)

21 23 22 22 22 22

0.31 0.28 1.41 0.87 1.35 0.10

0.76 0.78 0.16 0.38 0.18 0.92

AC C

Uncorrected Z -0.11 -1.30 -2.05 -2.12 -0.90 -0.69 -3.56 -2.13 -1.12 0.09 0.02 -0.33

Uncorrected P-value 0.92 0.19 0.04 0.03 0.37 0.50 ¥ <0.001 0.03 0.27 0.93 0.99 0.74

SC

RI PT

Mean (SE )

EP

10 12 16 18 21 34 36 40 42 45 58 60 How much do you want your favorite SSB right now? 10 12 16 18 21 34 36 40 42 45 58 60 How tired are you? 10 12 16 18 21 34 36 40 42 45 58 60 How irritable are you? 10 12 16 18 21 34

Cessation phase

TE D

How do you feel?

a

M AN U

Hours since midnight of phase

27

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

36 3.85 (0.43) 22 4.59 (0.66) 22 0.94 0.35 40 3.89 (0.42) 22 4.33 (0.69) 21 0.55 0.58 42 4.50 (0.60) 22 4.41 (0.55) 22 -1.01 0.31 45 4.39 (0.56) 22 4.59 (0.65) 22 0.24 0.81 58 4.15 (0.57) 22 3.55 (0.61) 20 -0.72 0.47 60 3.85 (0.43) 22 3.10 (0.55) 20 -1.08 0.28 ¥ Benjamini–Hochberg corrected P-value<0.05. a Responses were averaged across days 2-3 of usual intake (e.g., mean and SE for 10, 34, and 58 hours of usual intake phase are equivalent). b Uncorrected for multiplicity. c Huber-White sandwich standard errors estimated from generalized linear regression models that included indicators for each time point. Post hoc pairwise contrasts were used to compare responses during the cessation phase to responses at the same times during the usual intake phase. Questions and 10-point response scale anchors: (A) How do you feel? 1 (really bad) – 10 (really good); (B) How much do you want your favorite sugar-sweetened beverage right now? 1 (I don’t want at all) – (I really want); (C) How tired are you? 1 (not tired at all) – 10 (really tired); (D) How irritable are you? 1 (not at all irritable) – 10 (really irritable).

28