Early-life exposures to diet soda and aspartame are associated with autism in males
Data files
Sep 25, 2023 version files 231.36 KB
Abstract
Since its introduction, aspartame—the leading sweetener in U.S. diet sodas (DS)—has been reported to cause neurological problems in some users. In prospective studies, the offspring of mothers who consumed diet sodas/beverages (DSB) daily during pregnancy experienced increased health problems. We hypothesized that gestational/early-life exposure to ≥1 DS/day (DSearly) or equivalent aspartame (ASPearly: ≥177 mg/day) increases autism risk. The case-control Autism Tooth Fairy Study obtained retrospective dietary recalls for DSB and aspartame consumption during pregnancy/breastfeeding from the mothers of 235 offspring with autism spectrum disorder (ASD: cases) and 121 neurotypically developing offspring (controls). The exposure odds ratios (ORs) for DSearly and ASPearly were computed for autism, ASD, and the non-regressive conditions of each. Among males, DSearly odds were tripled among autism (OR = 3.1; 95% CI: 1.02, 9.7) and non-regressive autism (OR = 3.5; 95% CI: 1.1, 11.1) cases; ASPearly odds were even higher: OR = 3.4 (95% CI: 1.1, 10.4) and 3.7 (95% CI: 1.2, 11.8), respectively, among autism and non-regressive autism cases (p < 0.05 for each). ORs for non-regressive ASD in males were almost tripled but were not statistically significant: DSearly OR = 2.7 (95% CI: 0.9, 8.4); ASPearly OR = 2.9 (95% CI: 0.9, 8.8). No statistically significant associations were found in females. Our findings contribute to the growing literature raising concerns about potential offspring harm from maternal DSB/aspartame intake in pregnancy.
README: Early-life exposures to diet soda and aspartame are associated with autism in males.
https://doi.org/10.5061/dryad.cfxpnvx2p
Please see the attached READ ME file, which contains both a detailed data dictionary and also an 'exposure-variables map' worksheet, to describe which exposure variables were included in each of the tables and figures in our manuscript.
Description of the data and file structure
We have organized the variables in our dataset into the following categories, and ordered them in this sequence:
1. numeric study identifiers
2. demographic variables and other covariates
3. diagnosis- and regression-related variables
4. early-life exposures through maternal diet during pregnancy or breastfeeding:
a. computed daily exposure variables, dichotomized
b. computed minimum mean-daily-equivalent exposure variables, dichotomized
c. original retrospective dietary recall data recorded by biological mothers, for their pregnancy or breastfeeding periods
<br>
This READ ME file describes the meanings of values of numeric variables; missing data codes; and abbreviations used.
Sharing/Access information
This dataset stored and hosted by Dryad is the sole way of accessing the data used in all analyses in our published manuscript.
All the data used in our published manuscript were derived from the Autism Tooth Fairy Study, led by Raymond F. Palmer, Ph.D, Principal Investigator, and Lynne
Code/Software
This is an optional, freeform section for describing any code in your submission and the software used to run it.
Describe any scripts, code, or notebooks (e.g., R, Python, Mathematica, MatLab) as well as the software versions (including loaded packages) that you used to run those files. If your repository contains more than one file whose relationship to other scripts is not obvious, provide information about the workflow that you used to run those scripts and notebooks.
Methods
Participants
Recruitment and data collection for the Autism Tooth Fairy Study (ATFS) were performed through The University of Texas Health Science Center at San Antonio (UTHSCSA) from May 2011, through June 2014. The largest recruitment source (n=255) was the Interactive Autism Network (IAN), an Internet-based U.S. registry (1) which included >21,600 individuals with ASD, and >22,000 parents of individuals with ASD (2). Most families recruited through IAN had >1 child diagnosed with ASD (cases; n=178). Neurotypically-developing children (controls; n=77) identified through IAN recruitment included siblings of cases (n=61), and offspring of friends/associates of IAN parents (n=16) (3). Additional recruitment (n=101) was conducted through media outreach in San Antonio and South Texas (SA/STX), through which 57 cases, including 7 referred from local autism services, and 44 controls were enrolled. The ATFS thus included a total of 356 offspring: 235 cases and 121 controls. In addition to completing questionnaires regarding their children’s early-life exposures, ATFS parents provided one or more of their children’s shed deciduous teeth for compositional analyses, results of which are reported elsewhere (4).
The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of the University of Texas Health Science Center at San Antonio (UTHSCSA) (UTHSCSA IRB protocol number 11-313, approved on 12 May 2011). In early data collection, all parents provided written informed consent before completing self-administered hard-copy questionnaires. During later Internet-based enrollment and data collection, all parents reviewed an IRB-approved online study information sheet prior to completing online questionnaires through secure websites, in accordance with the protocol approved by the IRB of the UTHSCSA.
Materials
Demographic and Neurodevelopmental Data Collected
Parents provided demographic data about themselves and their households; the birth year and sex of each child; and whether each child had been diagnosed with either autism or autistic disorder, Asperger’s disorder, pervasive developmental disorder – not otherwise specified, or childhood disintegrative disorder. Children with any of these diagnoses were included as ASD cases. Parents were also asked whether each child had been diagnosed with any additional behavioral, developmental, and/or learning disabilities/disorders; non-cases with none of these diagnoses were included as controls.
Parents were asked, “Was there ever a time when your child used at least three words you could understand (besides mama or dada) on a daily basis for at least a month, and then seemed to stop talking for a while (at least a month) where he/she used no words that you could understand?” Cases where parents responded “no” were considered unlikely to have experienced a regressive form of autism or ASD and were thus categorized as “non-regressive” cases.
Early-Life Exposures to Diet Sodas, Other Diet Drinks, Aspartame, and Other NNSs
Biological mothers completed retrospective questionnaires on their intake of diet sodas (DS) and other diet drinks (DDother) during pregnancy/breastfeeding. For each child, they were asked, “While you were pregnant or breastfeeding your child, how often did you drink diet drinks containing artificial sweeteners? Please count diet sodas first, such as Diet Coke, Diet Dr. Pepper, and Diet Sprite, and then other diet drinks, such as Crystal Light, sugar-free Kool-Aid, Slim-Fast, and other ‘lite’ drinks”. Within each of these two beverage subcategories, mothers recorded the number of cans or bottles of DS and the number of glasses, cans, or bottles of DDother that they had consumed per time unit; whether this time unit was daily, weekly, monthly, annually, or never; and the specific brand(s) consumed. Mean intake of DS/day and DDother/day was calculated and summed to estimate total maternal intake of diet drinks/day (DDtotal/day) during pregnancy/breastfeeding.
Each mother was also asked, “While you were pregnant or breastfeeding your child, how many little packets of low-calorie sweeteners (such as Sweet ‘N Low, Equal or Splenda) did you use in your coffee, tea, or other foods and beverages? In answering the question, please keep in mind the number of drinks you had each day, and how many packets of sweetener you added to each drink. Also, please include the number of packets you used in cereal or other food”. Intake of the following three leading NNS packets was specifically requested: “Equal/Nutrasweet (blue)”, “Splenda (yellow)”, and “Sweet’N Low (pink)”; space was left for recording intake of other NNS packet brands. Maternal estimates of intake of these products during pregnancy and breastfeeding were part of an extensive questionnaire that gathered data on maternal and child exposures to a number of household products and other environmental exposures throughout pregnancy and the early life of the child.
Procedure
Case Definitions
Two case definitions were specified for analyses. The primary case definition of interest was autism disorder (autism); the secondary case definition was any ASD. In addition, subgroup analyses focused on offspring with non-regressive versions of each case definition, among whom the relative etiologic influence of intrauterine/neonatal exposures was expected to be maximized. In all analyses, controls were neurotypically developing offspring and – for sex-specific analyses – of the same sex as cases. Both sex-specific and pooled analyses were performed for each of the case definitions and exposures.
Exposure Variables
Maternal intake of ≥12 ounces (1 can) of DS/day during pregnancy/breastfeeding defined the primary, dichotomized exposure variable: daily early-life exposure to DS (DSearly). The secondary, dichotomized exposure variable – daily early-life exposure to aspartame (ASPearly) – was defined as maternal intake of ≥177 mg/day of aspartame during pregnancy/breastfeeding, from the sum of aspartame dosages in DS + DDother + packets consumed by the mother. Estimates of total maternal aspartame intake from these three sources were possible because mothers were asked to report the specific brands they had consumed. Dosages of aspartame and other NNSs used in each of the leading brands of DS had previously been published (5); for comparable products from other brands, NNS dosages were imputed according to the NNS types specified on product labels. Leading DS brands sweetened exclusively with aspartame contained ≥177 mg of aspartame/can (5); this dosage was therefore selected as the cut-point for ASPearly to identify offspring with daily early-life aspartame exposure comparable to that in 1 can of the leading diet colas (5).
For DDother, ingredient-labeling specified types of NNS used, but not dosages, which are proprietary. Based on limited available data, a 12-ounce serving of DDother sweetened exclusively with aspartame was estimated to contain approximately 100 mg of aspartame. A tabletop packet contains 37 mg of aspartame (6).
Statistical Analyses
Offspring born in 1984 or thereafter, with complete early-life DS exposure data and neurodevelopmental diagnostic data, were included in analyses. The proportions of offspring with DSearly and ASPearly exposures were separately calculated, by sex, for the following diagnostic groups: controls, all ASD cases, ASD cases without autism, autism cases, and the non-regressive version of each case definition. Differences in the proportions of exposed controls and cases, in each category, were calculated using the Pearson Χ2 statistic in Stata/IC 14.2 for Windows.
To assess associations between diagnostic status and the primary and secondary exposures under study – DSearly and ASPearly – multilevel mixed-effects generalized linear models (MEGLMs) were used to consider potential influences at two levels: the individual child (level one) and shared intrauterine and familial influences, through the mother (level two). Recruitment source was treated as a level-one predictor. A total of 356 children, including children from 79 sibships, were included in analyses. Models were fit using the meglm command in Stata/IC 14.2 for Windows.
Unadjusted odds ratios (ORs) for the primary and secondary exposures of interest – DSearly and ASPearly – were separately computed for each case definition, by sex, and for all participants pooled. Adjusted models included the following covariates: recruitment source (IAN vs. SA/STX); year of birth (range, 1985 to 2011), to adjust for secular trends in autism diagnosis; and three dichotomous demographic variables: ethnicity of the child (non-Hispanic white (NHW) vs. other), maternal education (college graduate vs. not), and household income (≥ $100,000/year vs. less). In pooled analyses, we additionally adjusted for sex of the child. For comparison purposes, we also calculated separate ORs for early-life exposures to ≥1 serving/day of aspartame specifically or of any NNS, from the combination of DS, DDother, and/or tabletop packets, with minimum serving size defined as one tabletop sweetener packet. The minimum daily dosage for these analyses would thus be 37 mg for aspartame, 36 mg for saccharin, and 12 mg for sucralose, for example. We also performed subgroup analyses for non-regressive ASD cases and non-regressive autism cases, using these same exposure variables and covariates. These models were all fit using the meglm command in Stata/IC 14.2 for Windows.
Sensitivity Analyses
To examine influences on maternal intake of these products, we compared the proportions of biological mothers of males who reported consuming DS, aspartame, or any other NNSs during pregnancy and/or breastfeeding, within dichotomized substrata of three demographic factors previously associated with cardiometabolic risk: income, educational level, and ethnicity. Among these six demographic subsets – mothers of offspring from households with incomes ≥ $100,000/year vs. less; mothers with ≥4 years of college vs. less education; and mothers of NHW children vs. children of other or mixed ethnicities – we compared the proportions of mothers whose mean total reported intake of diet products equaled or exceeded the dose in 1 packet/day of any NNS, 1 packet/day of aspartame, 1 DS/day, and 177 mg/day of aspartame. The significance of differences between sociodemographic substrata in the proportions of mothers who reported daily consumption of each of these diet products was assessed using the Pearson Χ2 statistic in Stata/IC 14.2 for Windows.
References
- The Interactive Autism Network: Research Overview: Kennedy Krieger Institute; 2023 [updated 2023. Available from: https://iancommunity.org/cs/ian_research/overview.
- Interactive Autism Network: Patient-Centered Outcomes Research Institute; 2023 [2023]. Available from: https://www.pcori.org/research-results/2015/interactive-autism-network#section_professional_abstract.
- Heilbrun LP, Palmer RF, Jaen CR, Svoboda MD, Miller CS, Perkins J. Maternal Chemical and Drug Intolerances: Potential Risk Factors for Autism and Attention Deficit Hyperactivity Disorder (ADHD). J Am Board Fam Med. 2015;28(4):461-70.
- Palmer RF, Heilbrun L, Camann D, Yau A, Schultz S, Elisco V, et al. Organic Compounds Detected in Deciduous Teeth: A Replication Study from Children with Autism in Two Samples. J Environ Public Health. 2015;2015:862414.
- Franz M. Diet Soft Drinks. Diabetes Self-Management: Rapaport Publishing, Inc.; 2010.
- Stein PJ. The Sweetness of Aspartame: A Biochemistry Lab for Health Science Chemistry Courses. Journal of Chemical Education. 1997;74(9):1112.
Usage notes
Please note that the second spreadsheet contain a data dictionary, which defines the meanings of numeric values assigned to categorical variables, and provides further explanatory notes, including the meanings of dichotomous variables.