Helpful or harmful? An examination of the relationship between low- and no-calorie sweetener exposure and dietary intake and health

Abstract

Background: Poor dietary patterns are a major contributor to chronic disease morbidity and mortality globally. Many governments and health authorities have implemented public health initiatives to improve dietary patterns, including by reducing added sugars intake. Replacing added sugars with low- and no-calorie sweeteners is becoming increasingly common, both in the food supply and by individuals. Although low- and no-calorie sweeteners approved for use are considered safe from a toxicological perspective, their long-term impacts on health and chronic disease risk are contested. Objectives: The objectives of this dissertation were to 1) summarize evidence from systematic reviews examining associations between consumption of low- and no-calorie sweeteners and chronic conditions and intermediate risk factors (Chapter 4); 2) develop a standardized approach to classify food codes used in nutrition surveillance in Canada as sources of low- and no-calorie sweeteners (Chapter 5); 3) estimate the proportion of individuals in Canada reporting the consumption of sources of low- and no-calorie sweeteners or tabletop sweeteners on a given day (Chapter 6); 4) characterize the nutrition profiles of sources of low- and no-calorie sweeteners (Chapter 6); and 5) assess diet quality in relation to the probability of consuming low- and no-calorie sweeteners among individuals in Canada (Chapter 7). Methods and results: Chapter 4 presents a rapid review of systematic reviews of randomized and non-randomized clinical trials and observational studies investigating associations between low- and no-calorie sweetener consumption and health outcomes and intermediate risk factors in adults and children. The search strategy returned 409 results, 25 of which met the eligibility criteria and underwent quality appraisal using the AMSTAR-2 assessment tool. Data from nine systematic reviews assessed as high- or moderate-quality were included in the narrative synthesis. Randomized controlled trials tended to report no association or beneficial impacts of consumption, whereas non-randomized trials and other observational studies observed no association or negative impacts on outcomes of interest, such as type 2 diabetes. There was limited evidence examining associations between consumption of sweeteners and diet quality. Chapters 5, 6, and 7 drew on 24-hour dietary recall data from the 2015 Canadian Community Health Survey—Nutrition Public Use Microdata File (2015 CCHS—Nutrition). In Chapter 5, a classification system was developed to categorize food codes used to code the 24-hour recalls as sources of sweeteners. First, food code descriptions were reviewed for confirmatory keywords, such as “low-calorie sweetener,” and possible source keywords, such as “diet.” Food codes with descriptions containing confirmatory keywords were categorized as sources, whereas those with descriptions containing possible source keywords progressed to step two. In step two, the descriptions for matching food codes in the U.S. Food and Nutrient Database for Dietary Studies were examined for confirmatory keywords. In step 3, the remaining possible sources were assessed using an online search of grocery store websites. A total of 76 food codes (1.5%) were classified as sources of low- and no-calorie sweeteners and 46 recipe codes used in the 2015 CCHS—Nutrition contained sources as ingredients. Steps 2 and 3 identified 41 food code sources not identified by Step 1 alone. In Chapter 6, the classification system was applied to recall data for individuals aged 1 year and above living in the Canadian provinces. The proportions reporting consumption of at least one source on a given day were estimated overall and by age, sex, household income, and self-reported diabetes status. Commonly reported sources were examined by age group, and the nutritional profiles of sources were examined relative to the 2019 Canada’s Food Guide. About one in 10 children aged 1-13 years (8.4%, 95% CI: 6.6, 10.2) and adolescents aged 14-18 years (10.7%, 95% CI: 7.6, 13.8) and one in five adults (19.8%, 95% CI: 16.2, 23.4) 19 years and above consumed sources of sweeteners on a given day. Among adults with diabetes, four in 10 (42.4%, 95% CI: 36.1, 48.7) consumed a source of low- and no-calorie sweeteners on a given day compared to approximately one in five (18.2%, 95% CI: 14.2, 22.0) who did not report having diabetes. Foods sweetened with low- and no-calorie sweeteners were more frequently consumed than beverages among children and adolescents (children: 5.5% (95% CI: 4.0, 7.0) vs. 3.1% (95% CI: 2.3, 3.9), respectively; adolescents: 6.2% (95% CI: 3.4, 9.1) vs. 4.1% (95% CI: 2.8, 5.4)), whereas adults consumed beverages more frequently than foods (10.9% (95% CI: 7.6, 14.2) vs. 7.7% (95% CI, 7.0, 8.4), respectively). On a given day, 4% of adults consumed tabletop sweeteners; estimates for children and adolescents were not reported due to small sample sizes. Diet sodas were the most frequently consumed beverage source across all three age groups. Yogurts were the most frequently consumed food source among children, whereas meal replacements, including protein powders, were the most frequently consumed food source among adolescents and adults. No food codes classified as sources of low- and no-calorie sweeteners were considered ‘foundational’ according to the 2019 Canada’s Food Guide Food Classification System. In Chapter 7, diet quality, measured using the 2019 Healthy Eating Food Index (HEFI-2019), was assessed in relation to the probability of consuming sources of low- and no-calorie sweeteners among respondents aged 2 years and older, overall and by sex and diabetes status. Higher HEFI-2019 scores, which represent closer alignment with the 2019 Canada’s Food Guide, were calculated based on simulated distributions of usual intakes estimated using the National Cancer Institute’s multivariate Markov chain Monte Carlo (MCMC) method. Variables (e.g. vegetables and fruit, free sugars) used to arrive at HEFI-2019 scores were modelled jointly with a variable indicating an individual’s probability to consume low- and no-calorie sweeteners on a recall day; probabilities of consumption were expressed as quintiles. There was weak evidence (P=0.08) of lower (-6.1 points, 95% CI: -13.1, 0.8) HEFI-2019 total scores among children aged 2 to 13 with the highest probability of consuming low- and no-calorie sweeteners (i.e. quintile 5) compared to children with the lowest probability. Among males 14 years of age and above, there was moderate evidence (P=0.04) of higher (4.6 points, 95% CI: 0.1, 9.1) HEFI-2019 total scores among those with the highest probability of consuming low- and no-calorie sweeteners compared to those with the lowest probability. There was no evidence of meaningful differences in HEFI-2019 total sores in relation to probability of consuming sources of low- and no-calorie sweeteners among females aged 14 years and above or among adults 19 years and above with and without diabetes. Conclusion: This dissertation makes several contributions to the evidence to inform policies related to low- and no-calorie sweeteners in Canada. First, the inconsistent associations between the use of sweeteners and health in the literature contribute to differing dietary guidance across health authorities. This lack of clarity on how and by whom low- and no-calorie sweeteners should be consumed emphasizes the importance of ongoing monitoring of population-level consumption patterns. Second, this dissertation presents a standardized approach to identifying sweetener sources in Canada’s national food composition database, enabling consistent classification across studies. Third, for the first time, evidence on population-level exposure to low- and no-calorie sweeteners from foods, beverages, and tabletop sweeteners in Canada and associations between diet quality and the probability of consuming low- and no-calorie sweeteners can inform healthy eating policies, such as Canada’s Healthy Eating Strategy.