Principal findings and possible explanations
Dietary sugar consumption is harmfully associated with multiple health outcomes across various measurements of exposure, including high versus low, never/low versus moderate/high, any versus none, or an extra increment of sugars per day (or week) versus none. We identified 73 meta-analyses and 83 health outcomes from 8601 unique articles, including 74 unique outcomes in meta-analyses of observational studies and nine unique outcomes in meta-analyses of randomised controlled trials.
Dietary sugar consumption had harmful associations with endocrine and metabolic outcomes, including changes in body mass index in children,53 changes in body weight,53 changes in body weight (one year),53 gout,111358 high density lipoprotein cholesterol,60 hyperuricaemia,1159 latent autoimmune diabetes in adults,30 low density lipoprotein cholesterol,60 metabolic syndrome,61 obesity in children,3 obesity in adults,1 serum uric acid,57 type 2 diabetes mellitus,6 liver fat accumulation,54 and muscle fat accumulation.54 In addition, harmful associations between dietary sugar consumption and cardiovascular outcomes were also observed, including coronary heart disease,10 cardiovascular disease,8 cardiovascular disease mortality,6871 hypertension in children and adolescents,69 hypertension in adults,70 myocardial infarction,66 change in systolic blood pressure in children and adolescents,69 and stroke.10 Significant harmful associations between dietary sugar consumption and a higher risk of cancer were observed for breast cancer,19 breast cancer mortality,18 hepatocellular carcinoma,18 prostate cancer,18 pancreatic cancer,72 overall cancer risk,18 and overall cancer mortality.18 Finally, harmful associations existed between dietary sugar consumption and all cause mortality,76 asthma in children,79 attention deficit/hyperactivity disorder,80 bone mineral density,81 dental caries,17 dental erosion,17 depression,77 non-alcoholic fatty liver disease,16 and intrahepatocellular lipids.78
In general, no reliable evidence shows beneficial associations between dietary sugar consumption and any health outcomes, apart from glioma (sugar sweetened beverages),18 total cholesterol (sugar sweetened beverages),60 type 2 diabetes mellitus (sucrose),62 and cardiovascular disease mortality (sucrose).71 However, these favourable associations are not supported by strong evidence, and the interpretation of these results should be done with caution. For the decreased risk of glioma, evidence for this came from only two cohort studies, and no studies have shown that sugar sweetened beverage consumption is a protective factor to lower the incidence of cancer. High sugar sweetened beverage consumption was associated with lower total cholesterol concentrations. However, subgroup analysis indicated that sugar sweetened beverage consumption was associated with higher total cholesterol concentrations in studies with sugar sweetened beverage consumption >750 g/day and studies involving adolescents. Therefore, potential confounders, including region, sugar sweetened beverage dose, sample size, and sex, should be considered in explaining the association between sugar sweetened beverage consumption and total cholesterol concentrations. In terms of the protective effect of sucrose intake on type 2 diabetes mellitus and cardiovascular disease mortality, we note that sucrose tends to be found more in solid foods than in sugar sweetened beverages, including grains and grain based products, fruit and fruit products, and sweetened dairy and dairy products.848586 These main sources of sucrose have shown beneficial associations (for example, whole grain cereals, fruit, and yogurt) with type 2 diabetes mellitus and cardiovascular disease mortality.878889909192 Therefore, the protective association between sucrose intake and type 2 diabetes mellitus and cardiovascular disease mortality may reflect important contributions from these other food sources rather than sucrose.6271 Further large scale, prospective studies are warranted to evaluate the association of sucrose intake with type 2 diabetes mellitus and cardiovascular disease mortality and to clarify the possible underlying mechanisms.
Our umbrella review showed harmful associations between dietary sugar consumption and a range of cardiometabolic diseases, especially weight gain, ectopic fat accumulation, obesity, and cardiovascular disease, which can largely be attributed to excessive consumption of fructose containing sugars. In response to the intake of large carbohydrates, fructose could enhance hepatic lipogenic capacity by inducing hepatic master transcription factors.939495 Moreover, an animal study found that dietary fructose could be converted to microbial acetate by the gut microbiota, which may enhance hepatic lipogenesis by supplying lipogenic acetyl-CoA independently of ATP citrate lyase.96 Intermediate products such as diacylglycerols generated during the process of lipogenesis may impair insulin signalling in the liver and peripheral tissues and then lead to insulin resistance.97 Subsequently, it may promote ectopic fat deposition in the liver and muscle.9899 Dietary fructose may also inhibit fatty acid oxidation in the liver by impairing mitochondrial size and function and acetylation of the rate limiting enzyme.100 A recent animal study showed that dietary fructose improves the survival of intestinal cells and increases the length of intestinal villus in mouse models, resulting in an expanded surface area of the gut and increased nutrient absorption and adiposity in mice.101 Furthermore, fructose contained in sugar sweetened beverages is suggested to likely induce the onset of obesity by reducing resting energy expenditure and promoting leptin resistance.102103 In addition, sugar sweetened beverages are associated with less satiety compared with solid food containing the same amount of calories, which may stimulate appetite and induce excessive calorie consumption, liver fat accumulation, and insulin resistance in the long term.104 This hypothesis is confirmed by several clinical trials conducted in healthy adults, which found that sugar sweetened beverage consumption results in more caloric intake and weight gain than artificially sweetened beverages.105106107 Additionally, a recent double blind, randomised controlled trial carried out in 94 healthy men suggested that consumption of sugar sweetened beverages containing fructose might induce a significant change in the low density lipoprotein particle distribution towards smaller, more atherogenic particles, partially mediating the associations of sugar sweetened beverage consumption with dyslipidaemia and cardiovascular disease.108
Another important mechanism to explain the associations between dietary sugar consumption and cardiometabolic diseases involves uric acid synthesis. Many studies have confirmed that excessive fructose consumption can promote uric acid synthesis by inducing degradation ATP to AMP, a substrate for uric acid production.109110111 Fructose phosphorylation in the liver uses ATP to convert fructose into fructose-1-phosphate and leads to phosphate depletion, which limits the regeneration of ATP from ADP. Then, ADP is converted to AMP and consequently induces the synthesis of uric acid.57 In addition, fructose induced hyperinsulinaemia and insulin resistance may also result in higher serum uric acid by reducing the excretion of uric acid.110112113 Hyperuricaemia is a precursor to gout.109110 The positive associations between gout, hyperuricaemia, and other cardiometabolic diseases, such as hypertension, type 2 diabetes mellitus, and cardiovascular disease, have been proposed for a long time.114115 Hyperuricaemia has been shown to precede the occurrence of type 2 diabetes mellitus and obesity.27 Mechanistically, hyperuricaemia could induce renal microvascular alteration, chronic sodium retention, reduction in nitric oxide concentrations in endothelial cells, and the activation of the renin-angiotensin system, which may account for the association between fructose containing sugar consumption and cardiovascular disease.114116117118
Until now, the evidence for the association between dietary sugar consumption and the risk of cancer has remained limited and controversial.27 In 2018 the World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) reported that evidence was limited for the associations between consumption of sugars and food containing sugars and the risk of colorectal cancer.119 However, at the same time, this report recommended reducing or avoiding sugar sweetened beverage consumption for the prevention of breast cancer.119 Evidence from this umbrella review supports the recommendations from the WCRF/AICR to some extent. In our study, although eight of the 25 cancer outcomes were identified as being positively associated with dietary sugar consumption (seven exposure factors were sugar sweetened beverages, and one was fructose), only evidence of hepatocellular carcinoma (sugar sweetened beverages) and pancreatic cancer (fructose) were rated as “low” quality because of the magnitude of effect or dose-response gradient, and the remaining outcomes were all rated as “very low” quality. As a result, caution is warranted when explaining the significant associations between dietary sugar consumption and some cancer risks.
The effect of dietary sugars on obesity might partly explain their association with the risk of cancer.21 As mentioned previously, dietary sugar consumption, especially sugar sweetened beverage consumption, is convincingly associated with the risk of obesity weight gain,1353 which in turn is regarded as a strong risk factor for various cancers.21119 Another pathway mediating the association between dietary sugar consumption and the risk of cancer might involve a high glycaemic index or glycaemic load. The glycaemic index has been associated with the risk of type 2 diabetes mellitus,120 which may be involved in carcinogenesis of the breast, prostate, liver, bladder, and endometrium.120121 Moreover, excessive fructose consumption might lead to intestinal flora disturbance and intestinal barrier deterioration, which promote the development of metabolic endotoxaemia, inflammation, and lipid accumulation, finally leading to colorectal carcinogenesis.20122123 A recent animal study showed that high fructose corn syrup intake could induce intestinal tumourigenesis in mice by expediting glycolysis and de novo lipogenesis. The mice treated with the syrup had a substantially increased tumour size and tumour grade, independent of obesity and metabolic syndrome.124 Considering the different mechanisms of site specific cancers, further prospective studies that explore the definite associations between sugar consumption and cancer risk for diverse cancer types and ethnic groups are warranted.27
On the other hand, dietary sugar consumption has also been shown to be negatively associated with some neuropsychiatric diseases, such as depression and attention deficit/hyperactivity disorder.7780 Several biological mechanisms might be involved in these associations. Data from an animal study showed that a high fructose diet might alter behaviour, hypothalamic-pituitary-adrenal axis function, and the hypothalamic transcriptome in male Wistar rats, inducing anxiety-like behaviour and depressive-like behaviour.125 Furthermore, sugar consumption has been suggested to stimulate the secretion of endogenous opioids in the nucleus accumbens and to stimulate the dopaminergic reward system.27 Evidence of sugar dependence in an animal model indicated that similarly to addiction to morphine and cocaine, rats with intermittent sugar intake had decreased concentrations of dopamine D2 receptor mRNA in the nucleus accumbens and showed the characteristics of addictive-like behaviours called sugar addiction.27126
In addition, the adverse association between sugar consumption and bone mineral density might be attributed to the increased loss of urinary calcium and imbalance in calcium homoeostasis induced by high sugar intake.127 As well as the negative effect of sugars, phosphate, acidity, and caffeine contained in sugar sweetened beverages are three other major factors that affect bone metabolism.81 We note that for the link between sugar sweetened beverages and bone mineral density, stratification analysis by gender showed a significant harmful effect of sugar sweetened beverages on bone mineral density in females but not in males.81 These diverse findings indicated that sugar sweetened beverage consumption had a more detrimental effect on female bone health than on male bone health because women generally have smaller bones and lower bone strength and are therefore more susceptible to osteoporosis.128 Moreover, the high acidity of sugar sweetened beverages is also thought to be an important factor in promoting dental caries and tooth erosion.129130131
Of the subgroup analyses conducted in this umbrella review, the most noteworthy is the stratification according to region, as several health outcomes showed a regional discrepancy, including overall cancer mortality, high density lipoprotein cholesterol, low density lipoprotein cholesterol, and total cholesterol. Potential reasons for these discrepancies may include regional differences in sugar consumption and culture. According to the report conducted in 2010 for the quantification of global, regional, and national consumption of sugar sweetened beverages in 187 countries, consumption among Asian countries was lower than that among European and American countries.33 The consumption of sugar sweetened beverages was highest in the Caribbean and lowest in East Asia and Oceania.33 In addition, cultural factors have been shown to potentially cause different dietary quality and health inequalities by affecting food preferences or choices.132 Regional cultural diversity in lifestyle and sociodemographic factors also plays an important role in dietary sugar consumption, which may partly explain the different relations between sugar consumption and disease risk in ethnically diverse populations.132133 On the other hand, subgroup analyses with adjustment for confounding factors should also be considered. High consumption of sugars, especially sugar sweetened beverages, may be a marker of an unhealthy diet and lifestyle.966 People who consumed sugar sweetened beverages more frequently were likely to ingest more total and saturated fat, carbohydrate, and sodium and less fruit, fibre, dairy products, and wholegrain foods.134135136137138 This dietary pattern was also associated with more frequent smoking and drinking, lower physical activity levels, and more time spent watching television.137138 Therefore, the role of these confounding factors should be taken into consideration when explaining the association between sugar consumption and burden of disease.
Strengths and weaknesses of study and in relation to other studies
This umbrella review first reported a comprehensive summary of the current evidence from previous meta-analyses of observational studies and randomised controlled trials for the association between dietary sugar consumption and all health outcomes. Given the high levels of dietary sugar consumption worldwide, this study has clinical and social significance for developing preventive strategies against excessive sugar consumption, especially for children and adolescents. This study was carried out on the basis of systematic methods in which independent literature searching, study selection, and data extraction by two authors were involved. If the data were sufficient, we reanalysed the risk ratio, odds ratio, weighted mean difference, or standardised mean difference with 95% confidence intervals through random or fixed effects models and evaluated the heterogeneity and publication bias for each included meta-analysis. Furthermore, we used three standard approaches, AMSTAR, GRADE, and evidence classification criteria, to assess the methodological quality (AMSTAR), strength (GRADE) and classification (evidence classification criteria) of evidence for each health outcome and to evaluate our confidence in the estimates. Interestingly, in our umbrella review, the GRADE rating of several health outcomes was not completely consistent with the results of evidence classification. As we know, evidence classification criteria are a completely objective classification standard, whereas the GRADE rating has a certain degree of subjectivity.139 Therefore, both the GRADE rating and evidence classification criteria should be considered when evaluating evidence and making recommendations.
Original studies included in meta-analyses used different methods of food intake investigation, including food records, 24 hour dietary recall, food frequency questionnaires, and dietary history. All of these are associated with an unavoidable measurement bias even if validated methods are used.3 This limitation is common to all major epidemiological studies carried out worldwide in this field.21 In addition, most studies focused on beverages pre-sweetened before purchase.9 For instance, in the Nurses’ Health Study, coffee with sugars was excluded from sugar sweetened beverages, which might affect the reliability of the association.137 Similarly, another limitation of our study was that we could not evaluate sugar intake in some foods that potentially contain sugars, such as chocolate and ice cream, because of a failure to extract data on sugar consumption. Furthermore, the types of sugar sweetened beverages and dosage of their consumption varied in the original studies. In this umbrella review, most meta-analyses produced summary effects from original studies that measured exposure to dietary sugars through the number of servings a day. However, in some original studies, the number of millilitres a day, grams a day, times a day, times a week, times a month, servings a week, or servings a month were used to estimate sugar consumption, which may partly explain the origin of heterogeneity in meta-analyses. Therefore, dose-response analysis and stratification analysis by sugar sweetened beverage types were unavailable for most outcomes owing to diverse measurements of sugar sweetened beverage consumption in the original studies. Consumption of sugars in sugar sweetened beverages is generally accompanied by the ingestion of some other chemical compounds, such as 4-methylimidazole,140141 pesticides,142143 artificial sweeteners,144 sodium benzoate,79 and sulfites,79 which may confuse the effect of sugars and therefore should be regarded as potential confounding factors.
We reviewed details of competing interest and funding disclosures from meta-analyses included in this umbrella review. Only two meta-analyses were funded by companies that produce sugar sweetened beverages.65145 Among them, the meta-analysis conducted by Wang and colleagues was selected for data extraction and is shown in summary tables.65 Therefore, caution is warranted when explaining the non-significant association between fructose intake and postprandial triglycerides. Another meta-analysis was not selected for data extraction,145 and the list of all meta-analyses not selected for data extraction and reanalysis are available if needed. We did not investigate the original studies included in each meta-analysis and therefore could not confirm whether these studies had a competing interest with companies associated with the sugar industry.42
The harmful association between dietary sugar consumption and multiple health outcomes observed in our umbrella review is supported by several large scale prospective cohort studies published in recent years. The first was a large prospective cohort study conducted using the results of the French NutriNet-Santé cohort (2009-17), which included 101 257 participants with an average age of 42.2.21 During the eight year follow-up period, a total of 2193 cases of cancer were reported, including 693 cases of breast cancer. A harmful association was found between sugar sweetened beverage consumption and the risk of overall cancer (hazard ratio 1.18, 1.10 to 1.27) and breast cancer (1.22, 1.07 to 1.39). No significant association was observed for sugar sweetened beverage consumption and the risk of prostate, colorectal, and lung cancer.21 In this umbrella review, however, the highest versus lowest sugar sweetened beverage consumption was associated with a 17% increased risk of prostate cancer, without a dose-response gradient. Notably, the non-significant association between sugar sweetened beverage consumption and the risk of colorectal cancer observed both in this study and in our umbrella review was inconsistent with another cohort conducted in women.20 In the Nurses’ Health Study II (1991-2015), the authors prospectively explored the association of sugar sweetened beverage consumption in adulthood and adolescence with the risk of early onset colorectal cancer among 95 464 women. A total of 109 cases of early onset colorectal cancer were confirmed during follow-up. Compared with women who consumed less than one serving a week of sugar sweetened beverages in adulthood, those who consumed two or more servings a day had a 118% higher risk of early onset colorectal cancer (risk ratio 2.18, 1.10 to 4.35). Each one serving a day increment of sugar sweetened beverage consumption was associated with a 16% (risk ratio 1.16, 1.00 to 1.36) increased risk of early onset colorectal cancer.20 In addition, another prospective cohort study showed that excessive consumption of sugars and sugar sweetened beverage during adolescence was significantly associated with the risk of colorectal adenoma (odds ratio 1.20, 1.04 to 1.39).146 Each one serving a day increase in sugar sweetened beverage consumption was associated with 11% (odds ratio 1.11, 1.02 to 1.20) and 30% (1.30, 1.08 to 1.55) higher risks of total colorectal adenoma and rectal adenoma, respectively.146 Given that the association between sugar consumption and colorectal cancer risk remains controversial, further well designed, large scale prospective studies are needed to clarify it.
The positive associations between sugar sweetened beverage consumption and the risk of mortality detected in this umbrella review were supported by a prospective cohort study of 118 363 people followed for 34 years in the US, during which time 36 436 deaths were documented.147 After adjustment for diet and lifestyle confounders, the consumption of two or more servings of sugar sweetened beverages a day was linked with a 21% (hazard ratio 1.21, 1.13 to 1.28) higher risk of total mortality, a 31% (1.31, 1.15 to 1.50) higher risk of cardiovascular disease mortality, and a 16% (1.16, 1.04 to 1.29) higher risk of cancer mortality.147 On the other hand, a prospective cohort study of 120 343 UK participants followed for 8.4 years confirmed the harmful association of added sugar consumption with the risk of type 2 diabetes mellitus.148 A dietary pattern high in added sugars was associated with a higher incidence of type 2 diabetes mellitus (hazard ratio 1.09, 1.06 to 1.12) after adjustment for confounders.148 Similar to their findings, we observed a strongly significant association between consumption of sugar sweetened beverages (one of the main sources of added sugars) and the risk of type 2 diabetes mellitus.