Compelling evidence linking sugary drinks with diabetesBMJ 2015; 351 doi: https://doi.org/10.1136/bmj.h4087 (Published 29 July 2015) Cite this as: BMJ 2015;351:h4087
- Edward W Gregg, chief, Epidemiology and Statistics Branch,
- Ann Albright, director
- 1Division of Diabetes Translation, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, GA, USA
- Correspondence to: E W Gregg
Global estimates of diabetes prevalence continue to outpace previous projections, as virtually all regions of the world continue to see a large and growing prevalence.1 Proposals by diabetes prevention experts for how to change these trends have included both structured programmes for adults at high risk for type 2 diabetes (the “high risk approach”) and environmental and food policies to help change risk factors in the whole population (the “population approach”).1 Promising programmes for those at high risk are already underway and need to be expanded,2 but complementary shifts in underlying risk factors in the broader population are also needed. The list of risk factors for type 2 diabetes is long, with dozens of demographic, behavioural, and genetic risk factors now identified.3 Unfortunately the list of simple, consensus based levers to change population risk factors and substantially influence the risk for type 2 diabetes remains short.
In a linked paper (doi:10.1136/bmj.h3576) Imamura and colleagues present a compelling case that beverages sweetened with sugar should be a major target for policies aimed at reducing the risk of diabetes.4 A systematic review of 17 prospective studies reported three findings; a higher intake of sugar sweetened beverages was associated with an increased incidence of type 2 diabetes, associations with juice consumption were weaker and less consistent, and the link between artificially sweetened beverages and diabetes incidence was largely explained by obesity, perhaps indicating that the association with risk of type 2 diabetes is due to obese people preferentially choosing artificially sweetened beverages.4
Each extra 250 mL serving (currently consumed by more than one third of the US population5) of sugar sweetened beverages, or soft drinks each day, was associated with an 18% increased risk of new diabetes. Important minority segments of the population—notably children and young adults—have even higher exposures and higher risks. Imamura and colleagues also found that the association between intake of sugar sweetened drinks and diabetes seems to be independent of obesity and thus can compound the effects on obesity, consistent with previous findings that sugar sweetened drinks have direct effects on glycaemia and insulin resistance. Furthermore, the finding that 9-12% of new cases of type 2 diabetes in the United States (and 20% of cases among those aged 20-44) are potentially attributable to sugar sweetened drinks, may be even more important for people in countries such as Mexico, selected parts of the Caribbean, Latin America, and South East Asia, where consumption has been shown to be much higher.4 6
Some risk factors do their damage in an epidemic by substantially increasing risk for the minority of people who have that risk factor (for example, smoking for cancer and cardiovascular disease; severe obesity for type 2 diabetes).7 8 Such factors lend themselves to targeted interventions confined to high risk populations; but other risk factors do their damage by increasing risk a modest amount in large numbers of people. These latter types can lend themselves to policy approaches that can efficiently change risk factors among large segments of the population. Conditions are ideal for a population targeted policy when a full cascade of clear evidence is in place: a causal relation between risk factor and health; ability to change health by reversing the risk factor; availability of low risk, low cost interventions to alter the risk factor; and a strong policy lever to implement the intervention efficiently.
Some, but not all, of this evidence cascade exists for sugar sweetened drinks. They singularly account for a large proportion of dietary added sugar and permit focused interventions.9 Modelling studies that incorporate additional downstream effects on obesity, diabetes, hypertension, cardiovascular disease, and cancers suggest that reducing consumption could have substantial effects on morbidity and mortality.6
It is, however, one thing to establish an association (which may be causal) between sugared drinks and incident diabetes in cohort studies and yet another to show that reducing consumption improves population health. Trials of interventions to lower consumption have shown reductions in body weight, but we are not aware of any studies with diabetes as an outcome.10 11 Furthermore, the policy levers that have been proposed by experts in obesity prevention and management—taxation or other pricing strategies, better product labeling, selected limits on sales and advertising, improved access to water and other lower calorie options12—have not been evaluated for their effects on diabetes prevention.13
Nevertheless, high consumption of sugary drinks—two to four times higher than it was in the 1980s—combined with clear associations between intake and risk of chronic diseases, leaves little room for doubt that action is warranted.5 6 The most compelling reason to gradually decrease intake of sugary drinks—to prevent obesity and diabetes in the first place—is difficult to test in traditional trials. The pressing questions to track are whether policies to reduce the intake of sugar sweetened beverages can make any dent in the diabetes problem, whether reducing sugary beverages is replaced by other substances that negate benefits, and what specific approaches will reduce consumption while lowering risk factors and improving health.
Action is already underway in the areas of relative pricing and reducing access in diverse international settings,14 15 and some evidence suggests that in the US the intake of sugar sweetened beverages, obesity, and diabetes prevalence have decreased, although the reasons for such progress remain unclear.16 17 Active policy development in the face of serious public health need would benefit from the design of rigorous natural experiments to quantify the magnitude of real health effects, compare different policy approaches, and monitor populations for compensatory changes in behaviour and unintended consequences.18 Most importantly, evidence from natural experiments in countries that adopt policies early could help others learn from their mistakes and refine their successes.
Cite this as: BMJ 2015;351:h4087
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following interests: none.
Provenance and peer review: Commissioned; not externally peer reviewed.