Diet soda, soda or water

A 2022 network meta-analysis (a method that compares several options across a set of trials) pooled 17 randomised trials of 1,733 adults with overweight or obesity who had or were at risk of diabetes. In the 12 trials (601 people) where low- or no-calorie sweetened drinks replaced sugar-sweetened drinks, body weight was 1.06 kg lower (95% CI −1.71 to −0.41), body mass index 0.32 lower, body fat 0.60 percentage points lower, and liver fat lower (standardised mean difference −0.42), with no evidence of harm. The authors rated the certainty for body weight moderate [1]. In the 6-month CHOICE trial, 318 adults with overweight or obesity replaced at least 200 kcal a day of caloric drinks with diet drinks or water, or made dietary changes of their choosing. Weight fell 2.5% with diet drinks, 2.03% with water and 1.76% with the comparison diet, with no significant difference between groups; the diet-drink group was more likely than the comparison group to reach 5% weight loss (odds ratio 2.29, 95% CI 1.05 to 5.01) [11].

A 2024 meta-analysis of 11 prospective studies (2,196,503 participants, 235,609 deaths, follow-up 7.0 to 28.9 years) found that replacing one serving a day of sugar-sweetened drinks with artificially sweetened drinks was associated with 4 to 6% lower all-cause and cardiovascular mortality [3]. In 118,363 US nurses and health professionals followed for up to 34 years, mortality rose step by step with sugar-sweetened drinks, to a hazard ratio (how much the death rate changed) of 1.21 at 2 or more a day against under 1 a month, mainly through cardiovascular deaths. Artificially sweetened drinks were associated with higher mortality only in the highest category (hazard ratio 1.04), and only in the women’s study, not the men’s [2]. A 2022 meta-analysis of substitution analyses linked diet drinks in place of sugary drinks with 11% less coronary heart disease (6 studies, 233,676 participants) and 4% lower total mortality (1 study, 118,363 participants) [7]. Per daily serving, sugar-sweetened drinks were associated with 8% higher all-cause mortality across 11 studies of 965,851 participants, evidence the authors rated high quality [4].

In the same 2022 network meta-analysis, 9 trials (974 people) compared diet drinks with water. There was no clear difference in body weight, body fat or liver fat. HbA1c, a measure of average blood sugar, was 0.21 percentage points higher with diet drinks (95% CI 0.02 to 0.40), and systolic blood pressure was 2.63 mm Hg lower (−4.71 to −0.55). Certainty for body weight was low [1]. Individual trials point in opposite directions. In a 1-year trial of 303 adults with overweight or obesity who already drank diet drinks, those assigned 710 mL a day of diet drinks maintained a 6.21 kg weight loss against 2.45 kg for those assigned water [5]. In a 24-week trial of 89 women with overweight or obesity who usually drank diet drinks (62 completed), those who switched to water lost 8.8 kg against 7.6 kg for those who kept diet drinks, with larger falls in fasting insulin and insulin resistance [6]. In the CHOICE trial, the water group had a significant fall in fasting glucose against the comparison group [11].

Few long-term studies compare diet drinks with water directly. A 2022 meta-analysis of 14 prospective studies (416,830 participants) modelled substitutions: diet drinks in place of water showed no adverse associations. Water in place of sugary drinks was linked to lower weight (0.10 kg a year), 15% less obesity (relative risk 0.85) and 4% less type 2 diabetes (0.96, 3 studies, 281,855 participants); diet drinks in place of sugary drinks were linked to 12% less obesity. The authors rated this evidence low to very low certainty [7]. A 2015 meta-analysis of 17 studies found each daily serving of artificially sweetened drinks associated with 25% more type 2 diabetes, falling to 8% after adjustment for body weight, and reported signs of publication bias and residual confounding for these drinks [12]. A 2019 systematic review of 56 studies found no difference between non-sugar sweetener use and non-use for most health outcomes, with confidence limited by small, short studies [13].

Across 11 prospective studies, the highest intake of artificially sweetened drinks was associated with 13% higher all-cause mortality (relative risk 1.13, 95% CI 1.06 to 1.21) and 26% higher cardiovascular mortality than the lowest, with no association for cancer deaths [3]. A 2021 dose-response analysis found a J-shaped curve for all-cause mortality: hazard ratios of 1.01 at 1 serving a day (not significant), 1.08 at 2 and 1.13 at 2.5 [4]. In 81,714 US postmenopausal women in the Women’s Health Initiative, followed for a mean 11.9 years, 2 or more diet drinks a day against under 1 a week was associated with 23% more stroke, 31% more ischaemic stroke, 29% more coronary heart disease and 16% higher all-cause mortality; 5.1% of the women drank that much [14]. In the Framingham Offspring study, daily artificially sweetened soft drinks against none were associated with 2.96 times the rate of ischaemic stroke and 2.89 times the rate of Alzheimer’s dementia, from 97 strokes and 81 dementia cases; sugar-sweetened drinks were not associated with either [15].

People often start drinking diet soda after gaining weight or being diagnosed with diabetes, so the illness can precede the drink, a pattern called reverse causation. A 2025 umbrella review (a review of earlier meta-analyses) compared two kinds of analysis. Simple comparisons of current intake linked low- and no-calorie sweeteners with more obesity, diabetes, stroke, and cardiovascular and all-cause mortality. Analyses of changes in intake, or of one drink substituted for another, linked them with lower body weight, less obesity and coronary heart disease, and lower cardiovascular and all-cause mortality. Certainty was generally moderate for the trials and very low for the long-term studies [8]. The 2024 mortality meta-analysis reported substantial differences between its studies (I² 66.3%) [3], and the 2015 diabetes meta-analysis found signs of publication bias for artificially sweetened drinks [12].

In July 2023 the International Agency for Research on Cancer (IARC) classified aspartame as possibly carcinogenic to humans (Group 2B), based on limited evidence for hepatocellular carcinoma, a type of liver cancer. The Joint FAO/WHO Expert Committee on Food Additives (JECFA) reaffirmed an acceptable daily intake of 0 to 40 mg per kg of body weight. With 200 to 300 mg of aspartame in a can of diet soft drink, a 70 kg adult would need more than 9 to 14 cans a day to exceed that intake, with no other source [9]. In the French NutriNet-Santé study (102,865 adults, median 7.8 years, sweeteners from all foods and drinks), higher artificial sweetener intake was associated with 13% more cancer overall (hazard ratio 1.13, 95% CI 1.03 to 1.25), 15% more for aspartame and 13% more for acesulfame-K; aspartame was associated with 22% more breast cancer [16]. In 103,388 adults from the same study, total artificial sweetener intake was associated with 9% more cardiovascular disease, aspartame with 17% more cerebrovascular disease, and acesulfame-K with 40% more coronary heart disease [17].

In a randomised trial of 120 healthy adults given saccharin, sucralose, aspartame or stevia for 2 weeks at doses below the acceptable daily intake, each sweetener changed the gut and mouth bacteria, and saccharin and sucralose significantly impaired the body’s glucose response; mice given the participants’ gut bacteria largely reproduced the individual responses [18]. In NutriNet-Santé, sucralose was associated with 31% more coronary heart disease (hazard ratio 1.31, 95% CI 1.00 to 1.71) [17]. Erythritol, a sugar alcohol sweetener, was studied through blood levels rather than intake: in patients undergoing heart evaluation, the highest quarter of blood erythritol was associated with 1.80 times (US, 2,149 patients) and 2.21 times (Europe, 833 patients) the rate of heart attack, stroke or death. Erythritol increased platelet clotting activity in laboratory tests, and in 8 volunteers a single dose raised blood levels for more than 2 days [19].

In May 2023 the World Health Organization issued a conditional recommendation against using non-sugar sweeteners to control body weight or lower the risk of noncommunicable diseases. It covers acesulfame-K, aspartame, advantame, cyclamates, neotame, saccharin, sucralose, stevia and stevia derivatives, applies to everyone except people with existing diabetes, and suggests foods with naturally occurring sugars, such as fruit, or unsweetened foods and drinks instead. WHO cited evidence of no long-term benefit for body fat and possible undesirable effects with long-term use, including higher risk of type 2 diabetes, cardiovascular disease and mortality in adults [10]. Researchers who weight bias-adjusted analyses more heavily conclude that the evidence supports sweeteners as a way to reduce sugar [8, 1].