Millions of people reach for diet soda as a seemingly safe swap for sugar-sweetened drinks, especially those managing blood sugar or trying to lose weight. A new narrative review paired with a meta-analysis of 21 randomized controlled trials now challenges that assumption, finding that non-nutritive sweeteners produced small but measurable increases in fasting insulin and HbA1c, a key marker of long-term blood-sugar control. The work, affiliated with the Food is Medicine Institute at Tufts University, adds fresh pressure to a debate already split between global and U.S. dietary authorities.
Why rising HbA1c signals from diet drinks matter right now
HbA1c reflects average blood-glucose levels over roughly three months, making it the standard gauge clinicians use to track diabetes risk and management. Even a small upward nudge in that marker, if confirmed across populations, could mean that a product marketed as metabolically neutral is instead pushing some users closer to prediabetic thresholds. The new meta-analysis focused specifically on trials that compared non-nutritive sweeteners against non-caloric controls such as water or unsweetened placebo, a design choice that strips out the confounding benefit of simply replacing sugar calories. By isolating sweetener exposure this way, the researchers reported harmful effects on glucose-insulin homeostasis including fasting insulin and HbA1c.
One plausible mechanism runs through the gut. Emerging research suggests that artificial sweeteners can alter the composition of intestinal microbiota, and individuals whose baseline gut communities already show low microbial diversity may be most vulnerable to metabolic disruption. If stored stool or blood samples from the 21 trials were re-analyzed and stratified by microbial alpha diversity, the HbA1c increase could turn out to be concentrated in that subgroup rather than evenly spread. No such stratification has been published, but the hypothesis aligns with animal and small-cohort human data linking sweetener intake to shifts in bacterial populations that regulate glucose metabolism.
Another possible pathway involves cephalic-phase insulin release, in which the brain responds to sweet taste with anticipatory hormonal changes even when no calories arrive. Repeated stimulation of sweet taste receptors without accompanying energy might, in theory, contribute to subtle insulin dysregulation over time. The current evidence base cannot disentangle these mechanisms, but the pattern of slightly higher fasting insulin alongside modestly elevated HbA1c is consistent with early-stage insulin resistance rather than dramatic spikes in day-to-day glucose.
Competing evidence from WHO and U.S. dietary reviews
The Tufts-affiliated findings do not exist in a vacuum. The World Health Organization issued a conditional recommendation against routine use of non-sugar sweeteners, drawing on its own commissioned systematic review and meta-analysis that examined both randomized trials and observational studies. The WHO guideline weighed biomarkers including fasting glucose, fasting insulin, and HbA1c, and concluded that the evidence did not support long-term metabolic benefits from these additives. Supporting evidence tables catalogued certainty ratings for each outcome domain, noting mixed signals across trial designs.
The U.S. Dietary Guidelines Advisory Committee took a different tack in its 2025 evidence review. Its systematic assessment of low- and no-calorie sweetened beverages and type 2 diabetes risk found mostly null effects on glycemic outcomes when diet drinks were compared with water. One exception stood out: the Madjd trial, an 18-month follow-up randomized clinical trial, reported that participants assigned to water showed better results on fasting insulin, HOMA-IR (a measure of insulin resistance), and two-hour postprandial glucose than those who continued drinking diet beverages. That single trial, documented under long-term beverage data, is one of the longest intervention studies in this space and carries outsized weight precisely because most other trials ran for only weeks or a few months.
The split between the WHO position and the U.S. advisory committee’s reading of the same evidence base reflects genuine uncertainty about effect size, study duration, and population differences. WHO treated observational data showing associations with type 2 diabetes as supporting evidence for caution; the U.S. panel leaned more heavily on the largely null short-term trial results. Both approaches are defensible, but they lead to different practical messages: one emphasizes precautionary reduction, the other accepts diet drinks as a tool for cutting sugar, at least in the near term.
Gaps the 21-trial meta-analysis cannot yet close
Several questions remain open. The full list of the 21 individual trials included in the new meta-analysis, along with trial-level effect sizes, is not yet available in public summary records. Without that granularity, outside researchers cannot assess whether the pooled HbA1c signal is driven by a handful of outlier studies or spread consistently across different sweetener types such as aspartame, sucralose, and stevia. Sweetener-specific effects matter because each compound interacts differently with taste receptors and gut bacteria.
Participant-level microbiome sequencing data from the trials cited by the U.S. advisory committee remain limited to narrative summaries rather than downloadable datasets. That gap blocks the kind of diversity-stratified reanalysis that could clarify whether low-diversity gut profiles amplify the metabolic response. Long-term HbA1c trajectories beyond the Madjd trial’s 18-month window are also absent, leaving open the question of whether the fasting-insulin and HOMA-IR differences persist, widen, or fade over years of habitual use.
There are also unresolved issues around dose and pattern of intake. Many randomized trials use fixed daily servings of diet beverages or capsules containing specific sweeteners, which may not reflect the sporadic but cumulative way people consume these products in the real world. It is unclear whether occasional diet soda use carries the same metabolic signature as multiple servings a day, or whether there is a threshold below which effects are negligible. Without that dose-response information, clinicians and patients are left with broad guidance rather than precise limits.
What this means for daily drink choices
For people who drink diet soda daily, the practical takeaway is not that every can is toxic, but that “free pass” assumptions are no longer supported by the most comprehensive analyses to date. The emerging picture is one of small but directionally adverse shifts in markers that matter for long-term metabolic health, superimposed on a background of uncertainty about who is most susceptible and how effects accumulate over years.
For individuals with diabetes or prediabetes, that nuance matters. Replacing sugary beverages with diet versions still avoids large, immediate glucose spikes and excess calories, which remain clear priorities. At the same time, the new findings suggest that water, unsweetened tea, and other minimally processed drinks should be the default, with diet sodas treated as occasional tools rather than cornerstone therapies. People who rely heavily on non-nutritive sweeteners might consider gradually cutting back, monitoring HbA1c and fasting insulin where possible, and paying attention to changes in appetite or cravings as they do so.
Policy makers and guideline writers face a tougher balancing act. Tightening recommendations against routine sweetener use, as WHO has done, may help shift population habits toward less sweetness overall, but it also risks pushing some consumers back toward sugar-sweetened products if appealing alternatives are not available or affordable. Maintaining a more permissive stance, as the U.S. advisory committee’s interpretation implies, preserves an option for sugar reduction but could underplay subtle metabolic downsides that only become apparent at scale.
Ultimately, the debate over diet sodas and other non-nutritive sweeteners is moving away from simple “good versus bad” framing toward a more conditional calculus: better than sugar in many short-term contexts, yet not metabolically inert, and potentially problematic for certain subgroups or over long horizons. As more detailed trial data and microbiome analyses emerge, that calculus may sharpen. For now, the safest bet for most people is to treat intense sweetness-whether from sugar or its calorie-free stand-ins-as something to dial down, not simply swap.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
