Does Drinking Diet Soda Increase Your Risk of Dementia?

Several large population studies have linked regular diet soda consumption to a higher risk of dementia, but at least one equally large study found no connection at all. A 2025 meta-analysis pooling results from seven prospective studies estimated that people who drank more artificially sweetened beverages had roughly a 42% higher risk of Alzheimer’s disease compared with those who drank fewer or none. That number sounds alarming, yet it comes with serious caveats about what these studies can and cannot prove. The gap between “associated with” and “causes” is especially wide here, and understanding why helps explain the real state of the science.

What the Population Studies Actually Found

The study that kicked off most of the public concern came from the Framingham Heart Study, published in 2017. Researchers tracked middle-aged and older adults over about a decade and found that those who reported drinking at least one artificially sweetened beverage per day had roughly triple the risk of ischemic stroke and Alzheimer’s disease compared to people who drank less than one per week. Interestingly, sugar-sweetened beverages showed no such association in the same analysis.¹ Those numbers got a lot of headlines, but the confidence intervals were wide, meaning the true risk could have been anywhere from modestly elevated to very elevated. The study also had a relatively small number of dementia cases, which makes any single estimate less stable.

A 2025 dose-response meta-analysis that pooled seven prospective cohort studies found a more moderate but still meaningful association. Higher artificially sweetened beverage intake was linked to about a 42% increased risk of Alzheimer’s disease. Notably, sugar-sweetened beverages also showed an increased risk of about 49% in that same analysis, which somewhat undercuts the narrative that diet soda is uniquely dangerous.²

But then there is the study that found nothing. A large 2025 analysis pooling data from multiple cohorts of older adults reported that neither sugar-sweetened nor artificially sweetened beverages were significantly associated with incident dementia at all. The hazard ratio for artificially sweetened beverages was essentially 1.00 per serving per week, about as close to “no effect” as you can get in epidemiology.³ This is not a minor study, and it directly contradicts the Framingham findings. One analysis from the UK Biobank found that drinking more than two artificially sweetened beverages a day was associated with a 20% higher dementia risk, but the confidence interval crossed 1.0 for that specific comparison, meaning the result was not statistically robust.⁴

So the large human studies do not agree. Some find a worrying signal, others find essentially nothing. The honest summary is that there may be a modest statistical association between heavy diet soda consumption and dementia risk, but we cannot be sure it is real, and we definitely cannot be sure it is causal.

The Reverse Causation Problem

One of the biggest methodological headaches in this area is reverse causation, which is the possibility that early, undetected brain changes lead people to change their diets before anyone knows they are on the path to dementia. A study of a large French cohort found that people who went on to develop dementia showed differences in their dietary patterns, including shifts toward snacking and processed foods, as much as a decade before their diagnosis.⁵ If subtle cognitive decline causes people to gravitate toward convenient, sweet beverages, then the association between diet soda and dementia might have the arrow of causation pointing in the wrong direction.

There is also straightforward confounding. People who drink a lot of diet soda are often managing conditions like obesity, type 2 diabetes, or cardiovascular disease. They may have switched from regular soda on a doctor’s advice. All of these conditions are themselves well-established risk factors for dementia. The best observational studies try to adjust for these factors statistically, but you can never be certain you have accounted for everything. Someone who drinks four diet sodas a day is probably living a different overall lifestyle from someone who drinks water and unsweetened tea, in ways that a questionnaire may not fully capture.

This does not mean the association is definitely spurious. It means that the kind of study needed to prove causation, a large randomized controlled trial assigning people to drink diet soda or a placebo for decades, has never been done and probably never will be. We are stuck interpreting observational data, and observational data in nutrition science has a long track record of generating associations that do not hold up.

Proposed Biological Pathways

Even though the human evidence is conflicting, researchers have identified several plausible routes by which artificial sweeteners could, in theory, affect brain health. None of these have been proven to cause dementia in humans, but they help explain why the hypothesis has not been dismissed outright.

The gut microbiome is probably the most actively studied pathway. Artificial sweeteners appear to change the composition of gut bacteria in ways that reduce populations of beneficial species, decrease the production of short-chain fatty acids (which help maintain the gut lining), and promote low-grade systemic inflammation.⁶ Chronic low-grade inflammation is implicated in many age-related diseases, including Alzheimer’s. The connection from sweetener to gut disruption to brain inflammation is biologically plausible, but each link in that chain still needs more evidence in humans.

A computational analysis identified over 100 molecular targets shared between common non-nutritive sweeteners and neuroinflammatory processes, with enrichment in pathways related to Alzheimer’s disease and inflammatory response.⁷ This is suggestive at the level of molecular biology, but computational pathway analysis generates hypotheses rather than confirming them. Finding that two things share molecular targets does not mean one causes the other in a living organism.

Vascular health is another avenue of concern. Rat studies have shown that artificial sweeteners consumed at levels within the accepted daily intake can impair the ability of blood vessels to dilate properly, a condition called endothelial dysfunction.⁸ The brain depends on healthy blood flow, and vascular dysfunction is closely tied to both stroke and vascular dementia. A broader review of the evidence has noted that certain sweeteners may contribute to systemic inflammation and disruptions in glucose metabolism.⁹

Insulin resistance is worth mentioning separately. A study of people with type 2 diabetes found that those who regularly used artificial sweeteners had substantially higher measures of insulin resistance compared to non-users.¹⁰ Insulin resistance in the brain has been linked to Alzheimer’s disease in its own right, which is part of why some researchers have called Alzheimer’s “type 3 diabetes.” But whether sweeteners cause insulin resistance or whether people with worse metabolic health simply reach for sweeteners more often is, again, unclear from observational data alone.

What Animal Experiments Show

Because you cannot ethically randomize humans into decades of diet soda consumption, much of the mechanistic evidence comes from animal studies. These have produced some striking results, but translating rodent findings to humans requires caution.

One of the more widely discussed experiments gave mice aspartame at doses equivalent to roughly two to four small diet sodas per day for up to 16 weeks. The mice developed significant deficits in spatial learning and memory. Even more provocatively, these cognitive effects appeared to be heritable: the offspring of exposed males also showed deficits, even though the offspring themselves never consumed aspartame.¹¹ That finding raised eyebrows, though it has not been replicated in humans, and mouse models of cognition do not map neatly onto human dementia.

On the other hand, a PET imaging study in humans found that aspartame consumed at normal dietary levels caused only a small, likely insignificant decrease in the rate at which amino acids crossed the blood-brain barrier. The researchers concluded that under typical consumption patterns, aspartame is unlikely to produce measurable changes in brain amino acid uptake.¹² This is an older study but still one of the few that directly measured what happens inside the human brain after consuming aspartame. It is reassuring for the specific mechanism it tested, though it says nothing about gut-mediated or vascular pathways.

Not All Sweeteners Behave the Same Way

One problem with lumping all “diet soda” together is that different products contain different sweeteners, and these appear to have different effects on the brain. A preliminary human study compared the cognitive effects of three substances: regular sugar (sucrose), sucralose, and steviol glycosides (derived from the stevia plant). After supplementation, the sucralose group showed significant decreases in overall memory, encoding memory, and executive function, along with changes in brain wave patterns. The sucrose group showed a decrease only in encoding memory. The steviol glycosides group showed no significant changes at all.¹³

This was a small, preliminary study, so it would be premature to declare sucralose dangerous or stevia safe based on it alone. But it highlights a point that the population-level studies mostly ignore: asking whether “artificial sweeteners” affect dementia risk may be too broad a question. Aspartame, sucralose, saccharin, and acesulfame potassium are chemically distinct compounds that are metabolized differently. The epidemiological studies usually ask participants about “diet soda” consumption without distinguishing which sweetener is in the can. If one sweetener is harmful and another is not, mixing them together in the analysis would dilute the signal from the harmful one and potentially create a false signal for the harmless one.

This is an area where the research badly needs to catch up. Most large cohort studies were not designed to track specific sweetener exposure, and reformulations by beverage manufacturers mean that what was in a can of diet soda in 2005 may not be what is in one today.

Does Switching to Coffee or Tea Help?

For people who are worried and wondering what to drink instead, there is some relevant evidence. A large UK Biobank study found that statistically replacing artificially sweetened beverages with coffee or tea was associated with a reduced risk of dementia, particularly among people who carried other modifiable risk factors such as high blood pressure or physical inactivity.¹⁴ This is a substitution analysis, which models what would happen if people swapped one beverage for another based on observed patterns. It is not the same as a randomized trial, but it does suggest that the beverage you replace diet soda with matters.

Coffee and tea both contain polyphenols and caffeine, which have shown neuroprotective properties in other research. Whether the benefit of switching comes from removing artificial sweeteners, gaining protective compounds from coffee and tea, or some combination is impossible to tease apart from observational data. But if you are looking for a practical takeaway, “drink more coffee or tea and less of anything sweetened” is one of the more consistent signals in the nutritional epidemiology of dementia, regardless of how you feel about artificial sweeteners specifically.

How to Think About Your Own Risk

If you drink a diet soda occasionally, the current evidence does not suggest you are meaningfully increasing your dementia risk. The associations that do show up in the literature are generally seen at higher consumption levels, usually daily or near-daily intake sustained over years. And even at those levels, the most recent and largest studies have not been able to confirm the association consistently.

The established, well-supported risk factors for dementia remain far more powerful than anything a can of diet soda is likely to contribute. Physical inactivity, poorly controlled blood pressure, untreated hearing loss, social isolation, smoking, excessive alcohol use, and poorly managed diabetes all have stronger and more consistent evidence linking them to cognitive decline. If you are worried about your brain health, addressing any of those factors is a much better use of your energy than agonizing over your beverage choice.

That said, drinking several diet sodas every day is not something any nutritional authority recommends for other reasons. The potential effects on gut health, insulin sensitivity, and cardiovascular function are concerning enough in their own right, even setting the dementia question aside. Water, coffee, and tea are cheap, widely available, and consistently associated with neutral-to-positive health outcomes. Diet soda is, at best, a compromise. Whether that compromise carries a real cost for brain health specifically is a question the science has not yet settled, and it may take another decade of well-designed studies to get a definitive answer.

When Research Gets Ahead of Itself

The diet-soda-and-dementia conversation is a case study in how preliminary findings get amplified. The original Framingham study was solid for what it was: a well-conducted observational analysis from a respected cohort. But the finding of a roughly threefold risk increase got reported as though it were established fact, when in reality it was a single analysis from one population with wide confidence intervals. The subsequent years of research have not reproduced that magnitude of effect. The meta-analytic estimate of about a 42% increased risk is both more modest and derived from a broader evidence base, yet it still cannot rule out confounding or reverse causation.²

Meanwhile, the null result from the 2025 JAMA Psychiatry analysis received far less attention, as negative findings in nutrition science usually do.³ The public discourse tends to remember the scary number and forget the study that found nothing. For a topic where the honest answer is “we are not sure,” that asymmetry in coverage creates a misleading impression of certainty that the data does not support.

Researchers studying this question face genuine methodological challenges that are unlikely to be resolved with more of the same study design. Observational cohorts cannot control for the hundreds of lifestyle differences between heavy diet soda drinkers and non-drinkers. Self-reported beverage intake is notoriously inaccurate. Dementia develops over decades, making it nearly impossible to pinpoint which exposures in midlife mattered. And the reformulation issue means that someone who drank diet soda in the 1990s was consuming a different product from someone drinking it now. Until study designs evolve to address these limitations, the headline should probably stay a question rather than an answer.