September 12, 2025
7 min read
Key takeaways:
- Mean consumption of low- and no-calorie sweeteners was 92.1 mg a day.
- Daily consumption was linked to memory, verbal fluency and global cognition declines.
- Alternatives like tagatose may mitigate these risks.
Consumption of low-calorie and no-calorie sweeteners was linked to accelerated rates of cognitive decline among middle-aged adults over an 8-year follow-up period, according to data published in Neurology.
Also, associations between consumption and cognitive decline were greater among people with diabetes, Claudia Kimie Suemoto, MD, PhD, associate professor, division of geriatrics, faculty of medicine, University of São Paolo, and colleagues wrote.
Diet beverages and other ultra-processed foods include aspartame and other low- and no-calorie sweeteners, which have been linked with faster rates of cognitive decline in middle-aged people. Image: Adobe Stock
“We had a previous work showing an association between ultra-processed food (UPF) and cognitive decline in the same sample, and this study was a follow-up to that research since several UPFs contain sweeteners,” Suemoto told Healio. “I also consumed a lot of sweeteners by drinking diet soda and adding sweeteners to my coffee.”
Claudia Kimie Suemoto
Suemoto and colleagues conducted an observational study in which they followed 12,772 civil servants aged 35 years and older at baseline (mean age, 51.9 years; 54.8% women; 43.2% Black/mixed race) who participated in the Brazilian Longitudinal Study of Adult Health for a median of 8 years.
Researchers assessed cognitive performance every 4 years in three study waves (2008-2010, 2012-2014, and 2017-2019).
Participants also completed a validated food frequency questionnaire with 114 items at baseline, including questions about seven low- and no-calorie sweeteners (LNCSs): aspartame, saccharin, acesulfame k, erythritol, xylitol, sorbitol and tagatose.
These sweeteners often appear in sugar-free UPFs such as diet beverages, yogurts, snacks, low-calorie deserts and milk-based beverages, the researchers said.
Researchers stratified daily LNCS consumption into first (0.02-37.2 mg), second (37.3-102.3 mg) and third (102.4-856.5 mg) tertiles of 4,257 participants each as well as into no/sporadic and daily consumption classifications.
Mean LNCS consumption was 92.1 mg/day. Sorbitol was the most consumed LNCS with a mean of 63.8 mg/day, followed by aspartame at 12.8 mg/day. Erythritol was the least consumed at 0.1 mg/day.
Participants in the tertile with the highest consumption were more likely to be older, women and white and have higher income and education, as well as greater odds for higher frequency of hypertension and uncontrolled diabetes. These participants were more likely to be physically active as well.
Higher rates
Rates of memory decline were 32% higher in the tertile with the highest consumption compared with the tertile with the lowest consumption, including betas of –0.008 (95% CI, –0.032 to 0.008) for the second tertile and –0.024 (95% CI, –0.04 to 0.004) for the third tertile.
Rates of verbal fluency decline were 110% higher for the tertile with medium consumption (beta = –0.024; 95% CI, –0.048 to –0.008) and 173% higher for the tertile with the highest consumption (beta = –0.04; 95% CI, –0.064 to –0.024).
Rates of global cognition decline were 35% higher for the medium consumption tertile (beta = –0.016; 95% CI, –0.032 to –0.001), corresponding to 1.3 excess years of aging, and 62% higher for the highest consumption tertile (beta = –0.024; 95% CI, –0.04 to –0.16), corresponding to 1.6 excess years of cognitive aging.
Participants with daily LNCS consumption also had accelerated declines in memory (beta = –0.018; 95% CI, -0.035 to 0), verbal fluency (beta = –0.029; 95% CI, –0.045 to –0.012) and global cognition (beta = –0.021; 95% CI, –0.034 to –0.009) compared with participants who did not consume LNCS or who consumed them sporadically.
Greater aspartame consumption was associated with faster rates of decline in memory, verbal fluency and global cognitive. Greater acesulfame k and erythritol consumption were associated with higher rates of memory and global cognitive decline. Tagatose consumption was not associated with any cognitive decline.
Specifically, among participants aged younger than 60 years, the tertile with the highest LNCS consumption had a faster decline in verbal fluency, with betas of –0.016 (95% CI, –0.04 to –0.008) for the medium consumption tertile and –0.04 (95% CI, –0.064 to –0.016) for the lowest consumption tertile.
Participants with the highest LNCS consumption in this age group also had faster declines in global cognition, with betas of –0.008 (95% CI, –0.024 to 0.008) for the medium tertile and –0.024 (95% CI, –0.04 to –0.008) for the lowest tertile.
Further, rates of decline in memory (beta = –0.001; 95% CI, –0.002 to 0), verbal fluency (beta = –0.001; 95% CI, –0.002 to 0) and global cognitive (beta = –0.0005; 95% CI, –0.001 to –0.0001) were faster in this age group with greater sorbitol consumption.
But greater tagatose consumption was associated with slower rates of decline in memory (beta = 0.064; 95% CI, 0.008-01.2) and global cognition (beta = 0.04; 95% CI, 0.008-0.08) in this age group.
Additionally, the researchers said they did not find any associations between cognitive decline and any individual sweeteners among participants aged 60 years and older.
“One of the most surprising findings for me was that the association between low- and no-calorie sweetener consumption and cognitive decline was only significant among participants younger than 60 years old,” Suemoto said.
Due to their higher risk for dementia and cognitive impairment, Suemoto said that she had expected this association to be more evident among older adults.
“Instead, our results suggest that midlife exposure to sweeteners may be particularly harmful, which is important because midlife is a critical period for setting the trajectory of brain health,” she said.
It is believed that the development of cognitive decline and dementia begins decades before noticeable symptoms manifest, Suemoto said, so midlife exposures may accelerate these processes and lead to long-term consequences.
“This emphasizes the importance of understanding dietary habits earlier in adulthood,” she said.
Among participants who did not have diabetes, those in the highest consumption tertile had faster declines in verbal fluency compared with medium (beta = –0.021; 95% CI, –0.042 to –0.001) and low consumption (beta = –0.036; 95% CI, –0.057 to –0.014).
These participants also had faster declines in global cognition compared with those with medium (beta = –0.008; 95% CI, –0.023 to –0.006) and low consumption (beta = –0.018; 95% CI, –0.034 to –0.002).
Among participants with diabetes, those with the highest consumption had faster declines in memory compared with medium (beta = –0.69; 95% CI, –0.093 to –0.006) and low (beta = –0.088; 95% CI, 95% CI, –0.148 to –0.028) consumption.
These participants also had faster declines in global cognition compared with medium (beta = –0.052; 95% CI, –0.096 to –0.008) and low (beta –0.063; 95% CI, –0.105 to –0.021) consumption.
Greater consumption of erythritol and xylitol were associated with faster rates of memory and global cognitive decline, and greater consumption of sorbitol was associated with faster rates of memory, verbal fluency and global cognition decline, among participants who did not have diabetes.
Greater consumption of aspartame and acesulfame k were associated with faster declines in memory, verbal fluency and global cognition, and tagatose consumption was associated with faster decline in executive function, among participants with diabetes.
Finally, saccharin and xylitol consumption were associated with faster declines in memory and global cognition.
Next steps
Based on these findings, Suemoto and colleagues concluded that individual and combined LNCS consumption were associated with faster declines in global cognition, memory and verbal fluency, with daily consumption accelerating these declines.
These findings also suggest that tagatose and other natural sweeteners or other options for sugar may mitigate these associations, with a need for further research to confirm these findings and investigate whether these alternatives may be effective.
However, Suemoto cautioned that this was an observational study, so the researchers could only discuss associations, not impact or causality. Still, she added, there are several explanations for why artificial sweeteners and cognitive decline may be linked.
“Some animal studies suggest that certain sweeteners, like aspartame, may trigger neuroinflammation or processes related to neurodegeneration,” Suemoto said. “Others indicate that sweeteners can alter the gut microbiota, which in turn may affect the blood–brain barrier and brain health.”
Memory and verbal fluency, which often are related to Alzheimer’s disease, were most affected in this study, she noted.
“These pieces of evidence point to plausible biological pathways, but more research is needed, including neuroimaging, replication in other cohorts, and clinical trials, before any firm conclusions can be drawn,” Suemoto said.
Instead of a causal conclusion, she continued, these findings should be interpreted as an early signal. Providers can use these results to prompt conversations about dietary habits that emphasize moderation in sugar and artificial sweetener use, she added.
“Encouraging a diet based on whole and minimally processed foods remains the safest approach for supporting long-term brain and overall health,” Suemoto said. “As further evidence emerges, particularly from studies with neuroimaging and replication in other populations, clinical guidance can be refined.”
Suemoto and her colleagues are continuing their research as well.
“We are performing neuroimaging (MRI) in the ELSA-Brasil cohort and expect to investigate the association of sweetener consumption and brain imaging,” she said.
Reference:
For more information:
Claudia Kimie Suemoto, MD, PhD, can be reached at cksuemoto@usp.br.
Perspective
There’s longstanding controversy about the health effects of artificial sweeteners, with concerns ranging from diabetes to cardiovascular health and even gut microbiome changes. This study is significant because it adds evidence that they may also be linked to cognitive decline, which hasn’t been well-studied before.
In clinical practice, we often see how conditions like diabetes and vascular disease more generally increase dementia risk. This study brings the dietary aspect into sharper focus, suggesting that sweeteners themselves could influence brain health beyond traditional risk factors.
Patients may wonder if using artificial sweeteners to control blood sugar is a good strategy for brain health. This study can allow doctors to more confidently guide patients toward approaches with stronger evidence like a Mediterranean-style diet and, even more importantly, maintaining regular exercise.
Future research should look for brain imaging correlates, track changes in sweetener intake over time and study whether the effects differ in people who already have memory or other cognitive impairment. That would help clarify both risk and potential mechanisms.
Ari B. Cuperfain, MD, MSc, FRCPC
Staff Psychiatrist, Centre for Addiction and Mental Health
Assistant Professor of Psychiatry, University of Toronto
Disclosures: Cuperfain reports no relevant financial disclosures.
