HD-OPM Archive
Topics › Nuts, fish, coffee, tea, cocoa & other foods

Sweetener harms in cohort studies largely reflect reverse causality, umbrella review says

Louis Nervegna · Jan 28, 2026 · ♥ 1 reactions · 2 comments · original Facebook post
Nuts, fish, coffee, tea, cocoa & other foodsHow to read research
AI summary & key facts

Louis, not promoting sweeteners, shares an umbrella review explaining why cohort studies show worse outcomes but RCTs show weight and glycemic benefits, attributing the gap to confounding and reverse causality.

My usual caveat on this topic: I’m not encouraging anyone to consume artificial sweeteners. There are better choices.

But it’s important to be factually accurate about what the preponderance of scientific evidence actually shows on this subject.

Keep two words in mind when you read the post below - reverse causality. This is the central problem with many observational studies on artificial sweeteners that appear to link them with worse health outcomes. After properly accounting for this reverse causality, these links tend to disappear.

#artificialsweetenersHD #dietsodaHD #reversecausalityHD

Do artificial sweeteners help you live longer?

A new umbrella review tried to reconcile why RCTs and observational studies have seemed to disagree. Many epidemiological studies show an increased risk of mortality, CVD, obesity, type 2 diabetes, and cognitive decline with artificial sweetener intake, but the shorter human randomized controlled trials show that artificial sweeteners tend to reduce body weight and improve glycemic control. So why do the studies disagree?

It boils down to study design. In these cohort studies that aren’t controlled or randomized, you have to consider differences in confounding variables as well as reverse causality. Example, is it that artificial sweeteners cause obesity and these other diseases or is it that people who become unhealthy are more likely to start using artificial sweeteners? Additionally, there can be differences in baseline characteristics. People who are already obese and eating too much may make diet attempts where they use more artificial sweeteners whereas people who avoid them might tend to be more fit by nature
What made this paper different is that it separated analyses into:

Once bias was properly adjusted for, the results were remarkably consistent with randomized controlled trials.
Across bias-adjusted analyses, intake of low- and no-calorie sweeteners was associated with:

In other words: when you account for the fact that people at higher cardiometabolic risk are more likely to choose diet products, the apparent harms largely disappear and the observational data line up with RCTs.
This doesn’t mean artificial sweeteners are magic.
It means methodology matters, and bias can completely flip conclusions if you don’t handle it correctly.
Study:
“Reconciling conflicting evidence on low- and no-calorie sweeteners and cardiometabolic health outcomes: an umbrella review using naïve and bias-adjusted methods”
PMID: 41428417

← newer: Two days of only oatmeal reprogrammed the gut microbiome in older: Harvard study implicates lithium deficiency in Alzheimer's p →