We Trusted the Wrong Sweetener

The first bar is from a brand you would recognize. It built its following on keto credentials and clean label positioning. The sweetener system is erythritol and stevia, which the packaging proudly notes are “naturally derived.” This bar is available at Whole Foods, Sprouts, every natural channel retailer that matters, and most conventional grocers besides. It represents the consensus position on responsible sugar replacement.

The second bar is from a smaller brand that reformulated around allulose two years ago.

Allulose is a rare sugar that occurs naturally in trace amounts in figs, raisins, and maple syrup. It provides about seventy percent of sucrose’s sweetness at roughly a tenth of the calories. It does not spike blood glucose or insulin. The FDA allows it to be excluded from total and added sugars on nutrition labels, which means a product sweetened primarily with allulose can carry “0g added sugar” on the front of pack while still delivering bulk and mouthfeel that stevia alone cannot provide.

What allulose does not do is replicate sugar perfectly across all applications. It browns less effectively than sucrose. It contributes bulk, but not identically. Formulators who expect a one to one swap will be disappointed. The ingredient works best as part of a blended system where its metabolic advantages can be captured while other components address functional gaps.

But by almost any safety or metabolic measure, it represents one of the more elegant sugar reduction tools available.

This bar cannot get into Whole Foods. Allulose is on their unacceptable ingredients list, grouped with aspartame and sucralose as if they were chemically or functionally equivalent.

I used to think this was simply a lag in retail standards catching up to ingredient science.

Then the Cleveland Clinic paper dropped.

In February 2023, researchers at the Cleveland Clinic published findings in Nature Medicine that should have reordered the entire conversation around sugar alcohols. Their study followed patients undergoing cardiac risk assessment and found that elevated blood erythritol levels were associated with approximately double the risk of major adverse cardiovascular events over three years. Heart attack. Stroke. Death.

The mechanistic work was equally concerning. When healthy volunteers consumed a standard serving of an erythritol sweetened beverage, their blood erythritol levels spiked more than a thousand fold and stayed elevated for days. In laboratory studies, these concentrations enhanced platelet reactivity and clotting potential.

A follow up study published in 2024 in Arteriosclerosis, Thrombosis and Vascular Biology directly compared erythritol to glucose in healthy subjects and found that erythritol increased multiple measures of platelet function while glucose did not.

This research came from one of the most respected cardiovascular institutions in the world. The mechanism is biologically plausible. The findings have been replicated across multiple study designs.

The populations most likely to consume erythritol sweetened products, people managing diabetes, obesity, or metabolic syndrome, are precisely the populations at highest baseline cardiovascular risk.

What would change my assessment is large prospective randomized controlled trials conducted at real world intake levels over meaningful timeframes. Until that evidence exists, the signal warrants serious consideration.

I read the paper the week it published and waited for the industry response.

I am still waiting.

The commercial incentive to protect erythritol’s market position has produced more deflection than engagement. The ingredient is embedded in thousands of SKUs positioned as healthy alternatives. Keto bars. Zero sugar chocolates. Better for you ice creams. Sugar free beverages. Reformulation would be expensive and complicated.

So the data sits there. Occasionally acknowledged in trade press. Rarely discussed in the conversations that actually shape product development and retail assortment.

Then there is the gastrointestinal reality that anyone who has formulated with sugar alcohols knows intimately. Erythritol is better tolerated than maltitol or sorbitol, but at the usage levels required to replace sugar in many applications, a meaningful percentage of consumers will experience bloating, gas, and laxative effects. I have seen product launches stumble because formulators optimized for taste and macros without adequately accounting for what happens when consumers eat a second serving.

Meanwhile, allulose, which has no comparable cardiovascular signal and causes minimal GI distress at typical usage levels, cannot get shelf space at the retailer that most loudly proclaims its commitment to ingredient quality.

The reflex response is to say that Whole Foods is simply behind the curve and will eventually update its standards. But the pattern is more troubling than a single retailer’s lag time.

Retail ingredient standards frameworks tend to codify cultural assumptions at a moment in time and then defend those assumptions long after the underlying evidence has shifted.

Consider high fructose corn syrup. Natural channel retailers banned it years ago as a statement of values. During that same period, many of these same retailers carried Karo corn syrup on their shelves and sold private label marshmallows made with corn syrup. The standard was never really about the molecule. It was about the signal the molecule sent.

High fructose corn syrup became culturally coded as industrial and artificial. Regular corn syrup, chemically similar and metabolically equivalent, carried no such stigma.

Allulose suffers from a different signaling problem. It sounds synthetic. It is produced through enzymatic conversion of fructose, which requires explanation. The fact that it exists naturally in foods consumers already eat gets lost in the unfamiliarity of the name.

Erythritol, by contrast, has had years to build the “natural sugar alcohol” positioning that makes it feel safe regardless of what the clinical data shows.

The irony deepens when you examine how erythritol actually reaches the market. It occurs naturally in small amounts in fruits, which is technically true. But the erythritol used in commercial products is produced through industrial fermentation of corn. The doses in a keto friendly chocolate bar or zero calorie beverage bear no resemblance to what anyone encounters eating whole foods.

The “natural” framing is a construction, not a description.

We trusted the sweetener with better marketing, not better evidence.

The European Food Safety Authority published an opinion in June 2025 stating they could not establish allulose’s safety as a novel food. Read the document carefully and you discover this was not because they found evidence of harm.

The applicant simply failed to provide sufficient data to complete the assessment. EFSA requested additional information on identity, production process, proposed uses, genotoxicity, and human data. The applicant did not respond.

Paperwork, not science, is keeping allulose out of Europe.

Meanwhile, allulose is approved and thriving in the United States, Japan, South Korea, Singapore, Mexico, and as of 2024, Australia and New Zealand. European consumers are locked out of an ingredient their counterparts elsewhere access freely, not because regulators identified a safety concern, but because an application stalled.

This matters beyond regulatory trivia because it shapes how brands build international portfolios. A company that reformulates around allulose for the North American market faces real complexity when expanding to the EU. Some will calculate that erythritol, despite emerging cardiovascular questions, is the easier path because it is approved everywhere and faces no retail restrictions.

Convenience wins over evidence, and the pattern continues.

None of this means erythritol should be pulled from the market or that every product containing it is dangerous. The research warrants concern, not panic.

It means we need a more honest framework for evaluating tradeoffs. We need to stop treating “natural” as a proxy for safe, “zero sugar” as inherently superior to reduced sugar, and retail acceptance as validation of ingredient quality.

It also means separating two questions the industry routinely conflates.

The first is ingredient safety. What does the evidence show about metabolic impact, cardiovascular risk, GI tolerance, and long term outcomes?

The second is portfolio and commercialization strategy. What can you get approved, what will retailers accept, what claims can you make, and what does your supply chain look like?

Both are legitimate concerns. They operate on different logics. When commercial convenience drives safety assessment rather than the reverse, you end up where we are now.

The fixation on novel ingredients has also obscured the continued utility of sweeteners with decades of safety data and predictable performance.

Isomalt rarely generates excitement at trade shows. Derived from sucrose through enzymatic conversion, it provides roughly half the sweetness of sugar with a glycemic index of 2. It excels in applications requiring bulk and crystalline structure: hard candies, compressed tablets, coatings. It has extensive safety documentation across global regulatory bodies accumulated over decades of use.

Its limitation is familiar. At high enough doses, some consumers will experience digestive discomfort. That is a formulation challenge, not a fundamental flaw.

Tagatose deserves similar reconsideration. It is a rare sugar structurally similar to fructose, providing about ninety two percent of sucrose’s sweetness at roughly 1.5 calories per gram. It has GRAS status in the United States and approval in the EU, Canada, and other markets.

Research from Tufts published in early 2026 demonstrated dramatically more efficient production methods using engineered bacteria to convert glucose directly to tagatose. That breakthrough could significantly reduce costs that have historically limited adoption.

The safety profile is well established.

Consumer research supports this reframe. Shoppers say they want less sugar. But at shelf, they choose products that deliver flavor and satisfaction. The gap between stated preference and revealed preference in sweetener research is enormous.

“50% less sugar” often outperforms “zero sugar” in purchase intent when paired with taste claims. Consumers have been burned by zero sugar products that failed to deliver on experience.

The best formulations I have worked on use intelligent blending. Allulose where the application allows. Isomalt where structure matters. Tagatose where its properties add value. Modest stevia or monk fruit for potentiation. Sometimes residual sugar, because eliminating the last ten percent rarely justifies the sensory compromise.

The goal is minimum effective dose, not zero at any cost.

Sweet proteins represent the most significant development in sweeteners in decades.

Brazzein, derived from the West African oubli fruit, is hundreds to thousands of times sweeter than sucrose. It triggers sweet receptors in ways that closely mimic sugar, then digests like dietary protein with no impact on blood glucose, insulin, or the gut microbiome.

Oobli received FDA GRAS confirmation in early 2024 and has partnered with Grupo Bimbo and Ingredion. When cost parity arrives, reformulation will follow.

Supply chains are shifting too. Monk fruit concentration risk remains high. Stevia continues evolving toward fermentation based glycosides. Elo Life Systems’ domestically produced mogroside V may reach market in 2026.

These are future tools reaching toward present relevance. But today’s decisions are still shaped by a consensus the evidence no longer supports.

The two protein bars are still on my desk.

One represents the path the industry chose.
The other represents the path the science suggested.

The erythritol bar has clean label credentials, wide distribution, and implicit retail endorsement. It also carries cardiovascular research the industry has largely chosen to ignore.

The allulose bar has a cleaner safety profile and metabolic advantages erythritol cannot match. It also faces retail exclusion driven by signaling rather than science.

For anyone navigating sweetener decisions, the rule of thumb is simple. Separate what the evidence shows from what the market currently accepts. These are not the same list. The gap between them is where both risk and opportunity live.

Build your formulation around the evidence.
Build your go to market around the reality.
Watch the distance between them.

That is where the correction will come.

Mark Haas is the founder and CEO of RegulateCPG, an AI-powered compliance infrastructure platform designed to democratize regulatory expertise for food and beverage companies. With 35 years of experience navigating food safety regulation, manufacturing operations and multi-jurisdiction compliance, Mark has formulated over 200 brands representing more than $2 billion in market value. His work spans conventional, plant-based and emerging protein technologies across FDA, USDA, CFIA and EU regulatory frameworks, with deep expertise in using sophisticated amino acid analysis and PDCAAS methodology to create litigation-proof label claims for alternative protein companies.

For more insights on using regulatory compliance as competitive advantage, visit regulatecpg.com or connect with Mark on LinkedIn.

Legal Disclaimer:
This article discusses regulatory strategy and compliance approaches but does not constitute legal advice. Companies should consult qualified food law attorneys and regulatory counsel for guidance on specific labeling decisions and regulatory interpretations applicable to their products.