Sugar Alternatives Rated: Stevia, Monk Fruit, Erythritol, Xylitol and Allulose

Global Watchdog is reader-supported, with no advertisers and no government entity to answer to. If you enjoyed this article, you can help keep it coming with a one-time donation or monthly support: Support on PatreonDonate on Ko-fi

17 min read

Last updated October 24, 2026

Sugar is the ingredient almost everybody loves and almost everybody is trying to cut. The substitutes promise the sweetness without the spike, and the shelves are full of them: stevia, monk fruit, erythritol, xylitol, allulose, and blends of all of the above wearing each other’s names. This guide rates the five big ones, and it has been rebuilt from the primary sources, because the first version repeated claims, and a few quotations, that could not be verified.

Two promises up front. First, the scores in this article rate taste, kitchen usefulness, and price. Nothing else. They are not health scores and they are not safety scores, and a high score on a sweetener with a hazard attached means it tastes good, not that the hazard went away. Second, where a sweetener carries real safety news, the news is printed in full, next to the sweetener it belongs to: a dog warning that cannot wait, two heart studies that deserve an honest read, and a World Health Organization guideline that is narrower than the headlines made it sound.

Last updated October 24, 2026: Rebuilt from primary sources (FDA, WHO, EFSA, and the journal papers themselves). Scores are now labeled as taste, use, and price only; the xylitol dog warning and the erythritol and xylitol cardiovascular evidence were added; product labels behind the picks were checked; unverifiable expert quotations were removed.

This article is general information, not medical advice. If you manage diabetes or any condition affected by sweeteners, your own clinician’s guidance comes before anything on this page.

Professor Winston Wants to Know

Our resident watchdog is taking the temperature of the room.

Which sweetener actually stuck in your kitchen?

One tap, no account or email needed. Results are shown after you vote.

Which sweetener actually stuck in your kitchen?

SteviaMonk fruitErythritol or another sugar alcoholPlain sugar, in smaller amounts

Cast your voteSee results

Want to say more? The comments below are open.

How the Scores Work

Each sweetener gets one score out of 10, and the score has a narrow job. It weighs how the sweetener tastes (including aftertaste), how it behaves in coffee, baking, and cooking, and what it costs per unit of sweetness. That is all it weighs. It does not weigh the evidence in the safety sections, because a number cannot carry that honestly: erythritol’s score says it is pleasant and cheap, and the section under it says what a 4,000-person study found, and both things are printed where you can see them. Anyone who collapses those into one number is selling you something.

Stevia: Score 7.5 for Taste, Use, and Price

Stevia’s sweetness comes from steviol glycosides in the leaves of a South American plant, and it is intense: purified extracts run 200 to 300 times sweeter than sugar, with a lingering finish some people read as licorice and some read as bitter. It holds up in drinks and yogurt, it does not brown or bulk like sugar in baking, and a little goes a long way, which keeps the cost per cup reasonable. That profile, good in a glass, awkward in a cake, is the 7.5.

The regulatory wording matters here, and this guide will not blur it. What the FDA has raised no objection to, through its GRAS notification program, is highly purified steviol glycosides, at least 95 percent pure. Whole stevia leaf and crude stevia extracts are not GRAS and are not approved food additives; the FDA maintains Import Alert 45-06 covering stevia leaves and crude extracts shipped as conventional food. So the accurate sentence is not “the FDA approved stevia.” It is: specific purified extracts passed through the GRAS process without objection, and the leaf itself did not.

Two label warnings. First, most retail “stevia” is a blend: the bulk is often erythritol or maltodextrin carrying a small amount of extract, which is why the erythritol section below matters even to stevia buyers. Second, blood-sugar behavior belongs to the specific product, not the plant on the box: a blend bulked with maltodextrin does not behave like the pure extract. Read the ingredient order, not the front panel.

Monk Fruit: Score 8.5 for Taste, Use, and Price

Monk fruit, luo han guo, is a small melon from southern China, and its mogroside extracts are roughly 150 to 250 times sweeter than sugar with the cleanest taste of the plant-based options: no licorice tail, no cooling effect, good heat stability for baking. On taste alone it is the strongest entry in this guide, and the 8.5 reflects it. The catch is price and blending: pure monk fruit extract is expensive, so most products are blends.

The blend question is the whole ballgame with monk fruit. Check the order on the label: Lakanto Classic and Truvia both list erythritol first, with the headline ingredient second, which makes them erythritol products wearing a monk fruit or stevia label. That is not a scandal, but it is information, and it is why this guide’s monk fruit pick is Lakanto’s “with Allulose” version, whose ingredient list is exactly two items, allulose and monk fruit extract, with no erythritol at all. If you are avoiding erythritol after reading the section below, that version and plain allulose are the shelf options that do it.

One correction of a common claim: the WHO’s 2023 sweetener guideline names stevia in its release. Monk fruit is not named; it falls, at most, under the guideline’s general class of non-sugar sweeteners. Anyone telling you the WHO singled out monk fruit, for praise or blame, is embellishing.

Check Price on Amazon: Lakanto Golden with Allulose (erythritol-free)

Erythritol: Score 7 for Taste, Use, and Price

Erythritol is a sugar alcohol made by fermenting glucose, about 70 percent as sweet as sugar, with a clean taste, a mild cooling sensation, almost no calories, and the best baking manners of the sugar alcohols. It is the bulk inside most stevia and monk fruit blends, including the two named above, and on taste, use, and price it earns its 7. Now the part that has to travel with it.

In 2023, researchers at the Cleveland Clinic reported in Nature Medicine that blood levels of erythritol were associated with later heart attack, stroke, and death across 4,139 patients in the United States and Europe: a discovery group of 1,157, a United States validation group of 2,149 (adjusted hazard ratio 1.80), and a European group of 833 (adjusted hazard ratio 2.21). In a small pilot, eight volunteers who drank 30 grams of erythritol, an amount in a sweetened drink or a pint of keto ice cream, carried elevated blood levels for more than two days. A follow-up study in Arteriosclerosis, Thrombosis, and Vascular Biology in 2024 put 30 grams of erythritol head to head with glucose in 20 volunteers and found the erythritol drink increased platelet reactivity, the clotting tendency the researchers were watching for.

Here is what the regulators said, because that is part of the record too. An FDA internal evaluation memo dated June 15, 2023 concluded that the observational studies were incapable of establishing causation, noted that many blood samples were drawn before erythritol was widely used in food, meaning the erythritol in those samples was largely made by the body itself, and did not question erythritol’s GRAS status. Europe’s EFSA, in December 2023, set an acceptable daily intake of 0.5 grams per kilogram of body weight per day, based on digestive effects, found that high consumers can exceed it, and found no established cause-and-effect link to cardiovascular disease. The honest summary: a replicated statistical association, a plausible platelet mechanism shown in small short studies, no proof of cause, and no regulator has pulled it. If you eat erythritol by the cupful daily, the prudent read of that record is to moderate. If it shows up in your stevia blend, that is a different exposure entirely, and the studies do not speak to it directly.

Xylitol: Score 7.5 for Taste, Use, and Price

Xylitol is a sugar alcohol found in birch bark and corn cobs, as sweet as sugar itself, with a clean taste and a genuine dental track record: it does not feed the bacteria that cause cavities, which is why it shows up in sugar-free gum and toothpaste. As a sweetener it is easy to like, and the 7.5 says so. It also carries the most serious household warning in this article, and the warning is not fine print.

Xylitol has a second file, and it belongs in the open beside the first. In June 2024, the European Heart Journal published a study from the same Cleveland Clinic group: across a discovery group of 1,157 patients and a validation group of 2,149, higher blood xylitol levels were associated with later heart attack, stroke, and death (adjusted hazard ratio 1.57 in the validation group), and in 10 volunteers, a 30-gram xylitol drink increased platelet reactivity. A larger conference presentation at the European Society of Cardiology congress in August 2026 pointed the same direction in two population cohorts, but a conference presentation is not a peer-reviewed paper, and this guide weighs it accordingly. Section by section, this is the same shape as the erythritol record, which is why the two are laid side by side next.

The Heart Question, Side by Side

Erythritol and xylitol arrived at the same uncomfortable place by the same road, and they deserve one shared, level-headed treatment rather than two panics.

What the studies are. Large observational cohorts, in the thousands, linking blood levels of each sweetener to later cardiovascular events, with hazard ratios roughly between 1.6 and 2.2, plus very small volunteer studies (8, 10, and 20 people) showing that a large single dose, 30 grams, raises platelet reactivity for a time. The cohorts establish association. They cannot establish cause, and both papers’ critics make the same structural point: the human body manufactures both substances internally, blood levels rise with age, illness, and metabolic trouble, and much of the sampled blood predates these sweeteners’ widespread use in food, so high blood levels may be a marker of the very conditions that cause heart events rather than a cause of them.

What the regulators say. The FDA’s internal review of the erythritol paper concluded the observational work was incapable of establishing causation and did not disturb erythritol’s GRAS status. EFSA set an intake limit for erythritol on digestive grounds and found no established cardiovascular cause and effect. Neither agency has restricted either sweetener.

What a reader should do with that. Three things. Do not treat either sweetener as proven poison; the evidence does not carry that weight. Do not treat either as exonerated; the association is replicated and the platelet signal is real in the lab. And weigh your dose: the exposure in these studies’ worrying end is large and daily, handfuls of sugar alcohol a day, which describes some keto and diabetic product habits and does not describe sweetening coffee with a blended packet. Dose is not a loophole. It is how toxicology works.

Allulose: Score 8 for Taste, Use, and Price

Allulose is the interesting one. It is a rare sugar, present in tiny amounts in figs, raisins, and wheat, that tastes and bakes like sugar because it is sugar, chemically, while the body absorbs little of it and excretes most of the rest. It carries about 0.4 calories per gram, a tenth of sugar’s load, and the FDA’s guidance excludes it from the Total Sugars and Added Sugars lines on labels and lets it be counted at that reduced calorie value. It has passed through the FDA’s GRAS notification program repeatedly without objection, across six notifications from different manufacturers.

The blood-sugar evidence is genuinely promising and honestly limited. A 2026 meta-analysis in the American Journal of Clinical Nutrition pooled 20 trials and 1,033 participants and found that allulose blunted the glucose rise after a meal or a glucose drink, a moderate-certainty finding, while finding no significant effect on HbA1c, the long-term blood sugar marker, in the trials that measured it. Short-term meal responses improving while the long-term marker does not move is a pattern that calls for more research, not for a victory lap. On safety: the doses in trials run into digestive limits, with stomach upset appearing above roughly half a gram per kilogram of body weight in a day, and the European Union has not authorized allulose, because EFSA concluded in 2025 that safety could not be established on an incomplete dossier. An incomplete file is not a finding of harm, and this guide reports it as exactly what it is: Europe asked for more data and has not gotten it yet.

It loses a step on price, which is what keeps the score at 8 rather than higher. This guide’s allulose pick is RxSugar, which is allulose and only allulose. Do not confuse it with similarly named blends: at least one sound-alike product mixes allulose with glucose, which is a different purchase entirely.

Check Price on Amazon: RxSugar Allulose

What the WHO Guideline Actually Says

In May 2023 the World Health Organization advised against using non-sugar sweeteners to control body weight, and the coverage mostly stopped there. The document itself is narrower and more conditional than the headline. The recommendation is explicitly conditional, built largely on low-certainty evidence. It covers the class of non-sugar sweeteners, and its release names stevia among them; it does not assess sugar alcohols, so erythritol and xylitol sit outside it, and allulose, being an actual sugar with reduced calories, is outside it categorically. It also does not apply to people with pre-existing diabetes, a group the guideline excludes by design.

Read precisely, the WHO is saying that swapping sugar for a non-sugar sweetener has not been shown to deliver lasting weight control, not that any of these substances was found dangerous in the guideline. My own view, as the person whose name is on this site, is that these products earn their keep as a transition: a way to walk a sweet tooth down from full sugar toward less sweetness overall, in smaller amounts, rather than a permanent one-for-one replacement at the same volume. That is an opinion built on kitchen experience, and I am labeling it as mine, because the WHO document does not say it and I will not put my words in their mouth.

Watchdog Approved

Winston’s Picks: the two shelf buys this guide stands behind, labels checked.

Lakanto Golden Monk Fruit Sweetener with Allulose
The erythritol-free version: allulose and monk fruit extract, nothing else on the label. The classic Lakanto blend is erythritol-first, so check which bag you are buying.

RxSugar Organic Allulose
Straight allulose, the rare sugar itself, for baking and drinks where real sugar behavior matters.

As an Amazon Associate, we earn from qualifying purchases.

Questions Readers Ask

Which sweeteners are safe to keep in a house with dogs?

Stevia, monk fruit, erythritol, and allulose carry no xylitol-style dog warning. Xylitol is the emergency: it is also labeled birch sugar, wood sugar, and birch bark extract, it hides in sugar-free peanut butter, and poisoning can start within 20 minutes. If a dog eats it, do not induce vomiting; call a veterinarian or animal poison control at once. Our companion article on dark chocolate carries the matching warning for theobromine, the dog hazard in chocolate: Is Dark Chocolate Actually Healthy? What the Evidence Says.

Did the WHO ban or condemn these sweeteners?

Neither. The WHO’s 2023 guideline is a conditional recommendation against using non-sugar sweeteners for weight control. It excludes sugar alcohols and allulose, and it does not apply to people with pre-existing diabetes. It is a statement about weight evidence, not a safety ban.

Does erythritol cause heart attacks?

That has not been shown. Large studies found an association between blood erythritol levels and later heart events, small studies found a platelet effect at a 30-gram dose, and the FDA’s internal review concluded the observational studies could not establish causation. The fair read is a signal worth moderating heavy use over, not a proven cause.

Why do stevia and monk fruit products contain erythritol?

Because the extracts are hundreds of times sweeter than sugar, a usable product needs bulk, and erythritol is the cheap, pleasant bulk of choice. That is why the ingredient order matters: Lakanto Classic and Truvia are both erythritol-first products. Blends bulked with maltodextrin can also behave differently on blood sugar than the pure extract, so judge the product, not the plant name.

Is allulose actually sugar?

Yes, chemically it is a rare sugar, and that is its advantage: it browns, dissolves, and bakes like sugar while the body absorbs little of it, at about 0.4 calories per gram. The FDA excludes it from Total and Added Sugars on labels. Large single doses can upset the stomach, and it is not yet authorized in the European Union, which asked for a more complete safety file.

If you came here from our look at high-fructose corn syrup around the world, the thread is the same: the sweetener question is never just sweetness. It is dose, labeling, and what the evidence actually says when you read it yourself. Tell us in the poll which one survived in your kitchen, and argue in the comments. The labels, as always, are where the truth is hiding.

Sources consulted for this rebuild include the FDA’s xylitol consumer warnings and poster, the FDA internal erythritol evaluation memorandum (June 15, 2023), Witkowski et al., Nature Medicine (2023), the erythritol platelet study, Arteriosclerosis, Thrombosis, and Vascular Biology (2024), the xylitol study, European Heart Journal (2024), EFSA’s erythritol opinion (2023) and allulose novel food opinion (2025), the WHO guideline on non-sugar sweeteners (2023), the FDA allulose labeling guidance, and manufacturer ingredient labels.