How we grade foods

Every grade on GoutSafe is built from published data, not opinion. Here is exactly how it works, and where the limits are.

Last reviewed July 18, 2026.

1. The data

Two things in food drive gout risk: purines, which break down into uric acid, and fructose, which raises uric acid production. For each food we store a value for both, per 100 grams, drawn from the USDA Purine Database, USDA FoodData Central, and peer-reviewed literature. We cover 539 foods today. The fructose figure counts free fructose plus the fructose released from sucrose (table sugar is half fructose), so a sucrose-based sweetener like maple syrup is not scored as if it were fructose-free.

2. We score a typical serving

Raw lab values are reported per 100 grams, but you eat servings, not 100-gram blocks. So for each food we scale to a typical serving and work out how much purine that delivers, and what share of the recommended daily purine intake that is. We use a daily target of about 400 mg, the figure set by the Japanese guideline for managing hyperuricemia and gout (reviewed in Nutrients, 2024). This is why a food that looks alarming per 100 grams, like a dried fish or a spice, can still be fine in the amount you actually eat.

3. The A to E grade

We combine the per-serving purine and fructose load with known triggers and proven benefits into a single risk score, then map that score to a plain grade from A to E.

• Triggers push a grade toward higher risk: alcohol impairs uric-acid excretion, and fructose raises uric acid, even when the raw purine number looks low. We flag added sugar, syrups, sweetened drinks and fruit juice. Juice is included as a category rather than by dose, because the cohort evidence tracks the drink and not the fruit: a daily glass of orange juice carried about 41% higher gout risk in women. Whole fruit is treated as the lower-risk form, but we are careful about why. It is not that fruit sugar is chemically gentler: matched gram for gram, whole apple raised uric acid as much as a fructose drink. It is that whole fruit fills you up before you reach a drink-sized dose. So a food page states the fructose in a serving either way, and only withholds a caution for whole foods under a modest per-serving load. Very sugar-dense foods (15 g or more per 100 g) cannot read as Safe regardless of purines.
  
  The fructose meter on a food page is scaled against ~12.5 g/day, and we call that a reference rather than a limit for a specific reason: no health authority publishes a fructose threshold for gout. Not the ACR, which conditionally advises limiting high-fructose corn syrup but attaches no number and rates the evidence very low. Not NICE, which does not mention fructose at all. When the European Food Safety Authority formally set out to define an upper limit for dietary sugars in 2022, it concluded that a level at which risk is not increased could not be identified, and declined to set one.
  
  So we derive one and show our working. Two reputable bodies converge on the same daily free-sugars figure: the WHO's conditional recommendation to keep free sugars under 5% of energy, which is 25 g on a 2,000 kcal day, and the American Heart Association's 25 g/day for women (36 g for men). Both were set for other reasons, dental caries and weight for the WHO, heart health for the AHA, not for uric acid. Our figures measure fructose, and sucrose and corn syrup are each about half fructose, so the fructose share of that 25 g is about 12.5 g. A drink carrying a whole day's worth in one glass reads Limit; about 5 g means it can never read Safe.
  
  One caveat we would rather state than bury, because it cuts against a tidy reading. WHO's "free sugars" deliberately excludes the sugar inside whole fruit and vegetables. Our number does not: it counts all the fructose a food delivers, wherever it came from. So our figure is measuring slightly more than the 25 g was written about. We count it that way on purpose, because your body does: sucrose is split into glucose and fructose before it is absorbed, and in the one trial that matched the dose, the fructose in a whole apple raised uric acid about as much as a fructose drink did. Counting an apple's own free fructose while ignoring its sucrose would be the arbitrary line, not this one. The practical upshot: read a meter near 100% as "that is a lot of fructose for one serving," not as "you have broken a WHO limit."
  
  Two things support that range from other directions. A single 355 ml can of soft drink raised uric acid as much as a 600 ml serving, with no measurable difference between them, so the response is effectively saturated at about one can. And in the cohorts, risk becomes clear at roughly one serving a day: about 45% higher gout risk for a daily soda in men, 74% in women, 41% for a daily glass of orange juice. Worth saying plainly, though: the rise from a single drink is small and temporary, around 0.2 to 0.4 mg/dL, and controlled trials that swapped fructose in without adding calories found no effect on uric acid at all. This is a signal about habits, not an alarm about one glass.
  
  One genuine conflict we will not paper over: controlled feeding trials found 100% juice lowered fasting uric acid, while prospective cohorts found it raised gout risk. We follow the cohorts, because incident gout is the outcome this site exists to answer for and fasting urate in healthy adults is a surrogate. Because alcohol's risk comes from the ethanol itself and not its purine content, every alcoholic drink is graded at least D, and regular beer, which adds malt purines on top, is graded E (low-alcohol beer grades D). Non-alcoholic versions are graded on their content like any other drink. (Choi 2004, Lancet, Choi & Curhan 2008, BMJ)
• Benefits are noted when research links a food to lower uric acid or less inflammation, such as cherries or low-fat dairy. We show the supporting study on the food's page. (Zhang 2012, Choi 2004, NEJM)
• Processing can cap a grade even when the purine number is low. Gout risk tracks how processed a food is, not just its purine content: people who eat the most ultra-processed food develop more gout than those who eat the least, likely because ultra-processed foods cluster with added sugar, salt, and weight gain, and because insulin resistance reduces how well the kidneys clear uric acid. So we classify every food on the standard NOVA scale (1 minimally processed to 4 ultra-processed), and an ultra-processed food, like potato chips or a frankfurter, cannot read above Moderate (C) however low its purine figure. We cap at Moderate, not Limit or Avoid, because this is a population association, not a direct flare trigger like organ meat or alcohol: the cap removes a false green light without overstating the risk. Drinks are the exception in the other direction, and they are graded by how much fructose a serving carries, not by how the drink was made: about 12 g or more reads at least Limit, 5 g or more can never read Safe. That deliberately puts a glass of 100% apple juice and a can of cola on the same rung, because they carry nearly the same sugar and the cohort evidence puts juice just below soda rather than a grade below it. It also means we size a drink by its real container, so a 20 oz sports drink is measured as a bottle rather than a small glass. Plant-based meat substitutes are graded a touch more gently (Safe, not Moderate): they are ultra-processed, but for gout they replace higher-purine meat. (UK Biobank 2024, Rheumatology, Rai 2017, BMJ, Choi & Curhan 2008, BMJ)

4. Confidence and what we don't know

Every food shows a confidence level, and it is derived from the evidence rather than assigned by us. It answers one question: how many times has this been measured? High means the value is replicated across three or more independent determinations. Moderate means it was measured, but on one or two samples, or on a sample count the source never recorded. Low is reserved for the small number of values inferred from composition rather than measured in the food at all. A food's overall confidence is never higher than its purine confidence, and if fructose is still pending the food says so separately.

This used to be a judgement about how well sourced a value looked, which quietly became a restatement of the source tier: values resting on a single sample read "high" simply because the source was good. Replication and source quality are different things, so the Sources panel now shows both side by side, a reliability tier (lab-verified, peer-reviewed literature, or estimated) and the sample count behind the number, including when that count was never recorded. A single determination is normal in food composition tables rather than a defect, so most foods sit at moderate.

Purine values are printed to three significant figures, so a measured 137.2 shows as 137 and a measured 0.4 stays 0.4 rather than collapsing to a zero that would claim the food contains no purine at all. Bands, grades and per-serving figures are all computed from the unrounded measurement. Where a number is measured on a concentrated form like skin-on fish or a dried ingredient, we flag it directly so it is not mistaken for an ordinary portion.

We also show where each number comes from. Every purine and fructose value carries its real source and a reliability tier on the food's page:

• Lab-verified (tier 1): direct laboratory measurement, such as the USDA Purine Database.
• Peer-reviewed literature (tier 2): published studies and international food tables.
• Estimated (tier 3): derived by composition rather than measured. For example, plain meat, seafood, and eggs contain no fructose, so we mark that value estimated instead of pretending it was tested in a lab.

That last flag matters: an honest "estimated" is more trustworthy than a number with no provenance at all. This is the difference between GoutSafe and the viral food advice that circulates without a single citation.

5. How we keep it honest

We do not publish "medically reviewed by" bylines, because we will not claim credentials we do not have. Instead we show our sources, our confidence, and our reasoning on every food, and we keep our claims conservative. If you spot an error, please tell us and we will correct it.