Where purine numbers come from

Search for the purine content of a food and you will find dozens of charts, almost none of which say where their numbers came from. Follow the ones that do, and they converge on a surprisingly small number of primary sources.

Knowing which source a number came from is not pedantry. It determines whether the value describes the food you are actually looking at.

The two databases behind this site

USDA and ODS-NIH Database for the Purine Content of Foods, Release 2.0 (2025). Published by the Agricultural Research Service of the U.S. Department of Agriculture with the NIH Office of Dietary Supplements. Release 2.0 covers 462 foods, 15 alcoholic beverages and 14 dietary supplements, with values per 100 g. It is the most rigorous English-language source available and it is free. Its practical limitation is coverage: it is built around foods eaten in the United States, so a large part of what is eaten in Asia is simply absent.

Dietary Guideline for Adults with Hyperuricaemia and Gout (2024 Edition). The Chinese National Health Commission's 2024 dietary guideline for adults with hyperuricaemia and gout. Its appendix 1 lists purine content for several hundred Chinese foods and ingredients — including preparations that appear nowhere in Western databases. It is also the source of the lower three cut lines used on this site: under 30 mg per 100 g, 30–75, 75–150 and from 150. Above that we follow the Japanese gout treatment guideline, which separates its high band at 200 mg and its extremely-high band at 300 — the NHC's top band alone would otherwise hold a quarter of the database across a nine-fold range.

Between them: 830 foods, 398 from the guideline and 432 from the USDA database.

The failure mode: matching names across languages

Here is the shortcut that goes wrong. An app wants a bilingual food library. It has one database in English. It generates the other language by translating the names, or it takes a Chinese food list and attaches USDA values by matching approximate names.

The result looks complete and is quietly wrong in places. "Bass" covers a dozen unrelated fish across cuisines. "Clam" maps onto species with materially different values. A Chinese preparation gets the number for a Western dish that shares an approximate translation. Every one of these produces a page that looks authoritative and states a figure belonging to something else.

Competitor analysis on the leading Chinese purine app found exactly this pattern — the same underlying USDA data, mapped onto Chinese names by approximation, which is the most plausible explanation for its recurring "the data is wrong" complaints.

So this database does the unglamorous thing instead: the two sources are kept separate and never merged. Where both measured something similar, both entries stay, each with its own value and its own citation. That has a visible cost — occasionally you will see two similar-looking foods with different numbers — and the cost is worth paying, because the alternative is a number you cannot trace.

Read the preparation, not just the name

Both sources record foods as measured, including how they were prepared, and the differences are large:

• Cod — 71 mg raw, 230 mg baked
• Anchovy — about 304 mg fresh, 1,109 mg dried
• Tofu — 100 mg for nigari tofu, 21.9 mg boiled three minutes
• Tea tree mushroom — 48 mg fresh, 293 mg dried

A chart that lists "cod: 71 mg" is not wrong so much as incomplete, and the incompleteness spans two entire bands of the scale. When a number is quoted without its preparation, treat it as a rough order of magnitude rather than a measurement. More on how preparation moves the numbers.

Total purines, and what that hides

Both databases report total purines: adenine, guanine, hypoxanthine and xanthine added together. This is the standard way to report and it is what makes cross-food comparison possible.

It also flattens a real distinction. The four bases are not metabolised identically — hypoxanthine and adenine appear to raise urate more efficiently than guanine, and there is evidence that yeast-derived purines are absorbed particularly readily. Two foods with the same total can plausibly behave differently. Nobody has a food database granular enough to act on that, so total purines is what everyone uses, including us. It is worth knowing what the single number is averaging over.

How to judge any purine chart you find

Four questions, in order of how much they tell you:

1. Does it name a source per entry? Not "adapted from various sources" at the bottom — a source on the row. Without this, nothing else can be checked.
2. Does it state the preparation? Raw, boiled, dried, canned. Absent this, the number could be off by a factor of three.
3. Does it state units? Milligrams per 100 g is standard. Charts quoting per serving without defining the serving are not comparable to anything.
4. Does it mix languages or cuisines without explaining how? A single tidy list covering both Chinese and Western foods, with no note about where each came from, is the signature of name-matching.

By that standard most of the top results for "purine food chart" fail at question one.

What to do with all of it

Use purine values for comparison — this preparation against that one, this portion against that portion. Do not use them as a prediction about your uric acid: most urate is produced by the body rather than absorbed, and diet alone moves it modestly compared with medication.

Every food page on this site shows the value, the preparation, the source and the serving arithmetic, and the same 830-food library ships inside the Breeze app for iPhone, where it works offline. Search the database, or read how the sources are used.