What Triggers US Food Recalls? We Analyzed 1,371 FDA Recall Events (2024-2026)

By
Alex Uspenskyi
October 16, 2026
Bar chart of US food recall events by cause, 2024-2026: undeclared allergens lead with 553 of 1,371 events

Last updated: October 2026

Between January 2024 and September 2026, FDA's enforcement database logged 1,371 food recall events. Four in ten were undeclared allergens, more than Salmonella and Listeria combined. Pathogen recalls were rarer but far more severe and wider-reaching. Here is what the data shows.

Key findings

  • Undeclared allergens were the top cause every year: 553 of 1,371 events (40.3%). Salmonella (174 events) and Listeria (141) together account for 315.
  • Pathogens are the most severe. Salmonella events were Class I in 92% of cases, E. coli/STEC in 76%, Listeria in 72%. Undeclared-allergen events were Class I in 35%.
  • Listeria events reach the most products. An average Listeria event involved 5.5 product records, against 2.2 for allergen events.
  • Foreign material is the fourth-largest cause (128 events, 9.3%), and none of those events were Class I.
  • Repeat recalls are common. 136 of 1,165 recalling firms (11.7%) had more than one recall event in the period.

Recall events by cause

CauseEventsShareClass I shareAvg. products per event
Undeclared allergen55340.3%35%2.2
Salmonella17412.7%92%3.5
Listeria14110.3%72%5.5
Foreign material1289.3%0%2.5
Chemical, heavy metal or toxin856.2%18%1.6
Other or unclear846.1%12%2.3
Labeling or misbranding795.8%5%2.2
Other pathogen or spoilage735.3%19%3.0
Process, sanitation or GMP332.4%0%4.0
E. coli / STEC211.5%76%2.6

Source: FDA openFDA food enforcement data, recall initiation dates 2024-01-01 to 2026-09-30. Analysis by IONI.

Events per year

There were 482 events in 2024, 569 in 2025 and 320 in January to September 2026. Undeclared allergens led in every year (184, 232 and 137 events). The 2026 figure covers nine months and may be revised as FDA classifies recent recalls.

Which allergens?

Where the recall reason names an allergen, the most frequently named were milk (126 events), soy (98), wheat (91), egg (48), sesame (43), peanut (40) and fish (32). Tree nuts named individually or as a group appear in about 85 events. One event can name several allergens, so these counts overlap.

What this does and does not tell us

The enforcement data records what was recalled and why, not the root cause inside the plant. It cannot tell us whether a missed allergen came from a label change, a changeover, a rework step or a supplier. What it does show is where the exposure is: allergen labeling accounts for far more events than any pathogen, while Salmonella and Listeria are the events most likely to be Class I and, for Listeria, to reach many products at once.

For a small manufacturer, a practical reading is to check the controls behind the biggest buckets: label and ingredient-statement verification before each run, allergen changeover and cleaning records, how rework is tracked back to its source lots, and whether you can trace a lot forward and back in minutes. If you want to test that last point, see our guide to recall management software and run a timed mock recall on one real lot.

Method and caveats

  • Source: openFDA food enforcement endpoint, records with a recall initiation date from 2024-01-01 to 2026-09-30. 3,857 product-level records were grouped by event ID into 1,371 events. Data pulled October 16, 2026.
  • Categories: assigned by keyword matching on the free-text recall reason, in a fixed order (allergen, Listeria, Salmonella, E. coli/STEC, other pathogen or spoilage, chemical or toxin, foreign material, labeling, process or GMP, otherwise other). Each event gets one category, so an event with several hazards is counted under the first match. About 6% of events could not be categorized.
  • Class and reason come from the first product record returned for each event.
  • Limits: openFDA enforcement data is not validated and is not a complete list of recalls. USDA-regulated meat, poultry and egg products are outside FDA's data. Firm counts use the recalling firm name as written, so name variants may split or merge firms.

Download the data

Download the event-level dataset (CSV, 1,371 events). It is free to use and cite. Credit: FDA openFDA data, analysis by IONI.

Disclosure: IONI makes food safety and production software for food manufacturers. This analysis uses public FDA data and does not depend on any IONI product.

Our Latest Blog Posts

Interested in Learning More?

Book a free consultation with our expert

Val Verbovetskyi, AI Expert at IONI

Val Verbovetskyi

AI Expert

Book a Call
Arrow icon

Download this article guide line!

Thank you! Download will start soon.
Oops! Something went wrong while submitting the form.
close mark