A clinic that has just had a load fail, or that is deciding whether to keep running an older autoclave for another year, can consult three public databases tonight: the FDA Manufacturer and User Facility Device Experience database (MAUDE), the FDA Enforcement Reports, and Health Canada's Recalls and Safety Alerts. All three are free, all three are searchable, and none of them can tell you whether a particular instrument set in your cupboard is sterile. What they can do is tell you whether the failure you are looking at has happened to somebody else, how the regulator classified it, and which field of your own record would let you match it.
The three disagree with each other constantly, and not because one is more accurate. They disagree because each one records a different event: a malfunction report, a recall action, or a Canadian recall notice. Reading them together is a skill with a small number of rules, and the rules are easier to see once you have the counts in front of you.
Two consumables do most of the daily work in the monitoring loop that these databases are, indirectly, about. A 24-hour readout steam biological indicator, 50 per pack produces the result that a recall question will eventually ask about, and indicator tape in a 3-roll set, 24 to 25 mm by 55 m is the part of the record an inspector reads off the outside of the pack. Neither is a substitute for the database question, which is a different question: what has already gone wrong elsewhere.
What is each database actually recording?
The single most useful thing to know before searching is what one row means in each system. In MAUDE a row is a report of an adverse event or a malfunction, filed by a manufacturer, an importer, or a user facility. In the Enforcement Reports a row is a recall action, one line per product and lot, with a classification attached. In Health Canada's database a row is a recall notice or safety alert, published in both official languages, with a category and a posting date.
| Question | FDA MAUDE | FDA Enforcement Reports | Health Canada Recalls and Safety Alerts |
|---|---|---|---|
| Unit of one record | One report of an event or malfunction | One recalled product and lot | One published recall notice or alert |
| Who files or publishes | Manufacturer, importer, or user facility | FDA publishes the firm's recall action | Health Canada publishes the notice |
| Severity field | Event type: death, injury, malfunction, other | Recall class: I, II or III | No class field in the search result; risk is described in the notice text |
| Date fields | Date of event, date received, date report sent | Recall initiation date, centre classification date, report date | Posting date, last updated date, and year facets |
| How it is searched | Fielded API query, for example by product code | Fielded API query, for example by recall reason | Full text site search with category and year filters |
| What it cannot tell you | Whether a recall followed the report | How many patients were exposed | Whether the firm was ever disciplined |
Two consequences follow. A count of MAUDE reports is not a count of incidents, because one incident can generate several reports and most malfunctions are never reported at all. A count of enforcement records is not a count of products, because one recall can span dozens of product lines and each one is a separate row. Reading a raw total as an incident rate is the most common mistake made with all three.
How much sterilization-related material is in each one today?
All of the readings below were taken on 18 November 2026, against the public endpoints, and each is reproducible with the query shown. The openFDA datasets stated a last update of 14 November 2026 at the time of the query. The purpose of the table is not to rank the three but to show how differently the same subject is sized by each.
| Database and query | Reading on 18 November 2026 |
|---|---|
| FDA Enforcement Reports, all device recall records | 40,535 records |
| FDA Enforcement Reports, reason for recall containing "steril" | 6,739 records |
| FDA Enforcement Reports, reason containing "steriliza" | 1,500 records: Class II 1,484, Class I 12, Class III 4 |
| Those 1,500 records by initiation type | 1,499 firm-initiated voluntary, 1 not stated |
| Those 1,500 records by status | Terminated 1,090, Ongoing 409, Completed 1 |
| Those 1,500 records by country of the recalling firm | United States 1,443, Italy 23, Japan 13, Switzerland 8, Canada 6 |
| FDA MAUDE, all device event reports | 26,849,547 reports |
| FDA MAUDE, product code FLE (steam sterilizer) | 1,530 reports |
| Those 1,530 reports by event type | Malfunction 907, event type blank 434, Injury 143, Other 45, Death 1 |
The product code is the field that makes MAUDE usable for a sterilizer question, and it is worth naming precisely: FLE is the code for a steam sterilizer under 21 CFR 880.6880, a Class II device. Two neighbours exist for the same family: FLF for an ethylene oxide sterilizer and KMH for a dry heat sterilizer under 21 CFR 880.6870. A clinic that searches a brand name instead of a product code will find a fraction of what is there, because brand names are entered inconsistently by reporters.
Health Canada's database cannot be queried the same way, and this is the methodological trap. Searching the site for "sterilizer" on 18 November 2026 returned 488 items, of which 483 were filed under health products, 4 under food and 1 under consumer products; the year facets ran from 50 items in 2018 to 87 in 2024. Searching for "autoclave" returned 7 items. Searching for the quoted phrase "medical device reprocessing" returned no results at all, while the same words unquoted returned 5,614, because the unquoted search matches the individual words anywhere in the page. A Canadian count is therefore only meaningful when the search string, the quoting and the category filter are all stated alongside it.
Do the three timelines agree?
They do not, and the divergence is informative. The table compares the number of FDA enforcement records whose reason mentions sterilization with the number of MAUDE reports under the steam sterilizer product code, by year.
| Year | FDA enforcement records (reason contains "steriliza") | MAUDE reports, product code FLE |
|---|---|---|
| 2019 | 238 | 91 |
| 2020 | 61 | 87 |
| 2021 | 36 | 81 |
| 2022 | 47 | 70 |
| 2023 | 55 | 66 |
| 2024 | 64 | 64 |
| 2025 | 77 | 71 |
| 2026 to date | 135 | 65 |
The enforcement series moves by a factor of six between 2019 and 2021, then climbs again in 2026. The MAUDE series barely moves, hovering between 64 and 91 reports a year for the whole period. A recall year is driven by which firms happened to run a review of their sterilization validation, not by whether more sterilizers failed; a report series is driven by how many events reached the reporting threshold. If a clinic used either series as a trend line for "how safe are sterilizers", it would draw the wrong conclusion from a real number, which is the most expensive kind of error to defend in an inspection.

Why do dental sterilizers barely appear in any of the three?
Because the databases are dominated by manufacturers and by hospitals, in that order. Of the 1,530 MAUDE reports under the steam sterilizer product code, 997 come from a single manufacturer's Mexican plant, followed by 38 from one tabletop manufacturer, 38 from another large supplier and 35 from a European plant. Searching the same product code for the small tabletop brands that sit in dental operatories returns 57 reports for one brand name and 5 for another, and nothing at all for several names that are common in Canadian dental offices.

Of the 1,500 enforcement records with a sterilization reason, 516 mention dental, handpiece, tray or instrument in the product description or the reason. That is a third of the total, and it is the honest answer to "is this a dental problem": sterilization failures do reach the recall record, but usually as part of a kit, a tray or an instrument set supplied by a manufacturer, not as an action against a dental office's sterilizer.
The reason is structural. A dental office is a user facility, not a manufacturer. Its reporting duty in the United States arises when a device may have caused or contributed to a death or serious injury, which a failed biological indicator followed by reprocessing does not usually produce. When a Canadian dental office's reprocessing breaks down, the record that results is normally a public health unit's infection prevention and control lapse report, published on the health unit's site, not an entry in any of the three databases discussed here.
Which field answers the question you actually have?
Most searches fail because the question was never sorted into one of four kinds. Each kind has one database and one field that answers it, and using the wrong one produces either a false reassurance or a false alarm.
| The question a clinic has | Where it is answered | The field that answers it |
|---|---|---|
| Has my model failed before? | FDA MAUDE | Product code FLE plus brand name |
| Was a lot or catalogue number recalled? | FDA Enforcement Reports | Recall number, product description, distribution pattern |
| Was there a Canadian action on this product? | Health Canada Recalls and Safety Alerts | Notice text and posting date, with the category filter |
| What do I produce if it happens to me? | None of the three | Your own load record, which is what a regulator will ask for |
The fourth row is the one that matters most and is the one the databases cannot supply. A recall notice can be matched to a clinic only if the clinic recorded the identifier the notice uses, on the day the item entered the building.
What should a clinic record so a notice can be matched to a load?
Six fields, all of which fit on one line of a log, and all of which appear in the recall record as well:
First, the sterilizer identifier, meaning the serial number or the asset number the clinic painted on the machine, because a notice names a model and sometimes a serial range. Second, the cycle date and the cycle number, because the notice gives a date range and a clinic needs to know which of its loads fall inside it. Third, the load contents by set name, so that a recalled pouch lot can be traced to the instruments that were inside it. Fourth, the biological indicator lot and the result, including the incubation start time, because a failed result is the event that triggers everything else. Fifth, the chemical indicator and the packaging lot, because packaging recalls are common and are the easiest to match. Sixth, the release decision and who made it, because an inspector's first question is who decided the load could be used.
The log itself is worth more than the database. A printable compliance log sheet covers the six fields above and can be filled in at the bench without a computer, which matters on the days when there is no time. A supply planner is the other half of the same discipline: knowing how many indicators and pouches a month of cycles consumes is how a clinic avoids the shortcut of skipping a test to stretch a box.
For questions that need a documented answer rather than a form, the compliance consultation channel returns a written answer that cites the regulation or standard it rests on, which is what an inspector is asking for when the question is about method rather than about product. Multi-site groups ordering indicators, pouches and tape across locations can use wholesale and multi-site ordering. Reselling? become a distributor.
What the numbers do not tell you
Three limits are worth stating plainly. First, under-reporting is large and unquantified: 434 of the 1,530 steam sterilizer reports in MAUDE have no event type recorded at all, and a report with no event type cannot be counted as an injury or as a malfunction. Second, the recall record is voluntary in almost every case: 1,499 of the 1,500 sterilization-related enforcement records are firm-initiated voluntary actions, which means the record shows what firms chose to recall, not what regulators found. Third, the Canadian record is thin for devices: 6 of those 1,500 enforcement records name a Canadian recalling firm, and Canada's own database is searched by text rather than by field, so absence of a hit is weak evidence of absence of a problem.
What the record does support is a narrow, defensible use: before buying or keeping a sterilizer, check the product code for the model family, read what the reports describe, and then check whether the clinic's own monitoring would have caught the same failure. That is a question a clinic can answer with a log and a biological indicator, and it is the only question in this article that a clinic controls.
Where does a clinic take this next?
Four things a clinic can act on rather than read: a printable compliance log sheet that records receiving dates, lot numbers and open-on dates and exports from the browser; the clinic supply planner that turns daily bench counts into cases per month; and, where a finding has to be documented for an inspector, Ask your compliance question which returns a written answer citing the regulation or standard it rests on.
Facility and multi-site buyers can also ask about wholesale and multi-site ordering for case pricing and account setup.
Related reading
- Compact steam sterilizers for the dental operatory: six models compared
- What a clinic does when a biological indicator fails
- How many biological indicators a dental hygiene clinic uses in a year
- Which records a mobile dental clinic keeps for infection control
- Sterilization pouch sizes: a working guide
Frequently Asked Questions
Is a MAUDE report the same thing as a recall?
No. A MAUDE report is a report of an event or a malfunction, and it can be filed by a manufacturer, an importer or a user facility. A recall is a separate action, published in the Enforcement Reports with a class attached. Many reports never lead to a recall, and many recalls are supported by reports that were filed months or years earlier, so the two counts cannot be added together or compared directly.
Why is the FDA count so much larger than the Canadian count?
Because the two systems count different things and are searched differently. The FDA endpoints are fielded databases that can be queried by product code or by words in the reason for recall, and the device recall record alone holds over forty thousand rows. Health Canada's public search is a full text search across recall notices, alerts and related pages, with category and year filters, so a word such as sterilizer returns a few hundred items rather than a fielded subset of recalls.
Does a failed biological indicator have to be reported to the FDA?
Not by itself. A user facility's reporting duty in the United States arises when a device may have caused or contributed to a death or serious injury, and a failed indicator that is caught before a load is released usually does not meet that threshold. The event still has to be handled internally: the load is quarantined, the cause is investigated, and the result is recorded. Whether a report follows depends on the harm, not on the failure.
Where do Canadian dental reprocessing failures appear if not in these databases?
Most often in a public health unit's infection prevention and control lapse report, published on the health unit's own website, after an inspection or a complaint. That report describes what was not followed, such as manufacturer instructions or provincial reprocessing guidance, and it is a public document with a date. It is a different forum from a recall database and from a college discipline decision, and a clinic should know which of the three an inspector or a patient is referring to.
Can I use these databases to decide whether to keep an older sterilizer?
Only as one input. The record can show whether the model family has a history of malfunctions or recalls, which is useful context before a capital decision. It cannot show the condition of the unit in front of you, which is established by its monitoring record, its maintenance history and its physical inspection. A sterilizer with a clean report history and a thin log is a worse risk than a sterilizer with an ordinary history and a complete log.
What is the single most useful field to record for recall matching?
The lot or catalogue number on the packaging, together with the date it entered the clinic. Recall notices are written against those identifiers, and a clinic that recorded them can answer a notice in minutes. A clinic that recorded only a product name has to reconstruct the answer from invoices, which is the step where most traceability breaks down.
Sources
U.S. Food and Drug Administration. openFDA device event endpoint, queried 18 November 2026 for product code FLE; dataset last updated 14 November 2026. Product code definitions from the openFDA device classification endpoint.
U.S. Food and Drug Administration. openFDA device enforcement endpoint, queried 18 November 2026 for records whose reason for recall contains the string steriliza, with classification counts and full record retrieval for the year, status, initiation type and country readings.
U.S. Food and Drug Administration. MAUDE search interface, the public front end for the manufacturer and user facility device experience reports.
U.S. Food and Drug Administration. Recalls, market withdrawals and safety alerts, the published recall actions that the enforcement dataset mirrors.
Code of Federal Regulations. 21 CFR 880.6880, steam sterilizer, Class II; 21 CFR 880.6870, dry heat sterilizer; 21 CFR 880.6860, ethylene oxide and chemical sterilizers.
Health Canada. Recalls and Safety Alerts advanced search, queried 18 November 2026 for the terms sterilizer and autoclave, with category and year facets as reported on the results page.
Health Canada. Recalls, advisories and safety alerts, the public database and its search guidance.
Public Health Ontario and the Provincial Infectious Diseases Advisory Committee. guidance on cleaning, disinfection and sterilization of medical equipment and devices in all health care settings, current provincial edition, the guidance against which Ontario reprocessing is inspected.
CliniEco Medical distributes clinic and laboratory consumables in Canada under MDEL #35334. Requirements that bind a specific practice remain the ones in its own provincial guidance, manufacturer instructions and written procedures.
Last updated: 18 November 2026.