BRENNEN

Industries / Health / Medical devices / Medical sterilization

Ethylene oxide

This work sterilizes packed devices in sealed chambers with ethylene oxide (EtO), a gas that kills microbes without heat but is toxic, flammable and causes cancer, and then airs the gas back out of the products.

6

Milestones achieved, from 1940 to 2019.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1940

    A vacuum, then the gas

    Carroll L. Griffith and Lloyd A. Hall were granted US patent 2,189,947 on 13 February 1940, filed in May 1937, for sterilizing with undiluted EtO after first drawing a vacuum on the load to pull out air and moisture. Most of their examples were spices, but the patent also covers surgical supplies such as sutures, which, it says, long high-heat steam treatment can damage.

    patents.google.com

    Credit

    Carroll L. Griffith and Lloyd A. Hall, Griffith Laboratories

  2. 1949

    Time, heat and moisture measured

    In November 1949 Charles R. Phillips and S. Kaye published a series of papers in the American Journal of Hygiene measuring how exposure time, gas concentration, temperature and moisture change how well gaseous EtO kills microbes. In 1964 Phillips's group reported that bacteria dried past a critical point partly resist the gas, more so when dried in organic matter.

    LessonMoisture decides whether the gas reaches the microbes, and the CDC now lists 40 to 80% relative humidity among the four settings an EtO cycle must hold, with a humidifying step at the start.

    ebi.ac.ukpmc.ncbi.nlm.nih.govcdc.gov

    Credit

    Charles R. Phillips and S. Kaye; by its 1960 paper, Phillips's group worked for the US Army Chemical Corps at Fort Detrick, Maryland

  3. 1984

    A 1 ppm worker limit

    On 22 June 1984 the US Occupational Safety and Health Administration (OSHA) limited workers' exposure to 1 part EtO per million parts of air (1 ppm), averaged over an 8-hour shift, after finding EtO a cancer, genetic, reproductive and nerve hazard. Above an "action level" of 0.5 ppm, employers must measure workers' air regularly and offer medical checks. The rule took effect on 21 August 1984.

    osha.gov

    Credit

    OSHA

  4. 1995

    Pure gas replaces a mix

    Most sterilizers had used a blend of 12% EtO and 88% chlorofluorocarbon (CFC), a gas added to stop the mix from burning. When the Montreal Protocol phased out CFCs to protect the ozone layer, plants moved to pure EtO, which burns at any concentration above 2.6% in air, and the National Fire Protection Association (NFPA) first published NFPA 560, its standard for EtO sterilization, in 1995. NIOSH counted 10 explosions at industrial EtO sterilizing and repackaging plants from 1994 to 1998, which killed one worker and injured 59.

    LessonRemoving a diluent that harmed the ozone layer left plants handling an explosive gas they were not used to, by the account of EtO makers' safety staff.

    csb.govstacks.cdc.gov

    Credit

    the parties to the Montreal Protocol; NFPA; NIOSH, the US EPA and the Ethylene Oxide Sterilization Association, which wrote the 2000 alert together

  5. 2004

    A safety step skipped

    On 19 August 2004 an explosion wrecked a sterilizing chamber and its thermal oxidizer, a burner that destroys the EtO in vent air, at Sterigenics's plant in Ontario, California. Four workers had minor injuries and the plant could not be used. The US Chemical Safety Board (CSB) found that a maintenance supervisor had used a password to skip the gas washes that clear explosive gas from the chamber, and that the company had not installed chamber monitoring that NIOSH and NFPA recommended.

    LessonBy the CSB's finding, the monitoring that would likely have prevented the blast was already in industry guidance but not installed.

    csb.govcsb.gov

    Credit

    CSB investigators; Sterigenics

  6. 2019

    FDA asks for less gas

    After a large EtO plant closed in February 2019, the FDA opened two public challenges on 15 July 2019, one for methods that avoid EtO and one for cutting EtO emissions, and chose 12 of 46 applicants. They included NovaSterilis (supercritical carbon dioxide) and Noxilizer (nitrogen dioxide) for the first, and Abbott, BD, Medtronic, Sterigenics, STERIS and three others for the second. From 25 November 2019 a pilot let makers of high-risk devices move to cycles with less EtO by a short report instead of a new approval; by August 2022 it covered 11 sites and 28 devices, and the FDA said some facilities had cut emissions by 20 to 35%.

    LessonAn FDA advisory panel found in November 2019 that paper instructions packed with a device hinder the gas and raise the amount of EtO a cycle needs, and the FDA urged makers to move instructions online where safe.

    fda.govfda.govfda.govcontent.govdelivery.com

    Credit

    the FDA's Center for Devices and Radiological Health; the challenge participants

Watch for

  1. Chlorine dioxide in factories

    Chlorine dioxide is a yellow-green gas that does not burn. ClorDiSys Solutions says it reaches into narrow tubes and openings in complex devices, does not cause cancer, and suits devices with built-in batteries, and that working groups are gathering data toward an ISO standard for the method.

    Achieved whenISO publishes a standard for chlorine dioxide sterilization of devices, or the FDA lists it as an established method.

    packagingdigest.com

    Claimed by

    Emily Lorcheim of ClorDiSys Solutions, at a TÜV SÜD webinar reported in June 2026

Open

  1. 01

    Sterilization changes, approval by approval

    A public, coded table of every FDA approval of a change to how a high-risk (Class III) device is sterilized. Each row gives the device, the date, the kind of change (a new sterilization site, a new contract sterilizer, an added chamber, a cycle change, less EtO, a step removed, a switch to another method), the method before and after where stated, and the sentence it was taken from. Yearly counts then show whether makers of high-risk devices have moved away from EtO since the FDA's 2019 challenges, or mostly moved from one EtO plant to another.

    The brief
    Credit · unclaimed
  2. 02

    Sterilizer explosions, logged

    An open log of fires, explosions and EtO releases at sterilizing sites worldwide since 1990, in factories and hospitals. Each entry gives the date, site, deaths and injuries, the step of the cycle, the equipment that ignited or released the gas, the root cause where an investigator named one, and its sources.

    The brief
    Credit · unclaimed

Open 01 · the brief

Sterilization changes, approval by approval

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public, coded table of every FDA approval of a change to how a high-risk (Class III) device is sterilized. Each row gives the device, the date, the kind of change (a new sterilization site, a new contract sterilizer, an added chamber, a cycle change, less EtO, a step removed, a switch to another method), the method before and after where stated, and the sentence it was taken from. Yearly counts then show whether makers of high-risk devices have moved away from EtO since the FDA's 2019 challenges, or mostly moved from one EtO plant to another.

Why it is still open

For each PMA supplement, the request a maker files to change an approved high-risk device, the FDA's database keeps a short approval statement. Through openFDA, 2,947 of these statements mention sterilization, 722 ethylene oxide, 356 a "sterilization site", 69 gamma and 11 electron beam (counted on 20 November 2026). Recent ones read "an additional ethylene oxide (EO) sterilizer" and "the removal of the ambient aeration step from the Ethylene Oxide (EO) sterilization process" (October 2026). The FDA reports its pilots only as totals, such as 11 sites and 28 devices by August 2022, and searches of the web and GitHub found no study that codes these approvals. The record has gaps the table must state: since 2019 some changes reach the FDA as reports that cite a sterilizer's master file rather than as supplements, and in November 2024 the FDA set a transitional enforcement policy for EtO site changes to Class III devices.

api.fda.govfda.govcontent.govdelivery.com

Built with what exists

the openFDA device PMA endpoint, which returns each supplement's type, reason and approval statement as JSON; a model that answers fixed-choice questions in bulk, such as Gemini Decision 1, to sort each statement into kinds of change (see the Note "A thousand questions, one pass"); DuckDB; two reviewers who know device regulation.

  1. First step

    Pull the 722 statements that mention ethylene oxide, code 100 by hand into seven kinds of change, and compare one model's codes with yours.

  2. Done looks like

    All 2,947 statements coded, with the quoted sentence for each. Two reviewers code 300 at random and agree with the table on at least 95%. Yearly counts by kind of change and by method are published with the code, and the years 2019 to 2022 are set beside the FDA's pilot totals, with any gap explained.

Open 02 · the brief

Sterilizer explosions, logged

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An open log of fires, explosions and EtO releases at sterilizing sites worldwide since 1990, in factories and hospitals. Each entry gives the date, site, deaths and injuries, the step of the cycle, the equipment that ignited or released the gas, the root cause where an investigator named one, and its sources.

Why it is still open

NIOSH's 2000 alert counts 10 explosions from 1994 to 1998. The CSB in 2006, and Reza and Christiansen of Exponent in 2013, studied the 2004 Ontario blast and looked back at earlier ones. Searches found no list after that. The question matters again: the EPA's 2024 rule adds limits on EtO in plants' room air, so more air goes to oxidizers and scrubbers, the kind of control the 1990s explosions passed through.

stacks.cdc.govaidic.itclearinghouse-umich-production.s3.amazonaws.comcsb.gov

Built with what exists

OSHA's public accident search and severe injury reports; CSB reports; France's ARIA and the EU's eMARS accident databases; state fire marshal reports and newspaper archives; an AI model that reads incident narratives and tags the cause with a quoted line (the Note "The jobs agents repeat"); a spreadsheet or DuckDB table published on GitHub.

  1. First step

    Search OSHA's accident database for "ethylene oxide", read the first 25 narratives that involve sterilizers, and tag their causes.

  2. Done looks like

    At least 40 incidents, each with a source a reader can open, including all 10 that NIOSH counted for 1994 to 1998. Two readers tag causes separately and agree on at least 90%. A count by year and by cause shows whether incidents changed after the 1997 and 2024 rule changes.