07 — Sterile, packed, aged

Sterilized, packed and aged.

+ The questionHow do sterilization, cleaning and time change what a device releases?

Sterilizing is a process like molding

The running device reaches its user clean but not sterile, and the user then cleans its mask, hose and humidifier tank at home. A single-use catheter, by contrast, reaches the patient sterile, after months or years in its package. A sterile device meets the sterilant in its package, after every step of Chapter 6, so the sterilant's changes add to those of the making. The catheter therefore carries what sterilization and storage leave in it, and the running device what years of cleaning leave.

Sterilization changes the material itself, as molding does. In a 2018 bench study, Cingolani and colleagues sterilized two commercial grades of a copolymer of lactide and caprolactone, a polyester the body slowly breaks down, by electron beam and by ethylene oxide. The electron beam's sterilizing dose was 25 to 30 kilograys, kGy, a measure of the radiation energy each kilogram absorbs. That dose cut the average molar mass of one grade to less than half, from about 120,000 to about 52,000 g/mol, and that of the other by nearly a quarter, from about 173,000 to about 133,000. Ethylene oxide gave both grades "a statistically relevant albeit small reduction", the first to about 96,000 and the second to about 159,000. The irradiated polymer then broke down in water about 3.5 times faster. One of the authors' affiliations was a device company, Industrie Biomediche Insubri.

Ethylene oxide goes in, and aeration draws it out

Ethylene oxide, EO, sterilizes as a gas, and a gas that reaches the microorganisms in a packed device also enters its materials. Absorbent materials such as cellulose and cotton, and some plastics, take up EO and hold it, as STERIS, a sterilization contractor, explains. After the cycle the device is aerated: the gas diffuses back out of the materials by the diffusion of Chapter 4, and heating speeds its release. STERIS gives 6 to 24 hours of aeration as typical in industry. Whatever stays is a residual: the sterilant, or a product of it, left in a material when the process ends. EO that meets free chloride ions forms ethylene chlorohydrin, ECH, and EO that meets water forms ethylene glycol, EG.

ISO 10993-7, the standard on EO residuals, sets allowable limits for EO and ECH and none for EG. Its third edition was published on 20 April 2026 and replaced the 2008 edition, with its 2009 corrigendum and 2019 amendment, which ISO lists as withdrawn. It applies to devices sterilized with EO, with two exceptions. Devices with no direct or indirect contact with the body or the user, such as in vitro diagnostic devices, lie outside it, and so do devices shown not to absorb or retain the two substances, which the standard limits to metal alloys and glass.

A laboratory measures the residuals in one of two ways: it extracts the patient-contacting part in water to simulate use, or it extracts the whole device exhaustively, and it then analyzes the residuals by gas chromatography. NAMSA, a contract laboratory, describes the method this way under the 2008 edition. Under the 2026 edition the extraction conditions follow from the patient population and the duration of use, as the limits do, according to DQS, an EU notified body.

The limit is an amount the patient may take in, so it depends on how long the contact lasts and on the patient. The 2026 edition starts from a tolerable intake, the daily dose per kilogram of body mass judged acceptable over a given duration of contact (Chapter 14). It sets one for each residual and for each contact category of Chapter 2. For EO its Table 1 sets 0.3 mg/kg a day for limited and for prolonged contact and 0.02 mg/kg a day for long-term contact, and for ECH 0.64, 0.27 and 0.029 mg/kg a day. The intake is smallest for long-term contact, because a dose repeated for years is held to a lower daily amount than a brief one: for EO it is a fifteenth of the value for limited contact. DQS adds that the intakes are derived with uncertainty factors.

Each device's limit is calculated from those intakes. The standard multiplies the intake by the patient's body mass and by a concomitant exposure factor, 0.2 by default, which Eurofins, a testing laboratory, describes as allowing for several devices used at once. The product is the most the patient may take in each day from the device. The standard then multiplies that daily amount by the days of contact and divides it by the number of devices used over them, which gives the device's allowable limit. The standard requires the lowest body mass among the patients the device is meant for, so the same device has a lower limit when it is meant for smaller patients. Eurofins adds that the 2026 edition uses a standard adult body mass of 60 kg, with lower body masses for children, infants and neonates "often resulting in stricter residual limits".

19 — Ethylene oxide taken up, then aerated out, against a limit
EO LEFT IN THE DEVICE EO CYCLE AERATION TIME allowable limit: adult smaller patients: lower each limit calculated from the intake with the lowest body mass of its patients (ISO 10993-7:2026) an absorbent material (cellulose, cotton) holds more, longer a plastic part residual below the limit for this patient group tolerable intake (mg/kg a day) limited prolonged long-term EO 0.3 0.3 0.02 ECH 0.64 0.27 0.029 ECH: EO + chloride EG: EO + water (no limit set) Schematic curves and limit lines. Tolerable intakes: ISO 10993-7:2026, Table 1. Uptake and aeration: STERIS (a sterilization contractor); limits by duration and patient group: DQS (a notified body) and Eurofins (a laboratory), 2026.
Released substanceLimit or thresholdFocal detail
Plate 19 — Ethylene oxide enters a device during the cycle and diffuses out during aeration. ISO 10993-7:2026 sets a tolerable intake for each contact duration and calculates each device's limit from it with the lowest body mass of its patients, so the same device has a lower limit when it is meant for smaller patients.

The withdrawn 2008 edition set fixed amounts per device instead, and they scale with the same contact categories. The 2026 edition notes that the earlier editions allowed 4 mg of EO and 9 mg of ECH for an adult of 70 kg in limited contact, up to 24 hours. In NAMSA's account of the 2008 edition, a device in prolonged contact, up to 30 days, could deliver an average of 2 mg of EO a day, and a permanent device an average of 0.1 mg a day and 2.5 g over a lifetime. The 2008 values show the structure of such limits, not values to design to.

EN ISO 10993-7:2026, the European adoption, became a harmonized standard under the EU's device regulation by an implementing decision published on 10 November 2026. A harmonized standard gives a presumption of conformity with the requirements it covers (Chapter 27). FDA has not recognized the 2026 edition: in November 2026 its list of recognized standards still names the 2008 edition, with its corrigendum and amendment.

Radiation breaks chains and joins them

Gamma rays, electron beams and X-rays sterilize by depositing energy in the material. Sterigenics, a contract sterilizer, gives 25 to 40 kGy as an example of a sterilizing dose, a range that takes in the electron-beam dose of the 2018 study. The energy ionizes the polymer's chains and leaves free radicals, the reactive species of Chapter 6's melt. A radical can join two chains, which is crosslinking, or split one, which is chain scission (Chapter 5). Under gamma radiation both happen at once, as Naikwadi and colleagues describe in a 2022 review of radiation-processed polymers. With oxygen present the radicals form peroxides and hydroperoxides, which lead to further scission, the reaction that antioxidants interrupt in a melt.

In the ratings Sterigenics publishes of how polymers tolerate radiation, the reaction that dominates depends on the polymer. In its account, scission lowers tensile strength and elongation, and crosslinking raises strength and lowers elongation. Polypropylene rates poor to fair unless stabilized, and the company recommends radiation-stabilized grades. PVC rates good, with crosslinking dominant, but yellows above about 30 kGy, a dose inside that example range. Silicone crosslinks and stiffens. Stabilizers called antirads, typically under 1 % of the material, act as antioxidants against the radicals that radiation makes. A part whose antioxidant was partly spent in its melt (Chapter 6) has less left for those radicals, so a radiation-sterilized part needs its antioxidant chosen with both steps in view.

Radicals can outlast the sterilization. Ultra-high molecular weight polyethylene, UHMWPE, is the bearing material of knee implants. Jahan and colleagues measured the radicals that radiation leaves in it by electron spin resonance, and in a 1996 meeting abstract they reported that these radicals "exist in shelf-aged UHMWPE tibial components for several years". In a 2003 bench study, Lu and colleagues reported that gamma irradiation "creates reactive free radicals" in UHMWPE and makes it more prone to oxidation. They also reported that packing it in vacuum limited oxidative degradation over three years of shelf aging. The package is therefore part of the process: a radical that meets oxygen on the shelf can go on cutting chains.

Steam cuts a polyester's chains

Steam sterilizes with heat and water together, and for a polyester the pair brings back the hydrolysis of Chapter 5, in which water splits the ester links of a chain. In a 2023 bench study funded by a Polish research grant, Kucharska-Jastrząbek and colleagues sterilized 39 printed bone implants of hydroxyapatite-filled polylactide, PLA, a polyester. Steam at 134 °C for 5.5 minutes deformed the implants and cut their average molar mass to less than half, from about 106,000 to about 41,000 g/mol, and steam at 121 °C for 16 minutes cut it to about 56,000. Ethylene oxide lowered it by about 2 %, and electron beam took it to about 78,000. The authors conclude that steam "should not be used" for these implants. They do not name a mechanism, and hydrolysis is this book's reading of their result.

Polycarbonate behaved differently in steam. In a 1996 trade-magazine article, Hong and colleagues reported bench tests of 10 specimens per condition, in which "hydrolysis did not occur during steam sterilization cycles lasting up to 120 minutes" at about 121 °C. All three authors were staff of the device maker Baxter Healthcare. The polycarbonate still lost ductility, for two reasons: heat alone aged it, as dry-oven heating showed, and foreign specks started cracks. The same article cites earlier work in which polycarbonate "is subject to hydrolysis by hot-water aging". That condition is like the one in the running device's polycarbonate chamber, which holds heated water night after night (Chapter 3).

20 — Three ways to sterilize, and what each does to the chains
ETHYLENE OXIDE — A GAS, THEN AERATION aeration small fall: one resorbable polyester grade, 119,609 to 96,000 g/mol PLA implants: 105,700 to 103,500 g/mol leaves residuals: EO; ECH where it meets chloride; EG where it meets water RADIATION — GAMMA, ELECTRON BEAM, X-RAY scission crosslink example sterilizing dose 25 to 40 kGy (a contract sterilizer) electron beam, 25 to 30 kGy: 119,609 to 51,814 g/mol same grade: less than half the molar mass radicals last for years on the shelf (UHMWPE) STEAM — HEAT AND WATER hydrolysis PLA implants, 134 °C, 5.5 min: 105,700 to 41,460 g/mol; implants deformed polycarbonate, about 121 °C, up to 120 min: no hydrolysis found (Baxter staff, 1996) Schematic; chain counts illustrative. Resorbable polyester (lactide–caprolactone copolymer): Cingolani and colleagues, 2018; PLA implants: Kucharska-Jastrząbek and colleagues, 2023; dose and radicals: Sterigenics (2018), Jahan and colleagues (1996); polycarbonate: Hong and colleagues, 1996 (Baxter Healthcare staff).
Device materialReleased substanceGas or fluid pathFocal detail
Plate 20 — Each sterilization route acts on the chains in its own way. In one grade of a resorbable polyester, electron beam cut the average molar mass to less than half while ethylene oxide reduced it less; steam at 134 °C cut that of a polylactide to less than half, and polycarbonate showed no hydrolysis in steam cycles of up to 120 minutes.

Heat stands in for time on the shelf

A sterile device carries an expiry date, which has to be set before years of real storage have passed. ASTM F1980-21 is the active guide to accelerated aging of sterile barrier systems and devices. It gives protocols that "model the possible effects of the passage of time on the sterile integrity" of the package. Heat speeds the reactions of aging, and the guide rests the date on "a conservative estimate of the aging factor (that is, Q10)", the factor by which a reaction speeds up for each 10 °C rise. A common choice is the "10 °C rule", a doubling for each 10 °C. In a 2019 theoretical analysis, Janting and colleagues describe it as an empirical rule often used for device polymers. Two of the three authors worked for companies, ConvaTec and the consultancy MedDevCon.

An activation energy sets how steeply a reaction speeds up with heat. The analysis by Janting and colleagues averages over the activation energies typical of polymer reactions, 0.1 to 0.9 eV, and on that average the 10 °C rule holds only at a reference temperature of 25 °C. The aging factor is therefore an estimate, and ASTM's guide treats the date it gives as "tentative" until real-time aging confirms it. Real-time aging is storage at the labeled conditions for the full period.

Solved example: how many days at 60 °C stand for a year at 25 °C
  • The rule. On the 10 °C rule, each 10 °C above the reference temperature doubles the rate of aging.
  • The temperatures. The reference is 25 °C, where the 2019 analysis finds that the rule agrees with its average. The oven runs at 60 °C, the top of the interval of about 25 to 60 °C that the same authors advise for testing device polymers, and the highest oven temperature of the 1998 protocol in Chapter 5.
  • The steps. 60 minus 25 is 35 °C, three and a half steps of 10 °C. Three doublings give 8, and the half step multiplies by the square root of 2, about 1.41, so the rate rises about 11.3 times.
  • The result. On the rule, 365 days at 25 °C correspond to about 32 days at 60 °C, since 365 divided by 11.3 is about 32.
  • Assumptions. One reaction that doubles for every 10 °C throughout, and no transition of the material, such as its glass transition, between 25 and 60 °C. A reaction with a higher activation energy speeds up more and one with a lower one less, so the 32 days may stand for more or less than a year.

An oven models what heat speeds up, and nothing else. ASTM states that extreme temperature and humidity challenges lie outside the guide, which refers them to another ASTM standard, and that accelerated data do not set the storage conditions on a label. A device in use meets more than heat: water drives the hydrolysis of Chapter 5, ozone from home cleaners attacks chains, and cleaning agents act on the surfaces night after night. On this book's reading, accelerated aging of a packed device models none of these, so a shelf life does not show how a mask or a hose ages in use.

21 — What accelerated aging models, and what it leaves out
WHAT AN OVEN MODELS 25 °C × 1 35 °C × 2 45 °C × 4 55 °C × 8 60 °C × about 11.3 10 °C rule: a doubling for each 10 °C (Q10 = 2) about 32 days at 60 °C stand for a year at 25 °C date tentative until real-time aging confirms it WHAT IT LEAVES OUT ozone from home cleaners attacks chains (Chapter 5) cleaning agents, night after night act on surfaces in use extreme temperature and humidity outside ASTM F1980-21's scope (another ASTM standard) Schematic. ASTM F1980-21 (active); the 10 °C rule as described by Janting and colleagues, 2019 (authors from ConvaTec and MedDevCon), who find it agrees with an average over activation energies of 0.1 to 0.9 eV only at a 25 °C reference. Ozone and cleaning as not modeled: this book's statement, not the standard's.
Device materialGas or fluid pathFocal detail
Plate 21 — Accelerated aging raises the temperature so that, on the 10 °C rule, about 32 days at 60 °C stand for a year at 25 °C, and the date stays tentative until real time confirms it. The oven models heat alone: ASTM places extreme temperature and humidity outside its guide, and on this book's reading the oven models neither the ozone nor the cleaning agents a device meets in use.

Cleaning at home is a process of use

The running device meets its longest process at home, where its user cleans it for years. In a consumer update of August 2024, FDA advises following the maker's schedule, cleaning most masks, hoses and humidifier tanks with mild soap and water, or with diluted vinegar where a maker recommends it, and air-drying them. The update describes two kinds of cleaner sold for these machines, ozone gas and UV light, and states that both can break down accessories such as hoses and masks, and that FDA has no data that either kind cleans inside hoses or leaves the equipment undamaged. For the foam of the recalled breathing machines, FDA's order of 10 March 2022 states that ozone cleaners "may exacerbate the breakdown of the foam" (Chapters 5 and 20).

Ozone also harms people: FDA's 2024 consumer update reports asthma attacks, headaches and breathlessness after ozone cleaning, and ozone inside the equipment may stay above safe limits for several hours. In November 2023 FDA issued a safety communication on SoClean's voluntary recall of its SoClean2 and SoClean3 ozone cleaners. The recall was a correction in the field, with a new manual and an adapter meant to keep ozone out of the machine. Because residual ozone can remain in the mask or tubing, the instructions add steps to air them.

The consumer update of August 2024 states: "There are currently no FDA cleared or approved devices for cleaning, disinfecting, or sanitizing CPAP machines." On 12 August 2024 FDA granted a De Novo request for SoClean 3+, an ozone add-on that reduces bacteria on certain masks and hoses after regular cleaning, not in place of it. On 10 October 2025 FDA cleared a later SoClean 3+ by 510(k) as the same kind of device. The sponsor's own summary of the 510(k) reports ambient and residual ozone below 0.05 ppm, the level that FDA's review of the De Novo takes from the agency's ozone regulation. It also reports that masks conditioned by the device were extracted and analyzed for what they released. The clearance is for reducing bacteria, so FDA's 2024 statement about cleaning holds as of its date.

In practice

For each part, list every process between assembly and the patient: the sterilization method with its dose or cycle and aeration, the package and storage time, and every cleaning agent the instructions allow. Take test articles after the harshest sterilization cycle and after aging, and for a reusable part after the cleanings of its useful life. Check EO residuals against the edition each market cites: EN ISO 10993-7:2026, harmonized in the EU, and the 2008 edition that FDA recognizes. For the running device, test the mask and hose after repeated cleaning as instructed, and name in the instructions the cleaners the device was not tested with.

+ What this chapter established
  • Sterilization changes a material as molding does: electron beam cut one grade of a resorbable polyester to less than half its average molar mass and another by nearly a quarter, and ethylene oxide reduced each less.
  • Ethylene oxide is taken up and aerated out, and ISO 10993-7:2026, harmonized in the EU but not recognized by FDA, limits residual EO and ECH by contact duration and patient group.
  • Radiation crosslinks and cuts chains through radicals that can last for years in UHMWPE on the shelf. Steam cut the average molar mass of printed PLA implants to about half or less, which this book reads as hydrolysis, and did not hydrolyze polycarbonate in cycles of up to two hours.
  • Accelerated aging lets heat stand in for time, and this book reads it as modeling neither ozone nor cleaning. FDA said in August 2024 that no device was cleared or approved for cleaning, disinfecting or sanitizing CPAP machines, and that month it granted a De Novo request for an ozone add-on that reduces bacteria after regular cleaning, not in place of it.
Created with Claude (Anthropic).
Educational reference material, not engineering, regulatory or clinical advice. Product and company names referenced are the property of their respective owners.