11 — Memory and distance

Allergy, irritation and harm at a distance.

+ The questionHow can a small dose matter for years, or far from where it touched?

Irritation answers each dose as it comes

A dose that leaves the skin unmarked at a first contact can raise a rash at a later one, because by then the immune system recognizes the chemical. Chapters 8 and 9 followed what a surface meets at the place of contact, and Chapter 10 how blood carries what forms there into the circulation. A small dose can reach beyond the place of contact in two ways: memory, when the immune system learns a chemical and answers it on a later contact, and distance, when a substance is absorbed and carried to tissues far from the contact. Irritation has neither, so it shows by contrast what memory and distance add.

Irritation is local damage that a substance does where it touches, and its size follows the dose at each contact. An irritant acts on the cells it reaches, so a larger amount, a higher concentration or a longer contact does more harm. No record of an earlier contact carries over, so each exposure is judged by its own dose. Devices are tested for irritation under ISO 10993-23, published in January 2021 and confirmed at its systematic review on 8 June 2026. The standard's Amendment 1 of 2025 adds further reconstructed human epidermis models, layers of human skin cells grown into a tissue, for testing in the laboratory, and Chapter 18 weighs them against the tests in animals.

Sensitization keeps a memory of the first contact

Sensitization is the process by which a first contact teaches the immune system to recognize a chemical, so that a later contact raises an allergy. The memory forms in a series of steps, which OECD's 2012 account of the pathway sets out for chemicals that bind to proteins. First the chemical, or a metabolite the body makes from it, forms a covalent bond, a bond of shared electrons, with the amino acids cysteine or lysine in proteins of the skin. Keratinocytes, the cells of the skin's outer layer, respond with inflammation. Dendritic cells, immune cells of the skin, are activated and migrate to the nearest lymph node. There T cells, white cells that each recognize one target, multiply.

These steps make up the induction phase. It starts at the chemical's first penetration of the skin, and it leaves behind T cells that recognize the chemical. A later contact with the same chemical starts the elicitation phase: those T cells meet the chemical again, and the skin shows the reaction, allergic contact dermatitis. In its 2014 reasoning on nickel, ECHA, the EU's chemicals agency, notes that repeated exposures can be more potent than a single one, and it links this both to nickel accumulating in the outer skin and to the skin's memory of the allergen. A cushion worn every night meets the same skin again and again, so that memory matters for years.

Memory also changes the dose that matters. In ECHA's reasoning the threshold for induction is generally higher than the threshold for elicitation, so a smaller dose can elicit than was needed to induce, which answers the puzzle of this chapter's first line. Dose here means dose per area: OECD's account names the total dose per area of skin, in micrograms per square centimeter, as the critical determinant of exposure. The same amount can stay below a threshold when it is spread thinly over the skin and exceed it when it is held under a small part.

28 — Sensitization in two phases, and the dose each needs
INDUCTION covalent bond to cysteine or lysine in a skin protein keratinocytes: inflammation dendritic cell activated, migrates lymph node T cells multiply ELICITATION allergic contact dermatitis T cells return first contact memory remains later contact, same chemical DOSE PER AREA OF SKIN (µg/cm²) induction threshold elicitation threshold generally higher (ECHA, 2014) after induction, a dose in this band can raise a rash Schematic; no threshold values. Pathway: OECD, adverse outcome pathway for skin sensitization (2012), which leaves nickel out. Dose scale: ECHA (2014).
Released substanceTissue and cellsLimit or thresholdFocal detail
Plate 28 — Sensitization runs in two phases. A first contact binds the chemical to skin proteins and leaves T cells that recognize it; a later contact with the same chemical raises allergic contact dermatitis, generally at a lower dose per area of skin than induction needed.

Nickel shows both phases across a population, although it departs from the pathway above. OECD's account leaves nickel out, because a receptor of the innate immune system, Toll-like receptor 4, has a central part in sensitization to it. Nickel is the most frequent cause of contact allergy worldwide, and about 8 % to 19 % of adults in the European general population are allergic to it, with a strong predominance among women. Ahlström and colleagues state both in a 2019 clinical review, and they credit the European regulation of nickel with reducing that prevalence.

That regulation allows articles in direct and prolonged skin contact to release at most 0.5 µg/cm² of nickel a week, as ECHA's 2014 document quotes it. The limit is a rate per area of skin, and it aims to protect both people already sensitized and people not yet sensitized. ECHA's document also proposes what prolonged contact means: contact of potentially more than 10 minutes on three or more occasions within two weeks, or of 30 minutes on one or more occasions within the same two weeks. ECHA prepared the proposal at the European Commission's request, and it is not binding. By that measure, a part worn for hours each night, such as the running device's cushion and headgear, is in prolonged contact from its first night, because a single night's wear exceeds 30 minutes.

Acrylates show the mechanism in devices themselves. In a retrospective study published in 2024, Ulriksdotter and colleagues at Lund University and Skåne University Hospital analyzed the patch-test records of a clinic in Malmö from October 2017 to October 2020. In a patch test, suspected allergens are held against the skin and the skin is read days later. Of 54 diabetes patients tested for suspected allergic contact dermatitis from their glucose monitors or insulin pumps, 34 reacted to isobornyl acrylate, IBOA, against 4 of the 1,840 dermatitis patients tested with it. The authors write that IBOA has been widely used in high concentrations in device glues, and that devices do not list their ingredients. Foundations funded the study, and none of its authors worked for a company.

Solved example: how much more often device patients reacted to IBOA
  • The device group. 34 divided by 54 is 0.630, so 63.0 % of the device patients reacted.
  • The comparison group. 4 divided by 1,840 is 0.00217, so about 0.2 % of the other dermatitis patients reacted.
  • The ratio. 63.0 divided by 0.217 is about 290, so a reaction to IBOA was about 290 times as frequent among the device patients.
  • What each share counts. Both groups were patients patch-tested for dermatitis, and the 54 were tested because a device was suspected. The 63.0 % is therefore the share of suspected device reactions with an allergy to IBOA, not the share of device users who become allergic.
  • The result. Allergy to IBOA concentrated among patients whose skin had reacted to a device, which is consistent with the authors' account of IBOA in device glues; the study reports the difference at p < 0.001.
  • Assumptions. Counts as the retrospective study reports them, with its own comparison group.

The skin tests reproduce the two phases

The animal tests for sensitization follow the two phases. The guinea pig maximization test is set out in OECD Test Guideline 406, adopted in 1992, and it uses at least 10 treated animals and 5 controls. Induction begins on day 0 with injections into the skin and continues on days 6 to 8 with a 48-hour patch. Some of the injections carry Freund's complete adjuvant, a mixture that strengthens the immune response. The challenge on days 20 to 22 applies the substance again under a 24-hour patch, and each site is scored about 48 and 72 hours later on a scale from 0, no visible change, to 3, intense redness and swelling. The two weeks between the phases give the memory time to form, so the test reads elicitation in an animal it has sensitized.

The local lymph node assay reads induction alone. The EU's test method B.42 describes it, in a text that replaced the method's earlier one in 2012. A substance is applied to the ears of mice on days 1 to 3, and on day 6 the lymph nodes that drain the ears are removed to measure how far their lymphocytes, the class of white cells to which T cells belong, have multiplied. A substance counts as a sensitizer when the multiplication reaches at least three times that in mice given only the vehicle, the liquid that carries the substance. The assay needs no challenge and no adjuvant, and the method states that it reduces the animals' pain and distress.

Devices are tested for sensitization under ISO 10993-10, whose fourth edition, published in November 2021, sets out in vivo test procedures, those run in living animals. Tests without animals now address single steps of the pathway. In a 2024 study of one of them, GARDskin, Jenvert and colleagues wrote that such methods had not yet been sufficiently qualified for devices, so device testing remained mainly in animals. SenzaGen, the company that developed GARDskin, funded the study and employed four of its six authors. Chapter 18 takes up these methods and asks where animal tests are being replaced and where they remain.

Systemic effects need absorption and time

Systemic toxicity is harm at a site distant from the contact, so it needs a route from one to the other. The 2017 third edition of ISO 10993-11 makes absorption and distribution its condition: the substance must enter the circulation and be carried to the tissue it harms. The route decides how much enters. A substance released into blood, as from the lines of Chapter 10, is distributed at once. One from the running device's gas pathway reaches the blood through the lungs (Chapter 19), and one from the cushion must first cross intact skin. The plasticizer of Chapter 4 traveled that distance: it was found in the lung tissue of patients transfused with bag-stored blood.

Time sets the second condition. Some harm follows a single exposure and some appears only after repeated exposure, so ISO 10993-11 sorts its studies by duration. Acute systemic toxicity covers adverse effects within 72 hours after a single exposure, or after several within 24 hours. Subacute studies cover repeated or continuous exposure between 24 hours and 28 days, subchronic studies part of the lifespan, usually 90 days in rodents, and chronic studies the major part of it, usually 6 to 12 months. A study shorter than the contact can miss an effect that needs the contact's length to appear. A mask worn every night for years sits at the long end of this scale (Chapter 2). Chapters 14 and 15 show how such a contact can be judged from what the device releases.

29 — From the contact site to a distant tissue, and for how long
SKIN (CUSHION, HEADGEAR) AIRWAY (GAS PATHWAY; CHAPTER 19) BLOOD (A BLOOD LINE; CHAPTER 10) must first cross intact skin distributed at once CIRCULATION a site distant from the contact absorption and distribution: the condition for systemic toxicity ACUTE effects within 72 h of a single exposure, or several within 24 h SUBACUTE repeated or continuous, 24 h to 28 d (14 to 28 d in most guidelines) SUBCHRONIC part of the lifespan, usually 90 d in rodents, at most 10 % of it in other species CHRONIC the major part of the lifespan, usually 6 to 12 months NOT TO SCALE a mask worn nightly for years (Chapter 2) Schematic. Durations: ISO 10993-11:2017, terms and definitions.
Released substanceTissue and cellsGas or fluid pathFocal detail
Plate 29 — Systemic toxicity needs a substance absorbed and carried to a site distant from the contact, and the route decides how much enters. ISO 10993-11 sorts its studies by duration, from adverse effects within 72 hours of a single exposure to the major part of a lifespan.

Damage to DNA has no assumed safe dose

Harm that works by overwhelming a cell's defenses and repair has a dose below which no effect is seen, because a smaller dose leaves the defenses enough to cope. For damage to DNA, called genotoxicity, the European Medicines Agency made the opposite assumption in its 2006 guideline on genotoxic impurities: compounds genotoxic in the body can damage DNA at any level of exposure. The guideline drew the consequence: without evidence of a threshold it is prudent to assume none and to extrapolate in a straight line from studies at high doses. The assumption follows from what DNA is: a change to it can be copied at every division of the cell, so one damaged molecule can be carried forward instead of cleared.

The guideline named mechanisms that do have thresholds, such as interference with the spindle that pulls chromosomes apart at division. It set its threshold of toxicological concern at an intake tied to an excess lifetime cancer risk below 1 in 100,000, which is a risk accepted, not a dose without effect (Chapter 16). The guideline was written for medicines, was valid until 31 January 2018 and has since been superseded, so it shows where the reasoning came from and is not a rule for devices.

For devices, ISO 10993-3 covers these effects. Its fourth edition, published in August 2026, replaces the 2014 edition, now withdrawn, and widens the standard's scope: its title changes from tests for genotoxicity, carcinogenicity and reproductive toxicity to the evaluation of genotoxicity, carcinogenicity, reproductive toxicity and developmental toxicity. Carcinogenicity is the capacity to cause cancer, and developmental toxicity is harm to the unborn or growing child. The new edition takes in the genotoxicity test guidance that ISO/TR 10993-33 held, according to ISO's catalogue. A 2026 summary by staff of the testing laboratory Eurofins adds that the edition gives chemical characterization a larger part and introduces considerations for endocrine disruptors. Chapter 18 describes the tests themselves.

Endocrine disruption is interference with the body's hormones. DEHP, the plasticizer that Chapter 4 follows from PVC into stored blood, is listed as a substance of very high concern under REACH, the EU's chemicals law. SCHEER, a scientific committee of the European Commission, issued guidelines on phthalates in devices in 2024, and they cover the phthalates that cause cancer, damage genes, harm reproduction or disrupt hormones. Chapter 27 sets out the EU rule.

30 — Four kinds of response, from the contact site to DNA
where it touched far from it DNA IRRITATION SENSITIZATION SYSTEMIC TOXICITY GENOTOXICITY, CARCINOGENICITY, REPRODUCTIVE AND DEVELOPMENTAL TOXICITY WHERE IT ACTS the contact site the contact site, through immune memory a site distant from the contact, after absorption DNA; cancer; fertility; the unborn or growing child HOW DOSE WORKS size follows the dose at each contact induction, then elicitation at a generally lower threshold single exposure to most of a lifespan: acute to chronic no threshold assumed by default for DNA damage DEVICE STANDARD ISO 10993-23:2021, confirmed 8 June 2026; Amd 1:2025 adds further reconstructed human epidermis ISO 10993-10:2021, edition 4 ISO 10993-11:2017, edition 3 ISO 10993-3:2026, edition 4, August 2026; takes in ISO/TR 10993-33 Status: ISO catalogue pages. No-threshold default: EMA guideline on genotoxic impurities, 2006, now superseded. Endocrine disruptors are considered in ISO 10993-3:2026, a testing laboratory's summary states.
Tissue and cellsFocal detail
Plate 30 — Four kinds of response reach further from the contact in turn: irritation stays where it touched, sensitization remembers, systemic toxicity needs absorption, and for damage to DNA no safe dose is assumed by default. Each has its part of ISO 10993.

The body's side completes the first half

Part IV completes the first half of this book's central question: why a plastic that is safe in a water bottle can harm a patient through a breathing circuit or a catheter. The plastic can be the same in all three. A bottle passes what leaves it to the gut, and a breathing circuit passes it to the lungs every night (Chapter 1). A catheter passes it straight into the blood, where clotting, platelets and complement meet the wall at once (Chapter 10). The liquid or gas decides how much leaves (Chapter 4). The tissue at the end of the route decides the response: memory in the skin, harm in distant organs, a change to DNA for which no safe dose is assumed by default, or a capsule at an implant (Chapter 9). The second half of the question begins in Part V, with the material's chemistry (Chapter 12).

In practice

For each part of the running device that touches skin, write down what it can give the skin and for how long. The silicone cushion and the headgear of polyurethane foam, nylon and polyester touch the face and head for hours each night for years (Chapters 2 and 3). Name what each may release, from the silicone's short chains and cure residues to the foam's catalysts and surfactant. Check any adhesive for acrylates and any metal that rests on skin for nickel release. Plan the evaluation of irritation under ISO 10993-23 and of sensitization under ISO 10993-10 for both parts, and carry whatever could be absorbed into Chapters 12 to 18.

+ What this chapter established
  • Irritation is local damage whose size follows the dose at each contact, and ISO 10993-23, confirmed on 8 June 2026, added further reconstructed human epidermis models in 2025.
  • Sensitization has two phases: induction leaves T cells that recognize a chemical bound to skin proteins, and elicitation raises the allergy on a later contact, generally at a lower threshold.
  • Nickel and the acrylates of device glues show the mechanism in patients, and the guinea pig maximization test reproduces both phases while the local lymph node assay reads induction alone.
  • Systemic harm needs absorption and is studied from a single exposure to most of a lifespan. For damage to DNA no safe dose is assumed by default, following the reasoning of a 2006 medicines guideline, and ISO 10993-3:2026 now evaluates it with cancer and reproductive and developmental harm.
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.