Something it cannot remove.
+ The questionHow does tissue respond to an implant or a particle?
The wound comes first
A sliver of plastic left in tissue can end up inside a capsule of fibrous tissue, a wall that the body builds around something it cannot remove. The capsule is the last stage of a sequence that begins with the wound. Placing an implant cuts or tears tissue, and the protein coat of Chapter 8 settles on the material at once, so the cells that answer the wound meet the coat. Sheikh and colleagues describe the sequence in a 2015 review of how macrophages and related cells respond to implanted materials.
The first response is acute inflammation, marked by neutrophils, a kind of white blood cell, and it usually resolves in less than a week. The stage that follows is called chronic inflammation, although it generally lasts no longer than two weeks. Acute inflammation that persists beyond three weeks usually indicates an infection. These periods belong to the wound, which heals. The material does not heal, and the response to the material itself, the foreign-body reaction, can last as long as the implant stays. Which cells carry that reaction, and what they can do, depends on the size of what they find.
Macrophages engulf, or fuse
Macrophages respond rapidly to an implanted material and become the dominant cells around it. A macrophage clears debris by phagocytosis: it wraps its outer membrane around a particle and draws the particle inside, where the cell can try to break it down. That works only for something the cell can wrap, so size decides which response a particle meets. In the review's account, particles under 10 µm across are taken in by phagocytosis, macrophages fuse together on materials between 10 and 100 µm, and anything larger is broken down outside the cells.
When a material cannot be engulfed by a single macrophage, several macrophages fuse into one foreign-body giant cell. The giant cell carries the nuclei of all the cells that formed it and can spread over more of the surface than any one of them. Mast cells, another kind of immune cell, release two signaling molecules, IL-4 and IL-13, as they discharge their granules, and in the laboratory either molecule induces the fusion. Giant cells may themselves cause chronic inflammation, with loss of the bone around an implant.
The surface shapes this reaction as it shaped the coat. Surface chemistry, form and topography determine the make-up and severity of the foreign-body reaction, so the reaction is a property of a part as made, not of its plastic in general. A smooth tube and a porous block of one polymer present different surfaces to the cells, and the cells meet each through its own coat.
A capsule walls off what stays
While the cells work at the material's surface, the wound around them heals. New tissue, called granulation tissue, fills the wound. It may be a precursor to the fibrous capsule, a layer of dense tissue that encloses the implant and separates it from the tissue around it. The capsule does not remove the material but walls it off. The review gives no time for the capsule to form.
How thick the capsule grows is one sign of how strongly tissue responds to a material. Reactive materials may produce a capsule extending 2 to 4 mm, Wallin and Upman wrote in a 1998 trade article on the implantation test. Both authors worked at NAMSA, a contract testing laboratory. The article described the edition of the standard used in the 1990s, so the figure shows the scale of a strong response, not a current criterion.
The airways sort particles by size
A device can also release pieces of itself. A slippery coating can separate from a catheter (Chapter 4). A foam can crumble as its chains are cut (Chapter 5), and parts that rub together shed particles as they wear. In tissue, such a particle meets macrophages, and its size decides which response it draws. In the air a patient breathes, the airways sort a particle before any cell meets it.
A model describes that sorting region by region. The International Commission on Radiological Protection, ICRP, published its human respiratory tract model in 1994 as Publication 66 and revised it in Publication 130 of 2015, which changed how the model clears deposited particles and treats their absorption. The model divides the airways into five regions: the front of the nose; the back of the nose with the pharynx and the larynx; the bronchi; the bronchioles; and the alveolar-interstitial region, where gas is exchanged. The model treats deposition as a series of filters, as Bair explained about 1991 in an overview written for the model's task group. Each region removes a share of the particles that reach it, and the share depends on the particle's size, the airway's dimensions, the rate of flow and how long the air stays.
Three mechanisms do the filtering: impaction, sedimentation and diffusion. Brown and colleagues describe them in a 2013 modeling study of particle size definitions, and Brown, the first author, was at the US Environmental Protection Agency. When the air turns a corner, a large particle keeps going straight and strikes the wall, and this impaction dominates for large particles in the nose, the larynx and the large bronchi. Smaller particles settle under gravity where the air moves slowly, and the smallest are carried to the walls by the random motion that Chapter 4 called diffusion.
All three mechanisms depend on how a particle moves in air, so the size that counts in the airways is its aerodynamic diameter, the diameter of a sphere with the density of water that settles at the same speed as the particle. Two cut points on that scale come from conventions adopted from CEN, the European Committee for Standardization. Under those conventions, half of particles of 10 µm pass into the chest as part of the thoracic fraction, and half of particles of 4 µm reach the gas-exchange region as part of the respirable fraction. Breathing through the nose filters far more than breathing through the mouth.
- The convention. The thoracic convention passes half of the particles with an aerodynamic diameter of 10 µm: 50 of every 100 breathed in.
- Through the nose. In the model Brown and colleagues ran for typical breathing through the nose, fewer than 20 % of 10 µm particles penetrate into the chest: fewer than 20 of the 100.
- Through the mouth. With breathing through the mouth the share can approach 40 %, about 40 of the 100, or about twice the share through the nose.
- The result. The convention's 50 is more than two and a half times the modeled share through the nose. The cut point names a size, and how many particles of that size reach the chest depends on how the person breathes.
- Assumptions. The shares are model outputs, not measurements in people, and "fewer than 20 %" is taken at its limit. The model describes ordinary breathing, not air delivered under pressure through a mask, as the running device delivers it.
Particles smaller than the cut points reach deeper. In a 2025 paper, Mintis and colleagues reproduced the ICRP model in software built by their own company. In their reproduction, particles larger than 10 µm were mostly filtered by the nose and upper airways, while up to half of particles of 0.1 µm deposited in the alveolar region. What a deposited particle then does depends on what it is made of and how much of it arrives. The breathing-gas standard, ISO 18562, sets its particle limits by size rather than by chemistry, and Chapter 19 gives them.
The 10 µm of the macrophage and the 10 µm of the thoracic fraction measure two things: a particle's width, and how it settles in air. They agree only for a sphere with the density of water. A fragment denser than water settles faster than a water drop of its width, so its aerodynamic diameter is larger than its width, and a lighter fragment's is smaller. One scale gives the other only when the particle's density and shape are known.
The test that reads the response
ISO 10993-6 is the part of the biological evaluation series that sets implantation test methods for the preclinical assessment of devices that breach skin or mucosal tissue. It reads local effects only and does not cover harm to the whole body, cancer, birth defects or genetic damage. Its fourth edition was published in April 2026. The same 1998 article by Wallin and Upman described how the test ran under the edition of the 1990s. The material was usually implanted in the muscle along a rabbit's lower spine, with tissue under the skin or bone as alternatives.
In that muscle, each animal carried four test implants and four implants of a negative control, a material expected to provoke little response, so the test material's response is judged beside it. The local response was assessed after 1, 4 and 12 weeks, or at 12, 26, 52 and 78 weeks in long-term studies, with three or more animals at each time. Set against the periods of Sheikh and colleagues' review, the first time comes when acute inflammation has usually resolved, and the later ones reach the macrophages, the giant cells and the capsule.
The 2026 edition gives examples of scoring systems for local effects in Annex F, which was Annex E in 2016, and adds annexes on the microscopic evaluation of tissue responses and on devices that contact peripheral nerves. ANSI, which sells the standard, describes these changes in its summary of the edition. In the EU, a Commission decision published on 10 November 2026 cites the 2026 edition as a harmonized standard under the medical device regulation (Chapter 27). As of November 2026, FDA recognizes the 2016 edition and not the 2026 one (Chapter 26).
The running device has no implant
None of the running device's parts breaches skin or mucosa, so none of them falls within the scope of the implantation test. Its foam is the part whose pieces can still reach the body. A foam that breaks down sheds particles into the air path (Chapter 5). For the recalled machines, FDA warned that pieces of foam could be breathed in or swallowed (opening). The running device's nasal mask delivers its air to the nose. As far as the model of ordinary breathing applies, a user who breathes through the nose has the nose's filtering between the foam and the chest, and one who breathes through the mouth has less of that filtering. Chapter 19 sets out how particles in breathing gas are tested, and Chapter 20 what was found for the recalled foam.
For each part that can shed particles, into tissue or into a gas pathway, record the sizes it can shed and whether they were measured as width or as aerodynamic diameter. Place each size on the scale its route uses: the macrophage's 10 and 100 µm in tissue, the cut points of 10 and 4 µm in the airways. For an implant, check that the test's later times reach the capsule and the giant cells, not only the inflammation. For the running device, assess the foam's particles as part of the gas pathway (Chapter 19), and record why no implantation test applies (Chapter 22).
- Implantation injures tissue, and the acute inflammation that follows usually resolves within a week, while the foreign-body reaction to the material itself can last as long as the material stays.
- Macrophages take in particles under 10 µm, fuse into foreign-body giant cells on materials of 10 to 100 µm and leave larger ones to be broken down outside cells, and a fibrous capsule may wall off what stays.
- The airways sort inhaled particles by aerodynamic diameter through a series of filters; the conventions pass half of 10 µm particles into the chest, while a model of nasal breathing passes fewer than 20 %.
- ISO 10993-6:2026 tests the local response to an implanted material and gives examples of scoring systems in Annex F; the EU cites it as harmonized, and FDA recognizes the 2016 edition.
Let's Talk