Biofouling Mitigation Strategies for Dermal Sensors

Biofouling Mitigation Strategies for Dermal Sensors

Biofouling is the body's reaction to a sensor it treats as foreign, and for a continuous glucose monitor (CGM) it is one of the main reasons a signal drifts or a sensor fails early [1]. The defenses that have reached marketed sensors are fairly plain: a slow-release steroid around one long-term implant, sensing chemistry built to tolerate the reaction, and wear periods counted in days. The more ambitious coatings are still animal research. Below is what we learned about each, and where the evidence stops.

What the body does to a sensor

Skin-worn CGM sensors are percutaneous, meaning they cross the skin: a thin filament passes through the epidermis and dermis into the tissue below, where it measures glucose in interstitial fluid, the fluid that bathes the cells [1]. Insertion makes a small wound. The foreign body response (FBR) starts with the ordinary wound-healing cascade [1].

Proteins get there first. Small protein fragments stick to the sensor surface, and Soto and colleagues note that this adsorption has been linked to a drop in sensor sensitivity of about 50% after implantation [2]. Inflammatory cells follow. Macrophages, and the foreign body giant cells they fuse into, gather at the surface and consume glucose and oxygen, so the sensor sees less of both than the tissue around it holds [2][3].

The last stage is a capsule: a dense collagen layer that persists for the device's lifetime, lowering sensitivity and adding lag [1]. Timing matters too. Anderson's review found that surface chemistry and texture shape the reaction most in the first two to four weeks after implantation, though the reaction never fully goes away [3].

Coatings that keep proteins off

The oldest idea is to make the surface hard for proteins to grab. Polyethylene glycol (PEG) surface treatment is a widely studied way to resist biofouling, and in rat studies PEG-containing hydrogels drew less fibrous encapsulation and fewer inflammatory cells than control materials [1].

Zwitterionic polymers are the newer candidate. A zwitterion carries both a positive and a negative charge; zwitterionic materials such as carboxybetaine and sulfobetaine hold water tightly and resist protein adsorption [4]. In a 2013 study, zwitterionic hydrogels implanted under the skin of mice resisted capsule formation for at least 3 months [5].

In 2018, an MIT-led team coated a commercial Medtronic CGM sensor with a zwitterionic polymer and tested it in mice and in non-human primates [6]. Without recalibration, uncoated sensors in mice read an average of 73% (plus or minus 36%) away from measured blood glucose, while coated sensors were off by 17% (plus or minus 11%), with six sensors per group [6]. The coated sensors also showed fewer signs of inflammation [6]. One detail we didn't expect: the team built its zwitterionic hydrogels on PEG crosslinkers, so the two coating families are less separate than they sound [6]. Both studies were in animals, not people; our page on zwitterionic polymer coatings covers the chemistry.

Texture is the third lever. When precisely made porous materials were implanted under the skin, pores of about 35 µm produced the thinnest, least dense collagen capsules, while larger or smaller pores did worse [1].

Why crossing the skin makes it harder

Here's what surprised us: a coating that helps a buried implant can do nothing for a sensor that crosses the skin. Porous foam sleeves thinned the capsule around sensors implanted fully under the skin, yet gave no benefit on sensors that passed through it, which the researchers blamed on micromotion, the constant small movement of a sensor anchored at the skin surface [1].

Stiffness adds to it. The dermis is much stiffer than the subcutaneous tissue beneath it (about 56 to 260 kPa against 0.12 to 23 kPa), so no one material matches both layers, and the mismatch creates shear forces that worsen the FBR [1].

Some sensors are dermal in the literal sense. The Biolinq Shine, authorized by the FDA through the De Novo pathway, uses an array of seven microelectrodes that sit in the papillary dermis, the uppermost layer of the dermis, and is worn for up to 5 days [7]. In the study the FDA reviewed, 84.7% of devices lasted the full 5 days [7].

Releasing a drug at the sensor

Dexamethasone is a glucocorticoid, a steroid that reduces inflammation. Taken systemically it suppresses immune defenses and raises the risk of infection, which is why researchers turned to releasing small amounts right at the sensor [1].

The Eversense 365 is the marketed example, a fluorescence-based sensor that a clinician implants under the skin for up to 1 year [8]. A silicone collar on the sensor releases 1.75 mg of dexamethasone acetate locally "to minimize local foreign body reaction," in the words of the FDA decision summary [8]. The same document says the system is contraindicated in people for whom dexamethasone or dexamethasone acetate may be contraindicated; whether that applies to you is a question for your care team [8].

Nitric oxide (NO) is the other molecule under study. Blood vessel lining cells release small amounts of NO to keep platelets from activating, and NO-releasing coatings have reduced the FBR in animals [2]. In a 10-day swine study, needle-type sensors with NO-releasing polyurethane coatings were more accurate than controls on days 1 and 3, and sensors with slower, longer release responded faster to glucose changes, so the benefit depended on release duration [2]. The same authors flagged a limit for any drug-release coating: controlling the release itself remains a major hurdle [2].

Designing the sensor to tolerate the reaction

Abbott took a different route. In a 2017 paper, two Abbott engineers wrote that a sensor's design and membrane chemistry must either minimize the FBR or keep it from interfering with the signal [9]. Their "Wired Enzyme" chemistry crosslinks the enzyme glucose oxidase into a polymer with an osmium mediator, a molecule that shuttles electrons to the electrode. The working electrode runs at only 40 mV, which limits interference from other substances in the tissue [9].

Sensitivity was lower on the first day, which they attributed to insertion trauma, then held steady from day 2 through day 14, which they read as negligible interference from the FBR [9]. That study used the original 14-day FreeStyle Libre sensor. Abbott's current US sensors, the FreeStyle Libre 2 Plus and 3 Plus, are cleared for up to 15 days [10].

Dexcom G7 sensors also use glucose oxidase, but they measure the hydrogen peroxide the enzyme produces; the current from that peroxide is proportional to the interstitial glucose level [11]. The Dexcom G7 is labeled for up to 10 days plus a 12-hour grace period [11], and the Dexcom G7 15 Day for up to 15 days plus 12 hours in adults 18 and older [12]. Short wear works as a mitigation strategy in its own right. A sensor worn 10 to 15 days never gets past the first two to four weeks that Anderson's review identifies as the period when surface properties shape the reaction most [3].

What is still unknown

In 2013, Nichols and colleagues concluded that physical and chemical changes at the sensor surface had not yet produced drastic improvements in glucose sensor function [1]. The animal results since then are encouraging, but the zwitterionic evidence we found comes from mice and non-human primates [5][6], and the NO study ran 10 days in swine [2]. For a year-long implant, the answer in FDA-reviewed labeling as of October 2026 is still a steroid collar [8]. Whether a coating alone can do that job in people is still an open question.