Non-Enzymatic Glucose Sensing Technologies

A non-enzymatic glucose sensor measures glucose without an enzyme doing the chemistry. There are two kinds. Electrochemical versions oxidize glucose directly on a metal electrode and read the resulting current [1]. Optical versions use a synthetic molecule that binds glucose and glows more brightly when it does [2]. Here's what surprised us: the one non-enzymatic CGM we found on the market belongs to the optical group. A 2024 review found that every current CGM sensor except the implantable Eversense relies on a glucose oxidase reaction [3], and the chemists whose 1990s work led to its chemistry wrote that Senseonics is the only company so far to market a diboronic acid CGM sensor [4].

Why anyone would skip the enzyme

Most people who wear a CGM wear an enzyme-based one; a 2024 perspective names Medtronic, Dexcom and Abbott as the main makers [4]. In the Dexcom G7, glucose oxidase oxidizes glucose from the fluid under the skin and produces hydrogen peroxide. That peroxide generates a current that rises with the glucose level [5]. Our article on CGM sensor technology mechanisms covers that design in detail.

Glucose oxidase earned its place by being cheap and selective [1]. It is also a protein, and proteins break down. A 2021 review calls long-term stability the main drawback of enzymatic sensors, "due to the inherent nature of enzymes" [1]. Remove the enzyme and, in theory, you remove the part that wears out.

Electrochemical sensing: oxidizing glucose on bare metal

The idea is older than any CGM. Walter Loeb first reported glucose oxidizing directly at a lead electrode in sulfuric acid, producing gluconic acid [1]. Modern versions apply a voltage to an electrode made of a catalytic metal such as platinum or gold, and the metal does the enzyme's job [1]. Reviewers sometimes call these devices the "fourth generation" of glucose sensors [1].

Surface area matters, because glucose has to touch the metal to react. Researchers spread platinum thinly over supports such as graphene and carbon nanotubes, which gives glucose more surface to reach and cuts the amount of costly platinum each sensor needs [1].

Three problems with bare metal

The 2021 review lists three main limits of oxidizing glucose on noble metals [1]:

• Slow reactions. Glucose oxidizes sluggishly on ordinary electrodes, which caps sensitivity.
• Poor selectivity. Several sugars oxidize in the same voltage range as glucose.
• Electrode poisoning. Ions, mainly chloride, reduce the metal's activity. Partly oxidized fragments of glucose, such as carbon monoxide, can block the surface too.

Strongly alkaline conditions keep chloride off the electrode and raise sensitivity [1]. That fix doesn't travel well into a person. Gold electrodes that work best in alkaline solution can't be used inside the body, the review notes [1].

Getting into whole blood

In 2012, a team at Seoul National University described what it called the first non-enzymatic electrochemical glucose monitoring in undiluted human whole blood and serum [6]. It used nanoporous platinum, a metal film full of tiny pores, behind a membrane tuned to let glucose through [6]. It kept its sensitivity for at least 7 hours of continuous operation and still worked after 30 days stored in blood and serum [6]. It even survived steam sterilization at 134 °C; unlike enzyme sensors, the authors wrote, it was barely affected by heat [6]. This was lab work, not a sensor worn by people.

Optical sensing: a synthetic molecule that grabs glucose

The optical route swaps the enzyme for a boronic acid, a molecule that forms reversible chemical bonds with glucose. Because the bond can form and break, the sensor can follow glucose up and down continuously [7].

The chemistry grew out of lab work in the 1990s. Czarnik's group found that a boronic acid compound built on anthracene lit up in the presence of glucose or fructose. It was the first abiotic (non-biological) fluorescent sensor for sugars in water, but it couldn't tell glucose from fructose [4]. Adding a second boronic acid, in work from Shinkai's laboratory, produced the first fluorescent sensor that preferred glucose over other simple sugars [4].

The human body exposed a new problem. The first sensors implanted in people lost their signal within a day, after the boronic acid was oxidized into compounds that no longer responded to glucose, most likely by the body's early inflammatory response [4]. Two fixes followed. A 3-nanometer platinum coating protected the boronic acids and stretched sensor life from hours to at least 6 months, and a silicone collar that slowly releases the anti-inflammatory drug dexamethasone acetate damped the body's response; the collar has stayed standard since [4]. The first fully successful clinical trial was published in 2014 [4]. It implanted 19 sensors in 12 people with type 1 diabetes for 28 days, and the mean absolute relative difference (MARD), the average percentage gap between sensor and reference readings, was 11.6% [8].

Eversense 365, the current version

As of October 2026, the current model is the Eversense 365. The FDA's 2024 review describes a fluorescent indicator grafted onto the sensor surface that binds glucose fully reversibly; binding raises the fluorescence, which the sensor's optical system measures [2]. A clinician implants the sensor under the skin. It is cleared for up to 1 year in adults 18 and older, and the sensor runs on radio-frequency power, sending its readings to an externally worn transmitter [2]. It succeeds the original 90-day Eversense, approved by the FDA in 2018, and the 180-day Eversense E3, approved in 2022 [4]. Our article on implantable fluorescence sensors covers the device itself.

Going enzyme-free didn't remove fingersticks. The Eversense 365 asks for 4 fingerstick calibrations after its 24-hour warm-up, then one a day through day 13, then one a week [2]. It has its own interference list too. The FDA summary says intravenous mannitol or sorbitol can falsely raise readings, and tetracycline-class antibiotics can falsely lower them [2]. Enzyme-based sensors have different weak spots: the 2024 review lists acetaminophen and vitamin C, usually above therapeutic doses, among the substances that can affect them [3]. How that bears on someone's choice of sensor, or on their medicines, is a question for their care team.

Stability against selectivity

The two approaches trade problems. Metal electrodes offer long-term stability but lower selectivity than enzyme sensors, because a bare catalyst struggles to oxidize glucose alone [1]. Boronic acids went the other way: chemists built glucose selectivity into the molecule, then had to win stability inside the body with a platinum coating and a drug-releasing collar [4].

What still stands in the way

For electrode-based designs, the 2021 review lists alkaline operating conditions in testing, costly materials, and the need for a protective surface layer to improve selectivity [1]. It adds that the sheer range of materials and combinations under study is itself a barrier to commercialization [1].

Biofouling, the buildup of proteins and cells on a sensor inside the body, is the other open problem. A 2025 review says it reduces sensitivity and selectivity, adds false signals, and shortens sensor life, and calls it a particular concern for non-enzymatic sensors meant for implantable CGMs [9]. We cover the body's side of it in biofouling and foreign body response mitigation. So far, direct oxidation on metal has strong lab results, like the 2012 whole-blood work [6], while the 2024 review we read lists no current CGM other than Eversense that works without glucose oxidase [3].