Non-Invasive Optical CGM in Pediatrics

Non-Invasive Optical Continuous [Glucose Monitoring](/) in Pediatrics

As of October 2026, no optical or other non-invasive glucose monitor is FDA-authorized for children. The FDA says it has not authorized, cleared or approved any smartwatch or smart ring that measures or estimates blood glucose on its own [1]. The CGMs children wear, such as the Dexcom G7 and the FreeStyle Libre 3 Plus sensor, are cleared from age 2 and read glucose in the fluid just beneath the skin through an inserted sensor [2][3]. One FDA-authorized CGM does read glucose optically: the implanted Eversense 365, which measures a fluorescent signal under the skin and is labeled for adults 18 and older [4]. For children, what exists is a short list of small studies, most of them in newborns.

Why a needle-free monitor matters most for the youngest

Newborns in hospital care have glucose checked with intermittent heel-prick or arterial blood samples [5]. A 2026 paper from Imperial College London points out that these draws hurt, that heel-prick pain has been linked to poorer neurological development, and that repeated sampling can mean transfusions to replace the blood taken [5]. A Japanese neonatal team made a similar case the same year, calling existing CGMs invasive and expensive [6].

The one non-invasive monitor the FDA cleared for children

It wasn't optical. The FDA approved the GlucoWatch Automatic Glucose Biographer for adults in March 2001, under its product code for non-invasive glucose instruments [7]. On August 26, 2002, the agency expanded the GlucoWatch G2 Biographer's label to children and teens aged 7 to 17. It was a prescription device for tracking trends, meant to supplement standard fingerstick testing rather than replace it [8]. The watch used reverse iontophoresis: a low electric current drew interstitial fluid out through the skin to a sensor patch on the surface [9].

Pediatric researchers, including the Diabetes Research in Children Network (DirecNet), studied it closely [10][11][12].

• Accuracy. In a 24-hour clinical research center study of 89 children and teens aged 3.5 to 17.7, the median gap between watch and lab glucose was 16%, and 60% of readings met the ISO criteria proposed at the time [10].
• Age. In a home study of 46 children, the watch tracked meter readings less closely in children under 7 than in older children, with correlations of 0.58 and 0.74 [11].
• Daily life. In a 6-month randomized trial of 200 children aged 7 to 17, adding the watch did not lower HbA1c or severe hypoglycemia. Use fell from 2.1 times a week in the first month to 1.5 in the sixth [12].

Here's what surprised us. Among the reasons given in that trial for wearing the watch less, skin irritation came first at 76%, ahead of skipped readings (56%), excessive alarms (47%) and inaccurate readings (33%) [12]. In February 2004 the FDA approved labeling that described skin pre-treatment products and a way to reduce irritation [13]. The approval was withdrawn on April 23, 2010 [7], and a 2022 review says the GlucoWatch products were discontinued after users complained of a burning sensation [9]. Needle-free did not mean easy on children's skin.

Optical studies in children so far

The clearest pediatric optical study we found comes from Kagawa University in Japan and appeared in Pediatric Research in June 2026 [6]. Using standard neonatal pulse-oximetry probes, the team recorded photoplethysmography traces (a light-based pulse signal) from 38 full-term babies on their second day of life. They estimated glucose from the timing gap between the oxygenated and deoxygenated hemoglobin waveforms and compared it with plasma glucose from a blood analyzer [6].

Only 30 of the 38 recordings met the signal-quality criteria. The optical index explained about half the variation in lab glucose (R² 0.51) [6]. Both the slope and the intercept of that relationship differed from published adult values, and the authors conclude that age-specific calibration is required before clinical use [6]. Hamamatsu Photonics supplied the prototype equipment [6].

Adult optical work hasn't reached children either. A 2025 clinical test of DiaMonTech's mid-infrared method, in which a laser makes glucose in the skin release a little heat that a detector picks up, enrolled 36 people aged 20 to 75 and used a table-top prototype [14].

Other needle-free approaches tested in children

Sweat. The Imperial College patch collects the small amount of sweat that skin makes for temperature control, without stimulating extra sweat. Across 17 samples from 15 premature babies, born between 23 and 38 weeks of gestation, skin glucose tracked blood glucose closely (R² 0.92); lactate did not [5].

Breath. When we searched ClinicalTrials.gov in October 2026 for non-invasive glucose studies open to people under 18, the two recruiting ones we found both test breath sensors at Indiana University: NCT06893341 for ages 12 to 19, with primary completion planned for December 2026 [15], and NCT07696273 for ages 7 to 25 [16]. Both look for compounds in exhaled breath that signal high or low glucose events [15][16].

Why children's skin changes the optical problem

Glucose makes up only 0.07 to 0.1% of blood plasma, against about 90% water, and water absorbs much of the incoming light, especially at mid-infrared wavelengths [17]. Each skin layer also has its own optical properties, which vary with layer thickness and composition [17].

Children's skin shifts those variables. Infant skin has a thinner epidermis and stratum corneum (the outermost layer) than adult skin, at least until the second year of life, and the infant stratum corneum holds more water [18]. The Kagawa result shows what that can mean in practice: for that method, the newborn glucose relationship didn't match the adult one [6]. We found no study measuring the same effect in toddlers or school-age children.

Skin tone is a separate variable. The 2022 review says absorption by melanin and other skin pigments will affect optical signals, mainly at visible and near-infrared wavelengths under about 1 micron, and that pigmentation's impact must be tested experimentally for each system [9]. The adult mid-infrared study made no selection for skin pigmentation, and every participant was white [14].

Motion during optical coherence tomography (OCT, a light-based imaging method) changes the signal and the accuracy of the glucose estimate [17]. The Kagawa team analyzed only stable 15- to 60-second stretches from each 5-minute recording [6]. Raman spectroscopy, which reads tiny shifts in scattered light, uses a high-intensity light source, and only about one in every million scattered photons carries its signal [17]. The 2022 review warns that a powerful laser poses a risk of pronounced skin damage [9]. For a device meant for thin infant skin, that safety question comes first.

Any method that reads tissue fluid also inherits its delay: glucose there trails blood glucose by roughly 5 to 15 minutes [17].

What this means for families now

The FDA tells people not to buy or use smartwatches or smart rings that claim to measure blood glucose [1]. Which monitor fits a particular child is a decision for the family and the child's care team. Our article on pediatric vs. adult calibration algorithms explains how the cleared sensors were tested in children, and non-invasive glucose monitoring technology covers adult devices.

The open question is whether any optical method can hold its calibration as a child grows, from newborn skin to teenage skin, and across skin tones. As of October 2026, the only pediatric optical glucose study we found compared readings with lab values in newborns, on one method, at one hospital [6].