Continuous glucose monitoring has transformed diabetes management, but the interstitial-fluid sensors that dominate the market remain expensive, invasive and poorly suited to young patients, a problem that is especially acute in low- and middle-income countries. A new systematic review published in BMC Biomedical Engineering argues that an unlikely alternative, glucose measured in sweat, has matured to the point where cheap, scalable and enzyme-free devices are approaching the accuracy thresholds that clinicians demand.
The review, conducted according to the PRISMA 2020 reporting guidelines, searched PubMed, Scopus, Web of Science and IEEE Xplore for studies published between 1 January 2016 and 20 November 2025. From 1,200 records screened, 150 reports were sought for retrieval and 140 assessed for eligibility; after exclusions for studies that were not sweat-based, not glucose-related, purely narrative or published outside the window, 25 peer-reviewed articles entered the qualitative synthesis. Twelve of these, the subset reporting mean absolute relative difference values, formed the quantitative analysis. The authors note candidly that the review protocol was not prospectively registered in PROSPERO, a limitation they disclose in the methods section.
The central trend the authors document is a decisive shift in sensing chemistry. Traditional glucose sensors rely on enzymes, biological catalysts that are exquisitely selective but fragile: they degrade over time, have short shelf lives and drift when exposed to temperature and pH changes. Non-enzymatic sensors instead exploit direct electrochemical oxidation of glucose using engineered nanomaterials, including MXenes, noble metal nanoparticles such as platinum and copper, and various metal oxides and sulfides. In the technical studies analysed, 54.5 percent of sensing modalities were non-enzymatic, a distribution the review interprets as a deliberate move toward thermostable, low-cost nanomaterials that sidestep the stability and expense problems of enzyme-based designs, goals that align closely with the constraints of resource-limited health systems.
To compare devices on more than raw accuracy alone, the authors developed a novel metric, the LMIC Scalability Score, an ordinal rating from 1 (low) to 3 (high). It is a weighted composite of three factors: material cost, favouring cheap and locally available inputs such as paper and carbon ink over noble metals and custom polymers; fabrication complexity, favouring high-throughput, low-infrastructure processes such as screen and inkjet printing over MEMS, 3D printing or clean-room fabrication; and compatibility with the purchasing power of low-resource settings. Two independent reviewers scored each study against pre-defined thresholds, resolving disagreements by consensus. The authors stress that the score is a relative engineering model of translational feasibility, not a precise economic measure of deployment cost.
Plotting reported accuracy against this score yields the review’s most striking finding, a measurable cost-accuracy trade-off. The best accuracy came from a complicated liquid-metal and 3D-printing fabrication route, which achieved a mean absolute relative difference of 6.5 percent but earned only the lowest scalability score. By contrast, devices built with scalable methods, an MXene paper chip and a screen-printed sensor, recorded high scalability scores with mean absolute relative differences of 9.2 and 10.1 percent respectively. The authors argue that the field is nearing a performance plateau in which the marginal accuracy gains of complex manufacturing are outweighed by the cost reductions and throughput advantages of scalable processes, bringing simple devices within reach of the clinically accepted benchmark of a mean absolute relative difference below 10 percent.
Accuracy alone, however, does not make a sweat sensor clinically viable. The deepest technical obstacle is physiological: glucose concentrations in sweat lag behind those in blood by roughly 5 to 30 minutes, a delay governed by transdermal transport kinetics and local sweat rate. The review formalises this with a first-order transfer function in which a time constant, dependent on individual physiology and sweat rate, relates sweat glucose to blood glucose. For real-time monitoring, this lag must be corrected mathematically, and the authors call for future devices to incorporate real-time sweat-rate measurement so the time constant can be adjusted dynamically. They also highlight work coupling sensors to microcontrollers for real-time data processing as a critical step toward lag compensation, even where no accuracy figure was claimed.
The pediatric dimension complicates matters further. Children generally produce less sweat, have thinner skin layers and have underdeveloped eccrine glands, and their changing hydration status, metabolic rate and thermoregulation alter sweat composition and analyte transport with age. Sweat glucose does correlate with blood glucose, provided physiological lag times and perspiration-rate changes are accounted for, but low sweat production and contamination risks add difficulty. The review warns that sensor metrics validated in adults cannot simply be extrapolated to children, and that the current literature contains a significant lack of pediatric data. Microfluidic sweat-collection mechanisms, adhesive performance and overall device architecture, the authors argue, must be designed specifically for this population.
Safety questions also temper the enthusiasm. Although nanomaterial-based sensors are becoming increasingly sensitive and stable, the review finds that very few of the included studies examined dermal toxicity, long-term skin compatibility or inflammatory responses to nanomaterial sensing layers, focusing instead on electrochemical performance and device integration. Children’s thinner epidermal barriers and developing integumentary systems may make them more susceptible to irritation or long-term nanoparticle exposure during continuous wear. The authors recommend that next-generation research incorporate standardised biocompatibility testing, including cytotoxicity, irritation and extended-wear assessments, before these materials are trusted against pediatric skin for prolonged contact.
The quality assessment reinforces a note of caution. Using QUADAS-2 for clinical studies and a tailored engineering appraisal for technical ones, the review found that while most studies carried low risk of bias in experimental design, few earned a high engineering-quality score, largely because of poor reporting of long-term stability or selectivity against common interferents such as ascorbic acid, uric acid and lactate. Stability was judged on whether a sensor maintained performance for at least seven days, and scalability on whether scale-up capability was substantively discussed or demonstrated. The gap between promising laboratory numbers and robustly characterised devices remains, on this evidence, substantial.
The review closes with a translational roadmap: dynamic time-lag compensation combining real-time sweat-rate measurement with mathematical modelling; pediatric-centred validation trials using age-stratified cohorts to confirm accuracy, usability and long-term comfort; standardised, quantitative scalability metrics incorporating production complexity, infrastructure requirements and supply-chain stability alongside material costs; and attention to ethical, safety and acceptability considerations across culturally diverse regions. On the authors’ reading, sweat-based continuous glucose monitoring has moved from theoretical concept to nearly imminent translation opportunity, with scalable, non-enzymatic platforms emerging as practical, low-cost candidates for the children with diabetes whom conventional technology has left behind.
Subject of Research: Non-enzymatic sweat-based glucose sensors for pediatric continuous glucose monitoring
Article Title: Non-enzymatic sweat sensors for pediatric continuous glucose monitoring: a systematic review of engineering readiness and the cost-accuracy trade-off
Article References: Bolaji, O. T., Ephraim, E. D., Adoghe, O. B., Daramola, D., & Kibuebu, E. F. (2026). Non-enzymatic sweat sensors for pediatric continuous glucose monitoring: a systematic review of engineering readiness and the cost-accuracy trade-off. BMC Biomedical Engineering, 8(1), Article 8. https://doi.org/10.1186/s42490-026-00109-8
Image Credits: AI Generated
DOI: 10.1186/s42490-026-00109-8
Keywords: continuous glucose monitoring, type 1 diabetes, sweat biosensors, non-enzymatic sensors, MXene, pediatric health, low- and middle-income countries, systematic review, mean absolute relative difference, screen printing, nanomaterials, medical devices
News Source: Gregory Coleman. (October 29, 2026). Sweat Sensors Move Closer to Needle-Free Glucose Monitoring for Children. Scienmag.
