Non-invasive measurement of glucose and lactate in thermoregulatory sweat in neonates and adults
Data files
Jun 29, 2026 version files 236.48 MB
-
Figure2.csv
44.11 MB
-
Figure3.csv
2.05 KB
-
Figure4.csv
3.50 KB
-
Figure5.csv
3.15 KB
-
Figure6.csv
1.06 KB
-
Figure6A.txt
186.31 MB
-
Figure7.csv
2.71 MB
-
Figure8.csv
3.33 MB
-
FigureSI10.csv
660 B
-
FigureSI2.csv
747 B
-
FigureSI4.csv
3.47 KB
-
FigureSI5.csv
3.47 KB
-
FigureSI6.csv
2.94 KB
-
FigureSI7.csv
285 B
-
FigureSI8.csv
147 B
-
FigureSI9.csv
442 B
-
README.md
4.19 KB
Abstract
Premature neonates are particularly vulnerable and therefore require rigorous monitoring of analytes such as glucose and lactate. Traditionally this is achieved through invasive, intermittent blood sampling, which is painful and may cause potential complications. To address this issue, we have developed a non-invasive perfusable skin sampling patch designed to collect analytes from low-volume thermoregulatory sweat without the need for external sweat induction. This can be combined with microfluidic biosensor analysis technologies to detect skin glucose and lactate levels. We were able to demonstrate the potential of this technology for non-invasive measurement of glucose and lactate in premature neonates. Our results showed that skin glucose levels correlated well with time-matched blood glucose measurements. Lactate skin measurements did not correlate with blood measurements, indicating a more complex relationship. Additionally, we successfully measured skin glucose levels in healthy adults, which correlated with blood glucose levels for each individual. In contrast with the neonate measurements, the relationship between blood and skin glucose varies between individual adults. We successfully carried out proof-of-concept experiments to detect glucose in real-time in both neonates and in healthy adults by connecting the sampling patch to a continuous analysis system. These results provide the groundwork for subsequent large-scale clinical trials intended to validate the system reliability and clinical applicability within healthcare contexts, particularly neonatal intensive care units.
Dataset DOI: 10.5061/dryad.rfj6q57pr
Description of the data and file structure
This dataset includes data required to replicate analysis in Science Advances manuscript (accepted).
Files and variables
- Figure2.csv: A) raw data for glucose, lactate null sensor response to a 3-point calibration, mean values for each calibration step for each sensor. B) raw data for glucose sensor current and null sensor current for 2 repeats of one skin sample. C) limits of detection for all sensors tested (n = 8) D) levels of glucose and lactate sampled from thermoregulatory sweat in neonates (n = 224 for glucose and 218 for lactate). E) blood values for all neonates monitored (n = 17).
- Figure3.csv: A and B) measured glucose and lactate levels in skin samples over time for two babies, along with feeding times and a one-off blood measurement. C and D) paired blood and skin values across all babies for glucose and lactate
- Figure4.csv: paired blood and skin values for 7 adults and corresponding times of samples
- Figure5.csv: A) all paired blood and skin values for 7 adults (from fig 4). B) multiple days for adult 1 (day 0, day 12, 28, and 61). C and D) raw paired values for day 12 and 28 alongside predicted blood values using day 0 relationship to predict blood values from sweat values
- Figure6A.txt: continuous glucose sensor current at 5 different calibration flow rates (2, 1, 0.5, 0.3, and 0.2 µl/min).csv
- Figure6.csv: B) mean measured glucose concentration at 5 different perfusion flow rates when the sampling patch is placed in a beaker of 20 µM glucose. C) calculated molecules of glucose in collected samples for both in vitro and on-skin experiment. D) mean measured glucose concentraton at 5 different perfusion flow rates when the sampling patch is placed on a healthy adult's forearm. E) concentration and collection time for each flow rate. F) concentration and delay time for each flow rate
- Figure7.csv: A) continuous glucose concentration (raw and smoothed with a Savitsky-Golay 2001-pt filter) on the forearm of a healthy adult and time-matched blood measurements. B) paired skin and blood glucose measurements for healthy adult.
- Figure8.csv: A) continuous glucose concentration (raw and smoothed with a Savitsky-Golay 2001-pt filter) on a neonate and a one-off blood measurement. B) paired skin and blood glucose values for all neonates (from fig 3 at 0.5 ul/min perfusion flow rate) and paired skin and blood glucose value for neonate in part C at 0.3 µl/min perfusion flow rate.
- FigureSI2.csv: measured glucose and lactate levels in skin samples over time for two babies, along with feeding times and a one-off blood measurement.
- FigureSI4.csv glucose levels measured from skin samples of seven adults in the same scale, correlated with blood glucose levels measurement at the same time.
- FigureSI5.csv: all paired blood and skin values for 7 adults (from fig 4).
- FigureSI6.csv: box plots comparing the blood glucose concentration and skin sample glucose concentration between neonates and adults monitored.
- FigureSI7.csv: raw paired values for day 61 alongside predicted blood values using day 0 relationship to predict blood values from sweat values.
- FigureSI8.csv: mean measured glucose concentration at 5 different perfusion flow rates when the sampling patch is placed on the arm of a healthy adult.
- FigureSI9.csv: mean measured flow rate of syringe pump in 4 different conditions at 3 different flow rates: 0.5, 0.3 and 0.2 µl/min. Conditions are 1) sampling patch with nothing connected to outlet in a beaker of glucose solution, 2) sampling patch with nothing connected to outlet on the forearm of a healthy adult, 3) sampling patch with microfluidic chip at outlet in a beaker of glucose solution and 4) sampling patch with microfluidic chip at outlet on the forearm of a healthy adult.
- FigureSI10.csv: averaging of online adult skin glucose monitoring into 20-minute time windows to mimic offline sample analysis (from Figure 7)
Code/software
Excel or NotePad