Previous research into alternative sites to place a flash glucose monitoring sensor have suggested that the upper thigh may be appropriate whilst, unlike with other continuous glucose monitoring sensors, the abdomen may perform poorly. The present study evaluated two other sites: the lower back (at the level of L4/5, halfway between the iliac crest and the vertebrae) and the upper chest (at the third intercostal space in the mid-clavicular line).
The accuracy of the FreeStyle Libre sensors at these locations were compared with the standard upper arm location over 14 days of usual wear, as well as, on day 3, during experimental conditions when blood glucose levels are expected to change rapidly: during a meal, during exercise and with lowering skin temperature. Twenty adults with type 1 diabetes participated in the study and wore the three sensors simultaneously.
The results of the free-living part of the study showed that the sensors were acceptable in terms of accuracy, with 98% of readings within zones A and B of the Clarke error grid compared with the upper arm location. However, only 65% and 74% of readings in the back and chest locations, respectively, were in the clinically accurate zone A compared with the arm location. In general, the back and chest recorded higher glucose levels in euglycaemic and hyperglycaemic conditions.
During the experimental conditions, compared with venous glucose measurements, the mean absolute relative deviation (MARD) was 8.2%, 11.0% and 11.3% in the arm, back and chest locations, respectively, after a standard meal. During exercise, the MARD was 13.1%, 18.6% and 15.0%, respectively, while during skin cooling it was 10.5%, 14.0% and 13.8%. During exercise, all three sites underestimated glucose compared with venous measurements and only approached acceptable accuracy 20 minutes after exercise finished; this was presumably due to the lag between venous and interstitial glucose levels.
In terms of acceptability, the chest site was preferred by 45% of participants, followed by the arm (40%) and then the back (15%). Half of participants rated the back as the least preferred site. The sensors commonly failed or fell off the back (five sensors in the first week of the study), despite the use of extra medical adhesive tape to secure them, contributing to their unpopularity.
The authors conclude that the chest and the lower back are clinically acceptable sites for flash glucose sensor placement if users wish to place them here, with the chest being more accurate and tolerable than the back site. Users should be informed that the sensors may read higher in the hyperglycaemic range and lower in the hypoglycaemic range, requiring fingerprick testing for confirmation at these levels.
Click here to read the study in full.
Amid accumulating evidence of benefits with new diabetes drugs, Vinod Patel asks whether the costs can be justified.
9 Nov 2023