Abstract
Background: Published research indicated that fluctuations in blood glucose concentration (BGC) change the biochemical balance of the retina and may alter the sensitivity of retinal cells. Low-contrast flicker was used in this research to measure the changes in sensitivity of the retina accompanying shifts in BGC. Methods: In five subjects without diabetes and four subjects with diabetes, simultaneous measurements of flicker thresholds, using a two-alternative, forced choice psychometric test, and BGC, were made every 15 to 30 minutes for several hours, while BGC was allowed to fluctuate spontaneously. Results: Flicker thresholds were found to rise and fall, generally tracking BGC in each subject. First-day flicker threshold measurements were used to "calibrate" the method for each subject, and subsequent days' measurements were used to infer blood glucose. The resulting inferred blood glucose values correlated with BGC measured from blood samples in the range of 70 mg/dL to 500 mg/dL with a maximum error of ± 33% in 97% of the measurements. Conclusions: The feasibility of using flicker sensitivity for monitoring blood glucose levels noninvasively is discussed. Improvements in the testing method can potentially reach ± 18% error margin with 1-minute measurements, according to computer simulations. Devices using a visual test would be inexpensive to manufacture and would need no disposable supplies. Planned research will assess the long-term stability of flicker threshold measurements and will determine the applicability of the technique in diabetic retinopathy, in hypoglycemia and in the presence of fatigue and inattention.
Get full access to this article
View all access options for this article.