Abstract
Purpose :
Many Type I Diabetes (T1D) subjects using real-time continuous glucose monitoring system (RT-CGM) may have an illusion of tight glycemic control. We hypothesize that T1D subjects who use RT-CGM systems may have an equal chance of having diabetic retinopathy (DR) in relation to subjects who do not use RT-CGM. The purpose was to determine the characteristics of subjects with T1D with and without DR and their use of RT-CGM systems.
Methods :
This analysis was conducted at Friends for Life Children with Diabetes (CWD) conference held annually in Florida. Doctors and health care professionals conducted a health and eye screening, collecting data on T1D subjects with and without RT-CGM systems. Participant demographics, diet, BMI, blood pressure, blood sugar, HbA1C and other data were collected. Subjects underwent tests for visual acuity, non-contact tonometery, non-mydriatic retinal imaging (Canon, CR2 18.1MP, Tokyo, Japan), and Optical Coherence Tomography Angiography (Optovue, Fremont, California). Software was used to manage captured images (Image Spectrum, Canon, Irvine, California). Images were read by on site health professionals as well as off-site consultation via tele-medicine following HIPAA and DICOM guidelines. Means, SD, TTtests, variance inflation factors (VIF) to test for multicollinearity, and regression analysis were calculated (significance p<0.05).
Results :
373 subjects were screened. 16 subjects (4.26%) had confirmed DR and 357 had no DR (95.71%). Overall, out of 210 subjects with confirmed RT-CGM status, 162 subjects (77.14%) used RT-CGM and 48 subjects (22.86%) did not use RT-GCM. Almost all subjects with DR were using RT-GCM (93.75%).The mean blood sugar and HbA1C were higher in the group using RT-GCM but this did not reach statistical significance (Table 1). VIF < 5 were accepted as not violating multicollinearity. Logistic regression was performed to determine predictors for the development of DR. None of the co-factors were significantly associated with developing DR on univariate regression (Table 2).
Conclusions :
Although T1D patients may have RT-CGM, their glycemic control was suboptimal. The majority of subjects with DR used RT-GCM. This finding overstates the importance of lifestyle and diet counselling with RT-GCM use. Misjudging how to care for T1D could possibly give a sense of false protection with a RT-CGM "band aid effect".
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.