Real‐World Accuracy of a Continuous Glucose Monitoring System after Radiologic Exposure

Abstract

Background. The increasing use of continuous glucose monitor (CGM) necessitates a review of variables that impact accuracy and interrupt use. Manufacturer recommendations include removing CGMs before diagnostic imaging, such as X‐ray and computed tomography (CT). Early removal and replacement of CGM components present financial, clinical, and psychosocial burdens to the wearer and interrupt optimal management of diabetes for pediatric patients who receive a total pancreatectomy with islet autotransplantation (TPIAT). The study’s aim was to evaluate the effect of scatter dose exposure during X‐ray or CT if the CGM remained intact but outside the field of view (FoV). Materials and Methods. Participants were followed through the first 3 months after TPIAT surgery, managed diabetes with an insulin pump and CGM, and were routinely exposed to diagnostic imaging. Participants’ CGMs were unshielded by a protective apron during any X‐ray or CT procedures for the duration of the study period, and the transmitter was collected after expiration or removal. Glucometer data was collected from hospital records and home glucometer downloads. Mixed models were used to analyze absolute differences between matched CGM and glucometer values, and Clarke error grid analyses (EGA) were performed. Scatter dose exposure was derived using anthropomorphic phantoms and calculated retrospectively. Results. A total of 14 patients (median 12.2 years, 64% female) received a median of five diagnostic imaging procedures with a median cumulative scatter dose of 559 µGy. The absolute difference between the CGM and glucometer values was not significantly associated with the cumulative scatter dose (p = 0.17) or time from TPIAT (p = 0.24) when analyzed in a mixed model. Regardless of scatter dose exposure, time from TPIAT, or glucometer, ≥98% of glucose values fell within zones A and B on EGA. Conclusion. Scatter dose exposure from diagnostic imaging did not affect the clinical accuracy of CGM values for the duration of transmitter use. Leaving CGM components in place when not in the FoV during diagnostic imaging successfully mitigated interruptions to use and undue burden or cost to participants.