Characterization of the "Transfusion-dependent Prediabetes" Using Continuous Glucose Monitoring

The selection of patients took place randomly among adult patients suffering from transfusion-dependent thalassemia (TDT) belonging to the Pediatric Hematology and Oncology Unit of the AOU University of Campania "L. Vanvitelli". Patients had a history of at least 10 years of therapy with red blood cell transfusions (transfusion frequency less than one months) and iron chelation therapy. Cases tested negative for diabetes mellitus screening tests (fasting glycemia, OGTT).

Some healthy volunteers were also recruited among medical resident and doctors. Major eligibility criteria for the volunteers were: caucasian; age ≥18 years; body mass index <25.0 kg/m2; no chronic illness or medications that might affect glucose metabolism; fasting glycemia<100 mg/dl; HbA1c <5.7% (39 mmol/mol); no positive islet autoantibodies; HOMA-IR <2.5. Female participants pregnant at the time of study enrollment were not eligible.

All procedures performed in this study were approved by the Institutional Research Committee of the University of Campania "Luigi Vanvitelli". All participants have provided written consent.

Demographic and Clinical characteristics A brief schedule was specifically designed and completed by clinicians (diabetologists or hematologists) to record demographic and clinical data, including age, gender, height, weight, duration of illness, and other medical conditions (family history of diabetes, endocrine disorders, cigarette smoking habit). Finally, the HOMA-IR was calculated as a parameter for evaluating insulin resistance. Possible missing data was obtained by reviewing medical charts

Study design

All participants (subjects and controls) applied a Medtronic Guardian4 sensor for seven days. At the end of the monitoring, the glucose sensor data was analyzed using the Carelink System software.

Glucose metrics considered for the analysis were glucose management indicator (GMI), mean glucose (mg/dl), glycemic variability as standard deviation (SD) and coefficient of variation (CV), time spent with glycemia greater than or equal to 200 mg /dl (TAR200%,median), time spent with blood glucose greater than or equal to 140 mg/dl (TAR140%, median), time spent with blood glucose less than 70 mg/dl (TBR<70%, median).

As it is described in the American Diabetes Association (ADA) guidelines ,2h-glucose value ≥ 140 mg/dl (7.8 mmol/L) and <199 mg/dl (11 mmol/l) during OGTT is diagnostic for prediabetes. Moreover, 2h glucose value ≥200 mg/dl(11.1 mmol/L) during OGTT or a random glucose value (in any time of the day without regard to time since previous meal) ≥200 mg/dl(11.1 mmol/L) are both diagnostic for diabetes mellitus.

During a continuous glucose monitoring (CGM) it is not easy to understand correlation between glucose values and meals. Furthermore, in this study, the evaluation of glucose values (during CGM) two hours after a controlled meal (75-g of oral glucose in OGTT) was not performed. Instead, the participants' glucose values were observed in real life. Therefore, although the number of participants who exceeded or equaled the value 140 mg/dl or the value 200mg/dl was recorded, in addition how long these patients remained with glucose values above these cut-offs (TAR140%; TAR200%)was registered, to make the observation of abnormal glycemic profiles more precise.

Finally, the value ≤6% of TAR 140%was used as an index of normality (as indicated in other studies on healthy volunteers).