Performance of Continuous Glucose Monitoring System Among Patients With Acute Ischaemic Stroke Treated With Mechanical Thrombectomy

Abstract

Aims

Glucose metabolism abnormalities are prevalent in acute ischaemic stroke (AIS) patients and are associated with poor prognosis. The continuous glucose monitoring (CGM) system can provide detailed information on glucose levels and glycaemic excursions. This study aimed to evaluate the feasibility and accuracy of CGM application in the acute phase of AIS patients.

Methods

This single-centre, prospective, and observational study consecutively enrolled patients with AIS with anterior circulation large vessel occlusion (AC-LVO) and received mechanical thrombectomy (MT) within 24 h of symptom onset. A user-retrospectively calibrated iPro2 CGM system was implanted right before the MT procedure started and removed on the fifth day after MT or at discharge. Fingertip glucose was measured as a reference. Accuracy evaluation included the Bland–Altman plot (with a proportion of CGM values within 15/15, 20/20 and 30/30), the absolute relative difference (ARD) and error grid analysis (EGA). The safety and glucose profiles were also evaluated.

Results

Of the 183 patients screened, 141 were included, with a median monitoring duration of 4.49 days. Compared to reference measurements, 3097 CGM readings were matched with a mean bias of −4.16 mg/dL. The proportions of sensor readings meeting the 15/15, 20/20 and 30/30 criteria were 64.55%, 76.07% and 87.21%, respectively. The overall mean and median ARD were 14.60% ± 14.62% and 9.77% (4.15, 20.00). EGA showed that 98.97%, 99.42% and 99.06% values fall within clinically accurate zones in Clarke, Parkes and continuous glucose EGA, respectively.

Conclusion

The CGM system was feasible, safe and accurate for in-hospital use among AIS patients who received MT.