Abstract
Type 2 diabetes mellitus (T2DM) is a major comorbidity of coronavirus disease 2019 (COVID‐19). Patients with T2DM had a significantly higher mortality and intensive glycemic control could improve the prognosis of the patients. 1 , 2 However, there is still a challenge for clinicians to control blood glucose effectively, especially in the intensive care unit. The FreeStyle Libre Flash Glucose Monitoring System (FSL‐CGM) (Abbott Diabetes Care) was commonly used for continuous glucose monitoring in outpatients. For the special group of COVID‐19 patients, we hope that the FSL‐CGM could reflect the glucose fluctuation in time and help the doctors provide personalized therapeutic schedule.
Seventeen adult COVID‐19 patients with hyperglycemia (fasting blood glucose ≥7.00 mmol/L) in ICU of Tongji Hospital Guanggu, Wuhan, were enrolled in this study during 10 February to 30 March 2020. This study was approved by the Huashan Hospital Ethics Committee. Given the urgency of the COVID‐19 pandemic and the specificity of the ICU patients, the informed consent forms were waived by the ethics board of the hospital. Male/female ratio was 11/6, nine with T2DM. The median duration of continuous glucose monitoring was 12 days and the simultaneous existence of venous blood glucose (VBG) and FreeStyle Libre sensor glucose (FLSG) at 6:00 am was 10 days. One hundred and seventy one paired VBG and FLSG values were obtained and corresponding blood biochemical indicators at the same time were also collected. Our error grid analyses showed acceptable clinical accuracy, with 97.1% of glucose readings falling into zones A (n = 85, 49.7%) and B (n = 81, 47.4%) in Clarke error grid analysis, 97.7% into zones A (n = 78, 45.6%) and B (n = 89, 52.0%) in consensus error grid analysis (Figure 1). However, the mean absolute relative difference (MARD) in our study was 22.4%, which was much higher than 11.4%, reported in diabetic outpatients, 3 and 14.0%, reported in a small‐sample study for patients with diabetes in ICU. 4
FIGURE 1.
Clarke (A) and consensus (B) error grid analyses comparing FreeStyle Libre sensor results to venous blood glucose. Different zones of the grid indicate the degree of risk caused by erroneous measurements. Values in zones A and B are considered clinically acceptable levels of agreements
The Bland‐Altman plot showed that FSL‐CGM tended to underestimate blood glucose (mean bias, −1.9 mmol/L; limits of agreement, −5.9 to 2.1 mmol/L), which was also been observed in previous studies. 4 , 5 As shown in Table 1, for 89.47% of paired samples Freestyle glucose underestimated the venous glucose, with the average Freestyle glucose 12.50% below venous glucose at low levels, 7.89% below at relatively normal levels, and 10.39% below at high glucose levels. The mean amplitude of glycemic excursion of FSL‐CGM was 4.07 mmol/L.
TABLE 1.
The difference between venous blood glucose and FreeStyle Libre sensor glucose
| VBG (mmol/L) | <4 mmol/L | 4‐8 mmol/L | >8‐10 mmol/L | >10 mmol/L |
|---|---|---|---|---|
| The number of paired VBG and FLSG values | 0 | 56 | 38 | 77 |
| Absolute value of the difference between VBG and FLSG | 0 | 1.50(0.96‐2.14) | 1.37(0.83‐2.41) | 2.72(1.65‐3.55) |
| The number of paired VBG and FLSG values (FLSG<VBG) | 0 | 49(87.50%) | 35(92.11%) | 69(89.61%) |
Note: Data are presented as median (25th percentile‐75th percentile) or n (percentage).
Abbreviations: FLSG, FreeStyle Libre sensor glucose; VBG, venous blood glucose.
The basic vital signs, blood routine indexes, coagulation indexes, inflammatory factors, liver function indexes, renal function indexes, blood ions, and myocardial enzymes were compared between zone A and B to identify potential substances influencing the accuracy of CGM. Several indicators might influence its accuracy, including mean corpuscular hemoglobin concentration, Cl−, Mg2+.
In summary, our results showed that the application of FSL‐CGM in ICU patients with COVID‐19 is safe and feasible. Although its accuracy is relatively lower than that in ordinary diabetic patients, the use of FSL‐CGM could be conducive to glycemia control, reducing the workload and minimizing the risk of infection among the medical staff. Our study still has several limitations, including a small sample size and wearing time of individual patients that was not long enough. We believe that our results potentially have some inspirations for FSL‐CGM use in COVID‐19 patients.
DISCLOSURE
The authors declare that there is no conflict of interest.
ACKNOWLEDGEMENT
The present study was supported by grants from the Shanghai Talent Development Fund Program (2018054, to Bin Lu). Weʼd like to thank the guidance on FSL‐CGM data analysis from Dr. Yufei Wang, Mr. Yunyang Sun, and Shanghai imedpower tech.
REFERENCES
- 1. Richardson S, Hirsch JS, Narasimhan M, et al. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID‐19 in the New York City area. JAMA. 2020;323:2052‐2059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Zhu L, She ZG, Cheng X, et al. Association of blood glucose control and outcomes in patients with COVID‐19 and pre‐existing type 2 diabetes. Cell Metab. 2020;31(6):1068‐1077.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Bailey T, Bode BW, Christiansen MP, Klaff LJ, Alva S. The performance and usability of a factory‐calibrated flash glucose monitoring system. Diabetes Technol Ther. 2015;17(11):787‐794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ancona P, Eastwood GM, Lucchetta L, Ekinci EI, Bellomo R, Martensson J. The performance of flash glucose monitoring in critically ill patients with diabetes. Crit Care Resusc. 2017;19(2):167‐174. [PubMed] [Google Scholar]
- 5. Aberer F, Hajnsek M, Rumpler M, et al. Evaluation of subcutaneous glucose monitoring systems under routine environmental conditions in patients with type 1 diabetes. Diabetes Obes Metab. 2017;19(7):1051‐1055. [DOI] [PubMed] [Google Scholar]