Performance of Subcutaneous Continuous Glucose Monitoring in Adult Critically Ill Patients Receiving Vasopressor Therapy

Ola Friman, NavidSoltaniMarcusLindPia ZetterqvistAnca BalintescuAnders PernerAnders OldnerOlav Rooyackers, and Johan Mårtensson

Diabetes Technology & Therapeutics DTT May 31 2024 published




Subcutaneous continuous glucose monitoring (CGM) may facilitate glucose control in the ICU. We aimed to assess the accuracy of CGM (Dexcom G6) against arterial blood glucose (ABG) in adult critically ill patients receiving intravenous insulin infusion and vasopressor therapy. We also aimed to assess feasibility and tolerability of CGM in this setting.


We included ICU patients receiving mechanical ventilation, insulin, and vasopressor therapy. Numerical accuracy was assessed by the mean absolute relative difference (MARD), overall, across arterial glucose strata, over different noradrenaline equivalent infusion rates, and over time since CGM start. MARD <14% was considered acceptable.

Clinical accuracy was assessed using Clarke Error Grid (CEG) analysis. Feasibility outcome included number and duration of interrupted sensor readings due to signal loss. Tolerability outcome included skin reactions related to sensor insertion or sensor adhesives.


We obtained 2946 paired samples from 40 patients (18 with type 2 diabetes) receiving a median (IQR) maximum noradrenaline equivalent infusion rate of 0.18 (0.08–0.33) µg/kg/min during CGM. Overall, MARD was 12.7% (95% CI 10.7–15.3), and 99.8% of CGM readings were within CEG zones A and B. MARD values ≥14% were observed when ABG was outside target range (6–10 mmol/L [108–180 mg/dL]) and with noradrenaline equivalent infusion rates above 0.10 µg/kg/min. Accuracy improved with time after CGM start, reaching MARD values <14% after 36 h. We observed four episodes of interrupted sensor readings due to signal loss, ranging from 5 to 20 min. We observed no skin reaction related to sensor insertion or sensor adhesives.


In our ICU cohort of patients receiving vasopressor infusion, subcutaneous CGM demonstrated acceptable overall numerical and clinical accuracy.

However, suboptimal accuracy may occur outside glucose ranges of 6–10 mmol/L (108–180 mg/dL), during higher dose vasopressor infusion, and during the first 36 h after CGM start.




Utdrag ur artikeln


Excessive glucose fluctuations and hypoglycemia are associated with mortality among intensive care unit (ICU) patients.1,2These complications are primarily associated with insulin therapy, requiring a meticulous and adaptive approach to glucose monitoring. Glucose control within the ICU relies on intermittent blood glucose measurements. The most recent ICU guidelines suggest at least hourly glucose measurements in critically ill adults on intravenous insulin during episodes of glucose instability.3However, the frequency of these measurements is subject to the constraints of nursing workload, resulting in periods of less frequent monitoring. This limitation hampers the early detection and attenuation of imminent hypoglycemia and glucose oscillations, thus impacting patient safety.

Subcutaneous continuous glucose monitoring (CGM) is a potential solution providing real-time feedback to health care providers, allowing a more proactive and timely approach to treatment decisions. Consequently, CGM-guided glucose control has the potential to reduce glucose variability, prevent hypoglycemia, and improve overall glycemic control in ICU patients. Indeed, implementation of CGM as standard for outpatient use has improved long-term glucose control and reduced the occurrence of hypoglycemia in individuals with type 1 diabetes.4Preliminary studies have suggested an acceptable accuracy of CGM in ICU settings.5–9However, these studies were limited by small sample sizes, with primary focus on COVID-19 patients, and involved comparative analyses against point-of-care meters, which are not recommended to use in the ICU.10–12Furthermore, there are concerns regarding impaired CGM accuracy in patients receiving vasopressor therapy due to reduced blood circulation to the skin where the CGM sensor is applied.13

Accordingly, we conducted a prospective, observational study to assess the accuracy of CGM against arterial blood gas glucose (ABG) in ICU patients receiving vasopressor and insulin infusions, overall and in relevant subgroups. Furthermore, we aimed to assess the feasibility and tolerability of CGM in this clinical context.



Key findings

In this prospective, observational study, we assessed the performance of CGM in a cohort of ICU patients with multiple organ failure receiving insulin infusion and a range of vasopressor doses. During more than 200 CGM-days, sensor interruptions due to signal loss were negligible. Compared with routine ABG, CGM demonstrated acceptable numerical and clinical accuracy, as reflected by an overall MARD <14%, and by 99.8% of readings contained within CEG safe zones A and B. Inferior accuracy due to underestimation of ABG was observed when ABG was >12 mmol/L (216 mg/dL), and when noradrenaline equivalent doses exceeded 0.10 µg/kg/min. Finally, accuracy improved over time, reaching MARD values <14% after ∼36 h.

Relationship with previous studies

With almost 3000 paired samples obtained during more than 200 CGM-days, this is the largest study of the performance of CGM against ABG in critically ill patients receiving vasopressor therapy. Contemporary studies using the same sensor technology demonstrated MARD values between 10.4% and 13.9%5–9,22which is consistent with the MARD of 12.7% observed in our study. Notably, we found 99.8% of values within Clarke error grid safe zones A and B, indicating better clinical accuracy than previously reported.5,7,9,22

However, direct comparisons with previous studies are complex due to several key distinctions. First, our study encompasses a broader ICU patient cohort, unlike prior research that predominantly focused on critical COVID-19 cases.5–9Second, there is variability in the reference glucose measurements among studies, with some lacking detailed methodology and others using a mix of arterial, capillary, and venous blood on portable glucometers. In contrast, we consistently measured ABG on a blood gas analyzer, which is the preferred reference modality in the ICU setting.10–12Finally, the interaction between vasopressor therapy and CGM accuracy has been underexplored up to date. While studies in critical COVID-19,8during and after coronary artery bypass surgery,22and in mixed ICU cohorts23,24imply that CGM accuracy is unaffected by vasopressor use, these studies did not account for vasopressor dosage. Our study contributes by examining accuracy across a wide range of vasopressor doses, showing a slight increase in MARD at noradrenaline equivalent doses above 0.10 µg/kg/min.

Regarding glucose level stratification and accuracy, earlier findings are mixed. Consistent with our results, Faulds et al. identified a U-shaped relationship with the lowest MARD values at glucose ranges of 4–10 mmol/L (72–180 mg/dL). Conversely, other studies evaluating Dexcom or Libre sensors have reported higher accuracy above 10 mmol/L (180 mg/dL) than below this threshold.7,9,24,25Notably, existing studies, including ours, have insufficient power to provide robust analysis of accuracy during hypoglycemic episodes.

Implications of study findings

Our findings imply that subcutaneous CGM is feasible and provides acceptable accuracy when compared with routine ABG values in the ICU setting. However, CGM values should be confirmed with ABG testing when CGM readings increase above 12 mmol/L (216 mg/dL), to avoid excessive hyperglycemia, and when noradrenaline equivalent infusion rates exceed 0.10 µg/kg/min. In our study, we deemed a MARD below 14% as acceptable, acknowledging that this cutoff is opinion based15and exceeds outpatient recommendations. The specific MARD requirements may vary depending on the intended application. Within the ICU setting, CGM systems may offer timely alerts when blood glucose levels show signs of trending beyond acceptable ranges. In instances where such alerts necessitate confirmation through arterial glucose measurements before adjusting insulin doses, a slightly lower degree of point accuracy may be deemed acceptable. Moreover, given that arterial blood gas analyses are routinely performed to evaluate additional biomarkers such as lactate, electrolytes, and partial pressure of oxygen, CGM values will be periodically cross-checked with arterial glucose measurements.


www red DiabetologNytt