What are the pitfalls of using continuous glucose monitoring (CGM) in patients with sepsis?

Pitfalls of CGM in Sepsis Patients

CGM sensors in critically ill sepsis patients are prone to false readings from catheter flushes, IV contamination, altered tissue perfusion, acid-base disturbances, and early sensor instability—all of which can lead to dangerous insulin dosing errors if not recognized and confirmed with arterial blood glucose measurements.

Critical Technical Pitfalls

False Low Readings from Catheter Flushing

• CGM sensors inserted in arterial or central venous catheters will transiently show falsely low glucose when flushed with saline or glucose-free solutions 1
• This can trigger inappropriate treatment for pseudohypoglycemia
• Always confirm rapid glucose drops with arterial blood sampling before treating

False High Readings from IV Contamination

• Sensors in peripheral or central veins show falsely elevated readings when contaminated with glucose-containing IV solutions 1
• This is particularly problematic in sepsis patients receiving multiple concurrent infusions
• Verify unexpected glucose elevations with arterial samples

Physiologic Accuracy Issues in Sepsis

Impaired Tissue Perfusion

Septic patients have multiple factors degrading CGM accuracy:

• Vasopressor use, particularly norepinephrine doses >0.6 µg/kg/min, independently reduces CGM accuracy 2
• Hypotension and shock slow glucose equilibration between blood and interstitial fluid
• Microcirculatory dysfunction, edema, and dehydration all impair sensor performance 3

Acid-Base Disturbances

• Each 0.1-unit decrease in arterial pH increases measurement error by 1.5 percentage points 2
• Acidosis (common in sepsis) significantly degrades accuracy
• The pH-accuracy relationship is approximately linear—more acidotic patients have progressively worse CGM performance 2

Early Sensor Instability

• CGM accuracy is substantially reduced during the first 50 hours after insertion, then stabilizes 2
• In sepsis patients requiring urgent glycemic control, this warm-up period is clinically problematic
• Predicted error decreases and accuracy improves only after this initial period

Rate/Trend Accuracy Limitations

• Rate and trend metrics have not been sufficiently tested in critically ill populations to provide definitive guidance 1
• Rapid glucose changes during sepsis resuscitation may not be accurately tracked
• CGM may lag behind actual blood glucose during dynamic clinical situations

Practical Safety Concerns

Reliability Issues

• CGM systems should display readings >95% of the time, but data gaps up to 30 minutes can occur 1
• Device malfunction occurs in 85.4% of CGM users over time 4
• In sepsis, missing data during critical periods (patient transport, procedures) creates dangerous blind spots

Confirmation Requirements

• 98% of CGM readings should be within 12.5% of reference standards, but only if used alone to guide insulin therapy 1
• The 2013 Critical Care consensus explicitly states CGM must demonstrate accuracy in the intended-use critical care population 1
• Always confirm out-of-range CGM values with arterial blood glucose before adjusting insulin 3

Clinical Implementation Pitfalls

Sampling Hierarchy Violations

• Arterial catheter blood is the first-choice sample in critically ill patients; capillary sampling introduces unacceptable errors 1
• CGM should supplement, not replace, arterial blood glucose monitoring in sepsis
• Never rely on CGM alone when vasopressors are running or perfusion is compromised

Interfering Substances

• CGM systems must demonstrate lack of interference from substances other than glucose, but new therapeutic compounds may not be tested 3
• Sepsis patients receive multiple medications that may interfere with sensor chemistry
• FDA 510(k) summaries don't cover all drugs used in critical care

Workload and Documentation

• While CGM may reduce nursing workload by 27.62% in some cases 5, implementation challenges include:
• Need to capture all results in electronic health records
• Requirement for confirmatory testing protocols
• Staff training on recognizing spurious data

Specific Recommendations for Sepsis

Use CGM only as an adjunct monitoring tool in sepsis patients, not as the sole guide for insulin therapy:

1. Always confirm CGM readings with arterial blood glucose when:

   • Norepinephrine dose >0.6 µg/kg/min
   • pH <7.30
   • Sensor wear time <50 hours
   • Rapid glucose changes occur
   • Any out-of-range value appears

2. Establish alerts for:

   • Rapid glucose drops (may indicate catheter flushing)
   • Rapid glucose rises (may indicate IV contamination)
   • Data gaps exceeding 30 minutes

3. Do not use CGM as the primary monitoring method during:

   • Active resuscitation with high-dose vasopressors
   • Severe acidosis
   • First 48 hours after sensor insertion in unstable patients

The evidence shows CGM can improve protocol adherence (73.91% vs 48.16%) and detect more hypoglycemic events in sepsis 6, but these benefits only materialize when clinicians recognize and work around the inherent limitations rather than blindly trusting sensor values in physiologically unstable patients.