Diagnostic Assessment and Classification of Dysglycemia in the ICU

1. Clinical Recognition of Dysglycemia

Early recognition relies on vigilant assessment of neurocognitive, autonomic, hemodynamic and metabolic clues that often overlap with critical illness manifestations.

Neuroglycopenic and Autonomic Signs

• Neuroglycopenia (glucose <54 mg/dL): Presents as confusion, delirium, seizures, or focal neurological deficits.
• Autonomic activation: Characterized by diaphoresis, tremor, tachycardia, and anxiety. These signs may be blunted by β-blockers or the systemic inflammatory response of sepsis.
• Sedation and underlying encephalopathy can mask these classic symptoms, necessitating a low threshold for glucose checks in critically ill patients.

Hemodynamic and Metabolic Indicators

• Hypoglycemia: Can lead to hypotension, decreased cardiac output, and arrhythmias due to myocardial substrate deprivation.
• Hyperglycemia: Causes osmotic diuresis, leading to intravascular volume depletion and significant electrolyte shifts, particularly of potassium (K⁺) and magnesium (Mg²⁺).
• Ketoacidosis: An anion-gap metabolic acidosis may occur even with moderate glucose levels, a condition known as euglycemic diabetic ketoacidosis (DKA), especially in patients on SGLT2 inhibitors.

2. Laboratory and Point-of-Care Diagnostics

Accurate glucose measurement is essential for guiding timely therapy. Clinicians must understand the strengths and limitations of various devices, especially in the context of shock, hypoperfusion, and other ICU-specific conditions.

Arterial vs. Capillary Sampling

• Capillary Point-of-Care (POC) Glucometers: While rapid, their accuracy diminishes in states of poor peripheral perfusion. During hypotension or hypoxia, fewer than 60% of readings are within 20% of the central laboratory value.
• Arterial Whole-Blood Analyzers (ABG): Show high agreement with central lab glucose measurements (r² >0.97) and are the preferred method in hemodynamically unstable patients.

Quality Control and Device Limitations

• Interfering Factors: Accuracy can be affected by hematocrit extremes, icterus (high bilirubin), lipemia (high triglycerides), and ambient temperature.
• Accuracy Standards: Per ISO 15197:2013, 95% of POC results must be within ±15 mg/dL of the lab value for glucose concentrations <100 mg/dL, or within ±15% for concentrations ≥100 mg/dL.
• Regular calibration, robust staff training, and protocol enforcement are critical to mitigate measurement errors.

Continuous Glucose Monitoring (CGM): Promise and Pitfalls

• Function: Measures interstitial glucose every 1–5 minutes, allowing for detection of trends and providing alarms for glycemic excursions.
• Limitations: There is a physiological lag of 5–15 minutes between blood and interstitial glucose. Sensor accuracy can be affected by edema, and frequent calibration is often required in the ICU setting.
• Current Status: Workflow integration remains largely investigational. Alarms must be confirmed with a blood sample before clinical action is taken.

3. Diagnostic Criteria and Classification

Precise thresholds and a clear classification system are necessary to guide the urgency of intervention and select appropriate glycemic targets. An admission HbA1c is invaluable for differentiating pre-existing diabetes from acute stress hyperglycemia.

ADA Definitions for Dysglycemia

Stress Hyperglycemia vs. Undiagnosed Diabetes

• An admission glucose >140 mg/dL with an HbA1c <6.5% suggests transient stress hyperglycemia.
• An admission glucose >140 mg/dL with an HbA1c ≥6.5% indicates likely undiagnosed diabetes that will require chronic management planning.
• Prognostic tools like the Stress Hyperglycemia Ratio (admission glucose / estimated average glucose from HbA1c) and the Glycemic Gap correlate with mortality.

Glycemic Domains and Prognostic Implications

The relationship between glucose levels and mortality in the ICU is not linear but follows a U-shaped curve. Both hypoglycemia and severe hyperglycemia are associated with worse outcomes.

• Target Range: For most critically ill patients, a target glucose range of 140–180 mg/dL is recommended.
• Glycemic Variability: High variability in glucose levels (measured by standard deviation or coefficient of variation) is an independent predictor of mortality.
• U-Shaped Risk: Both hypoglycemia (<70 mg/dL) and sustained hyperglycemia (>180 mg/dL) significantly increase mortality risk.

4. Risk Stratification and Severity Scores

Integrating mortality risk scores and inflammatory profiles can help personalize glycemic targets and allocate monitoring resources more effectively.

Pediatric Severity Scores (PRISM-3, PIM2)

• The Paediatric Index of Mortality (PIM2) score uses physiologic and clinical variables at admission to predict mortality risk in pediatric ICU patients.
• A higher score indicates a greater risk of complications from dysglycemia, suggesting a need for tighter monitoring and more aggressive management.

Biomarker Integration and Latent Class Analysis

• Biomarker panels (e.g., IL-6, IL-8, IL-10, TNFR-1) can identify patient subgroups with distinct inflammatory profiles, such as “hyperinflamed” vs. “hypoinflamed.”
• Recent pediatric randomized controlled trials suggest that hyperinflamed patients may derive a mortality benefit from tight glycemic control (80–110 mg/dL).
• Conversely, hypoinflamed patients may be harmed by such intensive targets, highlighting the importance of personalized therapy.

Prioritizing Tight vs. Moderate Control

• Moderate control (140–180 mg/dL) remains the standard of care for most adult ICU patients, as it effectively balances the benefits of avoiding severe hyperglycemia with the risks of iatrogenic hypoglycemia.
• Tighter targets (e.g., 110–140 mg/dL or lower) may be considered in select populations (e.g., some pediatric or cardiac surgery patients) but require a robust safety infrastructure, including frequent monitoring and validated protocols.
• Always reassess glycemic targets based on changes in clinical status, nutrition, and concomitant therapies (e.g., steroids, vasopressors).