In adult patients, how do the Dyks and Hannivoort three‑compartment models for target‑controlled infusion of dexmedetomidine differ, and when should each model be used?

Dyks vs Hannivoort Model for Dexmedetomidine TCI

The Hannivoort model should be used for target-controlled infusion of dexmedetomidine in adult patients when targeting plasma concentrations below 2 ng/mL, while the newer Dyks model must be used for higher target concentrations or prolonged infusions due to its incorporation of nonlinear pharmacokinetics.

Key Pharmacokinetic Differences

The fundamental distinction between these models lies in their assumptions about dexmedetomidine elimination:

• Hannivoort model (2017): Assumes linear pharmacokinetics with constant clearance regardless of drug concentration 1
• Dyks model (2022): Incorporates nonlinear elimination kinetics, recognizing that dexmedetomidine clearance decreases with increasing plasma concentrations 2

The Dyks model represents a critical advancement because it addresses a dangerous limitation of earlier models: they systematically underpredict actual plasma concentrations, especially at higher target levels 2.

Clinical Decision Algorithm

Use Hannivoort Model When:

• Target plasma concentration ≤ 2 ng/mL
• Short-term sedation (< 24 hours)
• Standard ICU sedation protocols
• Effect-site targeting for arousable sedation 1

At concentrations up to 2 ng/mL, both models produce similar predictions 2, making the simpler Hannivoort model adequate for routine use.

Use Dyks Model When:

• Target plasma concentration > 2 ng/mL
• Prolonged infusions (> 24 hours)
• High-dose protocols (approaching 1.5 μg/kg/hr)
• Loading dose administration
• Any scenario requiring precise control at higher concentrations

Critical Safety Considerations

The nonlinear pharmacokinetics identified by Dyks create significant safety implications 2:

1. Accumulation risk: At higher concentrations, decreasing clearance causes disproportionate drug accumulation
2. Loading dose hazard: Initial boluses may produce higher-than-expected concentrations, increasing hemodynamic instability risk
3. ICU dosing concerns: Current FDA-approved dosing regimens (0.2-0.7 μg/kg/hr, up to 1.5 μg/kg/hr) may lead to plasma concentrations exceeding predictions from linear models 2

Hemodynamic Monitoring Requirements

Both models require vigilant monitoring, but the Dyks model's recognition of concentration-dependent effects is particularly relevant given dexmedetomidine's profile 3:

• Bradycardia and hypotension are the most common adverse effects
• Loading doses can cause biphasic blood pressure response (initial hypertension, then hypotension)
• Risk increases with higher plasma concentrations due to nonlinear accumulation

Model-Specific Advantages

Hannivoort Model Strengths:

• Incorporates arousal effects on sedation depth 1
• Accounts for stimulation during assessment (170% increase in EC50 during MOAA/S scoring)
• Recognizes ambient noise effects (32% difference in required concentrations)
• Effect-site targeting capability for tighter sedation control 1

Dyks Model Strengths:

• Three-compartment nonlinear model with superior accuracy at all concentration ranges 2
• Includes covariates: age, sex, total body weight
• Prevents dangerous underprediction at higher targets
• Based on pooled data from 48 healthy subjects with stepwise increasing concentrations 2

Practical Implementation

For routine ICU sedation (target 0.2-0.7 μg/kg/hr maintenance):

1. Use Hannivoort model with plasma targets 1-2 ng/mL
2. Avoid loading doses in hemodynamically unstable patients 3
3. Monitor for arousal effects if using effect-site targeting 1

For deeper sedation or prolonged use:

1. Switch to Dyks model for any target > 2 ng/mL
2. Anticipate nonlinear accumulation with dose increases
3. Extend monitoring for delayed emergence (elimination half-life 1.8-3.1 hours in normal hepatic function) 3

For hepatic dysfunction:

• Both models require dose reduction
• Dyks model may be preferable given its superior handling of altered clearance 2
• Expect prolonged emergence regardless of model used 3

Common Pitfalls

1. Using Hannivoort model at high targets: This creates systematic underprediction and potential overdosing 2
2. Ignoring arousal state: Hannivoort model shows stimulation dramatically alters required concentrations 1
3. Assuming linear kinetics for long infusions: Accumulation accelerates nonlinearly beyond 24 hours 2
4. Overlooking covariate effects: Age, sex, and weight significantly affect both models' predictions

The evidence clearly demonstrates that model selection must be concentration-dependent, with the Dyks model becoming essential above 2 ng/mL targets where its nonlinear pharmacokinetic modeling prevents dangerous underprediction of actual drug exposure 2.