📄 medicine

Optimising propofol induction in elderly patients by constraining peak plasma concentration or infusion rate during Eleveld target-controlled infusions: an in silico simulation study

This in silico simulation study demonstrates that constraining either the peak plasma concentration or the maximum infusion rate during Eleveld target-controlled propofol induction in elderly patients significantly reduces plasma overshoot with only a minimal delay in reaching the desired effect-site target, offering a potential strategy to improve haemodynamic safety.

Original authors: George Zhong

Published 2026-10-20

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: George Zhong

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

When a patient undergoes general anesthesia, the goal is to guide them from wakefulness to a state of deep sleep safely and smoothly. To achieve this, anesthesiologists often use a computer-controlled system to deliver a precise amount of a drug called propofol. This system, known as a target-controlled infusion, acts like a sophisticated autopilot. Instead of a doctor manually pushing a syringe, the computer calculates exactly how fast to pump the drug to reach a specific concentration in the patient's bloodstream and, more importantly, at the site in the brain where the drug takes effect. The challenge lies in the timing: the drug must travel from the blood to the brain, and this journey takes time. If the computer tries to get the drug to the brain too quickly, it often has to flood the bloodstream with a massive initial dose. This creates a temporary spike in the amount of drug circulating in the blood, far higher than what is actually needed at the brain. For a healthy young person, this spike might be manageable, but for an elderly patient with a fragile heart, such a sudden surge can cause a dangerous drop in blood pressure.

This problem became particularly noticeable with a newer, highly accurate computer model used to guide these drug deliveries, known as the Eleveld model. While this model works well for people of all ages and body types, it has a specific quirk: it assumes the drug moves from the blood to the brain more slowly than older models did. To compensate for this slower journey and reach the desired level of sleep quickly, the computer calculates that it must deliver an even larger initial burst of the drug. This results in an even bigger spike in the blood concentration, which is risky for older adults who cannot tolerate sudden changes in their cardiovascular stability. The question researchers asked was whether they could slow down the delivery just enough to avoid this dangerous spike without making the induction of anesthesia take too long.

To find the answer, a researcher named George Zhong at Concord Repatriation General Hospital ran a series of computer simulations. He did not test real patients; instead, he created a digital representation of an 80-year-old man weighing 70 kilograms and standing 170 centimeters tall. He programmed this virtual patient to receive propofol using the Eleveld model, aiming for a specific level of drug concentration in the brain that would induce sleep. The standard approach, which serves as the baseline for comparison, involves the computer delivering a rapid, high-volume burst of the drug to reach the target as fast as possible. In this standard scenario, the computer delivered an initial dose of 94 milligrams over roughly 28 seconds. This method successfully reached the target brain concentration in 250 seconds, but it caused the drug level in the blood to spike to a peak of 14.1 micrograms per milliliter.

The researcher then tested two different strategies to see if they could prevent that high spike. The first strategy involved telling the computer to never let the concentration in the blood rise above a certain safe limit. The second strategy was to tell the computer to never pump the drug faster than a specific, slower rate. In both cases, the computer was allowed to take a little more time to reach the final target in the brain, but it was strictly forbidden from creating a massive surge in the blood. The results showed that these constraints worked remarkably well. By simply accepting a delay of just 68 seconds—meaning the patient reached the target sleep level in 318 seconds instead of 250—the researchers could drastically reduce the peak amount of drug in the blood.

When the computer was limited by the maximum infusion rate, the peak blood concentration dropped by 32 percent, falling to 9.6 micrograms per milliliter. When the computer was limited by the maximum blood concentration itself, the drop was even more significant, reducing the peak by 52 percent to just 6.7 micrograms per milliliter. Perhaps most importantly, these changes did not require the patient to receive significantly more or less total drug over the first 15 minutes. The total amount of propofol used remained almost identical across all three methods, differing by only a few milligrams. The primary difference was simply how that drug was delivered: either as a single, overwhelming flood, or as a more controlled, steady stream that avoided the dangerous peak.

The study suggests that for elderly patients, the trade-off is highly favorable. By accepting a delay of less than a minute, an anesthesiologist could potentially cut the risk of a dangerous blood pressure drop in half, without changing the total amount of drug the patient receives. The researcher notes that while these findings are based entirely on computer simulations, they provide a strong mathematical reason to believe that slowing down the induction process could protect vulnerable hearts. The study also highlights a potential neurological benefit. Rapid inductions that cause high drug spikes are linked to deeper levels of brain suppression, which can sometimes lead to confusion after surgery. A smoother, slower induction might preserve more natural brain activity during the transition to sleep.

Currently, many of the infusion pumps used in hospitals are programmed to prioritize speed above all else, automatically delivering the large initial bolus required by the standard settings. The researcher suggests that pump manufacturers could update their software to allow clinicians to set limits on how fast the drug is delivered or how high the blood concentration can rise. This would give doctors a simple tool to customize the induction for frail patients, balancing the need for speed with the need for safety. To help visualize this balance, the researcher even created a simple online tool that shows how changing the speed of delivery affects the peak drug levels, allowing clinicians to see the trade-off in real time. While real-world clinical trials are needed to confirm that these computer predictions translate to better outcomes for actual patients, the simulation offers a clear, practical path forward: a slightly slower start might just be the safest way to begin.

Technical Summary: Optimising Propofol Induction in Elderly Patients via Constrained Eleveld TCI

Problem Statement
Target-controlled infusion (TCI) systems utilizing the Eleveld propofol pharmacokinetic–pharmacodynamic (PK-PD) model are designed to achieve rapid effect-site concentrations ( $C_e$ ). However, the Eleveld model employs a relatively slow blood-to-effect-site equilibration rate constant ( $k_{e0}$ ), particularly when compared to models like Schnider. To compensate for this slow equilibration and reach the target $C_e$ quickly, the system calculates a large initial bolus, resulting in a significant transient overshoot in plasma concentration ( $C_p$ ).

In elderly or frail patients with limited haemodynamic reserve, this $C_p$ overshoot is clinically undesirable as it correlates with profound hypotension and myocardial depression. While mitigation strategies exist—such as gradual up-titration of the $C_e$ target, switching to plasma targeting, or limiting infusion rates—these approaches invariably delay the time to reach the desired effect-site concentration. The study addresses the need to quantify the trade-off between induction speed and the magnitude of plasma overshoot to identify strategies that preserve haemodynamic stability without excessively prolonging induction.

Methodology
The study conducted an in silico simulation using a representative 80-year-old male patient (70 kg, 170 cm). The simulation utilized the Eleveld TCI model with an opioid covariate included.

Target Parameters: The effect-site target ( $C_e$ ) was set at 3.0 µg·mL⁻¹. The propofol concentration was assumed to be 10 mg·mL⁻¹, with a maximum infusion rate cap of 1200 mL·h⁻¹.
Simulation Engine: Numerical integration of Eleveld PK equations was performed at 0.1-second time steps using MATLAB R2025b.
Strategies Evaluated:
1. Standard Effect-Site Targeting: Unconstrained induction serving as the reference.
2. Peak $C_p$ Limitation: Induction began in plasma targeting mode at maximum rate until a ceiling $C_p$ was reached. The rate was then adjusted to maintain this ceiling until the infusion was paused, allowing $C_e$ to rise to 3.0 µg·mL⁻¹ without overshoot.
3. Maximum Infusion Rate Limitation: Induction commenced at a constant, sub-maximal rate for a calculated duration, followed by a pause until $C_e$ reached the target.
Analysis: For each strategy, control parameters were varied to generate 20 induction trajectories with equilibration times ranging from the standard 250 seconds up to 900 seconds. Metrics analyzed included time to $C_e$ target, peak $C_p$ , and cumulative propofol dose over the first 15 minutes.

Key Results

Standard Regime: Achieved the $C_e$ target of 3.0 µg·mL⁻¹ in 250 seconds with an initial bolus of 94 mg. This resulted in a peak plasma concentration ( $C_p$ ) of 14.1 µg·mL⁻¹ and a cumulative 15-minute dose of 168 mg.
Impact of Constraints: Both constraint strategies significantly reduced peak $C_p$ $C_{p}$ with only modest delays in induction time.
- Specific Trade-off: Extending the time to target by 68 seconds (to 318 seconds) reduced the peak $C_p$ by 32% (to 9.6 µg·mL⁻¹) using infusion-rate limitation and by 52% (to 6.7 µg·mL⁻¹) using peak- $C_p$ limitation.
- Dose Efficiency: Despite the delays, the cumulative propofol dose over the first 15 minutes remained nearly identical to the standard regime (reductions of only 2–3%).
Mechanism: The constrained strategies effectively reduced the size of the initial bolus. For instance, the peak- $C_p$ limited strategy delivered a reduced bolus of 42 mg to reach a ceiling of 6.7 µg·mL⁻¹, maintained that level, and then paused, whereas the standard strategy delivered a 94 mg bolus immediately.

Significance and Claims
The paper claims that constraining either the maximum plasma concentration or the maximum infusion rate during effect-site TCI with the Eleveld model offers a viable method to reduce the magnitude of plasma overshoot. The author argues that this approach provides a pharmacokinetic basis for reducing haemodynamic perturbation during anaesthesia induction in elderly patients while maintaining reasonable induction speeds.

The study posits that these constraints may also have neurological implications, potentially reducing the risk of deep anaesthesia events (such as burst suppression) and postoperative delirium, which are associated with rapid inductions and high plasma overshoots in the elderly.

Limitations and Future Directions
The author explicitly states that this is a theoretical in silico study based on a single representative patient profile. While simulations across other demographics (obesity, underweight, female) yielded similar relative trade-off curves, the author cautions against generalizing these findings to clinical practice without validation. Crucially, the paper notes that peak $C_p$ was used as a surrogate for haemodynamic response; the complex relationship between infusion regimes and true in vivo cardiovascular response requires confirmation through future clinical studies. The author concludes that clinical trials are necessary to determine if these pharmacokinetic trade-offs translate into improved haemodynamic or electroencephalographic outcomes.

Technical Summary: Optimising Propofol Induction in Elderly Patients via Constrained Eleveld TCI

More like this