2025/09/28 by Matthew Hart, Stuart A. Ross, Justin Skowno · 1 voice · 1 citation
Medicine · Neuroscience · #Anesthesia and Sedative Agents #Intensive Care Unit Cognitive Disorders #Anesthesia and Neurotoxicity Research
paper · pdf · doi:10.1111/anae.70003
It was with sadness that we read the Prevention of Future Deaths report regarding the tragic death of Ms. Pamela Marking [1], in which concerns were raised about the use of total intravenous anaesthesia (TIVA) during rapid sequence induction (RSI), specifically regarding its perceived unsuitability for cases in which a RSI is indicated due to ‘slow onset of anaesthesia’. A slower induction when using TIVA for RSI was a concern shared by around two-thirds of respondents in a recent UK survey of anaesthetists [2]. There are many well-described clinical and environmental benefits to using TIVA. We believe that TIVA is not inherently unsuitable for use in cases requiring RSI; rather, safe use requires recognition of practical limitations and adoption of strategies to mitigate them. Mechanical and pressure limitations mean many commercially available devices cap flow at approximately 1200 ml.h-1, so a 20 ml propofol bolus takes 60 s to be administered, if delivered entirely by the pump. The same dose delivered manually into or close to the intravenous cannula takes just a few seconds. Studies of processed electroencephalography responses to different administration rates show a slower change in this marker of anaesthetic depth with slow infusions compared with a bolus [3]. For patients requiring RSI, the concern is that pump-limited delivery of the induction dose may delay loss of consciousness and increase the window for aspiration. Workarounds have been suggested. The guideline on safe practice of TIVA from the Association of Anaesthetists [4] suggests approaches can include co-administration of opioids; the administration of a manual bolus before commencing a TIVA infusion; or the use of alternative induction drugs. Each of these carries trade-offs and increases cognitive load. Seeking a practical solution, we worked with our local manufacturer (Arcomed, Artarmon, NSW, Australia) to develop an ‘RSI-TCI’ mode for its target-controlled infusion pumps. When programming the model, the anaesthetist pre-enters the intended manual bolus (e.g. propofol 250 mg). After routine checks and pre-oxygenation, a manual bolus is delivered rapidly. The target-controlled infusion is then started, which incorporates the bolus into the pharmacokinetic state and resumes infusing at an appropriate time to achieve the chosen target. If the delivered dose differs from the planned value, it can be edited immediately before starting the pump. Functionally, this mirrors a conventional RSI with a manual induction followed by maintenance, while keeping the model intact. We sought to model how this mode would perform compared with a conventional pump delivered induction. Using a publicly available implementation of the Eleveld PK–PD model [5], we simulated induction in a 24-year-old male (185 cm, 80 kg) with propofol 250 mg, given either solely by a pump at 1200 ml.h-1, or as a manual bolus over 5 s before starting a target-controlled infusion. The predicted effect-site concentration reached 2.4 μg.ml-1 at 75 s with pump-only delivery vs. 37 s with the manual bolus (Fig. 1). This approach requires no additional drugs and minimal extra steps. In our view, this reduces barriers to using TIVA for RSI. The key limitations include model dependence and that predicted effect site values are surrogates to clinical outcomes and may not reflect actual time to loss of consciousness in a real patient. In summary, a ‘RSI–TCI’ mode with a pre-programmed manual bolus offers a pragmatic solution for those who wish to preserve an intact pharmacokinetic model for accurate propofol delivery, while allowing the rapid induction clinicians expect from a manual bolus. It reduces the need for workaround strategies and is immediately implementable on systems that can incorporate a manual bolus into their target-controlled infusion model.