Challenges in adapting a conscious dog haemodynamic systems model to an anaesthetised dog protocol

Abstract

Previously¹, the cardiovascular-contractility systems model described by Fu et al.² was evaluated and extended using Simcyp™ Designer, a graphical interface platform and physiologically-based pharmacokinetic tool. Model outputs replicated published dog telemetry data¹. In the present study, the model was adapted for use in pentobarbital-anaesthetised beagle dogs, utilizing cardiovascular (CV) and exposure data for atenolol and atropine³. Model adaptations included removal of circadian rhythm, attenuation of baroreflex negative feedback, and development of a pharmacokinetic model in Simcyp™ Designer to accommodate intravenous dose escalation and hysteresis. Model simulations and experimental data for atenolol (0.3, 1, and 3 mg/kg/30 min) showed a decrease in dP/dt_max from 2300 to 1800 mmHg/s and mean arterial pressure (MAP) from 120 to 110 mmHg. The model predicts a decrease in heart rate (HR) from 110 to 85 bpm; however, bradycardia was not observed experimentally. For atropine (0.01, 0.03, 0.1 mg/kg/30 min), model outputs and experimental data displayed an increase in HR (124 to 164 bpm) and dP/dt_max (2650 to 3000 mmHg/s). While experimental MAP data decreased from 130 to 120 mmHg, model outputs predicted an increase from 105 to 130 mmHg. Several challenges were encountered during the development of the anaesthetised systems model. Firstly, the study design for the anaesthetised dog differs markedly from that of conscious dog telemetry. Secondly, the extent to which the baroreceptor reflex is attenuated by the anesthesia is unknown. Furthermore, the Fu et al. model² employs antagonist K_d values to predict haemodynamic effects, rather than in vivo EC₅₀ values. Finally, the decrease in HR predicted by the model following atenolol administration was not observed in anaesthetised dogs, possibly due to low resting cardiac sympathetic tone under the anesthesia. Despite these limitations, the adapted Fu et al.² model has the potential to maximise the use of anaesthetised CV dog data. Specifically: to interpolate CV effects at intermediate dose levels, derive threshold plasma concentrations for detectable CV effects, and predict the CV effects at doses which cannot be tolerated.

1. Mahmud et al. (2023) SPS Meeting, Brussels, Belgium

2. Fu et al. (2022) https://doi.org/10.1002/psp4.12774

3. Antic et al. (2024) https://doi.org/10.1016/j.vascn.2024.107497