The auditory-cardiac channel in the last days of life: a neurovisceral hypothesis for limiting opioid escalation

This is a hypothesis paper. No patient was studied and no data were collected. Nothing in it is clinical advice, and nothing in it should delay, reduce or substitute for established palliative care including adequate analgesia. In the last days of life, opioid infusion is titrated against observed distress. The signs being titrated against — restlessness, grimacing, tachycardia — are also the signs of opioid-induced neurotoxicity as metabolites accumulate in a patient whose clearance is failing. The bedside cannot distinguish too little drug from too much, and the humane default is to give more. The cost is that patients become unreachable days before they die. We propose that part of that escalation is driven by a self-reinforcing loop between autonomic arousal and perceived pain rather than by nociception alone, and that the loop is reachable through the subcortical auditory pathway, which remains functional when other sensory channels have failed. The modelled benefit is bounded. A minimal model of the loop indicates that damping autonomic arousal postpones the neurotoxic threshold by about 6 hours at modest strength, with a ceiling near 35 hours, and never prevents the crossing at any damping strength. An intervention that buys lucid hours is worth having; one presented as a way to avoid opioid escalation would be both wrong and dangerous. What is known is separated from what is proposed. Tone-based auditory responses persist in actively dying patients, and semantic processing survives unresponsiveness in disorders of consciousness — but no published study has demonstrated semantic processing in actively dying patients. That commonly conflated claim is treated here as the primary testable prediction, not as a premise. The molecular basis of opioid-induced neurotoxicity is likewise contested. Five falsification criteria are given, together with an explicit account of how the proposal could cause harm — principally by being read as a reason to withhold analgesia from a patient who cannot report pain. Every parameter in the model is illustrative; it is not a pharmacokinetic simulation and no output is a predicted dose or survival time.