Corridor-based relative severity assessment of the canine anaesthetic course

The Relative Severity Assessment (RELSA) score grades welfare severity based on the magnitude of multivariate deviations from a reference value. Many variables monitored during anaesthesia are homeostatic and become harmful when they leave the normal range in either direction. Building on the corridor concept of the PVS from human intensive care, we conducted a feasibility study to assess the application and behaviour of a corridor extension of RELSA in two canine anaesthesia datasets. Each homeostatic variable was given a normal band, and only its deviation was folded into the standard RELSA algorithm. Severity in this report denotes the number this algorithm returns, and that number has not been shown to represent clinical severity. We obtained two Beagle datasets, including one four-arm within-subject study, allowing a paired between-protocol comparison. This is an exploratory secondary analysis of existing data, with no a priori power calculation. Under the selected bands, the calculated score changed across the recorded anaesthetic course: low while awake, rising at the documented propofol induction, maintained, and resolving on recovery. The dominant departures were hypotension, hypothermia, and mild hypercapnia, sometimes coinciding with a single animal, though their magnitudes are band-dependent. Corridor RELSA produced a number that changed across the recorded anaesthetic course and represented bidirectional departures as a single continuous score. The between-protocol comparison is not interpreted as evidence about the protocols because the verdict it returns depends on the variable panel being scored. The approach is a domain extension and a small refinement of the human-ICU corridor concept.