CORE heat flux sensor tracks anesthetized dog temperature best when fur remains intact
Core temperature monitoring is routine in veterinary anesthesia, but it becomes harder to trust when blood flow shifts, skin is clipped, and body heat is being lost to the room. In a small pilot study of three healthy beagle dogs, a wireless, non-invasive heat flux device called CORE was compared with digital rectal thermometry during general anesthesia after two very different premedication strategies: acepromazine and dexmedetomidine.
The most clinically useful signal did not emerge from clipped skin in every circumstance.
In dogs given acepromazine, the sensor performed well whether it was placed over clipped or unclipped skin on the jugular region. In dogs given dexmedetomidine, agreement weakened on clipped skin but improved when fur was left intact, and the unclipped placement produced clinically acceptable Bland-Altman agreement with rectal temperature. For anesthetists trying to follow core temperature without relying on rectal access, that pattern matters because the drug-induced vascular state appears to change how well a surface-based heat flux system can estimate central temperature.
Why vascular tone changed the sensor’s behavior
The physiologic logic is straightforward. Acepromazine lowers systemic vascular resistance through alpha-1 blockade and tends to favor vasodilation, while dexmedetomidine activates alpha-2 receptors and produces peripheral vasoconstriction. Isoflurane adds its own burden by blunting thermoregulation and lowering vascular tone further. When peripheral blood flow changes, heat delivery to the skin changes too, and a surface sensor that infers core temperature from heat flux may be forced to estimate temperature from a signal that is either dampened or artificially shifted.
Fur changes the equation as well. In a clipped animal, the skin-sensor interface may be more direct, but clipping also removes insulation and may accentuate local heat loss. In an unclipped animal, fur can interfere with contact, yet the preserved natural barrier may stabilize the thermal environment around the sensor enough to improve performance in some hemodynamic states. The pilot data suggest that the balance between contact quality and insulation was not uniform across anesthetic regimens.
A small crossover study with repeated measurements
The study used three healthy male beagle dogs, each roughly 8 months old and weighing 8.5 ± 0.6 kg on average. All dogs were acclimated for seven days, examined before anesthesia, and enrolled in a crossover design with two anesthetic periods using different premedicants. One 5 cm section of fur over the left jugular vein was clipped before anesthesia, while a second CORE sensor was placed over an unclipped jugular area on the opposite side. Rectal temperature served as the reference standard.
Each dog received either acepromazine at 0.05 mg/kg intramuscularly or dexmedetomidine at 5 mcg/kg intramuscularly, followed by propofol induction and isoflurane maintenance. Mechanical ventilation, multiparameter monitoring, and active warming were all part of the protocol. Temperature data were collected every three minutes for up to 120 minutes after induction, with other anesthetic variables tracked every five minutes. The design was intentionally narrow, but it was structured to isolate how premedication and skin condition altered sensor performance during a controlled anesthetic period.
Correlation and agreement diverged by premedication
Under acepromazine, the CORE sensor showed a strong relationship with rectal temperature. On clipped skin, the repeated-measures correlation coefficient was 0.81 with P < 0.0001, and the intraclass correlation coefficient was 0.78. On unclipped skin, the repeated-measures correlation coefficient was 0.79 with P < 0.0001, and the intraclass correlation coefficient was 0.85. Those values support both correlation and reliability in a setting where vasodilation could make peripheral and core temperature signals behave more similarly.
Dexmedetomidine told a different story. On clipped skin, the repeated-measures correlation coefficient was only 0.30, despite statistical significance, and the intraclass correlation coefficient fell to 0.28, which signals poor reliability. When the sensor was moved to unclipped skin, the repeated-measures correlation coefficient rose to 0.48 and the intraclass correlation coefficient increased to 0.88. Bland-Altman analysis also found clinically acceptable agreement between rectal thermometry and the unclipped CORE method.
The shift was abrupt enough to be clinically visible, not just statistically interesting.
What the agreement data imply at the bedside
For anesthetized dogs, rectal thermometry remains a practical reference, but it is not always the easiest site to access and may be inconvenient during some procedures. A continuous, wireless method that can function without invasive placement would be attractive if it can provide dependable core estimates. The present data suggest that the CORE temperature sensor is not a universal substitute, but it may become a useful option when positioned correctly and interpreted in light of the drug used for premedication.
The strongest signal came from unclipped skin during dexmedetomidine anesthesia, which is counterintuitive only until the thermophysiology is considered. Peripheral vasoconstriction may alter how heat is transferred to a clipped site, while intact fur may help stabilize the microenvironment around the sensor and improve the model’s estimates. In acepromazine-treated dogs, where vasodilation is expected, the device behaved more consistently across skin conditions.
That combination suggests the sensor is not merely measuring temperature; it is responding to the interplay among skin contact, peripheral perfusion, and anesthetic pharmacology.
A pilot study with real limits
This is a preliminary study, and its size is the most obvious limitation. Three dogs cannot define performance across breeds, ages, body conditions, or surgical settings. The animals were healthy beagles, so disease states that affect perfusion or thermoregulation were not tested. The study also used a controlled anesthetic environment with active warming, standardized monitoring, and a limited time window, which may not mirror more chaotic operating room conditions.
Even with those constraints, the findings are clinically useful because they identify a physiologic interaction rather than a simple yes-or-no device failure. The sensor did not behave the same way across all conditions, and the pattern aligned with known effects of acepromazine and dexmedetomidine on vascular tone. For veterinary teams, that means the placement site and the chosen premedication should be considered together rather than separately.
The next step is larger validation in a broader canine population, with attention to different body coats, surgical procedures, and hemodynamic states. Until then, the clearest take-home is narrow: in this pilot, the CORE device aligned best with rectal temperature when placed on unclipped skin in dogs receiving dexmedetomidine, while acepromazine produced better agreement across both skin conditions.