Journal of veterinary emergency and critical care (San Antonio, Tex. : 2001) · 2025
Associations Among Various Physical Parameters and the Axillary-to-Rectal Temperature Difference in Dogs and Cats.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Martinez Alvarez Cristian, Thevelein Britt A L, Hodges Katie M, Weitzman Amy N, Koenig Amie, Brainard Benjamin M
Аннотация
OBJECTIVE: To compare the measurement of axillary temperature (AT) with rectal temperature (RT) in dogs and cats and to identify the influence of physical parameters on the difference between their measurement using a veterinary-specific thermometer.
DESIGN: Prospective study (2022-2024).
SETTING: University teaching hospital.
ANIMALS: A total of 106 dogs and 101 cats aged ≥4 months.
INTERVENTIONS: Body temperatures were measured contemporaneously using a veterinary-specific, calibrated, dual-thermistor axillary thermometer (AT) and a calibrated medical thermometer (RT). Data were evaluated for bias as a whole set and after stratification based on various physical parameters (species, body weight, haircoat length, body condition score [BCS]). A machine learning (ML) model was subsequently applied, and bias was reevaluated.
MEASUREMENTS AND MAIN RESULTS: A Bland-Altman analysis comparing the measurement of AT with RT showed a bias of 1.01°C (1.82°F) and 95% limits of agreement (LOA) -0.66°C to 2.68°C (-1.19°F to 4.83°F) in dogs and a bias of 0.23°C (0.42°F) and 95% LOA of -2.61°C to 3.08°C (-4.71°F to 5.55°F) in cats. Animals weighing <10 kg had a bias of 0.33°C (95% LOA: -2.29°C to 2.96°C), animals with BCS <5 had a bias of 0.09°C (95% LOA: -2.92°C to 3.11°C), and those with shorter haircoats had a bias of 0.49°C (95% LOA: -2.03°C to 3.00°C). The ML model overestimated RT when using AT and physical parameters in its algorithm (dogs: -0.57°C [95% LOA: -2.18°C to 0.99°C]; cats -0.84°C [95% LOA: -2.62°C to 0.93°C]).
CONCLUSIONS: Animals weighing <10 kg, with lower BCS, and shorter haircoats had less bias when AT was compared with RT. Cats exhibited a lower bias between AT and RT than dogs. ML models can be programmed to account for various physical characteristics, improving the predictive impact of AT for directly measured RT, especially in categories of animals where closer prediction does not already exist.
DESIGN: Prospective study (2022-2024).
SETTING: University teaching hospital.
ANIMALS: A total of 106 dogs and 101 cats aged ≥4 months.
INTERVENTIONS: Body temperatures were measured contemporaneously using a veterinary-specific, calibrated, dual-thermistor axillary thermometer (AT) and a calibrated medical thermometer (RT). Data were evaluated for bias as a whole set and after stratification based on various physical parameters (species, body weight, haircoat length, body condition score [BCS]). A machine learning (ML) model was subsequently applied, and bias was reevaluated.
MEASUREMENTS AND MAIN RESULTS: A Bland-Altman analysis comparing the measurement of AT with RT showed a bias of 1.01°C (1.82°F) and 95% limits of agreement (LOA) -0.66°C to 2.68°C (-1.19°F to 4.83°F) in dogs and a bias of 0.23°C (0.42°F) and 95% LOA of -2.61°C to 3.08°C (-4.71°F to 5.55°F) in cats. Animals weighing <10 kg had a bias of 0.33°C (95% LOA: -2.29°C to 2.96°C), animals with BCS <5 had a bias of 0.09°C (95% LOA: -2.92°C to 3.11°C), and those with shorter haircoats had a bias of 0.49°C (95% LOA: -2.03°C to 3.00°C). The ML model overestimated RT when using AT and physical parameters in its algorithm (dogs: -0.57°C [95% LOA: -2.18°C to 0.99°C]; cats -0.84°C [95% LOA: -2.62°C to 0.93°C]).
CONCLUSIONS: Animals weighing <10 kg, with lower BCS, and shorter haircoats had less bias when AT was compared with RT. Cats exhibited a lower bias between AT and RT than dogs. ML models can be programmed to account for various physical characteristics, improving the predictive impact of AT for directly measured RT, especially in categories of animals where closer prediction does not already exist.