Abstract
The purpose of this study was to determine if the temperatures obtained using a veterinary infrared (IR) thermometer agreed with a digital rectal thermometer in a group of research cats, half of which had transient fevers. The thermometers were weakly correlated (r=0.62). The mean difference was 0.13°F (0.07°C), and the limits of agreement were 2.6°F (1.43°C) and −2.5°F (−1.36°C), which were unacceptable for clinical purposes. The results of this study indicate that, while the IR thermometer was easy to use, it cannot be used interchangeably with the rectal thermometer.
Introduction
Measurement of body temperature is a routine part of the clinical assessment of a patient. Traditionally in veterinary medicine, rectal temperature measurement has been the most common method of obtaining a patient’s body temperature. However, obtaining a rectal temperature can be stressful and potentially injurious to the patient and veterinary staff, as well as time consuming and a source of cross-contamination.1–,3
Core body temperature can be defined as the blood’s temperature in the pulmonary artery or brain. When body temperature is measured rectally, accuracy and repeatability of the measurement can be affected by the distance from the upper thorax, muscular exertion, and the presence of feces.4 Conversely, the tympanic membrane is well vascularized by the internal carotid arteries that also perfuse the hypothalamus, the gland responsible for body-temperature regulation.4 Physiological temperature differences exist between different body sites in humans; tympanic and esophageal temperatures are considered to be equivalent to core body temperature, while rectal temperature may sometimes exceed core temperature or lag significantly from these other body sites.5 Using the tympanic membrane and the deep, or inner, external ear canal, infrared (IR) thermometry may be a promising alternative to rectal temperature measurement in veterinary medicine.
Infrared thermometry was introduced in the mid1980s and was originally developed for human use. This technology utilizes pyroelectric sensors to detect the heat emanating from the tympanic membrane and inner surface of the external ear canal to obtain what is considered to be a more accurate measurement of core body temperature.2,6 Infrared thermometers that were originally developed to measure the tympanic and inner external ear canal temperature in humans may give inaccurate or inconsistent results when used in animals because of the wide angle of view, the limited depth of insertion, and the difference in ear canal anatomy in animals as compared to humans.5–,7 Because the anatomy of the external ear canal in veterinary patients differs from that of humans, an IR tympanic thermometer designed for human use measures primarily the external ear canal temperature of the veterinary patient rather than the tympanic membrane temperature, thereby decreasing the accuracy of the device.6
Several IR thermometers have been developed specifically for veterinary use. Infrared thermometry provides the veterinarian with a noninvasive alternative to rectal temperature monitoring, resulting in a less stressful event for the patient, better patient compliance, and quicker results than provided by a rectal thermometer. The purpose of this study was to determine if the temperature measurement obtained using a specific veterinary IR thermometer agreed with the temperature measurement obtained using a digital rectal thermometer in febrile and afebrile research cats.
Materials and Methods
Animals
This study used 19 research cats (eight spayed females, 11 neutered males) ranging in age from 2 to 10 years, without gross clinical signs of ear disease or deformity in both ears. The cats were part of a colony of flea-allergic cats handled regularly and housed in this facility for ≥2 years. They were participating in a concurrent study in which nine of the cats received a therapy that caused transient fevers following administration; however, body temperatures returned to normal within 24 hours. The remaining 10 cats received the vehicle only and did not experience transient fevers.
This project was approved by the Institutional Animal Care and Use Committee, and all animals were cared for and housed in facilities according to the principles outlined in the Guide for the Care and Use of Laboratory Animals.8 The facilities maintained a room temperature of 70° to 76°F (21° to 24.4°C) and a 12-hour light/dark cycle. Water and food were available free choice.
Temperature Measurement
A commercially available IR thermometera was used in this study. Ear or IR temperature was defined as the temperature reading given by the IR thermometer when it detected IR energy emanating from the tympanic membrane and the surrounding inner external ear canal. The ear temperature of each cat was obtained using the IR thermometer immediately before or after rectal temperature measurement. Standard operating procedure and instructions specific for the IR thermometer were closely followed. The probe cover was changed between each temperature reading, and the optical window was inspected to make sure that it was clean and intact before each use. Ear temperature was measured until two short audible beeps were heard, indicating that an endpoint reading had been obtained. The IR thermometer registers an endpoint reading in one second and self-calibrates between each measurement by reading the temperature of the internal reference target before the arm is opened.b This thermometer can be adjusted to read temperatures in Fahrenheit (°F) or Celsius (°C), and for this study all temperatures were recorded as °F. This IR thermometer, although designed for veterinary use, does not provide a suggested temperature reference range for its use in the cat.
To measure rectal temperature, two identical digital thermometersc were used interchangeably. The digital thermometers, which read temperatures as °F only, were validated using a water bathd at 98.6°, 100.0°, 104.0°, and 108.0°F (37.0°, 37.7°, 40.0°, 42.2°C) and a reference mercury-in-glass thermometer certified by the National Bureau of Standards. The probe of the digital thermometer was cleaned with an alcohol sponge and coated with lubricante prior to each use. The thermometer was inserted a minimum of 3 cm into the rectum, and the rectal temperature was measured until an audible beep was heard, indicating that an endpoint reading had been obtained. The digital thermometers register an endpoint reading in >90 seconds according to manufacturer specifications.
Temperatures were measured with both thermometers at set times (2, 5, 8, 10, and 13 hours) after drug or vehicle administration over several treatments. The cats were not restrained as the ear temperature was obtained, and they were minimally restrained for the rectal measurement. The order of temperature collection (rectal versus ear) was random. The ear temperatures were measured in only one ear, which was random based on the position of the cat at the time. Five technicians measured temperatures; however, the same technician measured ear or rectal temperatures at the same time every week. Fever was defined as a temperature >102.5°F (39.17°C).9
Statistics
The temperatures were compared by calculating the difference in the temperature recorded by each thermometer (rectal minus IR). The mean and standard deviation (SD) were calculated, and the results were plotted.f At that time, the limits of agreement10 (mean±1.96 SD) were determined, and the results were evaluated for agreement and clinical acceptability. For the purposes of this study, a difference >1.5°F (0.83°C) was deemed outside the range of clinical acceptability (results of informal survey of internists, data not shown), which means if the thermometers differed, on average, >1.5°F, then the difference was too great to consider the thermometers interchangeable. All data was reported as mean±SD, unless otherwise specified.
Though the correlation coefficient (r) is often used in similar studies, it is not an appropriate analysis when comparing two methods of measurement.10 However, it is often expected by clinicians in these situations. Therefore, Pearson’s product-moment correlation coefficientg was used to evaluate correlation between the two thermometers in this study. An r ≥0.75 was considered positive (>75% of the time the thermometers report similar values).
Results
There were 555 times when temperatures were collected from both the rectal and IR thermometers. There were 15 times when the cats objected to repeated rectal temperature collection and a recording was not possible, but there were no times when the cats objected to the IR thermometer.
The maximum temperatures recorded by the rectal and IR thermometers were 105.1° and 106.5°F (40.61° and 41.39°C), respectively. The febrile cats did not exhibit clinical signs of fever. Some vomiting occurred at various times but did not appear to be related to time, treatment, or temperature.
The mean difference in temperature (rectal minus IR) was 0.13°±1.25°F (0.07°±0.69°C). The maximum differences were −4.7° and +4.4°F (−2.61° and +2.44°C). The 95% confidence interval was 0.03° to 0.24°F (0.02° to 0.13°C), which means the IR thermometer tended to be lower by 0.03° to 0.24°F (0.02° to 0.13°C). The limits of agreement (1.96 ×SD) were −2.32° to +2.58°F [Figure 1⇓] or −1.36° to +1.43°C. This was clinically unacceptable, as ±1.5°F (±0.83°C) was selected as the acceptable limits of agreement. Therefore, there is a lack of agreement between the two thermometers.
A plot of the difference between the rectal and infrared (IR) thermometer against their average temperature (°F), with the mean (——) and limits of agreement (mean±1.96 standard deviation [SD]) indicated (- - - -).
There was a weak positive correlation (r=0.62) between the two thermometers [Figure 2⇓]. When the thermometers were compared at temperatures considered to be febrile (>102.5°F or >39.2°C), the correlation was even weaker. When the rectal thermometer detected a fever (n=94), the r was 0.36; when the IR thermometer detected a fever (n=73), the r was 0.24.
Scatterplot showing a weak positive correlation coefficient (r=0.62) between rectal and IR temperatures (°F).
Of the 555 temperature readings with both thermometers, there were 94 observations of fever (>102.5°F or >39.2°C) with the rectal digital thermometer and 73 observations of fever with the IR thermometer. In the 94 cases where fever was noted by the digital thermometer, the IR thermometer differed by ±1.5°F (0.83°C) a total of 21 times, and no fever was detected by the IR thermometer 46 times. In the 73 fever observations by the IR thermometer, the rectal thermometer differed by ±1.5°F (0.83°C) a total of 12 times, and the digital rectal thermometer recorded a normal temperature 25 times.
Discussion
In this study using febrile and afebrile research cats, it was found that the IR thermometer designed for veterinary use and a digital rectal thermometer did not agree. The temperatures from the IR thermometer tended to be lower by 0.03° to 0.24°F (0.02° to 0.13°C). This does not imply that the IR thermometer is inaccurate. It is crucial to realize that body temperature varies by location,4 and these thermometers do not measure temperature at the same location, so they might be expected to differ. It is therefore important that the same type of thermometer be used consistently in each patient.
It is perhaps most critical that a temperature reading be accurate when an animal is febrile. This is important so the veterinarian can know not only the presence of a fever but also its severity. It was known that half of the cats from this study were receiving a therapy that would cause a transient fever. Therefore, there was a wide range of temperatures measured that allowed better assessment of the value of this thermometer. When the thermometers were compared at febrile temperatures, they differed more than when they were compared over all of the temperature measurements.
Some differences may be a result of operator error. Operator technique may have an important role in the performance of the IR thermometer. The probe must be positioned in the ear canal according to manufacturer’s directions to record the infrared emissions from the tympanic membrane. If the probe is not positioned properly, the skin temperature of the surrounding external ear will be measured instead of the tympanum and the inner external ear canal, resulting in a lower temperature measurement. Though the current study did not attempt to compare or control user variability, it has been documented elsewhere11 as a problem with this type of thermometer.
The differences observed between the two thermometers may represent a true physiological difference between body sites rather than operator error or inaccuracy on the part of the IR thermometer.5 In previous studies evaluating IR thermometry in animals, thermometers developed for human ear temperature measurement were used.1,6,12–,15 These thermometers have a conversion factor incorporated, which converts a tympanic reading to an estimated rectal or oral equivalent reading. These conversion factors are derived from comparative data determined through research in humans and may not be appropriate for animal subjects.5 The IR thermometer used in this study does not have a conversion factor incorporated; it records body temperature in the actual tympanic temperature-reading mode. The lack of a rectal or oral equivalent mode in the IR thermometer used in this study may cause a detectable difference in ear and rectal temperature measurements due to physiological body site temperature differences. Because the IR thermometer is used in the actual tympanic temperature-reading mode, an appropriate reference range for this body site in cats needs to be established.
It is also possible that the IR thermometer detected changes in body temperature before the rectal thermometer. It has been found in previous studies that changes in rectal temperature may lag significantly from the temperature detected at other body sites, such as the tympanum.5
Excitement and agitation from the rectal temperature measurement may have caused temperature elevation due to activity and muscular exertion resulting in a local temperature increase in the rectum.4,12 The IR thermometer may have registered a mild increase in temperature likely due to heightened activity; however, the tympanic membrane is not likely to be subject to a local increase in temperature due to muscular exertion.
Although many human clinical studies have been conducted to explore the use and limitations of IR thermometry, there have been only a few studies reporting the use of IR thermometry in dogs and cats.1,6,12–,15 In one feline study,12 it was found that the repeatability of tympanic thermometry was low and that inaccuracy of temperature measurement using this technique could be clinically significant. In the present study, there was no attempt to repeat readings at any one time point on any one cat, to average left and right ear canal temperatures, or to clean the ears before measurement. These practices would have defeated the assessment of the practicality and usefulness of the IR thermometer in the clinical setting.
In another feline study,1 an IR thermometer was compared to rectal mercury in glass thermometers in a clinical setting using feline inpatients and outpatients with various medical conditions. A human IR thermometer was used that automatically adjusted the reading to an equivalent oral reading in humans.1 Although that study found a correlation between temperatures taken from the ear and those taken from the rectum (r=0.995), most cats were without fever.1
Ease of use, quicker results, better patient compliance, and decreased cross contamination are frequently cited reasons for clinical use of tympanic IR thermometry in human and veterinary medicine.1–3,4,6,12,16 These features make IR thermometry a promising alternative in veterinary medicine as opposed to the traditional rectal temperature measurement. In the current study, there were no objections to repeated use of the tympanic IR measurement, while there were 15 occasions during which cats objected to repeated rectal temperature measurements.
Conclusion
This study was performed in a population of research cats, half of which were known to have transient fevers. This provided a wide range of temperatures over which to compare the veterinary IR thermometer to the standard rectal thermometer. The limits of agreement were not acceptable clinically, and the two thermometers did not correlate well. This does not imply that the IR thermometer is inaccurate, but that the two thermometers cannot be used interchangeably. Although IR thermometry is a promising alternative to rectal thermometry in cats, a reference range in cats for this body site and thermometer needs to be established.
Acknowledgments
The authors thank Heska, Inc., Fort Collins, Colorado for the financial support of the cat colony as well as Anthony Pilny, Jennifer Nerbonne, and Tara Clark for their technical assistance.
Footnotes
-
↵ a Model VT-100 Veterinary Infrared Tympanic Temperature Scanner; Advanced Monitors Corporation, San Diego, CA
-
↵ b Model VT-100 Veterinary Infrared Tympanic Temperature Scanner Instruction Manual 1999:1–12; Advanced Monitors Corporation, San Diego, CA
-
↵ c Model 1681 - BMWF, lot #9901 digital thermometers; Florida Medical Industries, Inc., Leesburg, FL
-
↵ d Model 939 Gyromax Reciprocating Water Bath Shaker; Amerex Instruments, Inc., Lafayette, CA
-
↵ e K-Y lubricant; Johnson & Johnson Consumer Products, Inc., Skillman, NJ
-
↵ f Microsoft Excel 97; Microsoft Corporation, Mountain View, CA
-
↵ g SAS System for Windows, version 8.2; SAS Institute, Cary, NC
