Tympanic Membrane Temperatures Compared to Rectal and Oral Temperatures Haitham Talo, MD*, Michael L. Macknin, MD*, Sharon
VanderBrug Medendorp, MPH†
One hundred thirty-seven children at the Cleveland Clinic Foundation were enrolled in a study the accuracy and acceptability of tympanic membrane temperatures taken with Thermoscan® to rectal or oral temperatures taken by IVAC® electronic thermometers. The mean age of the rectal/ear group was 1.2 ± 0.86 years (range 0.08- 5.0 years) with 22 females and 21 males. The mean age of the oral and ear group was 9.0 ± 4.24 years (range 3- 18 years) with 44 females and 50 males. Significantly large (p < 0.01) correlation coefficients were present between temperatures in the left ear and right ear (0.92), rectal and mean ear (0.77) and oral and mean ear (0.68). The temperature difference between mean ear and rectal measurements was -1.1° ± 0.51 °C and the temperature difference between the mean ear and oral measurements was -0.20° ± 0.59°C. A relative lack of correlation between ear and oral temperatures was noted with flat tympanograms, otoscopically diagnosed middle ear effusion, and with red-white or red tympanic membranes. However, small subgroup sizes (6-16) may have contributed to this lack of correlation. Parents and physicians were impressed by the 2.5 second speed of the new technique. However, younger children were nearly as likely to resist and cry with ear as with rectal temperatures. Parents were comfortable with the ease, cleanliness and safety of ear temperatures. While these data indicate a strong correspondence of ear temperatures with oral and rectal temperatures, parents were not as confident of ear temperatures’ accuracy. Further public education may be necessary to alleviate minor concerns about the accuracy of tympanic membrane temperature measurements.
comparing
Introduction Measurement of body temperature is
Thermoscan(R) is one such device developed by Thermoscan a
routine part of
clinical assessment, and, particularly in pediatric patients, accurate detection of body temperature provides useful clues about the severity of the illness. Core body sites have long been sought to obtain body temperature due to the role the central nervous system thermoreceptors (located in the hypothalamus) play in temperature homeostasis. Since the tympanic membrane shares the same vascular supply that perfuses the hypothalamus, it is considered an excellent readily accessible site for core temperature measurement. Several different devices which use an ear probe to determine tympanic membrane temperature have been invented, marketed and tested in vivo to assess their accuracy. 1,2,3,4
*The Cleveland Clinic Foundation, Department of Pediatric and Adolescent Medicine, †department of Biostatistics and Epidemiology.
30
Inc. (San Diego, California) to measure tympanic membrane temperature. This portable instant thermometer, which had not been studied in previously published literature, measures the body temperature by detecting the thermal infrared energy that is naturally emitted from the tympanic membrane without any direct contact with the membrane. This new way to detect temperature is fast, safe, non-invasive, clean and, if proven to be accurate, would provide an excellent alternative to rectal and oral
temperatures.
Objectives
and Methods
Objectives The objectives of this study were to: 1) Determine the equivalency of Thermoscan@ tympanic membrane tem-
perature
measurements and
IVAC® rectal and oral
tem-
peratures in pediatric patients, 2) to determine the effect of middle ear disease on tympanic membrane temperatures, and 3) to determine physician, parent and patient acceptance of tympanic membrane temperatures versus rectal and oral temperatures.
Table 1
· Descriptive Information
on
Temperature
Values
Methods Patients under age 18 years seen from October 25, 1990 18, 1990 at the Cleveland Clinic Foundation were consecutively enrolled in the study by the primary to November
investigator. Demographic data obtained on each patient included name, age, sex and clinic number. Clinical information obtained included: 1) TherrnoscanO tympanic membrane temperature (non-corrected) from both ears by the principal investigator blinded to results of oral or rectal temperature; 2) rectal or oral temperature (depending on patient cooperation) using an IV AC@ on predict mode performed by a pediatric nurse blinded to tympanic membrane temperature (rectal or oral temperature was taken before tympanic membrane temperatures in patients with even clinic numbers, and after tympanic membrane temperatures in patients with odd clinic numbers); 3) results of tympanograms performed by the principal investigator using a G.S.I. 27A Auto-tymp° tympanometer; 4) results of pneumatic otoscopy by the principal investigator; 5) physician, parent, and, if possible, patient evaluation of temperature taking methods. All thermometers used in the study were calibrated at the beginning and midpoint of the study according to manufacturer’s specifications. The study was approved by the Institutional Review Board and informed verbal consent was obtained.
Statistical
Analysis
Temperature values taken on the same patient by different methods were compared using the paired t-test.1 Characteristics of the two second temperature type groups (oral and rectal) were examined using t-tests and chi-square tests.5~6 Pearson correlation coefficients were used to 7 assess the relationship of the various temperature values. The parental evaluations of the three methods consisted of a rating system of speed, ease, cleanliness, safety, and accuracy from 1 (best) to 5 (worst). These characteristics were compared for the different methods using McNemar’ss teSt.8
Results A total of 137 children were enrolled in the study. Fortythree (31 %) children had rectal temperatures taken and 94 (69%) had oral temperatures taken. There were 22 (51 %) females and 21 (49%) males in the rectal group and 44
(47%) females and 50 (53%) males in the oral group. The rectal temperature group was significantly younger ( 1.2 ± 0.86 years) (range 0.08 - 5 years) than the oral temperature group (9 + 4.24 years) (range 3 - 18 years). Mean ear temperature was 36.6’ + 0.73°C, mean rectal temperature was 37.7° + 0.54°C, and mean oral temperature was 36.8’ + 0.76°C. The mean temperature difference between the mean ear and rectal was -1.1 ° + 0.51 °C, and between mean ear and oral was -0.20’ ± 0.59°C (Table 1). There were no significant differences in the discrepancy between ear and rectal or oral temperatures with temperatures < 38°C vs > 38°C. Temperature differences between ear and rectal or oral measurements tended to decrease in the 10 patients with rectal and five patients with oral temperatures > 38°C. This trend was not statistically significant, possibly due to the small number of febrile patients in the sample. Intra-patient correlation between left and right ear temperature measurements was extremely strong 0.915 (p < .owl). Because these measurements were not significantly different, they were averaged and used to compare to the oral and rectal temperatures. Both the rectal and oral temperatures were significantly correlated with mean ear temperature. Correlation for the ear and rectal temperatures was 0.765 (p < .O1 ). Correlation for the ear and oral temperatures was 0.682 (p < .01). When the correlation between temperatures was examined within categories of various diagnostic tests, weak correlation in the ear and oral group was found when there was a flat tympanogram ( R ear 0.53, L ear -0.19), otoscopically diagnosed middle ear effusion (R ear 0.64, L ear 0.32) and the presence of a red or red-white tympanic membrane (R ear 0.51, L ear 0.14). However, small numbers of patients (6-16) in the subgroups may have contributed to a lack of significant correlation. This lack 31
Table 2 · Parental Evaluation of
Temperature Taking Methods
of temperature correlation with the presence of middle ear disease did not occur between ear and rectal temperatures. Comparison of patient acceptance of temperature-taking methods as judged by the parent was unaffected by whether tympanic membrane or rectal and oral temperatures were performed first. Older children accepted both oral and ear temperatures, with only 9 out of 94 (9.6%) resisting or being afraid of each method. Younger children were less likely to accept either ear or rectal temperatures with the ear method having slightly higher acceptance. With ear temperatures, 3 out of 43 (7.0%) resisted and 11I out of 43 (25.6%) were afraid. With rectal temperatures, 4 out of 43 (9.3%) resisted and 13 out of 43 (30.2%) were
Tatale 3 · Parental Evaluation of
quick (1) easy (1) accurate (1) clean (1) safe (1 ) -
32
...
...
...
...
...
lengthy (5) difficult (5) not accurate unclean (5) unsafe (5)
(5)
afraid. The new device was well accepted by the physician involved in obtaining ear temperatures. Parents felt that the ear method was significantly quicker (p < .01) than rectal or oral methods. The physician recorded that the time it took to take the ear temperature was 2.5 ± 1.07 seconds. The time necessary to take oral or rectal temperatures ranged from 30-50 seconds according to an informal survey of nursing staff involved in the study. Parents felt that the ear, oral and rectal temperature methods were comparably clean and safe. Most parents thought the ear method was easier to use than the oral (p < 0.01) and rectal (p < 0.15) methods. In both the oral and rectal groups, there was some feeling that the ear method was not quite as
Temperature Taking Methods
accurate, but this was not statistically significant (p >
0.30)
(Tables 2,3). Discussion and
Summary
Temperature determination is an important factor in clinical assessment. The study demonstrated a reasonable correlation between tympanic membrane temperature obtained by Thermoscan@ and oral/rectal temperature taken by IVAC®. Unfortunately the number of febrile patients was not sufficient to accurately determine the performance of tympanic membrane temperature in febrile patients. However, Kenney et al.1 used FirstTemp TMT device on 964 patients, 30% of them were febrile. He showed good performance of the ear method in febrile patients (79% sensitivity, 74% specificity). Our results raised the concern in patients with middle ear disease about the relatively weak correlation between ear method values and oral method values. However, previous studies with up to 171 patients with middle ear disease demonstrated no such problems.9- ’I The weak correlation in our study may be due to the small number of patients in this group. Determining the influence of wax and exudate on the obtained ear temperature was one of the objectives of the study, but the recruited pertaining numbers of patients were insufficient to comment upon. Although another study9 showed no effect of these factors on the obtained ear
7. Pearson K. Notes on the history of correlation. Biometrika 1920; 13:25-45. 8. McNemar Q. Note on the sampling error of the difference between correlated proportions or percentages. Psychometrica 1947; 12:1537. 9. Kenney RD, Fortenberry JD, Surratt SS, et al. Evaluation of an infrared tympanic membrane thermometer in pediatric patients. Pediatr 1990; 85:854-8. 10. WeirMR,WeirTE. Are "hot" ears really hot? AJDC 1989; 143 :7634.
temperature. The effect of examiner variability on ear temperature values was not significant in our study because one investigator obtained all ear values. The effect of multiple observers was one of the disadvantages noted in other studies.3-9 In addition, the ear temperature method was well accepted by the general population and the physician, although further studies and education are needed to alleviate mild public concern in regard to accuracy of tympanic membrane temperature measurements.
References 1. Rhoads RA, Grandner J. Assessment of an aural infrared sensor for body temperature measurement in children. Clin Pediatr 1990;29:1125. 2. Green MM, Danzel DF, Praszkier H. Infrared tympanic thermography in the emergency department. J Emerg Med 1989; 7:437-40. 3. Ros SP. Evaluation of a tympanic membrane thermometer in an outpatient clinical setting. Ann Emerg Med 1989; 18:1004-6. 4. Temdrup TE, Allegra JR, Kealy JA. A comparison of oral, rectal, and tympanic membrane-derived temperature changes after ingestion of liquids and smoking. Am J Emerg Med 1989; 7:150-4. 5. Zar JH. Biostatistical Analysis. Prentice-Hall, 1984; 126-30. 6. Fisher RA. On the interpretation of chi-square from contingency tables and the calculation of p. J Royal Statistical Association 1922; 85:87-9.
33
VanderBrug Medendorp, MPH†
One hundred thirty-seven children at the Cleveland Clinic Foundation were enrolled in a study the accuracy and acceptability of tympanic membrane temperatures taken with Thermoscan® to rectal or oral temperatures taken by IVAC® electronic thermometers. The mean age of the rectal/ear group was 1.2 ± 0.86 years (range 0.08- 5.0 years) with 22 females and 21 males. The mean age of the oral and ear group was 9.0 ± 4.24 years (range 3- 18 years) with 44 females and 50 males. Significantly large (p < 0.01) correlation coefficients were present between temperatures in the left ear and right ear (0.92), rectal and mean ear (0.77) and oral and mean ear (0.68). The temperature difference between mean ear and rectal measurements was -1.1° ± 0.51 °C and the temperature difference between the mean ear and oral measurements was -0.20° ± 0.59°C. A relative lack of correlation between ear and oral temperatures was noted with flat tympanograms, otoscopically diagnosed middle ear effusion, and with red-white or red tympanic membranes. However, small subgroup sizes (6-16) may have contributed to this lack of correlation. Parents and physicians were impressed by the 2.5 second speed of the new technique. However, younger children were nearly as likely to resist and cry with ear as with rectal temperatures. Parents were comfortable with the ease, cleanliness and safety of ear temperatures. While these data indicate a strong correspondence of ear temperatures with oral and rectal temperatures, parents were not as confident of ear temperatures’ accuracy. Further public education may be necessary to alleviate minor concerns about the accuracy of tympanic membrane temperature measurements.
comparing
Introduction Measurement of body temperature is
Thermoscan(R) is one such device developed by Thermoscan a
routine part of
clinical assessment, and, particularly in pediatric patients, accurate detection of body temperature provides useful clues about the severity of the illness. Core body sites have long been sought to obtain body temperature due to the role the central nervous system thermoreceptors (located in the hypothalamus) play in temperature homeostasis. Since the tympanic membrane shares the same vascular supply that perfuses the hypothalamus, it is considered an excellent readily accessible site for core temperature measurement. Several different devices which use an ear probe to determine tympanic membrane temperature have been invented, marketed and tested in vivo to assess their accuracy. 1,2,3,4
*The Cleveland Clinic Foundation, Department of Pediatric and Adolescent Medicine, †department of Biostatistics and Epidemiology.
30
Inc. (San Diego, California) to measure tympanic membrane temperature. This portable instant thermometer, which had not been studied in previously published literature, measures the body temperature by detecting the thermal infrared energy that is naturally emitted from the tympanic membrane without any direct contact with the membrane. This new way to detect temperature is fast, safe, non-invasive, clean and, if proven to be accurate, would provide an excellent alternative to rectal and oral
temperatures.
Objectives
and Methods
Objectives The objectives of this study were to: 1) Determine the equivalency of Thermoscan@ tympanic membrane tem-
perature
measurements and
IVAC® rectal and oral
tem-
peratures in pediatric patients, 2) to determine the effect of middle ear disease on tympanic membrane temperatures, and 3) to determine physician, parent and patient acceptance of tympanic membrane temperatures versus rectal and oral temperatures.
Table 1
· Descriptive Information
on
Temperature
Values
Methods Patients under age 18 years seen from October 25, 1990 18, 1990 at the Cleveland Clinic Foundation were consecutively enrolled in the study by the primary to November
investigator. Demographic data obtained on each patient included name, age, sex and clinic number. Clinical information obtained included: 1) TherrnoscanO tympanic membrane temperature (non-corrected) from both ears by the principal investigator blinded to results of oral or rectal temperature; 2) rectal or oral temperature (depending on patient cooperation) using an IV AC@ on predict mode performed by a pediatric nurse blinded to tympanic membrane temperature (rectal or oral temperature was taken before tympanic membrane temperatures in patients with even clinic numbers, and after tympanic membrane temperatures in patients with odd clinic numbers); 3) results of tympanograms performed by the principal investigator using a G.S.I. 27A Auto-tymp° tympanometer; 4) results of pneumatic otoscopy by the principal investigator; 5) physician, parent, and, if possible, patient evaluation of temperature taking methods. All thermometers used in the study were calibrated at the beginning and midpoint of the study according to manufacturer’s specifications. The study was approved by the Institutional Review Board and informed verbal consent was obtained.
Statistical
Analysis
Temperature values taken on the same patient by different methods were compared using the paired t-test.1 Characteristics of the two second temperature type groups (oral and rectal) were examined using t-tests and chi-square tests.5~6 Pearson correlation coefficients were used to 7 assess the relationship of the various temperature values. The parental evaluations of the three methods consisted of a rating system of speed, ease, cleanliness, safety, and accuracy from 1 (best) to 5 (worst). These characteristics were compared for the different methods using McNemar’ss teSt.8
Results A total of 137 children were enrolled in the study. Fortythree (31 %) children had rectal temperatures taken and 94 (69%) had oral temperatures taken. There were 22 (51 %) females and 21 (49%) males in the rectal group and 44
(47%) females and 50 (53%) males in the oral group. The rectal temperature group was significantly younger ( 1.2 ± 0.86 years) (range 0.08 - 5 years) than the oral temperature group (9 + 4.24 years) (range 3 - 18 years). Mean ear temperature was 36.6’ + 0.73°C, mean rectal temperature was 37.7° + 0.54°C, and mean oral temperature was 36.8’ + 0.76°C. The mean temperature difference between the mean ear and rectal was -1.1 ° + 0.51 °C, and between mean ear and oral was -0.20’ ± 0.59°C (Table 1). There were no significant differences in the discrepancy between ear and rectal or oral temperatures with temperatures < 38°C vs > 38°C. Temperature differences between ear and rectal or oral measurements tended to decrease in the 10 patients with rectal and five patients with oral temperatures > 38°C. This trend was not statistically significant, possibly due to the small number of febrile patients in the sample. Intra-patient correlation between left and right ear temperature measurements was extremely strong 0.915 (p < .owl). Because these measurements were not significantly different, they were averaged and used to compare to the oral and rectal temperatures. Both the rectal and oral temperatures were significantly correlated with mean ear temperature. Correlation for the ear and rectal temperatures was 0.765 (p < .O1 ). Correlation for the ear and oral temperatures was 0.682 (p < .01). When the correlation between temperatures was examined within categories of various diagnostic tests, weak correlation in the ear and oral group was found when there was a flat tympanogram ( R ear 0.53, L ear -0.19), otoscopically diagnosed middle ear effusion (R ear 0.64, L ear 0.32) and the presence of a red or red-white tympanic membrane (R ear 0.51, L ear 0.14). However, small numbers of patients (6-16) in the subgroups may have contributed to a lack of significant correlation. This lack 31
Table 2 · Parental Evaluation of
Temperature Taking Methods
of temperature correlation with the presence of middle ear disease did not occur between ear and rectal temperatures. Comparison of patient acceptance of temperature-taking methods as judged by the parent was unaffected by whether tympanic membrane or rectal and oral temperatures were performed first. Older children accepted both oral and ear temperatures, with only 9 out of 94 (9.6%) resisting or being afraid of each method. Younger children were less likely to accept either ear or rectal temperatures with the ear method having slightly higher acceptance. With ear temperatures, 3 out of 43 (7.0%) resisted and 11I out of 43 (25.6%) were afraid. With rectal temperatures, 4 out of 43 (9.3%) resisted and 13 out of 43 (30.2%) were
Tatale 3 · Parental Evaluation of
quick (1) easy (1) accurate (1) clean (1) safe (1 ) -
32
...
...
...
...
...
lengthy (5) difficult (5) not accurate unclean (5) unsafe (5)
(5)
afraid. The new device was well accepted by the physician involved in obtaining ear temperatures. Parents felt that the ear method was significantly quicker (p < .01) than rectal or oral methods. The physician recorded that the time it took to take the ear temperature was 2.5 ± 1.07 seconds. The time necessary to take oral or rectal temperatures ranged from 30-50 seconds according to an informal survey of nursing staff involved in the study. Parents felt that the ear, oral and rectal temperature methods were comparably clean and safe. Most parents thought the ear method was easier to use than the oral (p < 0.01) and rectal (p < 0.15) methods. In both the oral and rectal groups, there was some feeling that the ear method was not quite as
Temperature Taking Methods
accurate, but this was not statistically significant (p >
0.30)
(Tables 2,3). Discussion and
Summary
Temperature determination is an important factor in clinical assessment. The study demonstrated a reasonable correlation between tympanic membrane temperature obtained by Thermoscan@ and oral/rectal temperature taken by IVAC®. Unfortunately the number of febrile patients was not sufficient to accurately determine the performance of tympanic membrane temperature in febrile patients. However, Kenney et al.1 used FirstTemp TMT device on 964 patients, 30% of them were febrile. He showed good performance of the ear method in febrile patients (79% sensitivity, 74% specificity). Our results raised the concern in patients with middle ear disease about the relatively weak correlation between ear method values and oral method values. However, previous studies with up to 171 patients with middle ear disease demonstrated no such problems.9- ’I The weak correlation in our study may be due to the small number of patients in this group. Determining the influence of wax and exudate on the obtained ear temperature was one of the objectives of the study, but the recruited pertaining numbers of patients were insufficient to comment upon. Although another study9 showed no effect of these factors on the obtained ear
7. Pearson K. Notes on the history of correlation. Biometrika 1920; 13:25-45. 8. McNemar Q. Note on the sampling error of the difference between correlated proportions or percentages. Psychometrica 1947; 12:1537. 9. Kenney RD, Fortenberry JD, Surratt SS, et al. Evaluation of an infrared tympanic membrane thermometer in pediatric patients. Pediatr 1990; 85:854-8. 10. WeirMR,WeirTE. Are "hot" ears really hot? AJDC 1989; 143 :7634.
temperature. The effect of examiner variability on ear temperature values was not significant in our study because one investigator obtained all ear values. The effect of multiple observers was one of the disadvantages noted in other studies.3-9 In addition, the ear temperature method was well accepted by the general population and the physician, although further studies and education are needed to alleviate mild public concern in regard to accuracy of tympanic membrane temperature measurements.
References 1. Rhoads RA, Grandner J. Assessment of an aural infrared sensor for body temperature measurement in children. Clin Pediatr 1990;29:1125. 2. Green MM, Danzel DF, Praszkier H. Infrared tympanic thermography in the emergency department. J Emerg Med 1989; 7:437-40. 3. Ros SP. Evaluation of a tympanic membrane thermometer in an outpatient clinical setting. Ann Emerg Med 1989; 18:1004-6. 4. Temdrup TE, Allegra JR, Kealy JA. A comparison of oral, rectal, and tympanic membrane-derived temperature changes after ingestion of liquids and smoking. Am J Emerg Med 1989; 7:150-4. 5. Zar JH. Biostatistical Analysis. Prentice-Hall, 1984; 126-30. 6. Fisher RA. On the interpretation of chi-square from contingency tables and the calculation of p. J Royal Statistical Association 1922; 85:87-9.
33
