D Myny, J De Waele, *T Defloor, S Blot, F Colardyn Department of intensive care, Ghent University Hospital, Ghent, Belgium *Faculty of medicine and health sciences, University of Ghent, Belgium
Correspondence to: Dries Myny, Getouwstraat 3, 8800 Roeselare, Belgium Tel: 09/240.64.12 Fax: 09/240.50.21
email: dries.myny@uzgent.be
SMJ 2005 50(1): 15-18
Abstract:
Background and Aims: Temperature measurement is a routine task of patient care, with considerable clinical impact, especially in the ICU. This study was conducted to evaluate the accuracy and variability of the Temporal Artery Thermometer (TAT) in ICU-patients. Therefore, a convenience sample of 57 adult patients, with indwelling pulmonary artery catheters (PAC) in a 40-bed intensive care unit in a university teaching hospital was used. Methods: The study design was a prospective, descriptive comparative analysis. Body temperature was thereby measured simultaneously with the TAT and the Axillary Thermometer (AT), and was compared with the temperature recording of the PAC. The use of vasoactive medication was recorded. Results and conclusions: Mean temperature of all measurements was: PAC: 37.1°C (SD: 0.87), TAT: 37.0°C (SD: 0.68) and axillary thermometer: 36.6°C (SD: 0.94). The measurements of the TAT and the PAC were not significantly different (mean difference: 0.14°C; SD: 0.51; p= 0.33); whereas the measurements of the PAC and the AT differed significantly (mean difference: 0.46°C; SD: 0.39; p< 0.001). Mean difference in PAC versus TAT analyses, between patients with vasopressor therapy (0.12°C; SD: 0.55), and without vasopressor therapy (0.19°C; SD: 0.48) was not statistically significant (p= 0.47). Conclusion:We can conclude that the temporal scanner has a relatively good reliability with an acceptable accuracy and variability in patients with normothermia. The results are comparable to those of the AT, but they do not seem to be sufficient to prove any substantial benefit compared to rectal, oral or bladder thermometry.
Key words: Body temperature, intensive care unit, temporal artery thermometer,pulmonary artery catheter, monitoring.
Introduction
Temperature measurement is a routine task of patient care, with considerable clinical impact, especially in the ICU. The ideal instrument should give a temperature close to the core body temperature, with an acceptable variability for everyday practice. It should be simple to use, fast and inexpensive. However at this moment, there is no perfect instrument for approximating core temperature in a non-invasive way.1 Recently a new technique of temperature measurement was developed. The Temporal scanner (Exergen Temporal scanner LXTA®, Exergen corporation, USA) is a non-invasive system of intermittent temperature measurement with advanced infrared technology and designed to measure the temperature at the skin surface.2
The objective of this study was to evaluate the accuracy and variability of the Temporal Artery Thermometry (TAT) in ICU patients, compared to the pulmonary artery catheter (PAC), which is the golden standard in temperature measurement. A prospective, descriptive comparative analysis was executed therefore.
Materials and methods
A convenience sample of 57 orally intubated patients, consecutively admitted at ICU during a five-month period (November 2001–March 2002), with an indwelling pulmonary artery catheter were included in this study. Age, gender and reason for admission to the ICU were recorded.
Data collection
Simultaneous temperature measurements were obtained from the PAC, the axillary thermometer and the TAT within three minutes. The time difference between temperature measurements on patients with repeated measurements was at least four hours. So, those measurements could be considered as independent measures. The measurements were performed by one of the authors (MD) and two, trained assistants. Axillary temperatures (AT) were measured with a digital electronic thermistor thermometer (Digital Classic® of Hartmann), according to the guidelines of the manufacturer. For the TAT, the Exergen Temporal Scanner LXTA® (Exergen corporation, USA) was used. Temperature is taken by slowly scanning the thermometer across the forehead over the temporal artery and behind the ear. The temporal artery is connected to the heart via the carotid artery, directly leading from the aorta. It offers constant blood flow. It is the only artery positioned close enough to the skin surface to provide access needed to take an accurate measurement. The thermometer measures the naturally emitted infrared heat from the arterial blood supply at about 1000 “frames” per second, locking in the highest temperature it senses.2 Pulmonary artery core temperature was measured using the temperature probe of a PAC (Thermodilution Catheter®, Baxter Health Care, Irvine, USA) and a cardiac output monitor (Edwards Lifesciences, Irvine, USA). Prior to the study, repeatability was tested in 26 consecutively admitted adult patients at the surgical ICU. A single investigator measured the temperature of the patient two times within one minute.
Statistical analysis
Statistical analysis of the data was performed with SPSS 10.0® (SPSS, Chicago, Ill., USA). The normality of the distribution of the temperature measurements of the PAC was confirmed by a Kolmogorov Smirnov-test, histogram and QQ-plot. The Pearson correlation coefficient and the Bland and Altman plot were used to perform intra- and inter-observer reliability. The Bland and Altman method was also used to compare TAT and AT versus PAC temperature.
MedCalc® (MedCalc, Mariakerke, Belgium) was used for Bland and Altman analysis.3
ANOVA, Tukey HSD test and student t-test was used to study the possible differences between the temperature measuring methods.
Statistical significance was defined as p-value < .05.
The device was supplied by the manufacturer for evaluation purposes. No other support was given for this study.
Results
The population of the pre-test (n=26) had a mean age of 63 years. Eighteen of them were men. The test-retest reliability was high (r=0.96; p=0.01). (Figs 1 and 2) In the study group, 34 out of the 57 patients were male. The mean age was 60 years (standard deviation (SD): 14.9). Thirty-two patients were admitted to the medical intensive care unit (MICU): 13 had sepsis and 19 had a variety of medical disorders. Twenty-five patients were admitted to the surgical intensive care unit (SICU): three trauma patients, 10 patients after a solid organ transplantation and 12 after other surgical interventions.Three hundred and eighteen measurements (106 triplicate measurements) were obtained; 29 patients were assessed once, 15 patients twice and 13 patients on three or more occasions.
The mean temperatures of the three measurement methods varied between 36.6 and 37.1°C (Table I) The three temperature measuring systems were significantly different (F=9.33; df= 2; p<0.001). Post hoc analysis revealed that the measurements of the TAT and the PAC were not significantly different, with a mean difference of 0.14°C (SD: 0.51; 95% CI: 0.04-0.23; p=0.33). The measurements of the PAC and the AT on the other hand (SD: 0.39; 95% CI: 0.39-0.54; p<0.001). The measurements of the TAT and the AT differed also significantly, with a mean difference of 0.36°C (SD: 0.58; 95% CI: 0.25-0.48; p= 0.01). Agreement between PAC and TAT was significantly better than between PAC and AT, while variability was comparable. The Bland and Altman plots confirm these results. (Figs 3 and 4) Peripheral vasoconstriction may influence the accuracy of the measurements with the temporal scanner, therefore we studied separately the measurements in patients with (n=33) and without (n=24) vasopressor therapy. The patients with vasopressor therapy had a mean offset of 0.12°C (SD: 0.55), whereas the other had 0.19°C (SD: 0.48). This was not statistically significant (p=.47).
Discussion
The mean difference of the temporal artery scanner was relatively good compared with the PAC readings, with mean offsets of 0.14°C, whereas the mean offset from the axillary thermometer was almost 0.5°C. The use of vasoactive medication was thereby not an influencing factor on the TAT values. This is very important in ICUs, because a lot of patients receive vasopressor therapy (33 of 57 patients in this study). The standard deviation was slightly higher for the TAT compared to the axillary thermometer with a SD of 0.51 versus 0.39. Variability is as important as accuracy because clinicians have to make decisions on single measurements. Those decisions-making can have implications on medical practice, like the time of starting empiric anti-infective therapy.
Overall accuracy is sufficient compared to the definition used by Giuliano et al4 and Robinson et al5 who defined accuracy as the mean difference from a standard of ± 0.3°C and a standard deviation ranging from 0.3°C to 0.5°C. The mean difference of the TAT in our study was good, the SD was just slightly higher. Our results, are comparable with the results of some other methods of temperature measurement in the literature. Stavem et al,6 using the Bland and Altman analysis of the tympanic thermometer the mean of the measurements in both ears, found a mean of difference of 0.45°C and a standard deviation of 0.38°C. The means of differences of the left and right ear were respectively 0.46°C (SD:0.49°C) and 0.45°C (SD:0.58°C). Rectal and oesophageal temperature were measured in the same study with means of differences of respectively –0.16°C (SD: 0.50°C) and –0.11°C (SD:0.26°C). The mean and standard deviation of TAT measurements in our study are closer to PAC recordings than tympanic measurements and are similar to rectal temperature recordings in the study of Stavem et al. TAT seems to perform better than the tympanic thermometer and has not the discomfort in patients like the rectal thermometer has.
Oral temperature measuring was studied by Giuliano et al4 with a similar mean difference of -0.15°C , but with a better standard deviation of 0.36°C). His study supports the hypothesis that oral thermometry is accurate and reproducible in orally intubated patients, when the user is focused on proper placement of the probe. The only method scoring clearly better than others is the urinary bladder catheter (UBC) with a mean of –0.04 (SD: 0.27), reported in a study of Nierman et al.7 Lefrant et al8 also concluded that the urinary bladder technique is the preferable method for temperature measurement in the most critically ill patients. The UBC was not taken in consideration in this study because it is too expensive for daily practice at this moment. The accuracy of the TAT can be considered as good and seems at least comparable with other available methods. The variability of the measurements of the instrument seems to be the weaker aspect. The instrument does not allow continuous monitoring of temperature what can be considered as the most important disadvantage.
Overall, the experience with TAT is limited, especially in adult patients. The study of Suleman et al9 contained only 15 adult patients and an other type of temporal scanner was used there. In a recent paper of Ostrowsky et al,10 the authors had the impression that temperature readings were systematically lower than the measurements with the oral/ rectal electronic thermometer systems they formerly used. There was no direct comparison between the two methods of measurement in the same patients in that study. Our findings seem to be in conflict with their results. Besides the study design used, a possible explanation of the difference with the results of the study of Ostrowsky et al,10 may be the fact that in the USA only scanning across the forehead over the temporal artery and not behind the ear is used. Sweat could thereby be a confounding factor. A limitation of our study is that the measurements are situated in normal range. The performance of the TAT at temperatures outside the normal range is not explored and may be less accurate. The accuracy of the device in patients with normothermia should not be extrapolated to patients in hypo- or hypertherm conditions. Another methodological weakness in this study, is that temperature in some patients have been measured more frequently than in others. Yet, in 44 patients (77%) only one or two measurements were obtained. In only 13 patients more than two measurements were taken. From this we assume that the impact of this confounder is limited but nevertheless present.
Conclusions
The temporal scanner has a relatively good reliability with an acceptable accuracy and variability in patients with normothermia. The results are comparable to those of the AT, but they do not seem to be sufficient to prove any substantial benefit compared to rectal, oral or bladder thermometry.
REFERENCES
1 Henker R, Coyne. Comparison of peripheral temperature measurements with core temperature. AACN Clin Issues. 1995; 6: 21-30.
2 Exergen Temporal Scanner®. Reference manual; Exergen corporation, Boston, USA
3 Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet. 1995; 346: 1085- 1087.
4 Giuliano KK, Scott SS, Elliot S, Giuliano AJ. Temperature measurement in critically ill orally intubated adults: A comparison of pulmonary artery core, tympanic, and oral methods. Crit Care Med. 1999; 27: 2188-2194.
5 Robinson J, Charlton J, Seal R et al. Esophageal, rectal, axillary, tympanic, and pulmonary artery temperatures during cardiag surgery. Can J Anaesth. 1998; 45: 317-323.
6 Stavem K, Saxhom H, Smith-Erichsen N. Accuracy of infrared ear thermometry in adult patients. Intensive Care Med. 1997; 23: 100-105.
7 Nierman DM. Core temperature measurement in the intensive care unit. Crit Care Med. 1991; 19: 818-823.
8 Lefrant JY, Muller L, Coussaye JE, et al. Temperature measurement in intensive care patients: comparison of urinary bladder, oesophageal, rectal, axillary, and inguinal methods versus pulmonary artery core method. Intensive Care Med. 2003; 29: 414-418.
9 Suleman MI, Doufas AG, Akca O. Insufficiency in a new temporal-artery thermometer for adult and pediatric patients.Anesth Analg.2002;95(1): 67-71.
10 Ostrowsky B, Ober J, Wenzel R, et al. The case of the cold thermometers. AJIC. 2003; 31(1): 57-59.