This article has Open Peer Review reports available.
A comparison of temporal artery thermometers with internal blood monitors to measure body temperature during hemodialysis
© The Author(s). 2018
Received: 22 February 2018
Accepted: 31 May 2018
Published: 14 June 2018
Thermometers that measure core (internal) body temperature are the gold standard for monitoring temperature. Despite that most modern hemodialysis machines are equipped with an internal blood monitor that measures core body temperature, current practice is to use peripheral thermometers. A better understanding of how peripheral thermometers compare with the dialysis machine thermometer may help guide practice.
The study followed a prospective cross-sectional design. Hemodialysis patients were recruited from 2 sites in Calgary, Alberta (April – June 2017). Body temperatures were obtained from peripheral (temporal artery) and dialysis machine thermometers concurrently. Paired t-tests, Bland-Altman plots, and quantile-quantile plots were used to compare measurements from the two devices and to explore potential factors affecting temperature in hemodialysis patients.
The mean body temperature of 94 hemodialysis patients measured using the temporal artery thermometer (36.7 °C) was significantly different than the dialysis machine thermometer (36.4 °C); p < 0.001. The mean difference (0.27 °C) appeared to be consistent across average temperature (range: 35.8–37.3 °C).
Temperature measured by the temporal artery thermometer was statistically and clinically higher than that measured by the dialysis machine thermometer. Using the dialysis machine to monitor body temperature may result in more accurate readings and is likely to reduce the purchasing and maintenance costs associated with manual temperature readings, as well as easing the workload for dialysis staff.
Measuring pre-dialysis body temperature is a routine part of care in many hemodialysis (HD) units. Although the gold standard for body temperature measurement is the temperature of central (core) blood, measurements of this parameter are not available in most clinical settings, and so temperature is typically measured at other sites, such as oral, axillary, tympanic, or temporal artery.
Modern HD machines are often equipped with an internal blood temperature monitor that can display core temperature readings. Despite this, common practice in many units is to record manual temperatures using peripheral thermometers. This practice is potentially unnecessary and may introduce error, given the known inaccuracy of peripheral temperature measurement .
In this study, we compared body temperature measured by HD machine thermometers with those measured by temporal artery (TA) thermometers. Our secondary objective was to assess the feasibility of replacing the currently used TA method with dialysis machine (DM) thermometers.
This prospective cross-sectional study is presented according to the STROBE guidelines for cross-sectional studies . This study was approved by the University of Calgary Conjoint Ethics Board.
All adult (≥ 18 years) patients with a diagnosis of end-stage renal disease, currently undergoing hemodialysis for at least one month at one of the two study sites in Calgary, Alberta, and able to provide informed consent were invited to participate. Patients were recruited between April and June 2017.
The primary outcome of the study was body temperature (°C). Other outcomes of interest were whether mean temperature varied depending on certain patient characteristics (self-reported gender, age, years on dialysis, diabetes, etc.) and thermometer preference of patients and nurses.
All patients treated in Calgary hemodialysis units are dialyzed using the Fresenius 5008 machine. This machine is equipped with an internal blood monitor that measures arterial temperature to two decimal places, which is transmitted to each unit’s electronic medical record system. Current practice locally is to monitor the patients’ temperature using a TA thermometer (Exergen Corporation, Massachusetts/United States). This device uses infrared technology to measure temporal artery temperature to 1 decimal place. Dialysis nurses and nurses’ aides collected the temperatures at the start of dialysis treatment on a best efforts basis. In some cases, temperature was collected later in the treatment, which was recorded as such. All data was collected during a single dialysis session per patient. Nurses or aides measured body temperature with the TA and DM methods simultaneously. Three measurements were taken for each method, 1 min apart (a total of 6 measurements). Participants were asked whether they preferred to have their temperature measured by the peripheral thermometer or directly by the HD machine. Nurses were also asked to submit anonymous surveys to the research assistants describing their preferences and willingness to use the dialysis machine reading.
Sample size calculation
A sample size calculation indicated that 90 patients would be required to yield 80% power to demonstrate a 0.3 °C difference between thermometer method pairs, assuming alpha = 0.05. The standard deviation of the difference between methods was assumed to be 1 °C, consistent with prior literature [3–5].
The mean difference and the 95% confidence interval (CI) of the two thermometer types (TA - DM) were calculated. In the primary analysis, we used the first pair of temperature measurements from each patient (one temperature using TA and one using DM), regardless of the time at which those measurements were made. A paired t-test was used to test the temperatures of each thermometer type. A Bland-Altman plot  was created to assess the level of agreement. Because dialysis treatment might alter body temperature, we also did a sensitivity analysis including only pairs of first temperatures that were obtained during the first 15 min of treatment initiation. We did a second sensitivity analysis that compared the mean value of the three temperatures within the first 15 min of treatment by each method (TA and DM).
Temperatures were compared between groups using t-tests (dichotomous variables) or ANOVA (categorical variables). A quantile-quantile plot was used to compare the distribution of temperature between thermometers.
Characteristics of hemodialysis patients
n = 94
Age; years (mean (SD))
Male: n (%)
Diabetes: n (%)
Years on dialysis (Median (IQR))a
1: n (%)
2: n (%)
Morning: n (%)
Afternoon: n (%)
Evening: n (%)
Temperatures as measured by temporal artery versus HD machine thermometers
Body temperature of hemodialysis patients measured by the temporal artery thermometer and the internal blood monitor of the dialysis machine for each dataset
Mean BT (°C)
Results were similar in the sensitivity analysis that compared only the first pairs of temperature measurement within the first 15 min of treatment (36.7 °C and 36.42 °C for TA and DM respectively, p < 0.001) (Table 2). Results were also similar in the second sensitivity analysis, which compared the mean of the first three measurements within the first 15 min of treatment (36.7 °C and 36.49 °C respectively, p < 0.001) (Table 2).
Preferences of patients and nurses
The large majority (95%) of patients did not express a preference about the thermometer type. Of the 12 nurses that completed a survey, 75% (n = 9) preferred using the dialysis machine instead of the forehead thermometer to measure patient body temperature. Reasons given for this included increased convenience and accuracy, reduced workload, and less perceived risk of transmitting infection between patients using the TA thermometer.
Epidemiology of body temperature among HD patients
There was no evidence that body temperature among HD patients statistically differed by the time of day. For example, median temperatures as measured by the HD machine were 36.24 °C during shift 1, 36.44 °C during shift 2, and 36.50 °C during shift 3) (p = 0.53). We also found no significant differences in median body temperature by gender (36.40 °C in males, 36.44 °C females; p = 0.68), age (36.49 °C in patients less than 65 and 36.36 in patients ≥65 years, p = 0.15), or years on dialysis (36.43 °C if < 6 years, 36.40 °C if ≥6 years, p = 0.72). The temperature of patients with diabetes (36.33 °C) was significantly lower than those without (36.52 °C) p = 0.02.
We found a statistically significant difference between body temperatures measured using temporal artery thermometers and core blood temperature measured by the dialysis machines. The mean difference between the methods was 0.27 °C, with the TA device reporting higher temperatures than the DM method. If one considers the DM method to be the gold standard, the limits of agreement (1.69 °C) exceed the recommended clinically acceptable difference of 0.5 , suggesting that the TA device may not be ideal for routine clinical use, given that a superior measurement is readily available.
To our knowledge, this is the first study comparing methods of temperature measurement in HD patients. Studies done in other populations have also reported a difference between temperatures measured using peripheral thermometers compared to those measuring core temperatures , [8–10]. A meta-analysis of mean differences in afebrile patients reported that temporal artery thermometers were 0.07 °C higher than the reference core temperature . However, this analysis included children and adults and a variety of core thermometer types, which could explain the lower discrepancy than found in our study. Interestingly, the meta-analysis found that TA thermometers underestimated temperatures among febrile patients (mean difference of − 0.19 °C). Taken together with our findings (overestimation of temperature among HD patients), these findings suggest that the temperatures obtained using TA thermometers may bias towards the mean (“normal”) temperature as seen in healthy populations. However, since all of the HD patients we studied were afebrile, this suggestion is speculative.
We found no association between body temperature and factors such as age, gender, years on dialysis, or time of day. This finding contrasts with previous research that showed a significant difference by gender  and time of day  in healthy individuals. Possible explanations for the discrepant finding could be related to kidney failure or because the participants in this study were older (IQR: 56–75 years). Patients with diabetes had significantly lower body temperature than those without, although the magnitude of the difference was small.
The findings from this study have potential implications for clinical practice. Considering core temperature as the gold standard, the temperatures measured by the dialysis machine thermometer were statistically different from those measured by the temporal artery thermometers. Although the magnitude of the difference was small, it might be clinically relevant in some scenarios. In addition, recording temperature from the dialysis machine may save small amounts of staff time compared to measuring it with a thermometer, especially if the data can be captured automatically by linkage between the dialysis machine and an electronic health record. In our study, although patients did not have a preference, most nurses preferred to measure temperature using the dialysis machine.
While this study achieved 80% power, included a representative HD population, and collected temperature measurements using two clinically relevant methods, it also has potential limitations. First, different nurses were involved in the data collection. Due to possible varying techniques, this could impact the precision of the temporal artery thermometers. However, this reflects real-world clinical practice and therefore is appropriate for understanding the practical implications of this study. Secondly, while we aimed to always obtain the pair of study measurements within the first 15 min of treatment initiation, this was not possible in 7 of 94 patients. Results were similar in our sensitivity analysis (done only in patients whose measurement was made within the first 15 min of treatment), suggesting that inclusion of these 7 patients did not affect our results. One could argue that a more stringent requirement (e.g. making all measurements within the first 30 s of dialysis) would have been preferable. However, although dialysis treatment may change true body temperature, there is no strong reason to think that it should differentially affect the temperature as measured by one technique vs. another. Nonetheless, this is a potential limitation of our study. Thirdly, these findings may not be generalizable to patients with significant access recirculation, which may affect the accuracy of core temperature measurement by the DM method. Furthermore, dialysate temperatures were not monitored in this study, which may have potential to influence the differences between the two thermometer types. However, these are unlikely to have affected our main conclusions. Finally, no febrile patients were included in this study and therefore the diagnostic performance of the two methods for detecting fever could not be compared.
The mean body temperature of hemodialysis patients as measured by a temporal artery thermometer was 0.27 °C higher than the mean core temperature measured by the dialysis machine. As core temperature is the gold standard, using the dialysis machine to measure body temperature in HD patients rather than an external thermometer may result in slightly greater accuracy while possibly also lowering staff workload.
The authors thank the patients involved in this study and the dialysis centre staff for their assistance in collecting the data for this study.
Availability of data and materials
The data that support the findings of this study are available from Alberta Health Services but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request according to the terms of relevant Alberta privacy legislation and regulation.
ML, RL, JM, and MT contributed to the conception or design; ML and BT acquired the data; ML, BT, NW, CT, and MT contributed to the analysis and interpretation of data; and ML and MT drafted the manuscript. All authors provided intellectual content of critical importance and approved the final version.
Ethics approval and consent to participate
This study was approved by the University of Calgary Conjoint Ethics Board. All participants provided written informed consent.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Niven DJ, Gaudet JE, Laupland KB, Mrklas KJ, Roberts DJ, Stelfox HT. Accuracy of peripheral thermometers for estimating temperature: a systematic review and meta-analysis. Ann Intern Med. 2015;163:768–77.View ArticlePubMedGoogle Scholar
- Institute of Social and Preventive Medicine (ISPM) - university of Bern. STROBE: Strengthening the reporting of observational studies in epidemiology 2009 https://www.strobe-statement.org/index.php?id=strobe-home. Accessed: 08 Nov 2017.Google Scholar
- Greenes DS, Fleisher GR. Accuracy of a noninvasive temporal artery thermometer for use in infants. Arch Pediatr Adolesc Med. 2001;155:376–81.View ArticlePubMedGoogle Scholar
- Hebbar K, Fortenberry J, Rogers K, Merritt R, Easley K. Comparison of temporal artery thermometer to standard temperature measurements in pediatric intensive care unit patients. Pediatr Crit Care Med. 2005;6:557–61.View ArticlePubMedGoogle Scholar
- Roy S, Powell K, Gerson LW. Temporal artery temperature measurements in healthy infants, children, and adolescents. Clin Pediatr. 2003;42:433–7.View ArticleGoogle Scholar
- Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res. 1999;8:135–60.View ArticlePubMedGoogle Scholar
- Hernandez JM, Upadhye S. Do peripheral thermometers accurately correlate to Core body temperature? Ann Emerg Med. 2016;68:562–3.View ArticlePubMedGoogle Scholar
- Geijer H, Udumyan R, Lohse G, Nilsagard Y. Temperature measurements with a temporal scanner: systematic review and meta-analysis. BMJ Open. 2016;6:e009509.View ArticlePubMedPubMed CentralGoogle Scholar
- Kiekkas P, Stefanopoulos N, Bakalis N, Kefaliakos A, Karanikolas M. Agreement of infrared temporal artery thermometry with other thermometry methods in adults: systematic review. J Clin Nurs. 2016;25:894–905.View ArticlePubMedGoogle Scholar
- Kimberger O, Cohen D, Illievich U, Lenhardt R. Temporal artery versus bladder thermometry during perioperative and intensive care unit monitoring. Anesth Analg. 2007;105:1042–7.View ArticlePubMedGoogle Scholar
- Mackowiak PA, Wasserman SS, Levine MM. A critical appraisal of 98.6 degrees F, the upper limit of the normal body temperature, and other legacies of Carl Reinhold August Wunderlich. JAMA. 1992;268:1578–80.View ArticlePubMedGoogle Scholar
- Usvyat LA, Kotanko P, van der Sande FM, Kooman JP, Carter M, Leunissen KML, Levin NW. Circadian variations in body temperature during dialysis. Nephrol Dial Transplant. 2012;27:1139–44.View ArticlePubMedGoogle Scholar