|Year : 2018 | Volume
| Issue : 6 | Page : 418-423
Accuracy of skin temperature over carotid artery in estimation of core temperature in infants and young children during general anaesthesia
CK Suhail, Nandini Dave, Raylene Dias, Madhu Garasia
Department of Paediatric Anaesthesiology, Seth GS Medical College and KEM Hospital, Mumbai, Maharashtra, India
|Date of Web Publication||11-Jun-2018|
Dr. Nandini Dave
C 303, Presidential Towers, L.B.S Marg, Ghatkopar West, Mumbai - 400 086, Maharashtra
Source of Support: None, Conflict of Interest: None
Background and Aims: Core temperature monitoring is essential in children under general anaesthesia as they are more susceptible to hypothermia. We aimed to use skin temperature over the carotid artery (Tsk-carotid) with correction factors (Cf) to estimate core temperature. Primary outcome measure was to assess the sensitivity of Tsk-carotid with Cf for detecting hypothermia. Secondary outcome measure was to assess the specificity of Tsk-carotid with Cf for detecting hypothermia. Methods: First consecutive 50 patients fulfilling the inclusion criteria were included in modelling group and next 60 in the validation group. In the modelling group, average estimation error between Tsk-carotid and Tnaso was calculated and Cf was derived by multiple regression analysis (body surface area to mass ratio, body fat %, room temperature, relative humidity and warm Gamgee). In the validation group, Cf derived was used to predict Tnaso using Tsk-carotid by the formula: Tnaso-predicted = Tsk-carotid + Cf. Bland–Altman plots were used to assess the agreement between Tsk-carotid with Cf and Tnaso in the validation group. Results: The sensitivity for detecting hypothermia with the use of Tsk-carotid and Cf was 100%. The final Cf derived was 0.064 × (room temperature) −2.65. Most of the measurements fell within 95% confidence limit of Bland–Altman plot; 95% confidence interval (0.504–[−0.451]).The specificity of this method was 11%. Conclusion: This method overestimated hypothermia in most cases and cannot be accurately used as a measure of core temperature monitoring perioperatively.
Keywords: Core temperature monitoring, infants, intraoperative hypothermia
|How to cite this article:|
Suhail C K, Dave N, Dias R, Garasia M. Accuracy of skin temperature over carotid artery in estimation of core temperature in infants and young children during general anaesthesia. Indian J Anaesth 2018;62:418-23
|How to cite this URL:|
Suhail C K, Dave N, Dias R, Garasia M. Accuracy of skin temperature over carotid artery in estimation of core temperature in infants and young children during general anaesthesia. Indian J Anaesth [serial online] 2018 [cited 2019 Jan 16];62:418-23. Available from: http://www.ijaweb.org/text.asp?2018/62/6/418/234024
| Introduction|| |
Infants under anaesthesia are at risk of hypothermia. Accurate and continuous measurements of intraoperative core temperature are therefore essential in this population so that their thermal status can be managed. Core temperature measuring sites recommended for clinical use are the tympanic membrane, nasopharynx, distal oesophagus, pulmonary artery, bladder and rectum. All these sites are generally invasive and may be associated with complications. Nasopharyngeal probes can cause epistaxis,, rectal probes may cause traumatic injury. Body surface temperature can be measured continuously and has near-zero health risk. The temperature monitoring site should be close to a major vessel or core tissues to reflect the core temperature. The carotid artery is one of the largest arteries running close to the skin surface. We hypothesised that skin temperature over the carotid artery (Tsk-carotid), with the addition of some correction factor (Cfs) (body mass, body surface area, body fat percentage and subcutaneous fat thickness) could be used as a surrogate method for core temperature monitoring. The core to skin temperature gradient would be less at this site due to less subcutaneous fat and body mass. Moreover, the skin over the carotid artery is easily accessible under anaesthesia. The aim of our study was to determine the accuracy of Tsk-carotid with Cf in the estimation of core temperature in infants and young children during general anaesthesia. The primary outcome measure was to assess the sensitivity of Tsk-carotid with Cf for detecting hypothermia. Secondary outcome measure was to assess the specificity of Tsk-carotid with Cf for detecting hypothermia.
| Methods|| |
This was a prospective interventional study. Following Institutional Ethics Committee approval, 110 infants and young children below 3 years of age undergoing any surgery under general anaesthesia with the duration of 60–90 min were included in this study. Participants were enrolled for the study during the pre-anaesthetic check a day before surgery after obtaining written informed parental consent. All children included were of the American Society of Anesthesiologists 1 or 2 status. Patients with medical condition causing difficulty in the placement of a nasopharyngeal temperature probe (e.g., choanal atresia, nasal polyp, deviated nasal septum, bleeding disorders, nasal bone fracture and epistaxis) or with any neurological conditions affecting thermoregulatory function (e.g., cerebral palsy) were excluded from the study. The first consecutive 50 children who met the inclusion criteria were assigned to the modelling group; subsequent 60 children were included in the validation group.
On the morning of surgery, after confirming nil per oral status, patients over 6 months of age were pre-medicated with injection midazolam 0.05 mg/kg intravenously (IV) in the pre-operative holding area. Those below 6 months of age received no premedication. Children were then wheeled into the operation theatre and monitors were attached. This included electrocardiogram and pulse rate on cardioscope, pulse oximetry, capnometry and non-invasive blood pressure. Injection fentanyl 2 μg/kg IV was administered, and anaesthesia was induced with incremental sevoflurane up to 8%. The airway was secured using the appropriate size endotracheal tube (after achieving neuromuscular blockade with injection atracurium 0.6 mg/kg IV) or laryngeal mask (after checking for adequate jaw relaxation). Following this, nasopharyngeal temperature probe (oesophageal/rectal temperature probe with 400 series thermistor of Mindray ® DPM6 monitor, Mindray DS USA, Inc.) was inserted at a depth equivalent to the distance between the tragus and nares of each patient. A skin temperature sensor with 400 series thermistor of Mindray® DPM6 monitor, Mindray DS USA, Inc., was placed on skin over carotid artery at exact location of maximum carotid pulsation by palpation on the right anterior side of the neck. All patients were warmed using heating mattress and in children <6 months warm Gamgee was also used. IV fluids used were according to the age of the child. Fluids were not warmed. For children, more than 6 months age, lactated Ringer's solution was used. For infants and children <6 months of age, 1% dextrose in lactated Ringer's solution was used. This was started at an infusion rate according to the expected surgical losses. Anaesthesia was maintained with air: oxygen (50:50) and sevoflurane maintaining end-tidal CO2 of 35–45 mm Hg using closed circuit on GE Datex Ohmeda Aestiva ®/5 anaesthesia machine. Caudal epidural block was administered depending on the surgery. Triceps, biceps, suprailiac and subscapular skinfold thickness were then measured , on the right side with a Harpenden™ skinfold calliper (Baker Gauges India Pvt. Ltd.). Triceps measurement was taken at midline in the back of triceps. The midpoint between top of the acromial process to the bottom of the olecranon process of ulna was taken. The skin was then pinched so that the fold would run vertically and measurement was taken. Biceps skinfold thickness measurement was taken in the arm anteriorly at its midpoint. For suprailiac skinfold thickness measurement, iliac crest was palpated, the skin that follows the natural fold which will follow a line approximately 1 cm above the anterior superior iliac crest was grasped, and measurement was taken. To measure the subscapular skinfold thickness inferior angle of the right scapula was palpated. The skin and subcutaneous adipose tissue directly below it was grasped and measurement was taken. Body fat percentage was estimated using age, sex and skinfold measurements. Body surface area was estimated using height and weight. Patients were then draped as per surgical routine exposing only the incision site.
In the modelling group, the average estimation error (Xo) between the Tsk-carotid and Tnaso was calculated for each individual using the formula: Xo = Tnaso– Tsk-carotid.
A stepwise multiple linear regression analysis was used to determine the contribution of the different characteristics such as body surface area to mass ratio, body fat%, room temperature, relative humidity and usage of warm Gamgee to the explanation of Xo values of each individual and a Cf model was derived. The model used was: Estimation error (Xo) = constant + B (body surface area [BSA]/Mass) + C (Body fat%) + D (Room temperature) + E (Relative humidity) + F (Gamgee).
In the above formula B, C, D, E and F are the regression coefficients representing the contribution of different physical characteristics in the prediction of estimation error (Xo).
In the validation group, the Cf was calculated for each individual using the model derived from the modelling group. The Cf was then used to predict Tnaso using Tsk-carotid in the validation group. Tnaso-predicted = Tsk-carotid + Cf.
Temperature measurement was done after attaining anaesthetic steady state which typically takes 10–15 min. and was performed every 10 min thereafter. Hypothermia was defined as nasopharyngeal temperature below 36°C.
Ambient temperature and relative humidity of the operation theatre were measured using Temptec™ thermo hygrometer, Scipro Technologies, Mumbai, India. On completion of the procedure temperature sensors were removed before emergence.
Sample size was estimated by a similar study where Tsk-carotid with a simple Cf provided a viable non-invasive estimate of Tnaso in young children under general anaesthesia. Sensitivity for detecting hypothermia was 88% using Tsk-carotid. With 95% confidence interval (CI) and 80% power, we got a sample size of 60 in the validation group.
The characteristics analysed in the modelling group were age, height, weight, body surface area, body surface area/mass, room temperature, relative humidity, duration of surgery, Tnaso and Tsk-carotid. Data obtained were expressed as mean ± standard deviation (SD). The proportion was expressed as percentage. Descriptive statistics were used to describe the data. P < 0.05 was considered statistically significant. Multiple regression analysis was used in the modelling group. Bland–Altman plots were used to assess the agreement between Tsk-carotid with Cf for predicting Tnaso in the validation Group by calculating mean bias, precision and limits of agreement. The acceptable limits of agreement for estimating core temperature were determined a priori as + 0.5°C. Sensitivity and specificity was calculated in the validation group. Statistical analysis was done using MS-Excel 2007 and GraphPadInStat V. 3.04 (GraphPad Prism Software, San Diego, CA).
| Results|| |
The analysis of patient characteristics in modelling group and validation group is presented in [Table 1]. There were no significant differences in the characteristics analysed except for relative humidity which was greater in the modelling group compared to the validation group.
|Table 1: Analysis of patient characteristics in modelling group and validation group|
Click here to view
Modelling group – The contribution of physical characteristics (B, C, D, E and F) to the estimation error was determined by multiple regression analysis as follows: (X0) = −2.646 + 0.8321 × (BSA/Mass) + 0.01932 × (Body fat%) +0.06418 × (Room temp) + 0.02203 × (Relative humidity) + 0.03134 × (Gamgee) [Table 2]. Each correlation coefficient (r) was calculated independently, without considering the other variables. B, C, E, F did not significantly correlate with the interindividual variability in Xo for Tsk-carotid. However, D (ambient room temperature) correlated positively with Xo (r = 0.06418, 95% CI 0.001–0.127, P = 0.045). The final predictive model for Xo and therefore the resultant Cf to predict Tnaso with the skin temperature over the carotid was: Cf = 0.064 × (room temperature) −2.65.
|Table 2: Multiple regression analysis for contribution of physical characteristics (B, C, D, E, F) to estimation error|
Click here to view
Validation group: Tnaso was predicted with the addition of Cf to the Tsk-carotid. Bland–Altman plot is depicted in [Figure 1]. The mean of difference between measurement of the two methods line was very near to 0 with no systematic deviation between the measured values of the two techniques. Most of the measurements fell within 95% confidence limit of Bland–Altman plot; mean = 0.0266, SD = 0.244, 95% CI (0.504–(−0.451)]; mean + 1.96 SD = 0.504; mean – 1.96 SD = −0.45. The width of agreement was 0.95 which is very narrow as compared to actual Tnaso values and thus acceptable in routine clinical practice. The scatter of measurements was uniform and almost symmetrical suggestive of no bias and close agreeability between Tsk-carotid and Tnaso.
|Figure 1: Comparison of Tnaso versus predicted Tnaso using Bland–Altman plot. Y-axis - diff = Difference between measurements by Tnaso and predicted Tnaso. X-axis - Mean = Mean of measurements (temperature in °C) by two methods; Tnaso-nasopharyngeal temperature|
Click here to view
Linear regression between mean difference and mean values of measurements showed no proportional bias and close agreeability between Tsk-carotid and Tnaso.
Hypothermia occurred in 15 of the total 60 patients in the validation group when measured by nasopharyngeal temperature probe. The predicted Tnaso detected hypothermia in 55 patients of the total sixty patients. The sensitivity for detecting hypothermia with Tsk-carotid was 100% with 95% CI (78.2%–100%). The specificity was 11.11% with 95% CI (3.71%–24.05%). The positive predictive value was only 27.27%.
| Discussion|| |
Accurate and continuous measurements of intraoperative core temperature are essential in children undergoing surgery under general anaesthesia so that their thermal status can be managed.
Our study demonstrates that Tsk-carotid with a Cf can provide an accurate non-invasive estimate of Tnaso in infants and young children undergoing surgery under general anaesthesia.
Among all the parameters, room temperature was found to be significant and the final Cf derived was (0.064 × (room temperature [°C] −2.641). There was close agreement between Tsk-carotid and Tnaso within the acceptable range of 0.5°C defined a priori for measuring core temperature irrespective of other factors such as relative humidity and warm Gamgee.
In a similar study published in 2013, the authors demonstrated that the skin temperature over the carotid artery can provide accurate non-invasive estimate of nasopharyngeal temperature in infants and young children undergoing elective lower abdominal surgery under general anaesthesia. Cf of + 0.52 was derived from an independent group of participants consisting of twenty patients. This Cf was used to predict nasopharyngeal temperature by adding it to skin temperature over the carotid in a validation group of 28 patients. 93% of all the predicted Tnaso were within ± 0.5°C of the actual Tnaso values. Furthermore, the Tsk-carotid method correctly identified the presence of hypothermia in 88% of all cases. The sensitivity and specificity for detecting hypothermia with Tsk-carotid was 88% and 91%, respectively. In our study, the skin temperature over the carotid with the Cf correctly predicted hypothermia in 100% of the cases. However, the specificity was low (11%); hence this method over diagnosed hypothermia in 89% of our patients. In a similar study published in 2016, authors compared and evaluated the correlation of skin temperature monitored over carotid artery in the neck to nasopharyngeal temperature in adults under general anaesthesia. They found that although Tsk-carotid highly correlated with nasopharyngeal temperature at all time intervals, Tsk-carotid was significantly lower than corresponding Tnaso which may misdiagnose hypothermia.
The authors of this study emphasised the need to demonstrate the robustness of this method of detecting hypothermia using a greater sample size of patients. They also mention the need for further studies to assess the efficacy of the Tsk-carotid in predicting core temperature in patients diagnosed with fever as well as during more prolonged surgical procedures where greater changes in core temperature occurs.
There have been other studies to estimate core temperature using non-invasive temperature monitoring. In one such study, a non-invasive infrared scanner that detects temperature on the skin of the forehead was used. The study was conducted in a population of 15 adults and 15 children who developed a fever after cardiopulmonary bypass. The temperature was recorded throughout recovery at 15 min interval with sensor touch thermometer and from the pulmonary artery (adults) and bladder (children). They concluded that non-invasive infrared forehead thermometer (sensor touch) in adults and paediatric patients was inefficient in detecting core temperature.
Another study compared forehead skin temperature measured using liquid crystal thermometer to oesophageal and rectal temperatures measured by thermistor probes in patients under general anaesthesia for coronary artery bypass grafting. Authors found that there was a significant difference between the skin temperature over the forehead and the temperature measured by oesophageal and rectal thermistor probes. They concluded that liquid crystal strip may be useful as a safe convenient method for routine monitoring of temperature trends during general anaesthesia in patients whose exact core temperature need not be continuously monitored. However, in infants, patients undergoing extracorporeal circulation, major abdominal or vascular surgeries or patients with a history of temperature regulation problem, temperature should be best monitored by a method which more exactly reflects core temperature.
We studied a larger sample size compared to the 2013 study thus representing a greater age with different morphological range. The possible limitation of our study was that we were not able to thermostatically maintain the ambient temperature and relative humidity. To overcome this, we had also taken relative humidity and room temperature in addition to body surface area to mass ratio, body fat percentage and the usage of warm Gamgee to derive the Cf. As children of a wider age range were studied compared to the 2013 study which included only infants, the effect of factors affecting the skin temperature in infants and older children (distribution of body fat and subcutaneous fat thickness) is likely to be different and may require different Cfs. Randomisation was not possible in this study as the first fifty patients were assigned the modelling group to derive the Cf which was then applied to the next sixty patients in the Validation group. This study included patients who underwent surgery for only 60–90 min. However, in long duration surgeries, the results of our study may not be applicable. Other significant variables like use of warm IV fluids, warming mattress was not included in this multivariate analysis. Since the specificity of this method of detecting hypothermia is low, most cases diagnosed as hypothermic will not be hypothermic.
| Conclusion|| |
The results of our study showed that Tsk-carotid with the addition of Cf overestimated hypothermia in most cases. This method cannot be accurately used to detect hypothermia in infants and young children undergoing surgery under general anaesthesia.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Kumar PA, Arora H. Potential aspiration hazard when using a nasopharyngeal temperature probe with disposable cover. J Clin Anesth 2006;18:556-7.
Sinha PK, Kaushik S, Neema PK. Massive epistaxis after nasopharyngeal temperature probe insertion after cardiac surgery. J Cardiothorac Vasc Anesth 2004;18:123-4.
Frank JD, Brown S. Thermometers and rectal perforations in the neonate. Arch Dis Child 1978;53:824-5.
Jay O, Molgat-Seon Y, Chou S, Murto K. Skin temperature over the carotid artery provides an accurate noninvasive estimation of core temperature in infants and young children during general anesthesia. Paediatr Anaesth 2013;23:1109-16.
Durnin JV, Rahaman MM. The assessment of the amount of fat in the human body from measurements of skinfold thickness. Br J Nutr 1967;21:681-9.
Durnin JV, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: Measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974;32:77-97.
Weststrate JA, Deurenberg P. Body composition in children: Proposal for a method for calculating body fat percentage from total body density or skinfold-thickness measurements. Am J Clin Nutr 1989;50:1104-15.
Haycock GB, Schwartz GJ, Wisotsky DH. Geometric method for measuring body surface area: A height-weight formula validated in infants, children, and adults. J Pediatr 1978;93:62-6.
WHO. Thermal Control of a New Born: A Practical Guide. Geneva: World Health Organisation; 1996.
Selvaraj V, Gnanaprakasam PV. Evaluation of skin temperature over carotid artery for temperature monitoring in comparison to nasopharyngeal temperature in adults under general anesthesia. Anesth Essays Res 2016;10:291-6.
] [Full text]
Suleman MI, Doufas AG, Akça O, Ducharme M, Sessler DI. Insufficiency in a new temporal-artery thermometer for adult and pediatric patients. Anesth Analg 2002;95:67-71.
Burgess GE 3rd
, Cooper JR, Marino RJ, Peuler MJ. Continuous monitoring of skin temperature using a liquid-crystal thermometer during anesthesia. South Med J 1978;71:516-8.
[Table 1], [Table 2]