TECHNOLOGY, COMPUTING,

SECTION EDITOR STEVEN J. BARKER

AND

SIMULATION

SOCIETY

FOR

TECHNOLOGY

IN

ANESTHESIA

A Comparative Study of Three Warming Interventions to Determine the Most Effective in Maintaining Perioperative Normothermia
Siew-Fong Ng, BHSN*, Cheng-Sim Oo, BHSN*, Khiam-Hong Loh, Yiong-Huak Chan, PhD, and Biauw-Chi Ong, MD
BHSN*,

Poh-Yan Lim,

BHSN*,

*Block 3 Major Operating Theatre, Singapore General Hospital; Clinical Trials & Epidemiology Research Unit, Ministry of Health, Singapore; and Department of Anaesthesia and Surgical Intensive Care, Singapore General Hospital, Singapore

Perioperative hypothermia poses a challenge because of its deleterious effects on patient recovery. The current practice of applying two cotton blankets on patients during surgery is thought to be less ideal than using reflective insulation or forced-air warming. We studied 300 patients who underwent unilateral total knee replacement and were randomized equally to three groups: (a) the two-cotton-blanket group, (b) the one-reflective-blanket with onecotton-blanket group, and (c) the forced-air-warming with one-cotton-blanket group. Tympanic temperature readings were taken before surgery in the induction room, on arrival at the recovery room, and at 10-min intervals until discharge from the recovery room. On arrival at the recovery room, the forced-airwarming group had significantly higher temperatures (adjusted for sex, age, and patients induction

room temperature) of 0.577C 0.079C (95% confidence interval [CI], 0.427 0.726; P 0.001) and 0.510C 0.08C (95% CI, 0.349 0.672; P 0.001) more than the reflective-blanket and two-cotton-blanket groups, respectively. The forced-airwarming group took a significantly (P 0.001) shorter time of 18.75 min (95% CI, 13.88 23.62) to achieve a temperature of 36.5C in the recovery room as compared with 41.78 min (95% CI, 36.86 46.58) and 36.43 min (95% CI, 31.23 41.62) for the reflectiveblanket and two-cotton-blanket groups, respectively. The reflective technology was less effective than using two cotton blankets, and the forced-air warming was most efficient in maintaining perioperative normothermia. (Anesth Analg 2003;96:1716)

ypothermia is a known consequence of surgery and anesthesia. It is a clinical state in which a human body is unable to generate sufficient heat to efficiently achieve and maintain optimum metabolic performance (1,2). Mild hypothermia is said to be present when a patients core body temperature is between 33.0C and 36.4C (1). In the operating room (OR), surgical patients are inadvertently exposed to many factors that may alter the thermoregulatory mechanism, resulting in postoperative hypothermia. Some of the factors include cold OR, IV fluids, antimicrobial skin preparations, and various forms of anesthesia (3). In addition, some patients may have factors that place them at high risk
Accepted for publication September 10, 2002. Address correspondence and reprint requests to Biauw-Chi Ong, MD, Department of Anaesthesia and Surgical Intensive Care, Singapore General Hospital, Outram Rd., Singapore 169608, Singapore. Address e-mail to ganobc@sgh.com.sg. DOI: 10.1213/01.ANE.0000039403.46033.2E
2003 by the International Anesthesia Research Society 0003-2999/03

for perioperative heat loss. These include surgery that exceeds 2 h in duration, young age, old age, trauma, intraabdominal or thoracic surgery, massive blood loss, and transfusion. Inadvertent perioperative hypothermia poses a challenge because of its deleterious effects on patient recovery. Mild perioperative hypothermia, between the temperatures of 33.0C and 36.4C, causes complications such as protein catabolism, changes in glucose metabolism, and hypokalemia and can affect glomerular filtration (4). It is also associated with impaired wound healing, coagulopathy, postoperative increase in oxygen consumption, decrease in drug metabolism, negative nitrogen balance, and respiratory distress (3,5). Severe hypothermia may even lead to bradycardia, atrial fibrillation, premature ventricular contractions, and ventricular fibrillation (6). Hypothermia causes more problems for a patient than just a mere unpleasant experience of feeling cold. Therefore, the ability to maintain normothermia during and after surgery at the recovery room poses a challenge and
Anesth Analg 2003;96:1716

171

172

TECHNOLOGY, COMPUTING, AND SIMULATION NG ET AL. MAINTAINING PERIOPERATIVE NORMOTHERMIA

ANESTH ANALG 2003;96:1716

should be addressed in todays clinical practice so as to improve the quality of patient care. We thought that the current practice of applying two cotton blankets during surgery on the patient is less ideal than reflective insulation or the forced-air warming method. Therefore, we embarked on this study with the aim of determining the most effective and cost-effective of the three warming interventions.

cotton blankets were then folded once over the trunk of the patient to make eight layers in thickness. The arm with the IV drip was extended, and the other arm was placed along the side of the operating table. The sides of the blankets were carefully tucked under the patient. The blankets covered the patient from the level of the iliac crests extending to both shoulders and the neck.

Materials and Methods

This study was a prospective, randomized, controlled trial conducted from February 2000 to October 2000 with the approval from the Ethics Committees. Threehundred patients scheduled for unilateral total knee replacement were equally randomized into three groups using the sealed-envelope method. The control group was the two-cotton-blankets group, which is the current practice. The other two groups comprised the interventions using the reflective blanket and the forced-air warming blanket. The patients were either ASA physical status of I or II, as determined by the anesthesiologists. The process of the study was explained to the patients, and written consent was obtained. The patients height and weight were taken using the same weighing scale. The environmental temperatures of the ORs and the recovery rooms were thermostatically adjusted to between 19C and 22C, respectively, by the facilities and plants engineering department of the hospital. Four nurse researchers were trained to ensure standardization in treatment applications and data collection. On the day of surgery, when the patient arrived at the OR reception, the researcher assigned to the case selected a sealed envelope containing a data collection form that specified the type of warming intervention to be administered. The patients initial temperature was taken on arrival at the OR induction room before the administration of anesthesia. The temperature of the OR was standardized and recorded daily (Table 1). The patients temperature was taken using an ear thermometer (Terumo Ear Thermometer EM20CH, Terumo Corporation, Tokyo, Japan) that had infrared radiation capabilities. All patients were brought into the OR 45 min before the commencement of surgery and donned in a patient operation gown. One cotton blanket (100% cotton, 180 222 cm, and 900 g) folded lengthwise to make two layers was used at the ward level to cover the subject when they were transferred onto the OR trolley.

Reflective-Blanket with One-CottonBlanket Group

On transferring the patient onto the operating table, a reflective blanket (Thermadrape, 4 4 ft Blanket T1300, Concepts Inc, Roanoke, TX) was applied in a prescribed manner below the existing cotton blanket. The reflective side of the blanket was placed facing upwards. The existing cotton blanket was folded once over the trunk of the patient to make four layers in thickness. The arm with the IV drip was extended, and the other arm was placed along the side of the operating table. The sides of the blankets were carefully tucked under the patient. The reflective and cotton blanket covered the subject from the level of the iliac crests extending to both shoulders and the neck.

Forced-Air-Warming-Blanket with One-CottonBlanket Group

On transferring the patient onto the operating table, a forced-air warming blanket (Bair Hugger, Model 540 Torso, Augustine Medical Inc, Eden Prairie, MN) was applied below the existing cotton blanket in a prescribed manner. The surface with small, perforated holes, which allowed for controlled warm air to filter through, was placed onto the patient. The lower edge of the warming blanket was placed at the level of the patients iliac crests and held in place by self-adhesive tape. The existing cotton blanket was folded once over the trunk of the subject to make four layers in thickness. The arm with the IV drip was extended, and the other arm was placed along the side of the operating table. The sides of the blankets were carefully tucked under the subject. The forced-air blanket and cotton blanket covered the patient from the level of the iliac crests extending to both shoulders and the neck. The forced-air warming blanket was then inflated by a forced-air warming unit (Bair Hugger, Model 505, Augustine Medical Inc) set at 38C. In all three groups, the patients laid on the operating table, which was lined with a warm watercirculating blanket (Gaymar-model MTA 4702, Gaymar Inc, Buffalo, NY) set at 37C, followed by a Mackintosh (Hospital Rubber Sheet, 2 0.9 m) and a trolley sheet (100% Cotton, no fluff, 230 124 cm, and 144 thread count). The cleaning solutions used and the surgical draping were standardized. The patient continued to receive the same warming intervention that

Two-Cotton-Blankets Group
On transferring onto the operating table, the nurse researcher placed in a prescribed manner another cotton blanket that had been folded lengthwise. Both

ANESTH ANALG 2003;96:1716

TECHNOLOGY, COMPUTING, AND SIMULATION NG ET AL. MAINTAINING PERIOPERATIVE NORMOTHERMIA

173

Table 1. Demographic Data Two cotton blankets (n 100) Age (yr) Height (cm) Weight (kg) Length of operation (h) Total time in recovery room (min) Anesthesia time (min) Operating room temperature (C) Patients temperature taken in induction room (C) Recovery room temperature (C)
Values are mean sem.

Reflective blanket (n 100) 66.61 0.73 151.84 0.80 62.66 1.29 1.41 0.094 128.15 5.10 200.22 5.18 18.74 0.19 36.53 0.051 22.28 0.052

Forced-air blanket (n 100) 66.27 0.91 150.52 1.62 65.11 1.17 1.28 0.038 126.70 5.04 203.60 4.61 18.86 0.056 36.50 0.052 22.20 0.051

65.83 0.63 151.59 0.90 63.20 1.05 1.29 0.026 129.53 4.28 202.63 4.41 19.07 0.099 36.54 0.047 22.19 0.066

was given during surgery until their discharge from the recovery room. After surgery, when the patient arrived at the recovery room, the ambient temperature of the recovery room was documented. The patients temperature, pulse rate, respiration rate, and blood pressure were taken and documented. Subsequently, the patients temperature was taken using the same ear thermometer at 10-min intervals until the patient was discharged from the recovery room. Based on a survival analysis calculation of 70% versus 90% achieving 36.5C by 1 h for the two-cottonblanket group and forced-air-warming/reflectiveblanket groups, respectively, 89 subjects in each group would be sufficient for a power of 90% and a twosided test of 5% (7). Factoring in a 10% attrition rate, 100 patients per group would be recruited. All statistical analyses were performed using SPSS version 8.0 (SPSS Inc, Chicago, IL). Differences in continuous variables were determined using analysis of variance, and associations between categorical variables were assessed using 2 or Fishers exact tests. A multiple regression analysis with patients temperature at the recovery room as the dependent variable and adjusting for sex, age, and patients temperature at the induction room was performed. Kaplan-Meier analysis (8) was performed on the time to achieve 36.5C within the first hour. Statistical significance was obtained with P 0.05.

Figure 1. Mean temperature of patients from induction room to recovery room for the three groups (two-cotton-blankets group, reflective-blanket group, and forced-air-warming group). Error bars indicate 95% confidence intervals (CIs).

Results
Table 1 shows the demographic data of our study. There were no statistically significant differences between the three groups in the variables. Figure 1 shows the mean temperature of patients from the induction room to the recovery room for the three groups. The average time from the induction room to the recovery room was approximately 1.5 h. The arrival temperature (mean sem) of the patients at the recovery room were significantly less than that

at the induction room for both the two-cotton-blanket (36.02C 0.06C, range, 33.7C37.3C, versus 36.54C 0.047C, range, 35.0C37.2C; P 0.001) and the reflective blanket (35.93C 0.063C, range, 34.4C37.8C, versus 36.53C 0.051C, range, 35.0C37.3C; P 0.001) groups but not the forcedair-warming group (36.49C 0.062C, range, 34.2C 38.4C, versus 36.5C 0.052C, range, 34.9C37.4C; P 0.790). Statistically significant differences (using multiple regression adjusting for sex, age, and patients induction room temperature) were found for patients temperature on immediate arrival at the recovery room: the forced-air-warming group had significantly higher temperatures of 0.577C (95% confidence interval [CI], 0.427 0.726; P 0.001) and 0.510C (95% CI, 0.349 0.672; P 0.001) more than the reflective-blanket and two-cotton-blankets groups, respectively. Defining no change and increase in temperature as positive responses, the forced-air-warming-blanket

174

TECHNOLOGY, COMPUTING, AND SIMULATION NG ET AL. MAINTAINING PERIOPERATIVE NORMOTHERMIA

ANESTH ANALG 2003;96:1716

group was significantly more effective in maintaining perioperative normothermia from induction room to recovery room compared with the reflective-blanket (odds ratio (9), 8.2; 95% CI, 4.216; P 0.001) and two-cotton-blankets groups (odds ratio, 4.5; 95% CI, 2.4 8.1; P 0.001), as given in Table 2. Shivering in the recovery room was observed in four patients in the two-cotton-blanket group, three patients in the reflective-blanket group, and one in the forced-air-warming group. There was no association between the type of warming intervention used and shivering (P 0.371; Fishers exact test). Interestingly, we found that all eight patients who shivered had temperatures between 35.2C and 37C, whereas there were nine patients with temperatures below 35.2C who did not demonstrate shivering. Figure 2 shows the mean temperature profiles for the 3 warming groups with respect to time including the 95% CI error bars. The rate of temperature increase with time was similar for the three groups. Performing a survival analysis (Kaplan-Meier log-rank test) on time to achieve 36.5C within the first hour, the forcedair-warming group has a mean sem of 18.75 2.48 min (95% CI, 13.88 23.62), which is significantly lower (P 0.001) compared with the reflectiveblanket group (41.72 2.48 min; 95% CI, 36.86 46.58) and two-cotton-blankets group (36.43 2.65 min; 95% CI, 31.23 41.62). Figure 3 shows the percentage of patients who achieved 36.5C at various time periods up to 60 min. The forced-air-warming group had a significantly larger percentage of patients who achieved 36.5C within the hour compared with the two-cotton-blanket group (75% versus 45%; P 0.001; odds ratio, 3.7; 95% CI, 2.0 6.9) and the reflectiveblanket group (75% versus 40%; P 0.001; odds ratio, 4.5; 95% CI, 2.5 8.2).

Discussion
In our study, we found that forced-air warming was significantly more effective than two cotton blankets or reflective blankets in maintaining perioperative normothermia. The forced-air warming blanket is significantly more expensive compared with the reflective blanket. At our institution, the forced-airwarming unit costs US $1470. Assuming a potential usable period of five years and a usage rate of one to two patients per day (365 patients per year), this would translate to less than $1 per use per patient. The disposable warming blanket costs US $22 apiece or US $23 per use for the forced-air warming method. It was hoped that the reflective blanket, at US $5 apiece, would be a better alternative than the current practice of covering patients with two cotton blankets. The cost of a piece of re-usable cotton blanket is US $5, and the total cost per use including laundry would be less than

$1. A number of studies have suggested the use of reflective technology as a prevention strategy. Ensminger and Moss (10) examined the use of reflective technology and suggested that it had properties to preserve body temperature by reflecting the patients radiant heat loss and also preventing convective heat loss to the cold OR. However, the results of the reflective blanket were not better than the two cotton blankets as we had assumed when we began this study. Although the rate of temperature increases over time was the same for all three groups, the two-cotton-blankets group achieved 36.5C within the first hour, with a mean sem of 36.43 2.65 minutes, whereas the reflective blanket group was 41.72 2.48 minutes. Our study was consistent with that of Borms et al. (11) who compared the efficacy of forced-air warming and reflective blanket in patients who had elective total hip arthroplasty. This study showed that reflective insulation was unable to maintain intraoperative normothermia. Recovery room stay costs US $20 per 15 minutes. The shorter time required to rewarm the patients in the recovery room with the forced-air-warming group (mean sem of 18.75 2.48 minutes) offsets the higher startup cost for the forced-air warming device. Our study showed that in the forced-air-warming group, normothermia in the OR was more effectively maintained, and a higher temperature was achieved upon arrival at the recovery room. Murat et al. (2) evaluated the efficacy of forced-air warming given to 51 patients during spinal surgery for scoliosis. Their study showed forced-air warming to be more effective in maintaining body temperature than when no warming devices were used. Bennet et al. (12) compared the efficacy of cutaneous active versus passive warming devices on 45 elderly patients undergoing elective hip arthroplasty. They found active skin surface warming provided a stable thermal homeostasis. This was also consistent with a study by Defina et al. (13) where patients using forced-air warming during surgery were more likely to arrive at the recovery room less hypothermic and had a shorter postoperative stay in the recovery room. It was noteworthy in our study that 25% of patients in the group receiving forced-air warming had core body temperatures more than 37.2C during their stay in the recovery room. Another 9% of patients in this same treatment group verbalized a feeling of overwarmth and discomfort, although their body temperatures were 37.2C. The treatment had to be discontinued for these 34% of patients. Further analysis of this cohort showed that three quarters of them had an initial body temperature, before the administration of anesthesia at the induction room, in the normothermia range of 36.5C37.2C. This was interesting because

ANESTH ANALG 2003;96:1716

TECHNOLOGY, COMPUTING, AND SIMULATION NG ET AL. MAINTAINING PERIOPERATIVE NORMOTHERMIA

175

Table 2. Percentage of Patients in Relation to Changes in Body Temperature from Induction Room to Recovery Room Increase in temperature (C) Two cotton blankets Reflective blanket Forced-air blanket 16 11 48 No change in temperature (C) 10 5 13 Decrease in temperature (C) 74 84 39

Figure 2. Mean temperature profiles for the three groups (twocotton-blanket group, reflective-blanket group, and forced-airwarming group) with respect to time in the recovery room in 10-min intervals. The error bars indicate 95% confidence intervals (CIs).

we did not find other studies reporting this hyperthermic effect. It raised a concern that one third of the cohort in the forced-air warming experienced a sense of over-warmth. There are limitations in our study. Prolonged surgery or surgeries that expose the pleural, the pericardial, or the peritoneal serous surfaces do result in dramatic evaporative heat loss. Our study focused only on unilateral total knee replacement surgery for standardization reasons. Further study could be extended to include open cavity operations so that the findings could be generalized. Because this study had more women who had total knee replacement surgery, our sex distribution was not equal. We had 251 women and 49 men (83% women). This is consistent with the general sex distribution of patients undergoing total knee replacement surgery in our institution over the last three years: there were 375 women and 72 men in 1999 (83% women), 419 women and 79 men in 2000 (84% women), and 606 women and 79 men in 2001 (84% women). Our recommendation is to determine the core body temperature of patients admitted to the recovery room. This will detect patients who are hypothermic so that they can be provided with the appropriate intervention. We hope that preventive measures taken to identify inadvertent perioperative hypothermia will

Figure 3. Percentage of patients achieving 36.5C within the first hour in the three groups (two-cotton-blanket group, reflectiveblanket group, and forced-air-warming group) with respect to time in the recovery room in 10-min intervals.

help to minimize the deleterious effects on patients postsurgical recovery. Recovery room personnel should not depend on the presence of shivering to determine whether patients are hypothermic because our results showed that patients who had temperatures below 35.2C did not necessarily shiver. We conclude that forced-air warming was significantly more efficient and cost-effective in maintaining perioperative normothermia during knee surgery. Although the cost of the forced-air warming may be more than that of two cotton blankets and a reflective blanket (US $23 versus US $1 versus US $5), this is compensated by the savings in reduced recovery room stay (US $20 per 15 minutes) and the cost of treating complications associated with postoperative hypothermia. The patients in the forced-air-warming group required 60 minutes compared with 150 minutes in the two-cotton-blankets group to reach normothermia in the recovery room. The performance of the two cotton blankets was better when compared with the reflective blanket.

176

TECHNOLOGY, COMPUTING, AND SIMULATION NG ET AL. MAINTAINING PERIOPERATIVE NORMOTHERMIA

ANESTH ANALG 2003;96:1716

The authors would like to thank the following at the Singapore General Hospital for their support and cooperation in this study: Department of Orthopaedic Surgery, Department of Anaesthesia and Surgical Intensive Care, OR Manager Ms Kwok Moon Hoe, Senior Nurse Manager Ms Gin Cheng Yam, and all the nurses and staff of the orthopaedic wards, OR, recovery room, anesthesia unit, physiotherapy department, and ambulatory surgery center.

References
1. Buggy DJ, Crossley AWA. Thermoregulation, mild perioperative hypothermia and post-anaesthesia shivering. Br J Anaesth 2000;84:61528. 2. Murat I, Berniere J, Constant I. Evaluation of the efficacy of a forced-air warmer (Bair Hugger) during spinal surgery in children. J Clin Anesth 1994;6:4259. 3. Sessler DI. Perioperative heat balance. Anesthesiology 2000;92: 578 96.

4. Dennison D. Thermal regulation of patients during the perioperative period. AORN 1995;61:82732. 5. Arndt K. Inadvertent hypothermia in the operating room. AORN 1999;70:204 14. 6. Villamiel LM. Help! This postanaesthesia care unit patient is hypothermic. J Post Anaesth Nurs 1990;5:759. 7. Machin D, Canpbell MJ, Fayers PM, et al. Sample size tables for clinical studies. Oxford: Blackwell, 1997. 8. Armitage P, Berry G. Statistical methods in medical research. 3rd ed. Boston, MA: Blackwell Science, 1994:477. 9. Armitage P, Berry G. Statistical methods in medical research. 3rd ed. Boston, MA: Blackwell Science, 1994:130. 10. Ensminger J, Moss R. Preventing inadvertent hypothermia: a success story. AORN 1999;70:298 301. 11. Borms SF, Englelen SLE, Himpe DGA, et al. Bair Hugger forcedair warming maintains normothermia more effectively than Thermo-Lite insulations. J Clin Anesth 1994;6:3037. 12. Bennet J, Ramachandra V, Webster J, Carli F. Prevention of hypothermia during hip surgery: effect of passive compared with active skin surface warming. Br J Anaesth 1994;73:180 3. 13. Defina J, Lincoln J. Prevalence of inadvertent hypothermia during the perioperative period: a quality assurance and performance improvement study. J Perianesth Nurs 1998;13:229 35.

Attention Authors! Submit Your Papers Online

You can now have your paper processed and reviewed faster by sending it to us through our new, web-based Rapid Review System. Submitting your manuscript online will mean that the time and expense of sending papers through the mail can be eliminated. Moreover, because our reviewers will also be working online, the entire review process will be significantly faster. You can submit manuscripts electronically via www.rapidreview.com. There are links to this site from the Anesthesia & Analgesia website (www.anesthesia-analgesia.org), and the IARS website (www.iars.org). To find out more about Rapid Review, go to www.rapidreview.com and click on About Rapid Review.