| ||||||||||||||
|
|
|||||||||||||
Department of Anesthesiology and Critical Care Medicine, The Johns Hopkins Medical Institutions, Baltimore, Maryland
Address correspondence and reprint requests to Christine G. Cattaneo, MD, Department of Anesthesiology and Critical Care Medicine, The Johns Hopkins Medical Institutions, Halsted 842, 600 N. Wolfe St., Baltimore, MD, 21287. Address e-mail to ccattaneo{at}jhmi.edu
| Abstract |
|---|
|
|
|---|
Implications: Body temperature should be monitored during spinal anesthesia because patients are at significant risk for hypothermia. Rectal temperature is a valid method of measuring core temperature, whereas other methods tend to underestimate true core temperature.
| Introduction |
|---|
|
|
|---|
Historically, body temperature has been monitored during general anesthesia (GA) to identify malignant hyperthermia. However, because regional anesthesia is not associated with malignant hyperthermia, temperature is often ignored in patients receiving this anesthetic technique (8,9). A recent survey found that only 33% of practicing anesthesiologists routinely monitor temperature during regional anesthesia, whereas 56% believe that temperature should be monitored routinely with regional anesthesia (8). Reasons for not monitoring temperature may include the misconception that regional anesthesia is not associated with a clinically significant risk for hypothermia, or the lack of convenient, reliable, and tolerable sites to monitor temperature in the awake or sedated patient. Although various skin-surface sites and methods are routinely used, no previous study has assessed the accuracy and precision of various temperature monitoring sites and methods during regional anesthesia. Because compensatory cutaneous vasoconstriction occurs above the level of regional block (10,11), and thermoregulatory vasoconstriction may also decrease skin temperature (12), it is possible that cutaneous temperature measured above this level is lower than core temperature (Tc) during regional anesthesia.
Thermoregulation is impaired during spinal anesthesia (SA) (13), epidural anesthesia (14), and GA (15). Studies comparing the incidence of hypothermia during regional and general anesthetic techniques report conflicting results. Some investigators report a similar risk of hypothermia with epidural anesthesia and GA (16,17), whereas others show an enhanced risk with epidural anesthesia compared with GA (18,19). Still another study reported a less frequent incidence of hypothermia with epidural than with GA (20). Whereas most comparative studies use epidural and GA, few studies have evaluated the relative risk of hypothermia with SA.
We therefore, determined the relative accuracy and precision of various temperature monitoring sites and methods during SA and GA. We tested the hypotheses that: 1) accuracy and precision of various temperature monitoring sites and devices are different during SA and GA; and that 2) patients receiving SA are at equal risk for perioperative hypothermia as those receiving GA.
| Methods |
|---|
|
|
|---|
= 0.05) and to detect a correlation coefficient (
) of 0.9. No patient had a history of significant cardiovascular or pulmonary disease. At the discretion of the anesthesia care team and the patients preference, patients received either SA (n = 16) or GA (n = 16). Patients were enrolled in the study consecutively until a total of 16 for each of the anesthetic techniques were studied. Percent body fat was measured by the infrared interactance method (21). Patients receiving SA were given IV premedication with midazolam 13 mg immediately before the spinal injection. Spinal anesthesia was started in the operating room (OR) with 20 mg bupivacaine (0.75%), 20 mcg fentanyl, and 200 mcg epinephrine, delivered through a 25- or 27-gauge Whitacre needle. A T3T8 sensory block level was achieved as detected by loss of pinprick sensation. The patients were sedated intraoperatively with additional IV midazolam (15 mg). Regression of spinal block in the recovery room was assessed by pinprick test every 20 min by trained postanesthesia care unit nurses. Duration of spinal block was defined as the time required for resolution to a level of L1.
General anesthesia included IV thiopental or propofol, an opioid (fentanyl, morphine or hydromorphone), a nondepolarizing muscle relaxant, and approximately 1 minimum alveolar anesthetic concentration of isoflurane. All patients receiving GA had tracheal intubation for the intraoperative course and were extubated in the OR. All patients had postoperative analgesia delivered IV by patient-controlled analgesia with hydromorphone.
In the preoperative holding area, two liquid crystal thermometry strips (LCT) (OmniTherm, Inc.; St. Louis, MO and Sharn, Inc., Tampa, FL) and a skin-surface thermocouple probe (Mon-a-thermTM; Mallinckrodt, St. Louis, MO) were placed on the forehead to monitor forehead skin temperature. An infrared tympanic thermometer (Thermoscan; San Diego, CA) was used to measure tympanic membrane temperatures.
On arrival in the OR, a thermocouple probe was placed in the aural canal by one of the investigators to measure Tc at the tympanic membrane. Placement was verified with the patient reporting detection of a scratching sound when the thermocouple wire was rubbed. Then, the probe was taped securely in place. Thermocouple probes were also placed on the skin in the axilla, 10 cm into the rectum, and near the patient, without contact to measure ambient temperature. Temperature measurements taken by the thermocouple probes were recorded at 2-min intervals by using the electric thermometer (Isothermex; Columbus Instruments, Inc., Columbus, OH) linked to a laptop computer, and then, stored on a hard disk. Forehead skin temperature taken by the LCTs and tympanic membrane temperature taken with the infrared thermometer were recorded at 15-min intervals by one of the investigators. Temperatures were monitored immediately after initiation of anesthesia and continued until discharge from the recovery room, approximately 23 h postoperatively.
Temperatures in the two anesthetic groups were compared before the induction of anesthesia, at 15-min intervals intraoperatively, and immediately before leaving the OR. Postoperatively, temperatures were compared on arrival in the recovery room, at 30-min intervals in the recovery room, and on discharge from the recovery room.
Patients in both anesthetic groups received similar thermal care both in the OR and in the recovery room as per protocol. All patients in both groups (GA and SA) received prewarmed (38°C) lactated Ringers solution and blood products through an in-line fluid warmer. Ambient OR temperature was 2123°C. No forced air warming devices were used at any time during the study. Patients had a single cotton blanket across their upper bodies intraoperatively and one or two cotton blankets over their entire bodies postoperatively.
Demographic variables were compared between the two groups by Students t-test. The incidence of postoperative core hypothermia in the two groups was compared by
2 analysis. By using tympanic temperature measured by thermocouple as the reference Tc, accuracy was defined as the difference from this reference (in °C) for each monitoring method. Precision was defined by linear regression of the reference Tc and temperature measured by each monitoring method. The Bland and Altman statistical method (22) was used to assess accuracy and precision of the temperature monitoring methods from the reference Tc. All determinations of accuracy and precision were made by using measurements taken in the OR at the end of the surgical procedure. Significance was defined as P < 0.05 and all data are given as mean ± SEM. Temperature differences of 0.5°C were defined a priori as clinically significant.
| Results |
|---|
|
|
|---|
|
|
|
|
|
|
|
|
| Discussion |
|---|
|
|
|---|
Given the proximity to the internal carotid artery, the tympanic membrane temperature, measured by thermocouple probe, is considered to represent true Tc. This does not imply that other methods of measuring tympanic membrane temperature are equally reliable for clinical use. Intraaural infrared thermometers are a relatively poor method for intraoperative temperature monitoring (23,24). In the current study, the infrared tympanic thermometer was less accurate and precise than either the rectum or the Omni LCT method.
Studies investigating the adequacy of LCTs for temperature monitoring have reported different conclusions. One study found LCTs adequate for monitoring an increase in temperature (25), whereas others found them either adequate (26,27) or inadequate (28) to follow body temperature trends perioperatively. However, one study (25) evaluated patients during cardiopulmonary bypass, which involved large Tc changes and therefore may not apply to patients with smaller Tc changes. Because patients receiving regional anesthesia are typically vasoconstricted above the level of sympathetic block (10,11), we hypothesized that skin surface monitoring on the upper body would underestimate true Tc. In the current study, skin-surface monitoring methods generally underestimated Tc during both SA and GA. This underestimation was especially apparent at the lower range of Tc. This may reflect the decreased cutaneous blood flow at lower Tc because of thermoregulatory vasoconstriction. Although rectal temperature was greater in accuracy and precision compared with other methods, rectal was the only method that appeared to overestimate Tc. This may reflect fecal insulation and delayed changes during the development of hypothermia.
Frank et al. (8) reported that body temperature is not regularly monitored during regional anesthesia, although most practitioners felt that it should be monitored. When body temperature is monitored during regional anesthesia, most clinicians report using either forehead skin surface with LCT devices (70%) or axillary skin temperature (40%) (8). It should be recognized that these techniques are likely to underestimate Tc, especially at lower Tc values, which may result in undue concern about hypothermia. The accuracy of LCT monitors is dependent on the built-in offset which is usually between 2°C and 4°C. This offset allows an estimate of Tc derived from skin temperature. Because surface warming methods may directly heat the skin temperature monitors, these methods of monitoring would be unreliable during active warming.
Previous studies have shown impaired thermoregulation with epidural anesthesia (14,16,29,30), SA (13,29,31,32), and GA (15,16,33). Vaughan et al. (17) reported a similar magnitude of postoperative hypothermia in patients receiving regional anesthesia and GA; however, that study did not specify which type of regional anesthesia (epidural or SA) was used, the level of regional block, or the specific surgical procedure. A later study reported an equivalent risk of postoperative hypothermia in patients receiving either epidural anesthesia or GA (10). Yet, other investigators have reported conflicting results. Epidural anesthesia was associated with greater hypothermia than GA in two studies (18,19), whereas GA was associated with greater hypothermia than epidural anesthesia in another study (20). We found an equal magnitude of intraoperative hypothermia in patients receiving either SA or GA for the same surgical procedure, and a longer duration of postoperative hypothermia in the SA group. The significant hypothermia with SA may be explained by the relatively high level of spinal block, which is associated with greater thermoregulatory inhibition (13,34).
The mechanism of impairment of thermoregulation during SA has been delineated in previous studies. In 1972, Roe and Cohn (31) described a rapid decline in Tc in patients receiving SA and postulated that this hypothermia was likely caused by vasodilation in the lower extremities and an impaired hypothalamic response with a lowered shivering threshold. Subsequent studies determined that hypothermia during SA is caused by both decreased shivering and vasoconstriction thresholds (32) and that the degree of thermoregulatory inhibition during regional block is proportionate to the number of dermatomes blocked (13,34). Given these mechanisms, the conclusions of previous investigations, and the results of the current study, it is clear that patients receiving regional anesthesia are at significant risk for hypothermia and should be monitored and treated accordingly.
We describe a significant difference in postoperative rewarming rates between SA and GA. On admission to the recovery room, patients in both groups had equivalent Tc. However, over the first 90 minutes postoperatively, patients in the GA group rewarmed significantly faster than patients in the SA group. A similar finding has been reported; however, the regional anesthetic technique and block height were not specified (17). Other investigators reported similar rewarming rates between patients receiving SA or GA (35) and similar rewarming rates in those receiving epidural anesthesia and GA (10). The difference in rewarming rates between these studies may be explained by the short duration of surgery in our study relative to the duration of the spinal block. Thus, the residual sympathetic block may have been a source of continuing cutaneous heat loss and slower postoperative rewarming in the SA group.
One limitation to our study was the lack of randomization for anesthetic technique. This is unlikely to have had a significant effect on our findings or conclusions because the two groups were similar for age, weight, percentage body fat, and duration of surgery, which are major determinants of thermoregulation and temperature changes. A possible second limitation is the assumption that tympanic temperature represents a valid reference Tc. Other investigators have proposed that distal esophageal temperature is more representative of Tc; however, it is well documented that esophageal and tympanic temperatures are closely correlated and reliable measurements of Tc (3638).
In conclusion, patients receiving SA are at a similar risk for developing perioperative hypothermia as are patients receiving GA during radical retropubic prostatectomy. Those receiving SA remain hypothermic longer into the postoperative course. These findings indicate that temperature should be monitored in patients receiving either SA or GA for major surgical procedures. Between the routinely used temperature monitoring methods, rectal temperature monitoring has the most accuracy and precision for temperature monitoring during regional or GA, whereas other methods tend to underestimate true Tc.
| Acknowledgments |
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
T. Kiya, M. Yamakage, T. Hayase, J.-I. Satoh, and A. Namiki The Usefulness of an Earphone-Type Infrared Tympanic Thermometer for Intraoperative Core Temperature Monitoring Anesth. Analg., December 1, 2007; 105(6): 1688 - 1692. [Abstract] [Full Text] [PDF] |
||||
![]() |
O. Kimberger, D. Cohen, U. Illievich, and R. Lenhardt Temporal Artery Versus Bladder Thermometry During Perioperative and Intensive Care Unit Monitoring Anesth. Analg., October 1, 2007; 105(4): 1042 - 1047. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Lawson, E. J. Bridges, I. Ballou, R. Eraker, S. Greco, J. Shively, and V. Sochulak Accuracy and Precision of Noninvasive Temperature Measurement in Adult Intensive Care Patients Am. J. Crit. Care., September 1, 2007; 16(5): 485 - 496. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Bock, U. Hohlfeld, K. von Engeln, P. A. Meier, J. Motsch, and A. J. Tasman The accuracy of a new infrared ear thermometer in patients undergoing cardiac surgery: [La precision d'un nouveau thermometre auriculaire infrarouge chez des patients de cardiochirurgie] Can J Anesth, December 1, 2005; 52(10): 1083 - 1087. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. G. Fritz, H. Hoff, M. Hartmann, W. Karzai, and K. R. G. Schwarzkopf The Effects of Urapidil on Thermoregulatory Thresholds in Volunteers Anesth. Analg., March 1, 2002; 94(3): 626 - 630. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. M. Frank Hypothermia After Vascular Surgery: Complications, Prevention, and Treatment Seminars in Cardiothoracic and Vascular Anesthesia, November 1, 2000; 4(4): 244 - 255. [Abstract] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|