Original Contributions

Postoperative Hemodynamic and Thermoregulatory Consequences of Intraoperative Core Hypothermia

Andrea Kurz, MD,* Daniel I. Sessler, MD,? Edith Narzt, MD,* Amir Bekar, MD,$ Rainer Lenhardt, MD,* Guenther Huemer, MD,8 Franz Lackner, MD” Department of and Intensive Care, University of Vienna, Vienna, Austria, and the Thermoregulation Research Laboratory, University of California School of Medicine, San Francisco, CA.

*Resident in , University of Vienna Study Objective: To evaluate the postoperative hemodynamic and thermoregulatory consequences of intraoperative core hypothermia. tAssociate Professor, University of Califor- Design: Prospective, randomized clinical trial. nia, San Francisco Setting: Operating room and postanesthesia care unit of a university hospital. *Research Fellow in Anesthesiology, Uni- Patients: 74 healthy, ASA status I, II, and III patients (average age 58 yrs) under- versity of Vienna going elective colon surgery. BAttending Anesthesiologist, University of Interventions: Patients were randomly assigned to be kept normothermic or =Z.YC Vienna hypothermic during surgery. Anesthesia was maintained with isoflurane, nitrous oxide, and fentanyl. Postoperatively, surgical pain was treated with patient-controlled anal- “Professor of Anesthesia and Intensive gesia (PCA) opioid. Care, University of Vienna Measurements and Main Results: An observer blinded to group assignment and Address correspondence to Dr. Sessler at core temperatures evaluated shivering, thermal comfort, surgical pain, heart rates the Thermoregulation Research Labora- (HRs), and blood p ressures (BPS) during the first six postoperative hours. Morphomet- tory, University of California, San Francis- ric characteristics, oxygen saturation, fluid balance, PCA-administered opioid, and co, CA 94143-0648, USA. visual analog pain scores were comparable in the two groups. Hypothermic patients felt Supported in part by a grant from the Max uncomfortably cold during recovery, and their postoperative core temperatures remained Kade Foundation, New York, NY, and the significantly less than in the normothermic patients for more than four hours. Periph- Joseph Drown Foundation, Los Angeles, eral vasoconstriction and shivering were common in the hypothermic patients but rare CA, and National Institutes of Health in those kept normothermic. HRs and BPS were comparable in the two groups. Grant GM49670. The authors do not con- sult for, accept honoraria from, or own Conclusions: These data confirm that the effects of intraoperative hypothermia on stock or stock options in any anesthesia- postoperative HR and BP are modest in relatively young, generally healthy patients. In related company. contrast, intraoperative hypothermia caused substantial postoperative thermal discom- fort, and full recovery from hypothermia required many hours. Delayed return to core Presented in part at the Annual Meeting of normothermia apparently resulted largely from postoperative thermoregulatory impair- the American Society of Anesthesiologists, San Francisco, October 15-19, 1994. ment. Received for publication April 12, 1994; re- vised manuscript accepted for publication Keywords: Anesthesia: postoperative care; blood pressure; heart rate; August 16, 1994. shivering; thermoregulation; vasoconstriction.

Journal of Clinical Anesthesia 7:359-366, 1995 0 1995 by Elsevier Science Inc. 0952-8 180/95/$10.00 655 Avenue of the Americas, New York, NY 10010 SSDI 0952-8180(95)00028-G Original Contribution5

Introduction pants had a history of thyroid disease,dysautonomia, Raynaud’s syndrome, or . Sur- Volatile anesthetics impair shivering, as evidenced by gery started at approximately 8:30 AM in nearly all of the the observation that shivering is rare during surgery, patients, and typically finished near noon. even in unparalyzed patients having core temperatures Patients were premeditated with 10 mg oral diaze- near 32°C. Similarly, all general anesthetics so far tested pam. General anesthesiawas induced by administration markedly reduce the core temperature, triggering ther- of 3 to 5 mg/kg of thiopental sodium, fentanyl 250 pg, moregulatory vasoconstriction. l-5 (Nonshivering ther- and vecuronium 0.1 mg/kg. The trachea was intubated, mogenesis is relatively unimportant in unanesthetized and the lungs were ventilated mechanically to maintain adults,“,’ and it does not contribute during general an- end-tidal CO2 partial pressure (P,,CO,) near 35 esthesia.*) Because thermoregulatory defenses are im- mmHg. Anesthesia subsequently was maintained with paired by general anesthesia, most surgical patients be- 60% nitrous oxide (N,O) and isoflurane. The isoflurane come hypothermic unless they are actively warmed.gp’0 concentration was adjusted as necessary to keep mean During the initial postoperative period, brain anes- arterial pressure (MAP) greater than 60 mmHg, and thetic concentrations usually decrease rapidly,” allow- additional fentanyl was given at the discretion of the ing re-emergence of thermoregulatory res onses in- B attending anesthesiologist. Additional vecuronium was cluding vasoconstriction and shivering.’ *I3 These administered as required to maintain 1 to 2 mechanical responses combine to decrease cutaneous heat loss,14 twitches in response to supramaximal electrical stimula- constrain metabolic heat to the thermal core,15 and in- tion of the ulnar nerve at the wrist. crease metabolic heat production.” As a result, core Patients were randomly assigned, via a computer- temperatures usually increase toward normal values generated random numbers table, to routine thermal once anesthesia is discontinued. management (hypothermic group, n = 35) or extra One factor limiting return to normothermia is simply warming (normothermic group, n = 39). (The patients that some patients develop enormous heat debts during were not informed as to which treatment group they surgery. Based on the reported specific heat of hu- were assigned.) The upper body of those assigned to mans,16a 2.5”C reduction in mean body temperature extra warming was covered with a disposable Bair Hug- corresponds to a debt of ~145 kcal in a 70 kg patient. ger@ forced-air cover (Augustine medical, Inc., Eden This is roughly the basal heat production for 2.5 Prarie, MN) that was connected to a Model 500 heater hours.” Even when vasoconstriction is combined with (Augustine medical, Inc.) set on “high” (=40”C).28 We” vigorous shivering, some time will be required to replace and others*’ have shown that this system maintains nor- this much heat. Additionally, impaired thermoregulato- mothermia, even during extensive surgical procedures. ry responses may limit full postoperative recovery of The other patients were allowed to become hypother- thermal equilibrium. Such impairment may result from: mic. In those patients whose core temperature ap- (1) residual volatile anesthetic;’ (2) opioids administered proached 34”C, forced-air warming was instituted to to treat surgical pain;la2* or, (3) a delayed effect of prevent further hypothermia. Ambient room tempera- intraoperative hypothermia. ture was kept at 21” to 22°C intraoperatively and 23” to Recovery of normal core temperature in the postop- 25°C postoperatively. erative period is a clinical priority because it may limit Intravenous (IV) fluids were administered at a basal some hypothermia-related complications. 13,23-26The rate of 10 to 15 ml/kg/hr. Additionally, 4 ml of crystal- extent to which the above factors contribute-and any loid was given to compensate for each estimated millili- potential interactions among them-cannot be predicted ter of intraoperative blood loss.Blood was administered from current thermoregulatory theories. Postoperative as necessary to maintain a hematocrit level of at least thermoregulatory responses also have the potential to 25% to 30%. Fluid was heated to 37°C in the patients adversely effect hemodynamic values durin recovery: assignedto extra warming, but not in those assigned to 18 shivering by increasing metabolic demand and vaso- routine thermal management. constriction by increasing peripheral vascular resis- Most patients participating in this study were rela- tance.*’ Accordingly, we evaluated the postoperative he- tively healthy (ASA physical status I or II). Only 2 pa- modynamic and thermoregulatory consequences of tients in each group were assignedto ASA physical sta- intraoperative core hypothermia. tus III. Only 3 of 35 hypothermic and 5 of 39 normothermic patients took antihypertensive medica- tions before surgery. No vasoactive drugs were used Materials and Methods during or after surgery. All studies were undertaken with approval of the Uni- When surgery was complete, the neuromuscular versity of Vienna Ethics Committee and informed con- block was antagonized by administration of neostigmine sent from the subjects. 2.5 mg and glycopyrrolate 0.5 mg. The tracheas were then extubated, and the patients immediately trans- ferred to the post-anesthesiacare unit, where they were Protocol monitored for six hours. Neither group was warmed We studied 74 ASA physical status I, II, and III patients during the recovery period. undergoing elective colon surgery. None of the partici- During the recovery period, oxygen (0,) was admin-

360 J. Clin. Anesth., vol. 7, August 1995 Postoperative consequences of hypothermia: Kurz et al. istered at an initial rate of 6 Wmin via nasal prongs. intense warmth. Surgical pain was similarly evaluated Additional O2 was given by mask if necessary to main- with a VAS: 0 mm on this scale indicated no pain and tain oxygen saturation as measured by pulse oximetry 100 mm indicated the most intense pain imaginable. (SpO*) greater than 95%. Administered O2 was then These qualitative assessmentswere made by an observer gradually reduced to 2 to 3 Urnin over the first few who was blinded to the patients’ group assignment and hours of recovery. IV fluids were administered at a rate core temperatures. Data were recorded at 20-minute in- of at least 2 ml/kg. Postoperative pain was treated with tervals. the opioid piritramid delivered via a patient-controlled analgesia(PCA) pump (background infusion of 2.2 mgl hr; bolus dose of 1.5 mg; 15-minute lockout period). Data Analysis (This p-agonist is approximately four times as potent as morphine sulphate.) The nurses and physicians admin- Morphometric data, anesthetic parameters, and initial istering fluid and managing pain were blinded to the and final intraoperative temperatures in the two groups patients’ group assignments and core temperatures. were compared using unpaired t tests. Postoperative data also were compared using two-tailed, unpaired t tests. As suggested by Matthews et al.,34 we limited the number of comparisons by evaluating differences only at l-hour intervals. All values are expressed as means f SD. A p-value less than 0.01 was considered statistically Ambient temperatures were measured using a thermo- significant. couple positioned at the level of the patient, well away from any heat-producing equipment. Core tempera- tures were measured at the tympanic membrane. The Results tympanic probe was inserted until the patients felt the thermocouple touch the tympanic membrane. Appro- There were no statistically significant differences in priate placement was confirmed when they easily de- morphometric characteristics or anesthetic management tected a gentle rubbing of the attached wire. The probe in the patients assigned to hypothermia and normo- was then taped in place, the aural canal was occluded thermia (Table I). Core temperatures before induction with cotton, and the external ear was covered with a of anesthesiawere similar in both groups (37.0 + 0.2”C gauze pad. Tympanic membrane temperatures correlate us. 37.1 * 0.3”C). well with distal esophageal temperatures in the periop- As per protocol, core temperatures at the end of sur- erative period.3z30 gery in the hypothermic group were significantly less Thermoregulatory vasoconstriction was evaluated us- than those in the warmed patients (34.4 2 0.4% US.37 2 ing forearm minus fingertip, skin-temperature gradi- 0.3”C, Figure I). Initial postoperative core temperatures ents.” The gradients were recorded from an arm not in the warmed patients were 36.8 + 0.4”C, and they having an IV cannula or blood pressure (BP) cuff. As in increased only slightly during the 6-hour recovery pe- most of our previous studies,‘- we considered a gradi- riod. In contrast, initial core temperatures were 34.7 2 ent exceeding 4°C to indicate significant vasoconstric- 0.5”C in those patients who were allowed to become hy- tion. This value was chosen because it indicates nearly pothermic, and they increased at a rate of =0.5”C/hr for maximal vasoconstriction and it was unlikely to result the first three hours of recovery, and at =O.S”C/hr for from inadequate vascular volume or poorly controlled surgical pain. Heart rate (HR) was monitored continuously using three-lead electrocardiography. SpO, was measured con- Table 1. Morphometric Characteristics,Anesthetic tinuously using pulse oximetry, and BP was determined Management, and Intraoperative Temperaturesin Patients oscillometrically at 5-minute intervals from the upper Assignedto Hypothermia and Normothermia arm (Model 68S, Hewlett Packard, Waltham, MA). We Hypothermia Normothermia used oscillometric rather than direct arterial BP mea- surements to minimize the artifact induced by thermo- Age W 59a 14 572 15 regulatory vasoconstriction.32Z33 End-tidal isoflurane Weight (kg) 70* 13 71% 13 concentration was measured with a Cicero infrared Height (cm) 168 + 10 17Ok 9 monitor (DrPger, Liibeck Germany). Gender (M/F) 17118 22117 Postoperatively, shivering was evaluated qualitatively, Duration of Surgery (hr) 3.5 rt 1.3 3.3 2 1.1 using a three-point scale: grade zero indicated no shiv- Intraoperativefluid (L) 3.4 f 1.0 3.5 f 0.9 ering, grade one indicated mild or intermittent shiver- Administeredfentanyl (Kg) 65Ok 190 720% 220 ing, grade two indicated moderate shivering, and grade End-tidal (isoflurane) (%) 0.6 k 0.1 0.6 k 0.2 three indicated prolonged, intense shivering. Thermal Final intraoperative core comfort was evaluated using a lOO-mm-long visual ana- temp (“C) 34.4 2 0.4 37 2 0.3* log scale (VAS), with 0 mm indicating intense cold, 50 *Only final intraoperative core temperatures differed significantly mm indicating thermal comfort, and 100 mm indicating in the patients given each treatment.

J. Clin. Anesth., vol. 7, August 1995 361 Originul Contributions

Intraoperative Postoperative 8 38 Gradient 4 37 PC) Core 0 ‘(ey 36 0 -4 35 q Normothennic 80 34 6o Thermal 33 40 Comfort (mm) 0 1 2 3 end o 123456 20 Time(h) Time(h) I 0 Figure 1. Core temperatures before induction of anesthe- 80 0 Normothermic siawere similar in both groups. Core temperatures changed = Hypothermic little in patients warmed with forced air but decreased Shivering 60 22.5% in the unwarmed patients. Core temperature in the G;;;y 40 \ unwarmed patients increasedat a rate of =0.5”C/hr for the (%) 20 first three hours (h) or recovery and at =O.S”C/hr for the subsequenttwo hours. In contrast, core temperatures in- 0 creasedonly =O.l”C/hr for three hours and then remained 0 2 4 6 constant in the patients who were warmed intraoperatively. Time (h) Core temperatures in the hypothermic patients remained Figure 2. All hypothermic patients demonstrated signifi- significantly different from thosein the warmed group un- cant thermoregulatory vasoconstriction (gradients > 4) til more than 4 hours after surgery. Error bars indicate during the first three postoperative hours (h). Subse- standard deviations; asterisks(*) indicate statistically signif- quently, gradients in most of these patients decreasedbut icant differences between the groups. never becamenegative. None of the normothermic patients had postoperative gradients exceeding 4”C, and the mean was2°C during the first 2 hours of recovery; subsequently, the subsequent 2 hours. Core temperatures in the two most of thesepatients developed negative (vasodilated)gra- groups thus remained significantly different from those dients. The hypothermic patients reported feeling unpleas- in the warmed group for more than 4 hours after sur- antly cold, and the sensation persisted for more than 2 gery * hours. Only a few of the warmed patients shivered at any Half of the hypothermic patients showed significant time during the recovery period. In contrast, most of the intraoperative vasoconstriction at a core temperature of hypothermic patients demonstrated grades 2 or 3 tremor 34.5 2 0.4”C, with an isoflurane concentration of 0.6% during the first postoperative hour. By the end of the first + 0.1%. In contrast, none of the normothermic patients elapsedhour, however, shivering was rare, even in the hy- vasoconstricted. pothermic patients. Error bars indicate standard devia- During the first three postoperative hours, all hypo- tions; asterisks (*) indicate statistically significant differ- thermic patients demonstrated significant thermoregu- encesbetween the groups. latory vasoconstriction. Subsequently, gradients in most of these patients decreased, but they never became neg- unpleasantly cold sensation persisted, with VAS values ative (which would indicate full vasodilation). None of reported by the hypothermic patients remaining signif- the normothermic patients had postoperative gradients icantly less than those reported by the normothermic exceeding 4°C. The mean gradient in patients assigned patients for two hours. Only a few of the warmed pa- to forced-air warming was 2°C during the first 2 hours tients shivered at any time during the recovery period. of recovery; subsequently, most of these patients had In contrast, shivering was common in the hypothermic negative gradients (F&WY 2). patients, most of whom demonstrated grades two or During the initial six postanesthetic hours, Sp02, three tremor at some time during the first postoperative fluid balance, and ambient temperatures were similar in hour. By the end of the second elapsed hour, however, the two groups (Table 2). VAS pain scoreswere near 50 shivering was rare, even in the hypothermic patients mm immediately after surgery, but they decreased to (Figure 2). ~30 mm after 1 hour, and to -10 mm after two hours. HRs and BPS were similar in the normothermic and Pain scoreswere virtually identical in the two groups at hypothermic patients throughout the six-hour recovery every postoperative measurement interval. Seven hypo- period. None of the hemodynamic differences were sta- thermic and 2 normothermic patients were unable to tistically significant or clinically important (Figure 3). provide pain scores during the first 1.5 hours of recov- ery. Discussion The patients who were assignedto normothermia re- ported significantly higher thermal comfort scoresthan As expected from previous studies,10z’4internal redistri- patients who were allowed to become hypothermic. This bution of body heat reduced core temperature ==l”C

362 J. Clin. Anesth., vol. 7, August 1995 Postoperative consequences of hypothermia: Kun et al.

Table 2. Postoperative Oxygen Saturation (SpO,) Fluid =35”~.~s Our current patients were given less N20 Balance, Analgesia, and Temperatures During the Initial Six (60%) and lessisoflurane (==0.6%) than those in our pre- Hours of Postanesthetic Recovery vious study, 35but they were also given fentanyl. Thus, it is not surprising that the vasoconstriction threshold in Hypothermia Normothermia these patients was similar. Core temperatures between 36°C and 37.5”C define spa,@‘c) 98 t 1 98 2 1 the normothermic range in typical populations. How- Administered fluid (L) 1.4 2 0.6 1.5 + 0.6 Urine output (L) 0.6 2 0.4 0.5 2 0.4 ever, the range of core temperatures tolerated without Administered piritramid triggering thermoregulatory responses (interthreshold (mg) 19 + 9 16 + 6 range) in a given individual at a given time is only Ambient temperature (“C) 25.1 ? 0.6 25.6 2 0.5 =0.2”C.“j This “setpoint” range varies diurnally, usually Initial core temperature having a minimum near 3:00 AM and a maximum, = 1°C (“C) 34.7 2 0.5* 36.8 + 0.4 greater, near 3:00 PM. ” Initial postoperative core tem- Core temperature after peratures in the warmed patients averaged 36.8 2 0.4”C. 6 hr (“C) 37.2 2 0.4 37.3 2 0.4 Although this value is well within the normothermic Note: *Initial postoperative core temperatures were less in the pa- range, it is likely that the lower boundary of the inter- tients assigned to intraoperative hypothermia. threshold range was near 37°C at noon when most of these patients finished surgery.37 It is thus not surpris- ing that core temperatures, even in the “normothermic” during the first hour of anesthesia and surgery. The patients, increased slightly during the initial hours of initial reduction in core temperature was comparable in postanesthetic recovery. both groups,14 but forced-air warming subsequently in- The initial autonomic thermoregulatory response to creased core temperature in treated patients whereas hypothermia is peripheral vasoconstriction, mediated core temperature continued to decrease in those not ac- primarily by arteriovenous shunts located in the fingers, tively warmed. Consequently, all patients assigned to toes, and nose.38The next response in infants is non- forced-air warming were normothermic at the end of shivering thermogenesis,3g but nonshivering thermo- surgery, despite having undergone large abdominal op- genesis does not occur in adults given isoflurane,’ and it erations in a cool environment. Such efficacy is consis- probably contributes little to thermoregulation in un- tent with previous studies showing that forced-air warm- anesthetized adults.‘js7 And finally, shivering is initiated ing transfers more heat than circulating-water or when core tern erature is sufficiently less than the “set- 3tr infrared heating,” and maintains intraoperative core point” range. Most hypothermic patients both vaso- temperature better than circulatin -water or inspiratory constricted and shivered. Despite the shivering and va- gas heating and humidification.“, 5’ soconstriction, core temperatures in the hypothermic As expected from previous studies, none of the nor- patients increased relatively slowly during the first five mothermic patients vasoconstricted during surgery. ‘,* postanesthetic hours. N20 (70%) and isoflurane (0.75% end-tidal) decreases Core temperatures in the hypothermic group re- the thermoregulatory threshold for vasoconstriction to quired more than 4 hours to reach values comparable to those in the patients kept normothermic during surgery. The relatively slow 0.5Whr increase in core tempera- 0 Normothermic ture suggests that thermoregulatory compensations were blunted by residual anesthetic or a time-dependent factor. Far more rapid increases in core temperatures have been observed under other circumstances,40 sug- gesting that prolonged hypothermia did not result sim- 180 ply from the magnitude of the heat debt, but rather from inadequate activation of thermoregulatory com- pensations. Consistent with this theory is our observa- tion that shivering stopped long before core tempera- ture returned to normal. At least three factors may have contributed to pro- longed postoperative thermoregulatory inhibition. The first is residual volatile anesthetic. The dose-response 50 1 I curve for thermoregulatory inhibition by isoflurane ’ ’ &rnz (hf ’ 6 appears linear at surgical doses.3*41Inhibition has not been specifically tested at the low concentrations typical Figure 3. Heart rates and blood pressureswere similar in during recovery. But if impairment is indeed a linear the normothermic and hypothermic patients throughout function of dose in this range, representative early post- the &hour recovery period. Error bars indicate standard operative anesthetic concentrations (e.g., 0.2% isoflu- deviations; none of the differences was statistically signifi- rane’*) might decrease the vasoconstriction threshold cant. ~0.6”C. Although we did not measure postoperative

J. Clin. Anesth., vol. 7, August 1995 363 Originul Contributions

end-tidal isoflurane concentrations, values exceeding ing from general anesthesia.44Little or no change in 0.2% surely did not persist beyond the first half hour of central BPS during thermoregulatory vasoconstriction is recovery. Such concentrations would not be sufficient to also consistent with the relatively small fraction of the explain the observed prolonged hypothermia in the un- cardiac output traversing arteriovenous shunts.45 warmed patients. In contrast to our findings of comparable HRs and Opioids represent an additional factor perhaps im- comparable or increased BPS during recovery from hy- pairing postoperative thermoregulatory control. Not pothermic anesthesia, Gorman et ~1.~~ reported de- only were opioids used as an anesthetic adjuvant, but the creased HRs and increased SBPs and radial arterial BPS patients were given systemic opioids postoperatively to in hypothermic postoperative patients. Important dif- treat surgical pain. Opioids cause a variety of species- ferences between these studies include the facts that specific and dose-specific effects in animals, most typi- Gorman et al. studied an elderly group of relatively sick cally hyperthermia at low doses and hypothermia at patients and that those patients were admitted directly to higher doses. 42 The thermoregulatory consequences of the intensive unit and quite likely were heavily sedated. opioid administration in humans remain unclear, but Consistent with thermoregulatory inhibition in the el- they certainly do not generally include hyperthermia. derlY35*47s48and impairment of shivering induced by Many opioids inhibit postanesthetic shivering,‘s** and opioid administration, 20,4glittle shivering was observed it is probable that these drugs also cause generalized even in their hypothermic patients. Nonetheless, HRs thermoregulatory impairment in humans. It is likely and BPS may have been greater in the hypothermic pa- that such impairment reduced thermoregulatory efli- tients than in the normothermic patients studied by Gor- cacy and, consequently, prolonged hypothermia in our man et al. because thermoregulatory vasoconstriction postoperative patients. Consistent with this mechanism had more hemodynamic effect in their elderly popula- is our observation that recovery of core normothermia tion. was considerably more rapid in a previous study in Although statistically significant differences in HRs which opioids were not administered.‘* and BPS were reported in some of the above studies, the A final factor potentially contributing to impaired differences usually were small. These data suggest that postoperative thermoregulatory responses is a pro- aggravated cardiac ischemia in hypothermic postopera- longed effect of intraoperative core hypothermia. The tive patients26 may be mediated by a mechanism more extent to which postoperative thermoregulatory control subtle than gross hemodynamic alterations. We did not is altered by intraoperative thermal disturbances re- look for myocardial ischemia in this study. Another lim- mains unknown. However, patients recovering from hy- itation of our study is that we did not evaluate plasma pothermic cardiopulmonary bypass frequently develop catecholamine concentrations, which might have pro- a postoperative fever that appears to be a compensation vided a better indication of hypothermia-induced for the period of intraoperative hypothermia.43 Our “stress.” data do not allow us to quantify the contribution-if All hypothermic patients demonstrated significant any-of prior thermal perturbations to postoperative thermoregulatory vasoconstriction postoperatively. regulatory impairment. Mean skin-temperature gradients exceeded 4°C for Hemodynamic responses were similar in the two more than 3 hours during recovery. Few of these pa- groups, with no statistically significant or clinically im- tients vasodilated fully (gradient

364 J. Clin. Anesth., vol. 7, August 1995 Postoperative consequences of hypothermia: Kurr et al. sia is thermoregulatory.‘* In that study, we identified impairment results, at least in part, from opioids given two tremor patterns, the most common having electro- to treat surgical pain. myographic patterns similar to those produced by cold- induced shivering in the absence of anesthesia. Far less common was a phasic pattern resembling pathologic clo- Acknowledgment nus. Although we did not evaluate tremor patterns in We thank Gepa-Med, Inc. Vienna, Austria, for donating the dis- this study, it is likely that both patterns contributed to posable Bair Hugger” forced-air covers that we used. the observed muscular activity. Measurements of O2 consumption would have quantified physiologic stress of shivering better than our qualitative assessments. How- References ever, such measurements have been reported previ- ously? 1. Sessler DI, Olofsson CI, Rubinstein EH, Beebe JJ: The ther- Sp02 values were comparable in the two groups, as moregulatory threshold in humans during halothane anesthe- would be expected since 0, administration was titrated sia. Anesthesiology 1988;68:836-42. to maintain SpOB greater than 95%. Nonetheless, Sp02 2. Sessler DI, Oloffsson CI, Rubinstein EH: The thermoregula- was easily maintained in all patients, suggesting that hy- tory threshold in humans during nitrous oxide-fentanyl anes- thesia. Anesthesiology 1988;69:357-64. pothermia and the resulting shivering did not signifi- 3. Steen R, Sessler DI: The thermoregulatory threshold is in- cantly predispose to hypoxemia in the relatively healthy versely proportional to isoflurane concentration. Anesthesiology patients we studied. (Thermoregulatory vasoconstric- 1990;72:822-7. tion per se slightly increases SpO,, but the increase is 4. Washington DE, Sessler DI, McCuire J, Hynson JM, Schroeder only -2%, which is not clinically important.5) These re- M, Moayeri A: Painful stimulation minimally increases the sults are consistent with our observations in hypothermic thermoregulatory threshold for vasoconstriction during enflu- young adults. ‘* They also are c onsistent with our previ- rane anesthesia in humans. Anesthesiology 1992;77:286-90. ous observation that desaturation is rare in hypothermic 5. Hynson JM, Sessler DI, Belani KG, et al: Thermoregulatory infants and children during the postoperative period, vasoconstriction during propofol/nitrous oxide anesthesia in even when they are given no supplemental 02.44 It re- humans: threshold and oxyhemoglobin saturation. Aneslh An- mains likely that patients having preexisting severe pul- alg. 1992;75:947-52. monary compromise would tolerate postoperative hypo- 6. Jessen K: An assessment of human regulatory nonshivering thermia and shivering less successfully, and at least one thermogenesis. Acta Anaesthesiol Stand 1980;24: 138-43. 7. Jessen K, Rab@l A, Winkler K: Total body and splanchnic ther- such episode has been reported.50 mogenesis in curarized man during a short exposure to cold. We have postulated that shivering per se aggravates Acta Anaesthesiol Stand 1980;24:339-44. pain simply because vigorous muscular activity will jar 8. Hynson JM, Sessler DI, Moayeri A, McCuire J: Absence of areas in which surgical incisions were made. It also is nonshivering thermogenesis in anesthetized adult humans. An- possible that hypothermia directly increases pain per- esthesiolog> 1993;79:695-703. ception, or that it alters the pharmacodynamics of opi- 9. Camus Y, Delva E, Just B, Lienhart A: Leg warming minimizes oids. Our hypothermic and normothermic patients re- core hypothermia during abdominal surgery. Anesth Analg. quired comparable amounts of opioid (administered via 1993;77:995-9. PCA) to maintain similar VAS pain scores. However, it 10. Hynson JM, Sessler DI: Intraoperative warming therapies: a remains likely that hypothermia delayed opioid clear- comparison of three devices. J Clin Anesth 1992;4: 194-9. ance and that plasma concentrations were greater in the 11. Mills P, Sessler DI, Moseley M, et al: An in vivo “F nuclear unwarmed patients. Since we did not measure plasma magnetic resonance study of isoflurane elimination from the opioid concentrations, we cannot directly evaluate the rabbit brain. Anesthesiology 1987;67: 169-73. extent to which postoperative hypothermia contributes 12. Sessler DI, Rubinstein EH, Moayeri A: Physiological responses to mild perianesthetic hypothermia in humans. Anesthesiology to surgical pain. 1991;75:594-610. In summary, we randomly assigned 74 patients to be 13. Just B, Delva E, Camus Y, Leinhart A: Oxygen uptake during kept normothermic or =2.5”C hypothermic during co- recovery following naloxone. Relationship with intraoperative lon surgery. Morphometric characteristics, SpO,, fluid heat loss. Anesthesiology 1992;76:60-4. balance, PCA-administered opioid, and VAS pain scores 14. Sessler DI, Moayeri A, Steen R, Glosten B, Hynson JM, were comparable in the two groups. Hypothermic pa- McCuire J: Thermoregulatory vasoconstriction decreases cu- tients felt uncomfortably cold during recovery, and their taneous heat loss. Anesthesiology 1990;73:656-60. postoperative core temperatures remained significantly 15. Belani KG, Sessler DI, Sessler AM, et al: Leg heat content less in the normothermic patients for more than four continues to decrease during the core temperature plateau in hours. Shivering was common in the hypothermic pa- humans anesthetized with isoflurane. Aneslhesiology 1993;78: tients, but rare in those kept normothermic. HRs and 856-63. BPS were comparable in the two groups. These data 16. Burton AC: Human calorimetry: the average temperature of the tissues of the body. J Nutr 1935;9:261-80. indicate that intraoperative hypothermia causes postop- 17. Stevens WC, Cromwell TH, Halsey MJ, Eger EI 2d, Shake- erative thermal discomfort and that full recovery from speare TF, Bahlman SH: The cardiovascular effects of a new hypothermia requires many hours. Delayed return to inhalation anesthetic, forane, in human volunteers at constant core normothermia appears to result largely from post- arterial carbon dioxide tension. Anesthesiology 1971;35:8-16. operative thermoregulatory impairment. It is likely that 18. Rodriguez JL, Weissman C, Damask MC, Askenazi J, Hyman

J. Clin. Anesth., vol. 7, August 1995 365 Original Contribulions

AI, Kinney JM: Morphine and postoperative rewarming in nitrous oxide/isoflurane anesthesia is lower in elderly than in critically ill patients. Circuhtion 1983;68: 123@6. young patients. Anesthesiology 1993;79:465-9. 19. Liem ST, Aldrete JA: Control of post-anaesthetic shivering. 36. Lopez M, Sessler DI, Walter K, Emerick T, Ozaki M: Rate and Can Anaesth Sot J 1974;21:50&10. gender dependence of the sweating, vasoconstriction, and 20. Kurz M, Belani KG, Sessler DI, et al: Naloxone, meperidine, shivering thresholds in humans. Anesthesiology 1994;80:780-8. and shivering. Anesthesiology 1993;79: 1193-201. 37. Sessler DI, Lee KA, McCuire J: Isoflurane anesthesia and cir- 21. Guffin A, Girard D, Kaplan JA: Shivering following cardiac cadian temperature cycles in humans. Anesthesiology 1991;75: surgery: hemodynamic changes and reversal. J Cardiothorac 985-9. VOX Anesth 1987; 1:24-8. 38 Hales JRS: Skin arteriovenous anastomoses, their control and 22. Pauca AL, Savage RT, Simpson S, Roy RC: Effect of , role in thermoregulation. In: Johansen K, Burggren W (eds): fentanyl and morphine on post-operative shivering in man. Cardiovascular Shunts: Phylogenetic, Ontogenetic and Clinical As- Acta Anaesth&ol Stand 1984;28: 138-43. pects. Copenhagen: Munksgaard, 1985;433-51. 23. Leslie K, Sessler DI, Bjorksten AR, Moayeri A: Mild hypother- 39. Dawkins MJ, Scopes JW: Non-shivering thermogenesis and mia alters propofol pharmacokinetics and increases the dura- brown adipose tissue in the human new-born infant. Nature tion of action of atracurium. Anesth Analg (in press). 1965;206:201-2. 24. Valeri CR, Khabbaz K, Khuri SF, et al: Effect of skin temper- ature on platelet function in patients undergoing extracorpo- 40. Giesbrecht CC, Sessler DI, Mekjavic IB, Schroeder M, Bristow GK: Treatment of mild immersion hypothermia by direct real bypass. J Thorac Cardiouasc Surg 1992; 104: 108-16. body-to-body contact. J Appl Physiol 1994;76:2373-9. 25. Sheffield CW, Sessler DI, Hunt TK: Mild hypothermia during isoflurane anesthesia decreases resistance to E. coli dermal in- 41. Washington DE, Sessler DI, Moayeri A, et al: Thermoregula- fection in guinea pigs. Acta Anaesthesiol Stand 1994;38:201-5. tory responses to hyperthermia during isotlurane anesthesia in 26. Frank SM, Beattie C, Christopherson R, et al: Unintentional humans. J Appl Physiol 1993;74:82-7. hypothermia is associated with postoperative myocardial isch- 42. Roscow CE, Miller JM, Pelikan EW, Cochin J: Opiates and emia. The Perioperative Ischemia Randomized Anesthesia thermoregulation in mice. I. Agonists. J Phurmacol Exp Ther Trial Study Group. Anesthesiology 1993;78:468-76. 1980;213:273-83. 27. Rowe11 LB: Cardiovascular aspects of human thermoregula- 43. Schmidt H, HauDmann R, Biihrer H, et al: Endotoxemia fol- tion. Circ Res 1983;52:367-79. lowing low-flow, low-pressure cardiopulmonary bypass [Ab- 28. Sessler DI, Moayeri A: Skin-surface warming: heat flux and stract]. Anesthesiology 1993;79:A308. central temperature. Anesthesiology 1990;73:218-24. 44. Bissonnette B, Sessler DI: Mild hypothermia does not impair 29. Kurz A, Kurz M, Poeschl G, Faryniak B, Red1 G, Hack1 W: postanesthetic recovery in infants and children. Anesth Analg Forced-air warming maintains intraoperative normothermia 1993;76: 168-72. better than circulating-water mattresses. Aneslh Analg 1993;77: 45. Hales JR, Fawcett AA, Bennett JW, Needham AD: Thermal 89-95. control of blood flow through capillaries and arteriovenous 30. Cork RC, Vaughan RW, Humphrey LS: Precision and accu- anastomoses in skin of sheep. Pjlugers Archiv 1978;378:55-63. racy of intraoperative temperature monitoring. Anesth Analg 46. Frank SM, Higgins MS, Breslow MJ, et al: The catecholamine, 1983;62:21 l-4. cortisol, and hemodynamic responses to mild perioperative hy- 31. Rubinstein EH, Sessler DI: Skin-surface temperature gradi- pothermia. Anesthesiology 1995;82:83-93. ents correlate with fingertip blood flow in humans. Anesthesi- 47. Wagner JA, Robinson S, Marino RP: Age and temperature ology 1990;73:541-5. regulation of humans in neutral and cold environments. J A# 32. Hynson JM, Sessler DI, Moayeri A, McCuire J, Schroeder M: Physiol 1974;37:562-5. The effects of preinduction warming on temperature and 48. Frank SM, Beattie C, Christopherson R, et al: Epidural versus blood pressure during propofol/nitrous oxide anesthesia. An- general anesthesia, ambient operating room temperature, and esthesiology 1993;79:219-28. 33. Hynson J, Moayeri A, Sessler DI: Femoral-to-radial arterial patient age as predictors of inadvertent hypothermia. Anesthe- pressure gradients during vasoconstriction, sweating, and siology 1992;77:252-7. propofol anesthesia [Abstract]. Anesthesiology 1993;79:A116. 49. Guffin A, Girard D, Kaplan JA: Shivering following cardiac 34. Matthews JN, Altman DC, Campbell MJ, Royston P: Analysis surgery: hemodynamic changes and reversal. J Cardiovasc Tho- of serial measurements in medical research. BMJ 1990;300: rat Anesth 1987; 1:248. 230-5. 50. Bay J, Nunn JF, Prys-Roberts C: Factors influencing arterial 35. Kurz A, Plattner 0, Sessler DI, Huemer G, Red1 G, Lackner F: PO, during recovery from anaesthesia. Br J Anaeslh 1968;40: The threshold for thermoregulatory vasoconstriction during 398407.

366 J. Clin. Anesth., vol. 7, August 1995