Adverse effects of hypothermia in postoperative patients

Adverse Effects of Hypothermia in Postoperative Patients Gu$ J. Slotman, MD, Providence, Rhode Island Erlca H, Jed, MD, Providence, Rhode Island Kenneth W. Burchard, MD, Providence, Rhode Island

Hypothermia is a frequent occurrence during major surgical procedures [I,2]. The treatment of profound accidental hypothermia, the short-term metabolic effects of mild and severe hypothermia, and the use of hypothermia as a therapeutic modality have been well described in the medical literature [3-10]. In the past decade, the most frequent indication for induced hypothermia has been cardiopulmonary bypass [3,7,II], and occasionally, it has been used as an adjunctive measure for the resection of large hepatic tumors [12,13]. Although the adverse effects of profound accidental hypothermia and the short-term metabolic sequelae of mild, incidental intraoperative hypothermia and induced hypothermia are well understood, the relationship of postoperative hypothermia to subsequent morbidity and mortality is unknown. The present study was undertaken to evaluate the effects of operative and postoperative hypothermia and the associated risk factors on postsurgical morbidity and mortality. Material and Methods

Charts of 100 consecutive patients admitted postoperativelyto the surgical intensive care unit of Rhode Island Hospital, Providence, Rhode Island were reviewed. Data collected included diagnosis, operative procedure, age, length of surgery, operative fluid requirements, emergency versus elective surgery, mortality, and postoperative complications. Temperature and blood pressure were recorded preoperatively, intraoperatively, at the end of the procedure (time 0), and 2, 4, and 8 hours postoperatively. From the Department of Surgery, Brown Unlvmsity, Rhode Island Hospital,

593 EddyStreet,Providence, Rhode Island. Requestsfor ml~ints~ouklbea~o~ad to Gus J. $1o~. MD, Rf~xJe IslandHospital, 593 Eddy Street, Providence, Rhode Island 02903. Presented st the 65thAnn~! Meetingof the New EnglandSurglcatSo-

ciety,DlxvllleNotch.New Hampshire,October12-14, 1984.

Vo~m 149, Ai)dl1989

Hypothermia was defined as an esophageal or rectal temperature of less than 97°F. Hypotension was defined as a systolic blood pressure of less than 100 mm Hg. Statistical analysis was carried out using the chi-square equation and the unpaired Student's t test where appropriate. Results

Of the 100 patients reviewed, 39 were male and 61 female. Seventeen patients died (I0 were male and 7 female). Fifty-three patients were older than 55 years of age and 47 less than 55 years of age. Thirtynine emergency procedures and 61 elective procedures were carried out. Operative procedures were gastrointestinal (72 patients), peripheral vascular (21 patients), urologic (3 patients), and others (4 patients). Five patients had a temperature of more than 101°F before surgery, with one postoperative death. No patients were hypothermic preoperatively. Emergency operations were more frequently gastrointestinal than were elective operations (33 of 39 emergency operations versus 22 of 61 elective operations, p <0.05). During surgery, 56 of the 73 patients in whom temperature was recorded (77 percent) were hypothermic (Table I). Core temperature was less than 97°F in 53 percent of the patients at time 0, in 39 percent 2 hours postoperatively, in 21 percent 4 hours postoperatively, and in 5 percent 8 hours postoperatively. The incidence of temperature less than 97°F was not significmltly different for gastrointestinal procedures compared with other operations at time 0 (48 percent versus 71 percent) or at 4 hours (15 percent versus .36 percent). Operative fluid requirements were significantly greater for patients who remained hypothermic after operation (1,589 4- 763 ml/hour versus 1,123 ~: 449 ml/hour at time 0, p <0.01). Patients appeared to achieve nor-

495

$1olman et al

TABLE I

Tempera~lure, Operative Fluid Requirements, and Length of Surgery

Temperature >97°F <97°F Operative fluid ~'equlrements (ml/h) >97°F <97°F Length of surgery (h) >97°F <97°F

During Surgew

Time 0"

2 h Postop

4 h Postop

8 h Post('p

17 (23%) 56 (77%)

31 (47%) 34 {53%)

45 (61%) 29 (39%)

66 (79%) 17 (21%)

89 (95%) 5 (5%)

1,123 4- 499 1,589 4- 763 t

1,048 + 438 1,483 4- 7161

t,159 :E 526 1,651 -~ 7201

1,169 4- 520 2,146 4- 733 t

3.16 4- 1.20 4,24 4- 1.58 t

3.50 4- 1.40 4.31 :E 1.63I

966 4- 466 1,230 4- 557 3.06 :E 1.t6 4.26 4- 1,541

3.61 4- 1.61 4.71 4- 1.401

3.78 4- 1.57 4,5b d: 1.33

• End of operation. p <0.01. t p <0.05.

mothermia at a rate inversely proportional to operative fluid requiro.ments, as these differences persisted through 8 hours (2,146 ~: 733 ml/hour for patients with temperatures less than 97°F versus 1,169 + 520 ml/hour for patients with temperatures greater than 97°F, p <0.01). Length of surgery (Table I) was significantly greater for hypothermic patients, from the operating room through 4 hours postoperatively. For emergency surgery compared with elective surgery, anesthesia time was shorter in the former (2.92 4- 1.44 hours versus 4.34 4- 1.3 hours, p <0.01), and operative pH was lower (7.32 4- 0.11 versus 7.40 40.07, p <0.05). Hypothermia intraoperatively and at time 0 was not significantly associated with morta]~ity (Figure 1). However, the mort~ity of patients who remained hypothermic at 2 hours was significantly increased compared with that of normothermic patients (7 of 29 patients died versus 2 of 45 patients, p <0.01). Mortality also significantly increased in relation to

SO % Mortality

70 60

$ pg7~

50 40 30 20 10

n

0 Ternpereture >9:"°<~,7° >~,7"<9r° >9~<96 >~¢<~¢>97°<~¢ Timepost-operative: OR 0 2 4 8 (Hrs) Figure I, Temperature and mortality during and after aurgery, mortality of those wffh temperaturesgreater than and less

fhan 9r° F durln~ e~pery and O, 2, 4, ana S Oou~ tmstap~RU~ is depicted.

496

hypothermia at 4 and 8 hours. The effect of a temperature of less than 97°F at 4 hours on mortality within other clinical groups is seen in Figure 2. When patient~ with intraoperative hypotension were excluded, mortality in the presence of hypothermia at 4 hours remained significantly increased (4 of 10 hypothermic patients died versus 2 of 54 patients with a temperature greater thala 97°F, p <0.01). Compared with normothermia, mortality with a body temperature less than 97°F was increased significantly in those who had emergency surgery, gastrointestinal procedures, and postoperative complications. Mortality was significantly increased in relation to hypotension at any time (Figure 3). Intraoperative hypotension was associated with a 36 percent mortality (5 of 14 patients) compared with a 7 percent mortality (4 of 60 patients) for normotensive patients (p <0.05). Mortality increased progressively with prolonged hypotension, reaching 75 percent in those with a blood pressure of less than 100 rnm Hg at 8 hours. Among these patients, the duration of hypotension was significantly longer in those who died (4.04 4- versus 1.55 -l- 1.21 hours, p <0.0I). Other clinical conditions associated with increased postoperative mortality are listed in Table II. These include patient age of more than 55 years, emergency surgery, operative fluid requirements of more than 1,500 ml/hour, and postoperative c o m p l i c a t i o n s . There were 68 postoperative complications occurring in 51 patients. One patient with terminal cancer who was admitted briefly to the intensive care unit was excluded from the complications and mortality analysis. These complications included cardiac arrhythmias (15 patients), acute respiratory failure (11 patients), wound infection or dehiscence (8 patients), urinary tract infection (8 patients), pulmonary atelectasis (7 patients), acute renal failure (6 patients), systemic sepsis (5 patients), pneumonia (2 patients), and other complications (6 patients). Operative fluid requirements were signiEcantly increased in patients with each of the mortality risk factors listed in Table HI. The interact,:on of the six

The An,l~rk~n Journal of ~j,rget'y

Adverse Effacts o{ Postoperative Hypothermla

.

.

.

.

.

.

.

.

.

.

EMERGENCY SURGERY

.

I

-- ....

.

°4

**

70

Mortality POST-OPERATIVE COMPLICAT|ONS

~P

p
~ >;oo

60

I

J * 40

.

GASTROINTESTINAL PROCEDURES 20

'°

OPERATIVE BP <

)00

rnmH~

0 {}P

** OPERATIVE BP~> 100 mmHg] ;:~

OPERAT|VE FLUIDS~' 15OO m|/hr

..

0

t

_ ,I_

20

i

p97 ° ~k p <0.05 vS T > 97 ~

} |

z

,~

t

80

Figure 2, Mortality In pat/et~ wllh hypothem)la ( temperature less than 94 ° F) at 4 hours, compared with I ~ t l ~ f s wilh temperatures greater than 97 ~F (clinical risk groups). BP = blood pressure; T = temperature,

~, Iot3 ,~1oo

> ~ o o <~Joo

2 - - ~ >1oo (%00)|oo

<1oo

~too ~|oo

F l ~ r e 3. The effoct of Intraoperatlvo postoperatlvo hypotermlon (blood p r e ~ r ~ tees than 100 mm Hg) on mo~alHy.

TABLE li

40 60 % Mortality

_

(mfn;-t~)

Clinical Conditions Associated With increased Postoperaltve Mortality Risk Factor

Mortality (%)

p Value"

Age > 5 5 yr Emergency surgery Operative fluids requirements > 1500 ml/h Po~top compllcatlons

26 32 45

<0.01 <0.001 <0.001

31

<0.001

* Versus absence of risk factor.

TABLE Ill

OperaUve Fluid Requirements and Mortality

Risk Factors

mortality risk factors is depicted in Table IV. Although patient age of more than 55 years, emergency surgery, and operative fluid requirements of more than 1,500 ml/hour significantly increased mortality for the entire study population, they did not significantly effect survival within most of the other mortality r~sk groups. Postoperative complications, however, were associated with significantly increased mortality for patients with or without all other mortality risk factors, except an age of less than 55 years and operative fluid requirements of more than 1,500 ml/hour. Operative hypotension was associated with increased mortality for those 55 years of age or ~:ounger who had emergency surgery and a body temperature of more than 97°F at 4 hours. An operative blood pressure of less than 100 mm Hg and a body temperature of less than 97°F at 4 hours did not effect mortality when combined, but they were both associated with significantly increased mortality for patients with postoperative complications. The clinical association of mortality risk factors is summarized in Table V. Compared with its absence, each risk factor was associated with a signficantly increased frequency of at least two other raortality risk factors. The number of such significant

Vo~me 149, April 1985

Operative Fluid Requirements (ml/h)

Rtc,k Factors Age > 5 5 yr <55 yr Emergency surgery Elective surge)'y Operative blood pressure <100 rnm Hg > 100 mm Hg Postop comp!lcations Yes No

1,425 4- 676" 1,024 4- 469 1,562 4- 758" 1,051 4- 383 1,819 4- 826" 1,085 4- 440 1,435 4- 827" 1,005 4- 375

° p <0.01.

associations ws not related to the mortalty rate for any individual risk factor. Comments Total body hypothermia has been described in the medical literature since the earliest times. The Hippocratic school in the fifth and fourth centuries BC recommended systemic hypothermia to treat tetanus and convulsions [3]. Since t h a t time, thera-

497

TABLE I V

lnteracllon of Postoperative Mortality Rlsk Factors: Effects on Suhroup Mortallly* Age

Emergency

Risk Factor

Y~'eryN~

Age >55 yr <55 yr Emergency surgery 9/24 flecilve surgery 5/30 Operatlve BP >I00 mm Ha 8/39t 7/16 I500 ml/h 9/41t <1500 ml/h Tat4h >97OF 6/36 <97OF 6/13 Postop complicattons Yes 13/35 NO 0120

...

...

tig

. .. . ..

3/15 Oi31

Operative BP
OFR milh ->1500 Yes

No

9/18 013

7/33 0146

T <97"F at 4 h Yes No

Postop Complii-stlons Yes No

9/24 3/15

...

...

0/42 3/5

4/28 9/12f

3/6 0140

8/16 4/12

1/32 1/1

5/31 5/6

3/13 0128

12/27 0/12

4/24 0/37

10115t 016

6/32 0146

6/9t 0120

6/31 0/35

.. .

...

... ...

' Data expressed as number who died/number at risk. t p <0.01 versus absence of risk factor in column heading. p <0.05 versus absence of risk factor In column heading. Temperature at 8 hours. BP = blood pressure; OFR = operative fluid replacement; T = temperature.

peutic hypothermia has undergone a cyclical evolution of reception and rejection, while becoming firmly established in the current medical literature ad a means of protecting the myocardium during cardiopulmonary bypass [7,14] and protecting normal liver tissue during major hepatic tumor resections [12,13]. Although not without complications, in these situations, profound hypothermia has been considered beneficial in reducing the incidence of postsurgical morbidity and mortality. The adverse effects of profound accidental hypothermia are also well-known. Mortality after severe hypothermia secondary to environmental exposure can exceed 80 percent, and is associated with cardiorespiratory arrest, as well as disseminated intravascular coagulation, among other complications [10,15-181. Additional clinical settings in which

TABLE V

profound hypothermia may occur include the metabolic sequelae of endocrine hypofunction, hypothalamic dysfunction, drug-induced hypothermia, and protein-calorie malnutrition [ B ] . In these situations, mortality is reported to be more closely associated with the underlying disease than with the magnitude of temperature depression. The phenomenon of inadvertent hypothermia during general surgical procedures has been described by several investigators. In most studies, operative tesnpi?ratured3pression has varied directly with age, length of anesthesia, gastrointestinal surgery, arid the administration of muscle relaxants as part of the anesthesia regimen [I ,2,5]. These demographic findings were similar to those of the present study in which elderly patients were more frequently hypothermic than younger patients. The length of

Mortalltv Rfsk Factors: Cllnlcal Assciclaflont (values expressed as percentages)

Risk Factor

Age >55 yr

Age >55 yr Emargency surgery Operative BP 1500 mllh T <97OF at 4 h Postop complications

45 30' 26 27' 63

...

Emergency SurgerY

Operative BP C100mmHg

OFR >1500ml/h

T <97OF at 4 h

Postop Complications

62

76' 57

31 41' 16 69t

70 80t 63t

577 37 71

93' 46 50' 40

69' 53' 29 36' 23

...

.,.

...

42t 857

...

47

...

p <0.05 versus absence of risk factor in column heading. t p <0.01 versus absence of risk factor in column heading. BP = blood presswe; O R = operative fluid requirements; 'l = temperature.

The Amerkan Journal oi S ~ J I O ~ ~

Adverse Effects of Postoperative Hypothermia

surgery and intraoperative fluid requirements were increased for patients with postoperative hypothermia compared with those with a body temperature above 97°F. However, in contrast to previous investigations, the incidence of hypothermia was not increased with gastrointestinal procedures, either in the operating room or 4 hours postoperatively. The most significant finding of this study is the association of prolonged postoperative hypothermia with increased mortality after surgery. To our knowledge this has not been previously reported. Several investigators have suggested that the most dangerous time for patients subjected to intraoperative hypothermia is the immediate postanesthesia recovery period when the thermostatic reflexes reappear [1,2]. The severe shivering response frequently seen in this situation can increase oxygen consumption several-fold [2], theoretically placing (~.lderly patients and patients with preexisting pulmonary or cardiovascular disease at significant risk. Rodriguez et al [19] have recommended the administration of neuromuscular blockh~g agents during this rewarming period in order to reduce caloric and metabolic demands and myocardial work. Jones and McLaren [20] have similarly recommended routine oxygen therapy for hypothermic patients in the initial two to three hours after operation in an effort to ameliorate the severe reductions in oxygen saturation associated with postoperative shivering. Our data suggest that these more immediate perioperative metabolic changes may not be related to subsequent morbidity and mortality. Rather, it appears that physiologic stress of sufficient magnitude to render patients hypothermic for many hours after surgery may be the prime determinant of mortality in critically ill surgical patients. As already discussed, previous studies have indicated that elderly patients more frequently become hypothermic than do younger patients [1,2]. This finding was confirmed in the present study. Such age-related effects are thought to be due to hypothalamic thermoregulatory failure, limited cardiopulmonary reserves [21,22], and protein-calorie malnutrition [21-23]. Although these mechanisms may be determinants of the increased incidence of postoperative hypothermia in older patients, they do not appear to significantly influence mortality. In this study, although the mortality risk factors of a body temperature of less than 97°F at 4 hours and an operative blood pressure of less than 100 mm Hg were significantly more frequent among patients over 55 years of age than in the younger group, they were associated with increased mortality only for patients less than 55 years of age. Elderly patients, then, are more likely to become hypotensive during surgery and hypothermic postoperatively, but mortality for them is not significantly increased with the presence of either risk factor. It may be that the overall in-

Volume 149, April 1985

crease in mortality risk for patients more than 55 years of age is simply not increased further by the addition of other mortality risk factors. These interrelationships are not clear from the data. In contrast, however, younger patients subjected to hypotension, prolonged hypothermia, or both appear to be less able to compensate for these risk factors, resulting in significantly increased mortality. In the final analysis, there are multiple determinants of postoperative morbidity and mortality, as evidenced by the six significm~t mortality risk factors established in this study. For the most part, these factors were associated with increased mortality for the study group as a whole and for selected subgroups which, in general, were selected for the absence of another significant risk factor. Possibleexceptions were operative hypotension and hypothermia 4 hours postoperatively, which were both associated with increased mortality for patients with postoperative complications. Persistent hypothermia and hypotension at any time appear to be relatively stronger determinants of mortality, compared with the other static risk factors, and may influence survival independent of the other predictive variables. Nevertheless, the data here do not repudiate the suggestion that the sickest patients with multiple risk factors die more often. Multivariable statistical analysis may be helpful in distinguishing those parameters most closely associated with mortality. Although a multifactorial analysis of these data may more clearly identify the interactions of determiDants of postoperative mortality, only operative hypotension and postoperative hypothermia of the factors discussed herein, are potentially preventable or reversible. Maintenance of intraoperative blood pressure is obviously a central concept in anesthetic and surgical management. Our results suggest that maintenance of operative normothcrmia, rapid postoperative rewarming, or both may also lessen morbidity and improve survival after surgery.

Summary The effect of intraoperative and postoperative temperature on morbidity, mortality, and other clinical risk factors was evaluated in 100 consecutive general surgical patients admitted postoperatively to a surgical intensive care unit. Hypothermia (temperature less than 97°F) was present in 77 percent of the patients intraoperatively, in 53 percent at the end of surgery, and in 21 percent at 4 hours. Mortality was increased with patient age greater than 55 years, emergency surgery, operative blood pressure less than 100 mm Hg, operative fluid requirements greater than 1,500 ml/hour, temperature less than 97°F at 2, 4, and 8 hours postoperatively, and presence of postoperative complications. Intraoperative fluid requirements were significantly greater

499

Slotman et al

for p a t i e n t s with m o r t a l i t y risk factors. P a t i e n t s over 55 y e a r s o f age w e r e m o r e o f t e n h y p o t e n s i v e a n d hypothermic than younger patients, but mortality was i n c r e a s e d o n l y for p a t i e n t s less t h a n 55 y e a r s o f a g e w i t h a t e m p e r a t u r e o f less t h a n 9 7 " F a t 8 h o u r s o r a n o p e r a t i v e b l o o d p r e s s u r e of less t h a n 100 m m Hg. M o r t a l i t y a f t e r g e n e r a l s u r g i c a l p r o c e d u r e s is i n c r e a s e d w i t h o p e r a t i v e h:~potension a n d p r o l o n g e d postoperative hypothermia. Hypothermic patients w i t h m o r t a l i t y risk f a c t o r s s h o u l d be a g g r e s s i v e l y rewarmed postoperatively. References

1. Goldberg MJ, Roe CF. Temperature changes during anesthesia and operations. Arch Surg 1966;93:365-9. 2. Dlenes RS. Inadvertenthypothermla in the operating room. Plast Reconstr Surg 1981;67:253-4. 3. Swan H. Clinical hypothermia: a lady with a past and some promise for the future. Surgery 1973;73:736-58. 4. Norwood WI, Norwood CR. Influence of hypothermla on Intracellular pH during anoxla. Cell Physlol 1982;12:62-5. 5. Vaughan MS, Vaughan RW, Cork RC. Postoperative hypothermta in adults: relationship of age, anesthesia, and shivering to rewarmlng. Anesth Analg 1981;60:746-51. 6. Reuler JB. Hypothermla: pathophysioiogy, clinical settings, and management. Ann Intern Med 1978;89(4):519-527. 7. Lederman RJ, Breuer AC, Hanson MR, et el. Peripheral nervous system complications of coronary artery bypass graft surgery. Ann Neurol 1982; 12:297-301. 8. Thomas BA, Hessel EA, Harker LA, Sands MP, DIIlard DH. Platelet function during and after deep surface hypothermla. J Surg Res 1981;31:314-8. 9. Hessel EA, Schmer G, Dlllard DH. Platelet kinetics during deep hypothermid. J Surg Res 1980;28:23-34. 10. Althaus U, Aeberhard P, Schupbach P, Nachbur BH, Muhlemann W. Management of profound accidental hypothermla with cardioresptratory arrest. Ann Surg 1982; 195:492-5. 11. PangLM, Stalcup SA, Upset JS, Hayes C,J,Bowman FO, Malllns RB. increased circulating bradyklnln during hypothermla and cardiopulmonary bypass in children. Circulation 1979;60: 1503-7. 12. Etn SH, Shandling S, Williams WG, Trusler G. Major hepatic tumor resection using profound hypothermla and circulation arrest. J Pedlatr Surg 1981;16:339-42. t3. Fortner JG, Shlu MH, Kinne DW, et el. Major hepatic resection using vascular Isolation and hypothermlc perfuslon. Ann Surg 1974;180:644-52. 14. Kohda Y, Tomtnaga R, Yoshltoshl M, Tokunaga K. Optimal myocardial and reparfusate temperature in global lschemla induced in rats. J Surg Res 1982;32:154-60. 15. Fruehan AE. Accidental hypothermla. Arch Intern Meal 1960; 105:219-29. 16. h~rrlwethar WD, Goodman RM. Severe accidental hypothermia with survival after rapid rewarmlng. Am J Med 1972;53: 505-10. 17. Weyman AE, Greenbaum DM, Grace WJ. Accidental hypothermla In an alcoholic population. Am J Med 1974;56: 13-21. 18. MahaJanSt., Myers TJ, Baldtnl MG. Disseminated Intravascular coagulatiOnduring rowarming following hypothermta. JAMA 1981 ;245:2517-8.

19. Rodrlguez JL, Welssman C, Damask MC, Askanaz,~J, Hyman At, Kinney JM. Physiologic requirements during rewarming: suppression of the shivering response. Crlt Care Med 1983;11:490-7. 20. Jones HD, McLaren CAR. Postoperative shivering and hypoxaemla a~e¢haiothane, nitrous oxide and oxygen anaesthesia.

500

Brlt J Anaesth 1985;37:35-41. 21. Heymann AD. The effect of incidental hypothermla on elderly surgical patients. J Gerontol 1977;32:46-8. 22. Bestow MD, Rawllngs J, Allison SP. Undernutritton, hypothermla, and injury in elderly women with fractured femur: An injury response to altered metabolism? Lancet 1983;1: 143-5.

23. Hillman H. Hypothermla and old age. Practitioner 1984;228: 285-8.

Discussion

G e o r g e H. A. Clowes, J r . (Boston, MA): Having studied hypothermia extensively in the days of early cardiac surgery, and having taken people down to 5°C and rewarmed them, I came to know a good deal about the difficulties that go with this. First of all, the circulation is unable to maintain itself at any temperature much below 90°F. People really don't rewarm themselves very well until they get up to about 92°F. T h e difficulty lies in the fact, and everybody knows it, that in a patient who is dying of septic shock, if he cools to a below normal temperature he will be dead in a few hours. So this brings up the question of what is the difficulty that has resulted in this? Sure, the patient can have evaporative heat 19ss during the operation, and, Dr. Slotman, you have reported that this really d i d n ' t make much difference in mortality. But if the patient is unable to rewarm himself, it seems to me that this represents a failure of metabolic energy production; he is unable to make enough heat to keep up with ambient heat loss. However, the difficulty either lies in the amount of circulatory output, or it's a metabolic problem caused by sepsis, mitochondrial difficulties, and so on. Dr. Slotman, do you have any information on these patients relative to cardiac output, metabolic acidosis, lactacidemia, or oxygen consumption? It seems to me t h a t the patient who goes 8 hours and still is cold is the one who is not using enough oxygen to rewarm himself or, for that matter, to take care of all of his other defense mechanisms. In that context, just before I finish, I want to say that in having had the pleasure of reviewing this paper, 32 out of 68 complications were infectious. We very well know from H u n t ' s work, Burke's work, and the work of lots of others that failure of the metabolic system is the best way to cause an infection. I raise these issues because it does not seem to me t h a t the proper treatment is outside rewarming; instead, it is to get the patient's circulation going, see that his oxygen intake is good, and so on. Gus J . S l o t m a n (closing): Dr. Clowes, I agree with you, and I have tried to relate t h a t all of the patients who became hypothermic had undergone very significant physiologic stress during surgery, and certainly there were multiple important factors. I cannot distinguish whether or not hypothermia postoperatively and sustained hypothermia were not indeed a result of the intraoperative stress which was mentioned in the manuscript. There isn't a good prospective study available yet, and perhaps some study of methods of specific rewarming might answer some of those questions since we can't control many of the other variables. As mentioned in the manuscript, there was no significant association with pH, at least for mortality or any of the risk factors other than the p H of 7.32 for emergency surgery and

The American Journalof Surgery

Adverse Effects of Postoperative Hypothermia

7.40 for elective surgery which was statistically significant, although I'm not sure how clinically significant it was. There were no differences in the gross parameters of arterial partial pressure of oxygen to fractional inspired oxygen ratios, and so on that could distinguish between these patients; there were insignificantnumbers of patients who arrived at the intensive care unit with hemodynamic

monitoring and also preoperative workup to allow a statistically valid analysis of ?.hat data. Certainly, rapid hemodynamic improvement in these patients would, more than likely, improve the overall metabolic and mortality situations. We, in this retrospective study, did not have sufficient data to answer all of those questions.

This discussion section has been abbreviated due to space limitati~Jtts. The full text can be obtained f r o m the reprints author listed on the title page.

Vol~ml 149, Aim'. 11185

501