Preoperative factors, such as age and body habitus, affect intraoperative hypothermia during general anesthesia. In a preliminary study, we developed a logistic model to retrospectively evaluate predictors of intraoperative hypothermia in patients who received major surgery. The following factors were selected to develop the model: Z = −15.014 + 0.097 × (Age) + 0.263 × (Height) − 0.323 × (Weight) − 0.055 × (Preoperative systolic blood pressure) − 0.121 × (Preoperative heart rate). By using this model, the probability of hypothermia can be estimated by applying the following for-mula: Probability = 1/(1 + e−Z). If an estimated probability of hypothermia was >0.5, the sensibility of prediction was 81.5% and the specificity was 83%. In the second study, the model was applied prospectively to other patients, and the validity of the logistic model was evaluated. The core temperature showed a significant decrease in patients with a probability >0.7, who were predicted to be hypothermic, and their thermoregulatory vasoconstriction threshold also showed a significant decrease, compared with the patients with a probability ≦0.3, who were predicted to be normothermic. We concluded that intraoperative hypothermia could be predicted from preoperative characteristics such as age, height, weight, systolic blood pressure, and heart rate.
Departments of *Anesthesiology and †Physiology, Kyoto Prefectural University of Medicine, Kyoto; and ‡Department of Anesthesiology, Yamanashi Medical University, Yamanashi, Japan
July 9, 2002.
Address correspondence and reprint requests to Toshihiro Kasai, MD, Department of Anesthesiology, Kyoto Prefectural University of Medicine, Kamigyoku, Kyoto 602-8566, Japan. Address e-mail to email@example.com.
Body temperature often decreases in anesthetized patients. Severe hypothermia tends to occur more often in longer operations, including abdominal or thoracic surgery, than in small procedures (1,2), and induces several adverse effects (2–4). Approximately one-half of patients develop hypothermia to a core temperature of <36.0°C (5), and approximately one-third of patients develop hypothermia to a core temperature of <35°C during surgery (6,7). Sessler (8) has suggested that the intraoperative core temperature should be maintained at >36.0°C. Forced-air warming, IV fluid warming (8), and preanesthetic warming (9–11) are effective in preventing intraoperative hypothermia. To predict intraoperative hypothermia during the preoperative period, we developed a logistical model to retrospectively evaluate preoperative risk factors of hypothermia in patients who received major surgery, and then assessed the validity of the model prospectively in other patients.
We selected 862 anesthesia records of patients (ASA physical status I or II) who received open abdominal surgery under general anesthesia with epidural blockade between April 1, 1999 and February 28, 2002. Patients whose minimal core temperature was >36.0°C during anesthesia (normothermia group;n = 200), and patients with a minimal core temperature <35.0°C (hypothermia group;n = 200) were enrolled in the retrospective study. We excluded cases in which blood transfusion was required or cardiovascular drugs were used to treat hypotension that resulted from excessive surgical bleeding. The duration of the operation and anesthesia, blood loss, and anesthetic management did not differ between the two groups.
Routine thermal care was performed on all patients in the operating room with the temperature adjusted to 22°–24°C. A circulating-water mattress, warmed at 38°C, was placed under the patient. The skin was covered with surgical drapes, and a heat-moisture exchanger was attached to the tracheal tube. IV fluids, warmed to near body temperature, were administered. The core body temperature was measured every 5 min by using a thermocouple (Mon-a-Therm; Mallinkrodt, St. Louis, MO) positioned adjacent to the tympanic membrane. The aural canal was occluded with cotton to prevent artifactual cooling by the ambient environment. Preoperative factors (Table 1) were measured on the ward before departure to the operating room.
Multiple logistic regression analysis with backward elimination (SPSS® 10.05J; SPSS Japan Inc., Tokyo, Japan) was conducted to develop a model for the prediction of intraoperative hypothermia. By using the preoperative factors (Table 1), the following factors were selected to develop the model:MATH
Positive and negative symbols before variate terms in the formula mean that increases in age and height, and decreases in weight, systolic blood pressure, and heart rate, are risk factors of intraoperative hypothermia. By using this model, the probability of hypothermia can be estimated by applying the following formula:MATH
If an estimated probability of intraoperative hypothermia was larger than 0.5, which was the best cutoff probability from the receiver operator characteristic curve analysis (12), the sensibility of prediction was 81.5% and the specificity was 83%.
With approval from the Human Research Committee of our institution, we assessed the validity of the above logistic model with preoperative factors in 50 patients (ASA physical status I or II) randomly selected from open abdominal surgery after informed consent was obtained. The anesthetic management was the same as in the retrospective study. None of the patients was taking medication that affected cardiovascular function or heat balance, or had a history of thyroid disease, dysautonomia, Raynaud’s syndrome, diabetes mellitus, or hypertension. The tympanic membrane temperature (core temperature) was measured by using a thermocouple (Mon-a-Therm). The peripheral vasoconstriction was quantified by using the forearm minus finger skin-surface temperature gradients (13,14). The thermoregulatory vasoconstriction threshold was considered to be the tympanic membrane temperature at the time when a forearm minus fingertip skin-temperature gradient exceeded 0°C. The model was applied to each patient, and the probability of intraoperative hypothermia was estimated preoperatively.
Obtaining the best sensibility in the above study, patients with a probability >0.7 were predicted to be hypothermic during anesthesia (n = 14), and patients with a probability ≦0.3 were predicted to be normothermic (n = 17). The core temperature showed a significant decrease in patients with a probability >0.7 throughout the study. The thermoregulatory vasoconstriction threshold (35.3° ± 0.5°C) in the patients with a probability >0.7 also showed a significant decrease, compared with 36.1° ± 0.5°C in the patients with a probability ≦0.3 (Table 2).
The results of the present study are compatible with previous reports that elderly patients become hypothermic because of a lower vasoconstriction threshold than younger patients (15), that thin patients become more hypothermic with more profound redistribution hypothermia than obese patients (16), and that patients with high preoperative systolic blood pressure become less hypothermic than those with normal preoperative systolic blood pressure (17). Although intraoperative factors affect thermoregulation during longer operations, preoperative patient characteristics are cardinal factors for intraoperative hypothermia, which may cause serious complications postoperatively. By using this predictive formula from simple preoperative patient information, positive thermal care, including prewarming, can be administered to patients with a high probability of hypothermia and this should reduce the occurrence of severe hypothermia and minimize the occurrence of adverse effects.
1. Sessler DI. Perioperative thermoregulation and heat balance. Ann NY Acad Sci 1997; 813: 757–77.
2. Flores-Maldonado A, Guzman-Llanez Y, Castaneda-Zarate S, et al. Risk factors for mild intraoperative hypothermia. Arch Med Res 1997; 28: 587–90.
3. Kurz A, Sessler DI, Lenhardt R. Perioperative normothermia to reduce the incidence of surgical-wound infection and shorten hospitalization. N Engl J Med 1996; 334: 1209–15.
4. Frank SM, Fleisher LA, Breslow MJ, et al. Perioperative maintenance of normothermia reduces the incidence of morbid cardiac events. JAMA 1997; 277: 1127–34.
5. Vaughan MS, Vaughan RW, Cork RC. Postoperative hypothermia in adults: relationship of age, anesthesia, and shivering in rewarming. Anesth Analg 1981; 60: 746–51.
6. Frank SM, Beattie C, Christopherson R, et al. Epidural versus general anesthesia, ambient operating room temperature, and patient age as predictors of inadvertent hypothermia. Anesthesiology 1992; 77: 252–7.
7. Morris RH. Operating room temperature and the anesthetized, paralyzed patient. Arch Surg 1971; 102: 95–7.
8. Sessler DI. Mild perioperative hypothermia. N Engl J Med 1997; 336: 1730–7.
9. Sessler DI, Schroeder M, Merrifield B, et al. Optimal duration and temperature of prewarming. Anesthesiology 1995; 82: 674–81.
10. Glosten B, Hynson J, Sessler DI, McGuire J. Preanesthetic skin-surface warming reduces redistribution hypothermia caused by epidural block. Anesth Analg 1993; 77: 488–93.
11. Hynson JM, Sessler DI, Moayeri A, et al. The effects of preinduction warming on temperature and blood pressure during propofol/nitrous oxide anesthesia. Anesthesiology 1993; 79: 219–28.
12. Fletcher RH, Fletcher SW, Wagner EH. Clinical epidemiology: the essentials. 3rd ed. Baltimore: Williams & Wilkins, 1996.
13. Rubinstein EH, Sessler DI. Skin-surface temperature gradients correlate with fingertip blood flow in humans. Anesthesiology 1990; 73: 541–5.
14. Frank SM, Shir Y, Raja SN, et al. Core hypothermia and skin-surface temperature gradients: epidural versus general anesthesia and the effects of age. Anesthesiology 1994; 80: 502–8.
15. Ozaki M, Sessler DI, Matsukawa T, et al. The threshold for thermoregulatory vasoconstriction during nitrous oxide/sevoflurane anesthesia is reduced in the elderly. Anesth Analg 1997; 84: 1029–33.
16. Kurz A, Sessler DI, Narzt E, et al. Morphometric influences on intraoperative core temperature changes. Anesth Analg 1995; 80: 562–7.
17. Kasai T, Hirose M, Matsukawa T, et al. Preoperative blood pressure and intraoperative hypothermia during lower abdominal surgery. Acta Anaesthesiol Scand 2001; 45: 1028–31.