Predictors of Hypothermia during Spinal Anesthesia

After obtaining institutional review board approval and written informed patient consent, 44 patients scheduled for radical retropubic prostatectomy were enrolled. Exclusion criteria were preoperative fever (> 38.0°C), hypo- or hyperthyroidism, or any dysautonomia syndrome. Percentage body fat was measured using the infrared interactance method over the biceps skin fold. 15 

The anesthetic technique consisted of 1 or 2 mg midazolam administered intravenously at arrival in the operating room, and a lumbar intrathecal injection of 18–22 mg bupivacaine, 0.75%, in 7.5% dextrose solution with 20 μg fentanyl. Blood transfusions were administered to maintain hemoglobin concentration above 10 g/dl. Blood was given through an in-line warmer and warmed to 38°C. Crystalloid solutions were preheated to 37°C in a warmed cabinet and given without in-line warming. No other warming device was used. All patients were covered with one layer of paper surgical drapes and one layer of cotton blanket positioned over the thighs and calves. In addition, one layer of cotton blanket was placed over the chest and arms. Spinal block level was assessed by testing for the loss of pinprick sensation just before skin incision, approximately 20–25 min after spinal injection. Ambient temperature in the operating room was not controlled by protocol.

All temperatures were measured using thermocouple probes (Mon-a-therm; Mallinckrodt Medical, St. Louis, MO) and recorded every 1 min by an electronic data-acquisition system (Iso-thermex; Columbus Instruments, Columbus, OH). Core temperature was measured at the tympanic membrane with a cotton-tipped probe inserted to a depth that elicited a slight scratching sound reported by the patient. The external auditory canal was packed with cotton for insulation. Ambient temperature also was measured using thermocouples placed near the patient but away from any heat-generating equipment. The ambient temperature was averaged over the entire intraoperative course, using measurements taken at 5-min intervals. Monitoring was begun just before the spinal injection. After admission to the postanesthesia care unit (PACU), temperature monitoring was continued for 3 h. From the electronically recorded temperature data, temperatures taken at specific time points were analyzed. These times were before the spinal injection and 15, 30, and 45 min afterward; at the time of prostate removal; 15 and 30 min after prostate removal; at the end of surgery; and in the PACU at 0, 15, 30, 45, 60, 90, 120, 150, and 180 min after admission.

Clinical variables that were assessed as predictors of core temperature at admission to the PACU were duration of surgery, ambient operating room temperature, body mass, body mass index (kg/m²), percentage body fat, age, and height of spinal blockade.

Changes in core temperature over time were analyzed by analysis of variance with repeated measures. Univariate predictors of core temperature at admission to the PACU were determined by simple linear regression. Multivariate predictors were determined by multiple linear regression with backward elimination, 16retaining all variables in the model with P > 0.12. P < 0.05 defined significance, and all data are expressed as the mean ± SD.

The demographic data are given in table 1. All patients were male, with a normal distribution of age and morphometric characteristics for a patient population undergoing radical prostate surgery.

During the initial 45 min after the spinal was administered, core temperature decreased rapidly to more than 1°C below the preoperative baseline (fig. 1). By the end of the surgical procedure, mean core temperature was 1.5°C below the preoperative baseline value. Mean core temperature at admission to the PACU was 35.1 ± 0.6°C (range, 33.6–36.3). Rewarming was slow in the postoperative period, in which patients were allowed to rewarm passively.

Predictors of core temperature at admission to the PACU were identified by univariate and multivariate analyses (table 2). Duration of surgery, ambient operating room temperature, body mass, body mass index, and percentage body fat were not significant predictors of hypothermia. A high level of spinal blockade and advanced age were predictors of low core temperature by both univariate and multivariate tests. Figure 2shows the associations of age and level of spinal blockade with core temperature at admission to the PACU.

The regression equations describing the effects of age and block level were, respectively, as follows: core temperature (°C) = 34.37 + (0.15 · dermatomal block level), and core temperature (°C) = 36.72 − (0.03 · age in yr). The effect of increasingly high block level, therefore, was a 0.15°C decrease in core temperature for each additional dermatomal level. For each year of increasing age there was a 0.03°C decrease in core temperature.

Of the five surgeons whose patients were involved in the study, one routinely used a smaller incision (≈ 10–12 cm) compared with the others (≈ 20–25 cm). Seven of the 44 patients were operated on by this surgeon. Demographically, these patients were no different from the other patients in age, height, weight, duration of surgery, ambient temperature, amount of blood loss, or height of spinal block. Mean core temperature at admission to the PACU was similar between patients of this surgeon (34.9 ± 0.5°C) and those of the other surgeons (35.1 ± 0.6°C;P = 0.77).

Although advanced age and high level of spinal blockade were significant predictors of core hypothermia during spinal anesthesia, other variables (ambient temperature, body habitus, amount of blood loss, and duration of surgery) were not predictors of changes in core temperature. High levels of spinal blockade are known to decrease the core temperature threshold for shivering, 17but no previous study has shown a high level of blockade to be associated with a greater magnitude of core hypothermia in the clinical setting. As with general anesthesia, 18advanced age is a risk factor for hypothermia during spinal anesthesia, but in contrast to general anesthesia, body habitus and ambient operating room temperature (within the range reported) were not predictors of hypothermia.

The correlation between high-level spinal blockade and low core temperature during spinal anesthesia is consistent with the known physiologic effects of spinal anesthesia. Both vasomotor tone and shivering are inhibited below the level of the spinal block through sympathetic and somatic neural blockade, respectively. 19The greater the proportion of the body that is blocked, the more impairment of thermoregulatory function is to be expected. In support of this is the study by Leslie and Sessler, 17which showed a decreased shivering threshold in proportion with the level of spinal blockade. These investigators showed that core temperature threshold was decreased by 0.06°C for each dermatomal level that was blocked. Our study confirms the findings of Leslie and Sessler 17in the clinical setting by demonstrating high block level to be a risk factor for core hypothermia. In our analysis, for each additional dermatomal level of spinal blockade there was a decrease of 0.15°C for predicted core temperature at admission to the PACU.

It is conceivable that factors associated with higher levels of blockade caused increased hypothermia and that blockade level, per se , was not the cause. For example, fluid and blood requirements are sometimes greater with higher block levels. Because fluids and blood were warmed to body temperature before administration, however, the chances of such a confounding effect are minimal. In addition, there was no relation between blockade level and actual fluid and blood requirements.

The magnitude of core hypothermia in the current study was similar to that reported in previous studies during general anesthesia. 8,9,20Some studies that have compared core temperature during general and regional anesthesia indicate a similar incidence and magnitude of hypothermia. 4,21Thus, although there are different physiologic mechanisms for thermoregulatory inhibition during general and regional anesthesia (central vs.  peripheral nervous system effects, respectively), patients are at significant risk for hypothermia during spinal anesthesia, especially if the level of spinal block is high.

According to a recent survey, it is not common practice to monitor body temperature during regional anesthesia. 5Reasons for this include the absence of risk for malignant hyperthermia and the lack of a convenient and reliable site for temperature-probe insertion. Hypothermia therefore often is neither recognized nor treated in patients receiving regional anesthetic techniques. 6Even if patients are awake, they are unlikely to report the sensation of cold because there is an “apparent” warm input from the portion of the body below the level of neural blockade. 22Often the first manifestation of hypothermia occurs postoperatively, when the spinal block is resolving and patients feel cold and start to shiver. Unfortunately, in this circumstance hypothermia often is treated rather than prevented.

Ambient temperature was not a predictor of hypothermia during spinal anesthesia, despite the fact that it is a known predictor during general anesthesia. 13It is possible that the effect of ambient temperature was underestimated in the current study because the range of ambient temperature may not have been great enough to show an effect. The range in this study, however, was no different than the range in previous studies using general anesthesia. 23Further studies are needed to show the ambient temperature required to prevent hypothermia during spinal anesthesia. Body habitus and duration of surgery also were not predictors of hypothermia. It may be that the intense vasodilation associated with spinal blockade is accompanied by rapid and significant heat loss, which results in hypothermia regardless of the other factors (ambient temperature, body habitus, and duration of surgery).

In the current study we did not find a relation between surgical incision size and core temperature. Although there was no randomization or control group, it appears that heat loss through the incision represents a small proportion of total heat loss in patients receiving spinal anesthesia. A larger sample size is required to determine the effects of incision size with greater certainty.

High-level spinal blockade has known complications such as respiratory and hemodynamic compromise. In addition to these complications, there is an increased risk of hypothermia. Clinical implications are to limit the spinal block height to the level that is necessary for the surgery and to carefully monitor and control body temperature in older patients and in those patients with high-level blocks during spinal anesthesia.