Comparison of peri-operative core temperature in obese and non-obese patients


Correspondence to: J. Reinaldo Cerqueira Braz


Our aim was to compare peri-operative core temperatures and the incidence of hypothermia in obese and non-obese women with active forced-air warming. Twenty female patients scheduled for abdominal surgery were allocated to two groups according to body mass index. Ten obese (30.0–34.9 kg.m−2) and 10 non-obese (18.5–24.9 kg.m−2) women received forced-air warming on their lower limbs. At the end of surgery, the mean (SD) core temperatures were 36.7 (0.5) °C in the obese group and 36.0 (0.6) °C in the non-obese group (p < 0.001). Only in the non-obese group was there a significant decrease in the intra-operative core temperature values (p < 0.001). The incidences of intra-operative hypothermia were lower in the obese group (10%) compared with non-obese group (60%; p = 0.019). In the postoperative recovery phase, the mean (SD) core temperature data were higher in the obese group than in the non-obese group (36.2 (0.4) vs 35.6 (0.5) °C, respectively (p < 0.001)). In conclusion, obese female patients have higher peri-operative core temperature and a lower incidence of hypothermia compared with non-obese female patients during abdominal surgery with active forced-air warming.

Unintentional peri-operative hypothermia (core temperatures between 34.5 and 35.9 °C) often leads to adverse outcomes, including cardiac events secondary to sympathetic nervous system activation [1], surgical-wound infections and prolonged hospitalisation [2], coagulopathy and increased blood loss [3], impaired drug metabolism, delayed postoperative recovery period [4, 5] and shivering [6]. Therefore, maintaining peri-operative normothermia reduces morbidity [1–3] and the use of peri-operative warming devices has become routine.

The prevalence of obesity has increased markedly worldwide in recent years [7]. In clinical practice, body mass index (BMI) is used to estimate the degree of obesity, which is classified in three levels: grade 1 (BMI from 30.0 to 34.9 kg.m−2), grade 2 (35.0–39.9 kg.m−2) and grade 3 or morbid obesity (≥40.0 kg.m−2) [8]. The prevalence of grade-1 obesity is higher among both female and male obese populations than are the other grades [8]. Obese patients are more likely to vasoconstrict in cooler environments [9], have reduced heat redistribution from core to peripheral tissue after induction of anaesthesia induction the amount of redistribution hypothermia is inversely proportional to the percentage of body fat, and better thermal insulation than the general population [10]. It has been reported that the intra-operative core temperature is maintained with an acceptably low incidence of hypothermia (9%) on admission to the postanaesthesia care unit (PACU) in grade-3 obese patients when active forced-air warming had been used during bariatric surgery [11]. Intra-operative warming strategies have minimised, but do not prevent, unintentional peri-operative hypothermia in non-obese patients [12–14].

We tested the null hypothesis that peri-operative core temperatures are the same in grade-1 obese female patients as in non-obese female patients during general (total intravenous, TIVA) anaesthesia, with background active forced-air warming during open abdominal gynaecological surgery.


This study was approved by the Human Research Ethics Committee of the Botucatu Medical School. After providing written informed consent, 22 female patients (ASA physical status 1–2) aged from 26 to 55 years, and who were scheduled for elective open abdominal gynaecological surgery (hysterectomy or oophorosalpingectomy) of at least 120 min duration, were enrolled in this study. Patients’ heights and weights were measured during pre-operative evaluation. Patients were not studied if they were <18 or >65 years old, classified as ASA 3 or poorer, had a fever or endocrine disease, had a BMI > 35.0 kg.m−2 or <18.5 kg.m−2, or between 25 and 29.9 kg.m−2 and were taking vasoactive drugs.

Patients were allocated to two groups according to BMI. All surgeries were performed in an operating theatre with laminar airflow. Ten grade-1 obese and 10 non-obese (BMI from 18.5 to 24.9 kg.m−2) consecutive patients received active forced-air warming with a specific blanket on the lower limbs from a device (Bair Hugger®, model 750; Arizant Healthcare, Eden Prairie, MN, USA) set to deliver forced-air at 43 °C. The two groups received intravenous fluids that were maintained at operating theatre temperature. In all groups, the warming period lasted from 1 min before anaesthesia induction until the end of surgery. If any patient’s core temperature was >37 °C, the use of the warming device was discontinued.

In the operating theatre, and after 8 h of fasting, an 18-G peripheral intravenous line was inserted into the right-arm cephalic vein of non-premedicated patients. Lactated Ringer’s solution at 10−1.h−1 was administered using a two-channel infusion pump (Anne®; Abbott Laboratories, Abbott Park, IL, USA). Standard clinical monitoring was used, including: electrocardiography (ECG); non-invasive arterial blood pressure; arterial oxygen saturation; cerebral state index (CSI) (Danmeter®, Biometer International, Odense, Denmark); neuromuscular blockade monitoring by train-of-four count at the adductor pollicis (TOF-Guard®; Biometer International); and temperature monitors using a two-channel electronic thermometer (model G750; Mallinckrodt, St. Louis, MO, USA). All temperatures were measured using thermocouple probes (Mon-a-therm®; Mallinckrodt Medical, St. Louis, MO, USA). The operating theatre temperature was measured using a thermocouple probe placed near the patient, but away from any heat-generating device; the operating theatre temperature was maintained at 21–23 °C. Temperatures were measured at the oral and oesophageal sites, before and after the induction of anaesthesia, respectively. Oral temperature was measured using a probe inserted in the posterior sublingual pocket close to the lingual artery, which is a branch of the carotid artery and reflects changes in core temperature [15], and the oesophageal temperature was measured using a probe inserted into the lower oesophagus, after the induction of anaesthesia.

TIVA Anaesthesia was induced with a target-controlled infusion (TCI) of remifentanil 50 μ−1 using a computer-controlled infusion pump (Alaris PK model®; Cardinal Health, Rolle, Switzerland) programmed with the Minto adult pharmacokinetics model for remifentanil [16]. Two minutes later, a TCI of propofol 1% was administered using a computer-controlled infusion pump (Diprifusor®; Fresenius Vial, Brezens, France) programmed with the Marsh adult pharmacokinetics model for propofol [17]. The initial target remifentanil plasma concentration was set at 6−1 and the propofol plasma target concentration was set at 4 μ−1 to induce loss of consciousness. Rocuronium 0.6−1 was administered to facilitate orotracheal intubation. Target propofol and remifentanil concentrations were maintained between 35 and 4–7 μ−1, respectively, until the end of surgery as adjusted to a CSI of 40–50 and haemodynamic responses. All patients had a hydrophobic heat and moisture exchanger (HME) with a bacteriological and viral filter (Pall BB100; Pall Corporation, East Hills, NY, USA) placed between the Y-piece of the breathing circuit and the tracheal tube. In our anaesthesia service, the use of a HME with filter in the breathing circuit is mandatory in all general anaesthetics. Patients’ lungs were mechanically ventilated using the volume-controlled mode of the Primus® anaesthesia workstation (Dräger Medical, Lübeck, Germany) with a tidal volume of 8−1 of 60% oxygen in air with a fresh gas flow of 1 l.min−1 and a respiratory rate necessary to maintain normocapnia.

Postoperative analgesia was provided 15 min before the end of surgery with 2 μ.kg−1 fentanyl, 100 mg tramadol and 100 mg ketoprofen intravenously. Residual neuromuscular blockade was reversed with 30 μ−1 neostigmine and 10 μ−1 atropine, if necessary. Tracheal extubation was performed after the full reversal of neuromuscular blockade. All patients were immediately transferred to the PACU.

In the PACU, patients were covered with a full-body blanket up to the shoulders connected to a forced-air warming device (Bair Hugger model 750) set to 43 °C until their oral temperature reached 36 °C. The PACU temperature was maintained at approximately 25 °C. In the PACU, shivering was evaluated by a blinded, independent observer as absent, mild (when only detected using ECG artefacts), or severe (when clinically obvious). A minimum oral temperature of 36 °C was required for discharge of the patients from the PACU.

The operating theatre and core temperatures were recorded in patients before anaesthetic induction (baseline) and at 15, 30, 60, 90 and 120 min after anaesthetic induction and at the end of surgery. The type of surgery, operative time (from skin incision until wound closure), total volume of intravenous fluids, consumption of propofol and remifentanil and tracheal extubation time were recorded. In the PACU, oral temperature data were recorded every 15 min for 60 min. Hypothermia was defined as core temperatures <36 °C. The numbers of patients in the groups who developed hypothermia at the end of surgery and after one hour in the PACU, and shivering after arrival in the PACU, were recorded.

The minimum sample size was calculated based on data found in the literature [18, 19] and assuming mean differences (SD) of 0.5 (0.4) °C between the core temperatures of the groups as clinically significant. A minimum of 10 patients in each group was necessary to detect this difference using a two-tailed test with the probability of a type-1 error (α) of 0.05 and a probability of a type-2 error (β) of 0.2 (80% power). Statistical analysis was performed using the Statistical Package for the Social Sciences software (Version 6.0; SPSS, Inc., Chicago, IL, USA). Normal data distribution was confirmed using the Lilliefors test. Characteristics, type of surgery, baseline core temperature, operative time, total volume of intra-operative infused fluids and total doses of propofol and remifentanil were compared between the two groups using ANOVA. Continuous data between groups were compared using ANOVA for repeated measurements, followed by Tukey’s test to investigate differences at different times in each group. Pearson’s coefficient was used to detect correlation between oesophageal temperature and BMI data of both groups together at the end of surgery. The proportion of patients who developed hypothermia was determined using chi-squared analysis. For all analyses, a value of p < 0.05 was considered to be statistically significant.


Twenty-two obese and non-obese female patients were enrolled in the study, but one obese patient was removed as a result of surgical cancellation and critical data were missing from one non-obese patient; the remaining 20 patients were included in the statistical analyses with 10 female patients in each group (Fig. 1).

Figure 1.

 Enrollment and allocation of study patients. BMI, body mass index.

The groups did not differ according to age (p = 0.73) or height (p = 0.41), but as expected, they differed according to weight and BMI (p < 0.001) (Table 1). The type of surgery, baseline oral temperature, operative time, total volume of intravenous fluids, total remifentanil dose and extubation time were similar between groups (Table 2). The total propofol dose was higher in obese group than in non-obese group (p < 0.001) (Table 2). The operating theatre temperature values were maintained between 21 and 23 °C without statistically significant differences between groups (data not shown).

Table 1. Characteristics of the patients in obese and non-obese groups. Values are as means (SD).
 Obese (n = 10)Non-obese (n = 10)
  1. BMI, body mass index.

Age; years43 (6)40 (9)
Height; cm157 (5)160 (6)
Weight; kg78 (7)60 (6)
BMI; kg.m−231.6 (1.3)22.9 (1.7)
Table 2. Intra-operative data in the obese and non-obese groups. Values are number or median (IQR [range]). Operative time = from skin incision until wound closure; time for extubation = from cessation of anaesthetic administration until tracheal extubation.
 Obese (n = 10)Non-obese (n = 10)p value
  1. AH, abdominal hysterectomy.

Type of surgery
Operative time; min199 (175–248 [125–260])163 (155–207 [150–295])0.42
Fluids infused; ml2300 (2000–3475 [1800–4500])1900 (1600–2500 [1500–5000])0.26
Basal core temperature; °C36.6 (36.5–36.9 [36.5–36.9])36.5 (36.3–36.6 [36.2–37.0])0.89
Consumption of propofol; mg1819 (1327–2193 [1300–2461])1271 (1119–1557 [818–1978])0.003
Consumption of remifentanil; mg2.85 (2.34–3.62 [1.99–3.90])2.40 (2.22–3.33 [1.94–4.83])0.45
Time for extubation; min6.0 (4.75–8.5 [3.0–10.0])6.5 (5.0–8.5 [4.0 13.0])0.98

The core temperature data were significantly higher over time in the obese group compared with the non-obese group (p = 0.008; Fig. 2). Only in the non-obese group, the intra-operative core temperature values declined (p < 0.001; Fig. 2). A significant positive correlation between BMI and core temperature data was observed at the end of surgery (p = 0.008; r = 0.57).

Figure 2.

 Mean (SD) intra-operative core temperature data in obese (inline image) and non-obese groups (inline image). Error bars are SD. The values were significantly different between the groups (p = 0.008). Only in the non-obese group there was a significant decrease in the values over time (p < 0.001).

The incidences of hypothermia in the groups at the end of surgery were 10% in the obese group and 60% in the non-obese group (p = 0.019).

Active forced-air warming was discontinued in one patient in the obese group for 30 min during surgery. None of the patients experienced surgical complications or required blood transfusions.

In the PACU, oral temperature data in the obese group was higher than in the non-obese group over time (p < 0.001; Fig. 3). All patients of the obese group were normothermic as against 30% of the patients in the non-obese group with oral temperatures <36 °C after 1 h in the PACU; however, these differences in proportions were not statistically significant (p = 0.06). Discharge of the unwarmed non-obese patients from the PACU was delayed by about 30–45 min. No patient in any group shivered in the PACU.

Figure 3.

 Mean (SD) oral temperature data in obese (inline image) and non-obese (inline image) groups in PACU. The values in the obese group were higher than non-obese group (p < 0.001). The values increase significantly only in the non-obese group over time (p < 0.001).


The main findings of this study are: first, the peri-operative core temperature values were higher in the obese group compared with non-obese group (such that the incidence of intra-operative hypothermia in the obese was lower); and second, only in the non-obese group was there a significant decrease in the intra-operative core temperature.

Obese individuals have body fat with low thermal conductivity that reduces heat loss from the skin and minimises hypothermia [9]. They have higher leptin levels, but they have higher peripheral resistance to the actions of this protein instead [20]. Leptin is secreted by adipocytes and has several physiological roles that include increasing activity of the sympathetic nervous system, which stimulates energy expenditure in brown adipose tissue and increases metabolic rate and, therefore, body heat [21]. Thus, obese patients exhibit less heat redistribution from core to peripheral tissue after anaesthetic induction [10]. Reduction in intra-operative core temperature in grades-2 and -3 obese patients may be inversely proportional to BMI [10] and in our study, there was also a positive correlation between BMI and core temperature in both groups. A previous study compared changes in core temperature during laparoscopic and open bariatric surgery when an upper warming blanket was utilised in grade-3 obese patients [22]. The authors observed that intra-operative core temperature values increased significantly in both groups in relation to the baseline values (after anaesthetic induction). However, in this study, the authors verified that intra-operative hypothermia was a common event in both groups with an incidence about 40%. In contrast, the incidence of hypothermia was lower (10%) in the obese group in our study, when we utilised a lower body warming blanket. Studies suggested that heat transfer with forced air warming over the lower body is higher than over the upper body, because the former cover a larger area of the body surface [23–25].

However, as demonstrated in our study, the use of active skin-surface warming was more effective in obese than in non-obese patients. Other studies also demonstrated that intra-operative active forced-air warming does not prevent unintentional peri-operative hypothermia in non-obese patients [14, 19]. These findings reinforce the need for additional intra-operative thermal care in normal or underweight (BMI < 18 kg.m−2) patients.

In our study, as normothermia was required to discharge the patients, 30% of the unwarmed non-obese patients had delayed discharge from the PACU. None of the patients from either group shivered in the PACU; this may be because of propofol [26], fentanyl [27] and tramadol [28] that have residual effects on the thermoregulatory control centre, including reduced shivering. Besides, the postoperative forced-air warming over the skin could also have prevented shivering. In addition, the shivering threshold is a full degree lower than the vasoconstriction threshold [29]. Consequently, the hypothermic patients did not shiver because of they had core temperature higher than the shivering threshold.

One limitation of our study is that we chose only women as our study population. Female patients generally begin to respond to cold at 0.3 °C higher core temperatures than male patients [30].

In conclusion, grade-2 obese female patients have a higher peri-operative core temperature and a lower incidence of hypothermia compared with non-obese female patients during abdominal surgery with active forced-air warming under intravenous anaesthesia. These findings are perhaps reassuring for temperature control in the obese, but they raise questions about the optimum methods of temperature management in the non-obese, for which additional research is needed.


LAF was granted a scholarship from CAPES. FAK was granted a scholarship from PIBIC/CNPq. This study received financial support from FAPESP (grant number 2008/53271-3). CAPES, PIBIC/CNPq and FAPESP are Brazilian governmental agencies dedicated to promoting scientific research.

Competing interests

No competing interests declared.