In this study, the incidence of postoperative hypothermia was 25.7% in patients immediately upon arrival at the PACU. After active forced-air warming for patients with hypothermia, the incidence of hypothermia was 13.9% at discharge from the PACU; to some extent, active forced-air warming might be an effective method to rectify hypothermia. In addition, only 8 patients experienced shivering in the PACU. High weight and BMI, high baseline body temperature on the morning of surgery and low estimated blood loss were protective factors for postoperative hypothermia, while advanced age, nonopen (laparoscopic and thoracoscopic) surgery, and long duration of surgery were risk factors for postoperative hypothermia.

The previously reported incidence of postoperative hypothermia in the PACU varied widely, ranging from 4% to 56.7% [19, 20], probably in part because of differences in patients and surgeries, which were included in these studies. Previous studies suggested that patients undergoing major-plus or nonsuperficial surgery were vulnerable to perioperative hypothermia [16]. Therefore, we paid closer attention to these MT populations undergoing relatively major and complicated abdominal, thoracic and spinal surgeries and excluded those patients undergoing endoscopic or superficial surgeries, such as breast and thyroid surgeries. In our study, the incidence of postoperative hypothermia was 25.7% on arrival at the PACU, which was similar to the incidence in one previous study [17].

We found that high body weight and BMI ≥ 25 were protective factors against postoperative hypothermia, which was consistent with previous findings [17, 21]. A total of 202 patients were included in our study, and 68.8% of the patients’ BMIs were < 25, perhaps because most MT patients were frail with poor nutritional status. An age older than 65 years was another risk factor for postoperative hypothermia, as suggested in previous studies [16, 17]. Elderly patients, particularly those who are over 65 years old, have less effective regulatory ability of their central nervous system and an increased susceptibility to hypothermia [9]. As mentioned above, many MT patients possess risk factors for hypothermia, including poor nutritional status, low BMI and advanced age; thus, it is necessary for us to take active measures to prevent hypothermia in this population.

We also found that high baseline body temperature on the morning of surgery could effectively protect patients from hypothermia, which was consistent with previous studies [17, 21, 22]. Thus, raising the baseline body temperature by active prewarming should be an effective strategy to prevent perioperative hypothermia. Several studies have also suggested that prewarming with a forced-air warming system is an effective, simple and convenient method to prevent hypothermia [23]. However, most forced-air warming systems need to be purchased at patients’ own expense because of the reimbursement policy in China, which might be a barrier to generalizing these devices to more patients.

It was not surprising that a long duration of surgery was a risk factor for hypothermia, which was also consistent with some previous reports [21, 22]. A longer duration of surgery increased the time that the patient was exposed to the ambient environment and resulted in a longer duration of anesthesia and the receipt of more unwarmed CO2 in the laparoscopic or thoracoscopic surgery. In this study, low estimated blood loss was associated with decreased postoperative hypothermia, which was consistent with one prior study of advanced ovarian cancer surgery [24]. However, another study indicated that blood loss was not associated with hypothermia. This inconsistency might be due to the different volumes of blood loss in different studies [17, 22].

With regard to the type of surgery, some previous studies suggested that laparoscopic surgery was not associated with a higher incidence of perioperative hypothermia than open surgery [25, 26]. However, we found that nonopen surgery (laparoscopic and thoracoscopic surgery) increased the incidence of postoperative hypothermia, which was similar to the finding in one previous study [17]. A large amount of cold, dry CO2 was continuously insufflated into the body during the laparoscopic or thoracoscopic surgery, which then could cause heat loss and a lower body temperature. A meta-analysis also revealed that during laparoscopic abdominal surgery, the core body temperature was significantly lower in the cold CO2 groups than in the heated, humidified CO2 groups [27]. Therefore, as one previous study indicated, perhaps the use of warm and humidified CO2 for peritoneal insufflation was a safe, feasible and cost-effective intervention to prevent hypothermia [28].

After 30 min of care in the PACU, we found that GEA, one type of anesthesia, was associated with increased postoperative hypothermia in univariate logistic regression. However, GEA was not significantly associated with postoperative hypothermia in multivariate logistic regression, perhaps because of the small sample size of only 52 patients in the GA group. In our hospital, we often perform epidural anesthesia during general anesthesia for abdominal and thoracic surgeries. One previous retrospective study also demonstrated that epidural anesthesia was associated with intraoperative hypothermia in advanced ovarian cancer surgery [24]. The mechanism by which epidural anesthesia results in hypothermia might involve impairing central thermoregulatory control and preventing vasoconstriction and shivering in blocked areas [15]. Therefore, further randomized controlled studies are required to establish the precise relationship between epidural anesthesia and hypothermia in MT surgery.

In this study, only a total of 8 patients experienced shivering, which was less than the numbers of patients reported in previous studies [21, 29]. However, the core body temperatures of patients with shivering were not all below 36 ℃. Shivering traditionally is attributed to hypothermia, but it is not always thermoregulatory. As a previous study recommended, shivering should be treated by active warming [30]. In fact, shivering disappeared in almost all the patients by active forced-air warming without supplementary medical therapy in our study. To a certain extent, we found that advanced age and male sex showed little association with a longer duration of shivering. Thus, the relationship between shivering and age and sex can be investigated in further prospective studies.

There were some limitations in this study. First, only 30 patients received intraoperative temperature monitoring, perhaps because of lack of awareness of perioperative temperature management among many anesthesiologists. Thus, it limited possible findings of this study, and we could not determine the incidence of intraoperative hypothermia and the accurate relationship between postoperative hypothermia and intraoperative hypothermia. We will pay more attention to perioperative temperature management and investigate the incidence of intraoperative hypothermia and their risk factors in patients undergoing MT surgery in the future study. Second, we did not record the ambient temperature in the OR and PACU, even though there was a monitoring panel in the OR. Because there was no temperature monitoring in the PACU, we could not record the room temperature in the PACU. However, the room temperature in the OR and PACU was controlled by the central control system according to the specified standard in our hospital. Third, statistical analysis of risk factors associated with hypothermia at time point 2 and 3 was based on the facts that hypothermia was rectified by active forced-air warming, and the results of the two time points might be not as accurate as the time point 1. However, to some extent, these results also helped us acquire potential risk factors related to hypothermia for future studies in more detail.