Risk factors for intraoperative hypothermia in patients undergoing thoracic spinal surgery: A retrospective cohort study

Abstract

Objective

The present study aimed to develop a clinical prediction model for intraoperative hypothermia in patients undergoing thoracic spinal surgery, with the objective of supporting decision-making in clinical care and anesthesia management.

Methods

A total of 600 patients were enrolled in this retrospective cohort study. The dataset was randomly divided into training (70% of all data) and validation (30% of all data) sets using sample function in R language. Based on the results of the multivariable binary logistic regression analysis, a nomogram was developed and evaluated using R language. Calibration was performed by comparing the predicted and actual probability curves using the ‘rms’ package. The ‘ggDCA’ package was used to construct decision analysis curves for evaluating clinical utility.

Results

A least absolute shrinkage and selection operator binary logistic regression model was constructed to identify 5 potential risk factors from the 14 variables previously identified. These five risk factors, including duration of anesthesia, preoperative body temperature, body mass index, number of diseased vertebrae, and reason for surgery, were then validated using logistic regression models. The concordance index was calculated as 0.919.

Conclusions

Tumor-related spinal surgery, multilevel procedures, prolonged surgical duration, low preoperative temperature, and low body mass index were associated with a higher risk of intraoperative hypothermia. These factors may help guide targeted warming strategies in patients undergoing thoracic spinal surgery.

Keywords

Intraoperative hypothermia thoracic spinal surgery general anesthesia risk factors perioperative temperature management

Introduction

Intraoperative hypothermia is commonly defined as a core body temperature <36°C, a threshold widely adopted following seminal work by Sessler and colleagues in the 1990s that characterized its incidence, mechanisms, and clinical implications under anesthesia.¹ A multicenter study conducted in China has reported a prevalence rate of 44.5% among 3126 cases.²

Thoracic spine surgery often involves long surgical duration, prolonged exposure of the body cavity, considerable blood loss, and substantial fluid replacement, all of which increase the risk of hypothermia. General anesthesia is commonly administered to patients undergoing spinal surgery.³ Under unprotected condition, the patient’s body temperature can reportedly drop by 1°C–3°C after 1 h of general anesthesia administration. Infusing 1000 mL of normal temperature liquid can lower the patient’s body temperature by 0.3°C, whereas infusing 200 mL of 4°C stock blood can lower the patient’s body temperature by 0.25°C. In addition, the special low-temperature environment in the operating room and the thermoregulatory disorder caused by anesthesia drugs can lower the patient’s body temperature. Evaporation of skin antiseptic solution, loss of inherent body heat due to surgical operation, and infusion of room temperature fluids can contribute to hypothermia (core body temperature <36°C) in perioperative patients.⁴

The incidence of intraoperative hypothermia can reportedly reach 50%–70%. In mild hypothermia, body metabolism slows down, oxygen consumption decreases, and tolerance to injury increases.⁵ However, severe and prolonged hypothermia can result in complications, including decreased immune function, impaired coagulation, prolonged drug metabolism, and increased risk of incisional infection.

Therefore, assessing the risk of perioperative hypothermia and adopting effective thermal management to prevent or alleviate perioperative hypothermia are crucial for patient safety. Notably, the number of diseased vertebrae reflects the extent of spinal involvement and the reason for surgery, such as tumor resection compared with degenerative or traumatic conditions, and is strongly associated with greater surgical complexity, longer surgical duration, and increased blood loss. These factors significantly promote heat loss and impair thermoregulation, supporting their relevance as predictors of intraoperative hypothermia in thoracic spine surgery.⁶ However, there is a lack of effective clinical assessment protocols for intraoperative hypothermia occurrence during spinal surgery. Therefore, the present study was designed to develop a clinical prediction model for hypothermia in patients undergoing thoracic spine surgery, with the aim of supporting clinical decision-making, care, and anesthesia use.

Patients and methods

Patients enrollment

In this single-center, retrospective cohort study, patients were selected consecutively from electronic medical records based on pre-specified inclusion and exclusion criteria. Patients who underwent thoracic spine surgery at the local hospital from January 2023 to January 2024 were consecutively enrolled. Patients with multiple surgical sites were excluded. In addition, patients whose condition was complicated with lung, heart, or endocrine system dysfunction were excluded. This study was conducted in accordance with the ethical principles of the Declaration of Helsinki (1975), as revised in 2013. The study protocol was approved by the Institutional Review Board of The First Affiliated Hospital of Soochow University. The requirement for written informed consent was waived due to the retrospective nature of the study and use of anonymized data. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.⁷ All patient data were deidentified to ensure anonymity.

Surgery setting

All thoracic procedures were completed in standard operating rooms with laminar flow. Room temperature was set at 23°C and humidity at 50%–60%. General anesthesia was administered using a Datex-Ohmeda anesthesia machine. Intraoperative irrigation fluid was heated to 38°C using a thermostat. All intravenous fluids were heated to 38°C using a 3 M Ranger 24200 Infusion Heating System, and an inflatable thermal blanket (model 750, 3 M Bair Hugger; USA), with temperature set at 38°C, was used to cover the patient’s body. Body temperature was measured intraoperatively using a BRAun ThermoScan 6023 ThermoScan thermometer, and nasopharyngeal temperature was measured intraoperatively every 1 min using a disposable temperature-sensing probe. Intraoperative hypothermia was defined as intraoperative nasopharyngeal temperature <36°C.

Data collection

Data concerning demographic information, surgical procedure, laboratory tests, and medical history were collected for further analyses. Demographic information included age and body mass index (BMI). Data on surgical duration, anesthesia duration, blood loss, volume of blood transfusion, and total input and output were also recorded. Blood levels of albumin and hemoglobin as well as white blood cell counts were recorded. In addition, the reason for surgery and number of affected vertebrae were recorded.

Embellishment and validation of predicting models

In the present study, the entire dataset was randomly divided into a training set (70% of all patients) and an internal validation set (30% of all patients) using the sample() function in R software (version 4.3.2; R Foundation for Statistical Computing; Vienna, Austria).

Candidate predictors were first screened using least absolute shrinkage and selection operator (LASSO) logistic regression implemented using the ‘glmnet’ package (version 4.1.8). Ten-fold cross-validation was applied to select the optimal tuning parameter (λ) that minimized prediction error, and variables with non-zero coefficients in the LASSO model were retained for further analyses.

These selected variables were then entered into a multivariable binary logistic regression model to identify independent risk factors for intraoperative hypothermia. Based on this final regression model, a clinical prediction nomogram was constructed using the ‘rms’ package (version 8.1.0) and visualized interactively using the ‘DynNom’ package (version 5.1).

Model performance was assessed in both training and internal validation sets in terms of discrimination, calibration, and robustness. Discrimination was quantified using the concordance index (C-index), whereas calibration was evaluated by comparing predicted probabilities against observed outcomes through calibration curves generated using the ‘rms’ package. To further assess model stability, 10-fold cross-validation was performed using the caret package. All statistical tests were two-sided, and a p-value <0.05 was considered statistically significant. Continuous variables were compared using unpaired Student’s t-test (normally distributed data) or Mann–Whitney U test (non-normal distributions).

Results

Sampling and data collection

Six hundred patients were enrolled in the present study. Of these, 24 did not meet the inclusion criteria and were excluded due to lack of data integrity. The remaining 576 patients were randomly assigned to a training set (n = 403, 70%) and a validation set (n = 173, 30%). The process of patient selection and allocation is illustrated in Figure 1.

Figure 1.

Flowchart of patient selection and allocation. In total, 600 patients were assessed for eligibility. Of these, 24 were excluded; 20 did not meet the inclusion criteria, and 4 declined to participate. The remaining 576 patients were randomly allocated into a training set (n = 403) and a validation set (n = 173) in a 7:3 ratio. The training set was used to establish a predictive logistic regression model and calculate the C-index. The validation set was used to assess the precision and reliability of the model and calculate the C-index. C-index: concordance index.

Comparison of different variables between patients with normal temperature and those with hypothermia in the training dataset

Among the 403 patients in the training set, 111 developed hypothermia, whereas 292 maintained normal body temperature, as listed in Table 1. Patients with normal body temperature had significantly shorter surgical duration (p < 0.001), shorter anesthesia duration (p < 0.001), lesser blood loss (p = 0.005), lower blood transfusion volume (p = 0.021), lower total input (p = 0.003), lower total output (p < 0.001), higher preoperative body temperature (p < 0.001), higher hemoglobin levels (p = 0.028), and higher BMI (p = 0.002) compared with those who developed hypothermia. In addition, the number of diseased vertebrae was significantly lower in the normal temperature group (p < 0.001), and tumor surgery was markedly more common among patients with hypothermia (p < 0.001). No statistically significant differences were observed in age (p = 0.068), albumin levels (p = 0.053), or white blood cell counts (p = 0.428) between the two groups.

Table 1.

Comparisons between hypothermia and normal groups in terms of potential risk factors based on the training dataset.

	Normal	Hypothermia	t/X²	p (overall)
	n = 292	n = 111	t/X²	p (overall)
Age	53.1 (14.5)	56.0 (14.5)	1.831	0.068
Surgical duration	222 (97.4)	266 (109)	3.891	<0.001
Anesthesia duration	250 (99.4)	295 (114)	3.931	<0.001
Blood loss	271 (337)	393 (396)	3.071	0.005
Blood transfusion volume	207 (411)	327 (483)	2.501	0.021
Total input	1526 (780)	1804 (862)	3.105	0.003
Total output	596 (545)	831 (578)	3.810	<0.001
Preoperative body temperature	36.7 (0.33)	36.4 (0.31)	7.512	<0.001
Albumin	39.2 (6.77)	38.0 (4.73)	1.659	0.053
Hemoglobin	134 (66.8)	124 (20.9)	2.459	0.028
White blood cell	10.5 (37.2)	8.56 (11.5)	0.527	0.428
BMI	23.8 (3.72)	22.6 (3.19)	2.993	0.002
Number of diseased vertebrae:			180.5	<0.001
1	112 (38.4%)	8 (7.21%)
2	152 (52.1%)	23 (20.7%)
3	28 (9.59%)	43 (38.7%)
4	0 (0.00%)	30 (27.0%)
5	0 (0.00%)	7 (6.31%)
Reason for surgery			67.21	<0.001
Non-tumor	285 (97.6%)	78 (70.3%)
Tumor	7 (2.40%)	33 (29.7%)

BMI: body mass index.

LASSO regression model for the selection of potential risk factors

Using the training dataset of 403 patients, a LASSO binary logistic regression formula was developed, as shown in Figure 2, and employed to identify 5 potential risk factors, including duration of anesthesia, preoperative body temperature, BMI, number of diseased vertebrae and reason for surgery, among previously identified 14 variables.

Figure 2.

LASSO binary logistic regression modeling to identify potential risk factors. (a) Ten-fold cross-validation was used to select the optimal parameter (λ) of the LASSO model with the lowest criteria. (b) Plot of partial likelihood deviation (binomial deviation) versus log(λ). Vertical dashed lines are drawn at the optimal values using the minimum criterion and the 1 SE of the minimum criterion (1-SE criterion). LASSO: least absolute shrinkage and selection operator; SE: standard error.

Furthermore, logistic regression analysis of these five factors filtered using LASSO regression showed that they were all statistically significant, as listed in Table 2. Tumor surgery (odds ratio (OR) = 10.02), a greater number of diseased vertebrae (OR = 18.39 per level), and lower preoperative body temperature (OR = 0.09 per centigrade) were strongly associated with an increased risk of intraoperative hypothermia. Lower BMI was also a significant risk predictor of intraoperative hypothermia (OR = 0.88 per unit).

Table 2.

Multivariate logistic regression analysis identified five statistically significant risk factors.

	Coef	SE	Wald Z	p-value	OR (95% CI)
Intercept	84.6472	20.9406	4.04	<0.0001	–
Anesthesia duration	−0.0116	0.0025	−4.64	<0.0001	0.99 (0.98–1.00)
Preoperative body temperature	−2.3626	0.5720	−4.13	<0.0001	0.09 (0.03–0.30)
BMI	−0.1230	0.0552	−2.23	0.0258	0.88 (0.80–0.98)
Number of diseased vertebrae	2.9119	0.3561	8.18	<0.0001	18.39 (9.12–37.08)
Reason for surgery	2.3043	0.6173	3.73	0.0002	10.02 (2.96–33.92)

BMI: body mass index; CI: confidence interval; OR: odds ratio; SE: standard error; Coef: coefficient.

Notably, longer anesthesia duration was associated with a slightly reduced risk of intraoperative hypothermia (OR = 0.99 per minute, p < 0.001). Although this inverse relationship may seem unexpected, more complex thoracic spine procedures, such as those performed for tumors or involving multiple vertebral levels, could prompt anesthesiologists to use warming measures more actively and monitor patients more closely. These intensified efforts during prolonged surgeries may help preserve core body temperature and reduce the risk of intraoperative hypothermia.

Subsequently, a nomogram model was established to visualize the contribution of each risk factor to hypothermia occurrence, as plotted in Figure 3.

Figure 3.

Nomogram of risk predictors based on logistic regression analysis. The intraoperative hypothermia nomogram was developed using the training dataset; the duration of anesthesia, preoperative body temperature, BMI, number of diseased vertebrae, and reason for surgery are illustrated in the figure.

Calibration curve and C-index

As shown in Figure 4, the calibration plot illustrated that the curve of the nomogram is similar to that of the ideal prediction, with a C-index of 0.919 (95% confidence interval (CI): 0.882–0.956).

Figure 4.

Calibration curves of the nomogram prediction using the training dataset. The apparent line indicates the performance of the self-training. The bias-corrected line indicates the performance of the model trained by repeated self-sampling, which corrects overfitting.

Decision curve analysis of the nomogram using the training dataset

As the threshold probability increases, the net benefit of the model decreases. However, our model that used the training dataset performed better in all threshold probability scenarios, except for the case where the threshold probability was very small, as illustrated in Figure 5.

Figure 5.

Decision curve analysis for the intraoperative hypothermia nomogram. The Y-axis indicates net benefits. The thin solid line represents the hypothesis that hypothermia occurred in all patients. The thick solid line indicates the hypothesis that no patient developed hypothermia.

Comparison of different variables between patients with normal temperature and those with hypothermia in the validation dataset

Among the 173 patients in the validation set, 42 developed hypothermia, and 131 maintained normal body temperature, as listed in Table 3. Patients with normal body temperature had significantly shorter surgical duration (p = 0.021), shorter anesthesia duration (p = 0.014), lower total output (p = 0.025), higher preoperative body temperature (p < 0.001), and lower number of diseased vertebrae (p < 0.001) compared with those who developed hypothermia. Additionally, tumor surgery was significantly more common in the hypothermia group (p < 0.001). No statistically significant differences were observed in the age (p = 0.364), blood loss (p = 0.102), blood transfusion volume (p = 0.642), total input (p = 0.059), albumin level (p = 0.299), hemoglobin level (p = 0.969), white blood cell count (p = 0.251), or BMI (p = 0.388) between the two groups. No difference was found between training and validation group in terms of the aforementioned 15 variables, as listed in Table 4.

Table 3.

Comparisons of variables between the hypothermia and normothermia groups in the validation set.

	Normal	Hypothermia	t/X²	p (overall)
	n = 131	n = 42	t/X²	p (overall)
Age	53.5 (14.4)	55.8 (13.8)	0.894	0.364
Surgical duration	221 (102)	276 (139)	2.774	0.021
Anesthesia duration	250 (107)	309 (138)	2.875	0.014
Blood loss	290 (321)	403 (403)	1.867	0.102
Blood transfusion volume	246 (431)	287 (512)	0.511	0.642
Total input	1546 (807)	1929 (1204)	2.134	0.059
Total output	646 (567)	945 (779)	2.702	0.025
Preoperative body temperature	36.7 (0.33)	36.4 (0.32)	5.154	<0.001
Albumin	41.3 (31.5)	38.3 (3.87)	0.601	0.299
Hemoglobin	128 (20.3)	128 (20.7)	0.039	0.969
White blood cell	13.9 (68.0)	7.05 (2.71)	0.653	0.251
BMI	24.0 (5.63)	23.2 (4.40)	0.765	0.388
Number of diseased vertebrae			75.53	<0.001
1	53 (40.5%)	1 (2.38%)
2	65 (49.6%)	10 (23.8%)
3	10 (7.63%)	14 (33.3%)
4	3 (2.29%)	13 (31.0%)
5	0 (0.00%)	3 (7.14%)
6	0 (0.00%)	1 (2.38%)
Reason for surgery			21.50	<0.001
Non-tumor	127 (96.9%)	31 (73.8%)
Tumor	4 (3.05%)	11 (26.2%)

BMI: body mass index.

Table 4.

Comparison of the 15 variables in the training and validation datasets.

	Test	Train	t/X²	p (overall)
	n = 403	n = 173	t/X²	p (overall)
Status			0.662	0.477
0	292 (72.5%)	131 (75.7%)
1	111 (27.5%)	42 (24.3%)
Age	53.9 (14.5)	54.0 (14.3)	0.137	0.890
Surgical duration	235 (102)	234 (114)	0.022	0.983
Anesthesia duration	263 (105)	265 (118)	0.212	0.835
Blood loss	305 (357)	317 (345)	0.386	0.696
Blood transfusion volume	240 (435)	256 (450)	0.416	0.682
Total input	1603 (812)	1639 (930)	0.469	0.657
Total output	661 (563)	719 (636)	1.091	0.299
Preoperative body temperature	36.6 (0.34)	36.6 (0.35)	0.653	0.516
Albumin	38.9 (6.29)	40.5 (27.5)	1.156	0.429
Hemoglobin	131 (58.1)	128 (20.3)	0.696	0.487
White blood cell	9.96 (32.2)	12.3 (59.2)	0.599	0.632
BMI	23.4 (3.62)	23.8 (5.36)	0.789	0.430
Number of diseased vertebrae			3.935	0.585
1	120 (29.8%)	54 (31.2%)
2	175 (43.4%)	75 (43.4%)
3	71 (17.6%)	24 (13.9%)
4	30 (7.44%)	16 (9.25%)
5	7 (1.74%)	3 (1.73%)
6	0 (0.00%)	1 (0.58%)
Reason for surgery			0.221	0.753
Non-tumor	363 (90.1%)	158 (91.3%)
Tumor	40 (9.93%)	15 (8.67%)

BMI: body mass index.

Comparison between the training and validation datasets

C-index of the validation set

The external validation C-index was 0.947 (95% CI: 0.878–0.960), indicating that the nomogram had good discrimination and good predictive abilities for predicting the probability of intraoperative hypothermia.

Discussion

As a computational tool based on regression modeling, a nomogram visualizes complex mathematical formulas graphically, allowing clinicians to quickly assess a patient’s condition using a specific model.⁷ It is an integrative model that helps promote individualized, precision medical care. Nomograms have been shown to achieve higher accuracy and be more personalized than traditional scoring systems in predicting the prognosis of certain diseases. They enable personalized prediction based on the values of each factor and are now widely used in studies on disease diagnosis and prognosis evaluation. Our results have shown that the nomogram established using the training dataset demonstrated satisfactory predictability; this was proven via external validation performed using a validation dataset. The C-index calculated using the validation dataset was 0.947.

The risk factors identified in our model include tumor etiology, involvement of multiple spinal levels, low preoperative body temperature, and low BMI. These factors were associated with a higher likelihood of intraoperative hypothermia, although all patients underwent comprehensive thermal management. At our institution, every patient underwent proactive warming measures such as prewarming, forced air warming, administration of warmed intravenous fluids, and continuous monitoring of core temperature. The fact that these variables remained significant despite standardized care suggests that they reflect underlying physiological or procedural challenges that routine interventions do not fully overcome.⁸ Patients with these characteristics may therefore benefit from more individualized or intensified warming strategies to maintain intraoperative normothermia.³

Intraoperative hypothermia during spine surgery under general anesthesia is common. Mild hypothermia can protect organs during hypoperfusion;⁹ in most cases, it will cause a series of stress reactions in the patient’s body, which exerts adverse effects on their recovery.¹⁰ If the core body temperature continues to decrease, the patient’s platelet function is inhibited, resulting in reduced coagulation activity and increased intraoperative bleeding.¹¹ This will also extend the action time of most anesthetic drugs and delay postoperative wakefulness.¹² After the patient is out of the effect of anesthesia, the stress response in the body is activated and a large number of catecholamines such as norepinephrine (NE) and epinephrine (E) are synthesized and secreted, resulting in excitation of the patient’s sympathetic nervous system, which may cause hemodynamic changes, including changes in the heart rate, systolic blood pressure, and diastolic blood pressure.¹³ Complications such as chills, arrhythmias, and hypotension occur after the core body temperature drops; these not only compromise patient safety but also lead to prolonged hospital stays and increased treatment costs.¹⁴ Therefore, it is crucial to strengthen perioperative temperature protection for patients undergoing spinal surgery.

The nomogram suggests that lower BMI, low preoperative body temperature, higher number of diseased vertebrae, and spinal tumor surgery are key risk factors for intraoperative hypothermia in patients undergoing spinal surgery. In this study, all patients were administered general anesthesia, which could have inhibited the hypothalamic thermoregulatory central function of patients. This has a certain impact on the protection of the patients’ core body temperature, which increases the outward diffusion of internal heat in patients and makes them prone to hypothermia.¹⁵ Higher number of diseased vertebrae and spinal tumor surgery have also been regarded as potential risk factors for intraoperative hypothermia. This can be explained by the fact that it takes more time to operate on patients with a greater number of affected vertebrae or those who require tumor dissection. Therefore, the extension of anesthesia time makes patients more susceptible to hypothermia. The use of different thermal insulation measures during general anesthesia can delay the outward diffusion of the patient’s core body temperature; however, with time, the core body temperature of patients who have undergone conventional heating only will gradually decrease, whereas that of patients for whom comprehensive thermal insulation (intravenous heating combined with circulating water temperature blanket) has been performed is effectively protected.^16–18

The predictive model in this study also suggests that preoperative hypothermia is a risk factor for intraoperative hypothermia. Previous studies have shown that preoperative warming can prevent an intraoperative drop in the core body temperature.¹⁹ Therefore, the patient’s body temperature can be increased in advance before the surgery, which improves the stability of their peripheral body surface temperature and allows a certain amount of heat storage.²⁰

BMI is a measure of a person’s weight relative to their height. People with a high BMI usually have more body fat and/or muscle mass. Body fat acts as insulation, helping retain heat, whereas muscle generates heat more efficiently. Therefore, individuals with higher BMIs may have better insulation and heat production, which can reduce the risk of intraoperative hypothermia. Another reason for the protective effect of high BMI is that in patients with obesity, when the core body temperature is lowered, body fat maintains heat balance by triggering early vasoconstriction.²¹ In addition, patients with a high BMI have a lower average body surface area and lower heat dissipation.

One notable finding of our multivariable analysis is the inverse association between anesthesia duration and intraoperative hypothermia, with each additional minute of anesthesia duration linked to a 1% reduction in risk. This result appears counterintuitive at first glance because longer procedures are expected to increase heat loss. In our clinical practice, however, surgeries that are expected to be lengthy or complex, such as tumor resections or multilevel spinal reconstructions, are routinely managed with more intensive temperature control, starting from the time of anesthesia induction. These measures often include prewarming, continuous forced air warming, infusion of warmed intravenous fluids, and frequent core temperature monitoring. Therefore, patients with longer surgical durations may actually maintain better thermal stability than those with shorter surgical duration wherein such interventions are used less consistently. This observation suggests that surgical duration by itself is not a reliable predictor of hypothermia when comprehensive warming strategies are applied proactively throughout the procedure.

There was no significant difference between the two groups in terms of intraoperative blood transfusion, intraoperative blood loss, and total input and output, which can be explained by the use of prewarming systems. Unlike in previous studies, age was not included as a risk factor in this study. Studies have shown that older patients are more prone to intraoperative hypothermia due to poor thermoregulation, less subcutaneous fat, slow blood circulation, low metabolism, poor sensitivity to temperature changes, and lower threshold for regulating vasoconstriction under general anesthesia.²²

The study has some limitations. It was a single-center study with a limited sample size, which may have affected the representativeness of the results. In addition, all patients in our cohort received aggressive and standardized thermal management as part of routine care at our institution. This included prewarming, forced air warming, fluid warming, and continuous temperature monitoring. Such high-level perioperative temperature control may not be available in other settings, especially in hospitals with fewer resources or less specialized protocols. Therefore, the incidence of hypothermia and the strength of the identified risk factors might differ in other centers, limiting the generalizability of our nomogram. Although our sample size was sufficient to detect moderate effect sizes for major predictors, rare outcomes or subgroup analyses may be underpowered.

Conclusion

In this retrospective cohort study, tumor-related spinal surgery, multilevel involvement, prolonged surgical duration, low preoperative body temperature, and low BMI were associated with an increased risk of intraoperative hypothermia. These factors may help identify patients who could benefit from enhanced thermal management during surgery. Further prospective studies are needed to confirm these findings.

Footnotes

Acknowledgments

Language editing assistance was provided by Qwen (Tongyi Qianwen), an AI-powered large language model developed by Alibaba Cloud.

Author contributions

All authors meet the ICMJE criteria for authorship. Y.W. contributed to conceptualization, methodology, formal analysis, writing–original draft, and visualization; X.Y. was responsible for data curation, investigation, software, and validation; A.P. provided resources and oversaw project administration and supervision; F.J. assisted in methodology, data collection, and clinical support; X.Z. contributed to writing–review and editing, critical revision, and interpretation of results; and X.L. supervised the study, acquired funding, and approved the final version for publication. All authors have read and approved the final manuscript.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declaration of conflicting interests

The author(s) declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

ORCID iD

Xiaowei Lin

References

Habib

Arifuzaman

Elbeltagi

, et al. Peri-operative hypothermia in trauma patients: a retrospective cohort analysis at a busy district general hospital within the National Health Service (NHS). Cureus 2024; 16: e74979.

Torossian

Bräuer

Höcker

, et al. Preventing inadvertent perioperative hypothermia. Dtsch Arztebl Int 2015; 112: 166–172.

Zhou

, et al. The warming management measures may need to be further enhanced during scoliosis correction and internal fixation surgery: a retrospective cohort study. Ther Hypothermia Temp Manag 2025; 15: 124–133.

Tzabazis

Miller

Dobrow

, et al. Delayed emergence after anesthesia. J Clin Anesth 2015; 27: 353–360.

Wang

, et al. Strategies for perioperative hypothermia management: advances in warming techniques and clinical implications: a narrative review. BMC Surg 2024; 24: 425.

Lin

Chen

Liu

, et al. Long-term trends in the burden of fracture of vertebral column in China: a comprehensive analysis from 1990 to 2021 based on the global burden of disease study 2021. Eur Spine J 2025.

Von Elm

Altman

Egger

; STROBE Initiativeet al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med 2007; 147: 573–577.

Zhang

Gao

, et al. Characteristics and risk factors of intraoperative hypothermia in adults: a multicenter prospective observational clinical study. J Clin Med 2025; 15: 31.

Yang

Zhao

Liu

LJ.

Mild hypothermia in improving multiple organ dysfunction after cardiac arrest. World J Emerg Med 2010; 1: 196–200.

10.

Huang

Chen

Wang

, et al. A study on the efficacy of acupuncture combined with erector spinae plane blockade in patients undergoing thoracoscopic lung lobectomy. Medicine (Baltimore) 2026; 105: e47325.

11.

Bjertnæs

Næsheim

Reierth

, et al. Physiological changes in subjects exposed to accidental hypothermia: an update. Front Med (Lausanne) 2022; 9: 824395.

12.

Liao

Chen

, et al. Bilateral thoracic paravertebral block for robotic hepatectomy without hepatic portal occlusion for a cirrhosis patient: a case report. Front Oncol 2026; 15: 1684794.

13.

Kamijo

Masuda

Nishikawa

, et al. Cardiovascular response and stress reaction to flumazenil injection in patients under infusion with midazolam. Crit Care Med 2000; 28: 318–323.

14.

Echeverria-Villalobos

Todeschini

Stoicea

, et al. Perioperative care of cannabis users: a comprehensive review of pharmacological and anesthetic considerations. J Clin Anesth 2019; 57: 41–49.

15.

Sessler

DI.

Temperature monitoring and perioperative thermoregulation. Anesthesiology 2008; 109: 318–338.

16.

Karapınar

Kocatürk

Cİ.

The effectiveness of sterile wound drapes in the prevention of surgical site infection in thoracic surgery. Biomed Res Int 2019; 2019: 1438793.

17.

Chataule

Hazarika

Jain

, et al. Preoperative forced-air warming strategy: is it effective in averting intraoperative hypothermia in elderly trauma surgical patients? Cureus 2022; 14: e29305.

18.

De Lira

CRN

Akutsu

RCCA

Coelho

, et al. Occupational stress and body composition of hospital workers: a follow-up study. Front Public Health 2024; 12: 1459809.

19.

Kasai

Hirose

Matsukawa

, et al. The vasoconstriction threshold is increased in obese patients during general anaesthesia. Acta Anaesthesiol Scand 2003; 47: 588–592.

20.

Jdidi

De Bisschop

Dugué

, et al. Optimal duration of whole-body cryostimulation exposure to achieve target skin temperature: influence of body mass index-a randomized cross-over controlled trial. J Physiol Anthropol 2024; 43: 28.

21.

Sari

Aksoy

But

The incidence of inadvertent perioperative hypothermia in patients undergoing general anesthesia and an examination of risk factors. Int J Clin Pract 2021; 75: e14103.

22.

Huang

, et al. Incidence of unintentional intraoperative hypothermia and its risk factors in oral and maxillofacial surgery: a prospective study. J Perianesth Nurs 2023; 38: 876–880.