Computer-assisted searches were conducted in databases such as PubMed, The Cochrane Library, Web of Science, Embase to collect RCTs on AW devices for hypothermia in cesarean section patients during the perioperative period, with the search period being from the inception of the database to October 2024. Search Strategy: The search strategy was determined based on the (patient, intervention, comparison, outcome) PICO principle, with main search terms including Cesarean Section, Post Cesarean Section, and Heating. All searches combined subject headings with free-text words, and the search strategy was finalized after several pre-searches. Expanded Search: Hand searches and gray literature database searches were conducted, searching for relevant databases of ongoing trials (clinical trial registries), databases related to clinical practice guidelines, public search engines, and tracing references. Hand searches of published journals and searching for relevant conference information were performed to comprehensively collect clinical trials that meet the inclusion criteria. The specific search history is shown in Supplementary Table 1.

Literature was screened based on the pre-established inclusion criteria, and duplicate documents in the databases were removed by Zotero (version 7.0.0; Corporation for Digital Scholarship, Vienna, VA, USA). The titles and abstracts of the obtained literature were read to preliminarily exclude documents that clearly did not meet the inclusion criteria. Full texts of potentially eligible documents were reviewed, and data was extracted from documents that met the inclusion criteria. If there was any doubt about a document, a decision was made after contacting the original author to obtain relevant information or after analysis and evaluation. The “snowball” method was used to trace the references cited in the included documents. All steps were independently conducted by two evaluators, with cross-checking. Disagreements were resolved through discussion or by seeking the opinion of a third party.

The methodological quality of the included studies were evaluated using the “Risk of Bias Assessment” tool recommended by the Cochrane Handbook 5.0, which includes: ① Randomization process, the method used to generate the random sequence; ② Allocation, the implementation of the allocation concealment; ③ Blinding, whether blinding methods were used; ④ Completeness of outcome data; ⑤ Selective reporting of study results; ⑥ Presence of other sources of bias. The assessment results were identified as “low risk”, “high risk”, or “unclear”. The evaluation levels were divided into three categories: A (all items are at low risk), B (some items are at low risk), and C (no items are at low risk). Studies with a grade of C were excluded.

Data were analyzed using the Review Manager 5.4. (Cochrane Collaboration, Oxford, UK) Software for Meta-analysis. Because not all studies reported every endpoint, the number of trials and participants contributing to each meta-analysis varies; we reported the sample size per outcome. Binary data such as the incidence of hypothermia and shivering were expressed using the relative risk (RR) and 95% confidence interval (CI). Heterogeneity among studies was assessed using the chi-squared test and the I² statistic. When there was statistical homogeneity among the studies (p $>$ 0.05, I² 50%), a fixed-effects model was used for analysis. If there was statistical heterogeneity among the studies (p 0.05, I² $>$ 50%), a random-effects model was used for analysis. When heterogeneity existed among the studies, subgroup analysis was conducted after identifying the possible reasons. For continuous variables such as core body temperature, when the units of measurement are not the same, the standardized mean difference (SMD) was used as the metric, which was calculated as the difference between the estimated means of two groups divided by the average standard deviation, eliminating the influence of units. For continuous variables with the same units of measurement, the weighted mean difference (WMD) was used as the summary outcome measure. Sensitivity analyses were conducted to identify the sources of heterogeneity, and funnel plots were used to test for publication bias. Symmetry in the funnel plot indicates the absence of publication bias, while asymmetry suggests its presence.

A total of 440 articles were initially identified. After a step-by-step screening process, 12 RCTs [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22] were ultimately included, involving a total of 1184 patients. The literature screening process and results are shown in Fig. 1.

Fig. 1.

Literature search flow chart.

The basic characteristics of the included studies are presented in Table 1 (Ref. [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22]). The included literature encompasses studies from the United States, the United Kingdom, Germany, Canada, South Korea, and other countries, with a total of 1184 patients across the included studies, of which 591 were in the experimental group and 593 in the control group. 9 articles provided detailed reports on randomization methods, 11 articles described the concealment of the allocation sequence, 6 articles described how blinding was implemented for the study subjects and interventionists, and 4 articles introduced blinding for outcome assessors. The specific results of the risk of bias assessment for the included studies are shown in Fig. 2. In this study, grade rating was used for analysis. The results of the analysis (Table 2) suggested that the results of surgery/core temperature at the time of PACU entry, incidence of hypothermia, and incidence of shivering were of low quality.

Fig. 2.

Risk bias of graph.

Funnel plots and the Egger test were used to evaluate the publication bias of this study (Figs. 7,8,9). The results showed that core temperature at the end of surgery/core temperature at the time of PACU entry (p = 0.742), incidence of shivering (p = 0.136) suggested that the possibility of publication bias is small. The funnel plot of Incidence of Hypothermia (p = 0.001) was asymmetric, so that publication bias was more likely. The trim-and-fill results of this study suggest that there is a publication bias in the incidence of shivering and hypothermia, which does not affect this conclusion.

Fig. 7.

Funnel plot of the meta-analysis of core temperature at the end of surgery/core temperature at the time of post-anesthesia care unit (PACU) entry.

Fig. 8.

Funnel plot of the meta-analysis of incidence of hypothermia.

Fig. 9.

Funnel plot of the meta-analysis of incidence of shivering.

In this study, Meta regression analysis was conducted according to heating method, starting time, anesthesia method, temperature measurement site, and country. The analysis results (Table 3) suggested that the heating method, starting time, anesthesia method, temperature measurement site, and country were not the source of heterogeneity.

The constancy of body temperature is fundamental for maintaining various physiological functions. Effective monitoring and regulation of body temperature are one of the most important measures to ensure the success of surgery and anesthesia and to reduce postoperative complications [23]. The findings of this study indicate that the use of AW devices effectively maintains core body temperature in cesarean section patients during and after surgery, with significantly higher temperatures observed at the conclusion of surgery and upon arrival in the PACU compared to the control group. This helps reduce the incidence of perioperative hypothermia, improves postoperative recovery, and lowers the risk of complications such as shivering. Therefore, the application of AW devices in cesarean section surgery has clinical significance. It significantly enhances patients’ temperature management outcomes, thereby improving postoperative comfort and overall recovery. Cesarean section surgeries often use spinal anesthesia, and in special circumstances, general or local anesthesia may be used. Spinal anesthesia can lead to heat loss in patients during surgery, mainly through vasodilation and redistribution of blood, while also impairing peripheral and central thermoregulatory centers [24]. Therefore, during major or prolonged surgeries under spinal anesthesia, patients may experience severe hypothermia. The results of this meta-analysis show that the core temperature of the experimental group after surgery/upon entry to PACU is higher than that of the control group, and the incidence of hypothermia is lower than that of the control group, consistent with the results of the study by Mackeen et al. [25]. AW devices mainly provide additional heat for active insulation, with air warming and fluid warming being the two most common types. Air warming increases skin temperature through peripheral conduction and convection, while fluid warming increases skin temperature through peripheral and central conduction, which is related to the distribution of fluid throughout the body in a short time via blood circulation [6, 26]. However, in a systematic review by Dumont et al. [27], the effectiveness of FAW devices in preventing perioperative hypothermia in patients undergoing joint replacement surgery was evaluated, and the results showed no difference in the incidence of hypothermia with or without the device. There are several reasons for these observations. In this systematic review, only six RCTs with a total of 287 patients were included, which is a relatively small sample size and may not yield accurate results. Although the effect size for core body temperature (MD = 0.26 °C) was statistically significant, its clinical significance may be limited. While this change achieved statistical significance, a 0.26 °C temperature variation may be insufficient to produce a clinically meaningful improvement in patient outcomes in clinical practice. Therefore, caution is warranted when interpreting these findings regarding their practical relevance in clinical applications.

Shivering is one of the most common complications of spinal anesthesia, often manifesting as rapid rhythmic contractions of the skeletal muscles, which is a compensatory response of the body to hypothermia [27]. Shivering can increase a patient’s oxygen consumption, leading to myocardial ischemia and arrhythmias in patients with underlying medical conditions [28]. The mechanism of shivering is likely related to factors such as the distribution of core temperature heat to the periphery, heat loss, and a lower threshold for response. Low temperatures suppress the thermogenic center and excite cold-sensitive neurons, causing muscle tremors to generate heat and maintain thermal balance, thereby inducing shivering [29]. The results of this study show that the incidence of shivering in the experimental group during the perioperative period was significantly lower than that in the control group, consistent with the findings of the study by Pu et al. [30]. The body temperature regulation of cesarean section patients during surgery can be influenced by multiple factors. To maintain normal body temperature during surgery, the key is to perform active and PW measures simultaneously and strengthen heat management, so as to ensure that the body temperature is maintained at a normal level [31]. AW measures are important for maintaining the normal body temperature of patients, reducing the incidence of hypothermia, and preventing shivering. Zhao et al. [32] used a warming blanket with an intermittent warming mode to keep cesarean section patients warm, providing a stable core temperature while avoiding skin damage. Although PW has been convenient and easy to use, it is limited in its effectiveness in improving hypothermia in patients, mainly relying on the patient’s body temperature for warmth. The insulating effect of surgical drapes and blankets during surgery can also be affected by wet fluids. Compared to the effect of size on core body temperature, the reduction in hypothermia (RR = 0.69) and shivering (RR = 0.54) holds greater clinical significance. Hypothermia and shivering are common complications during cesarean section. Controlling these factors not only enhances patient comfort but also promotes postoperative recovery. Therefore, these findings more directly reflect the clinical efficacy of warming interventions, particularly their contribution to reducing hypothermia and shivering.

The results of this study showed significant heterogeneity. Subgroup analyses were conducted to explore the sources of this heterogeneity, but they did not significantly alter the heterogeneity of the results. There are certain limitations to this study, as it did not include more outcome measures, such as patient thermal comfort and the incidence of adverse outcomes. Further research is needed on various AW devices to explore the optimal heating device and to assess whether the use of multiple devices in combination offers more advantages than a single mode. However, differences in blinding methods and allocation concealment exist in the studies, which may introduce certain biases, particularly when evaluating the effectiveness of active and passive heating devices. Such biases are more common in thermal insulation trials, as operators and patients may be unable to completely avoid cognitive or behavioral biases. We acknowledge this limitation and recommend that future studies have strict controls over blinding and allocation concealment to enhance the credibility of the research and the reliability of the results. Although we assessed publication bias using funnel plots, the results revealed asymmetry in certain outcomes. This asymmetry may stem from publication bias, particularly when studies with negative or non-significant results are underreported. Furthermore, funnel plot asymmetry may also be associated with small sample sizes or methodological differences. Therefore, while we have made every effort to assess and control for bias, it is important to note that this limitation may exert some influence on the conclusions of this study.