Intelligent thermoregulation strategy versus routine thermal management on myocardial injury after noncardiac surgery during gastrointestinal tumor resection in elderly patients (SAGE): a multicenter, parallel-group, superiority trial

Abstract

Background

Inadvertent perioperative hypothermia is associated with adverse cardiovascular outcomes, but adequately powered trials using high-sensitivity troponin endpoints is lacking. The SAGE trial evaluates whether an intelligent, closed-loop thermoregulation strategy reduces myocardial injury after noncardiac surgery (MINS) in high-risk elderly patients.

Methods

This prospective, multicenter, parallel-group, randomized controlled superiority trial conducted at seven tertiary hospitals in China. Participants aged ≥65 years scheduled for elective gastrointestinal tumor resection under general anesthesia are randomly assigned (1:1) to intelligent closed-loop warming (target 36.8 °C, upper limit 37.0 °C) or routine thermal management (active warming only if core temperature ≤35.5 °C). The primary outcome is the incidence of MINS within the first 2 postoperative days , adjudicated using high-sensitivity troponin thresholds per AHA criteria. Secondary outcomes include surgical site infection, transfusion requirement, hypothermia burden, emergence delirium, delayed emergence, and postoperative shivering. With 695 participants per group (total 1,390, accounting for 10% dropout), the trial has 80% power at a one-sided significance level of 0.05 to detect a 6 absolute percentage point reduction in MINS (12% to 6%) with a superiority margin of −2 percentage points.

Discussion

The SAGE trial is designed to generate new evidence on the relationship between intraoperative hypothermia exposure and postoperative cardiovascular complications. It will evaluate whether closed-loop targeted normothermia reduces myocardial injury and will explore hypothermia exposure thresholds that may inform individualized perioperative management for geriatric surgical patients.

Trial Acronym

Intelligent Thermoregulation Strategy for Advanced-Age Elders undergoing gastrointestinal tumor resection (SAGE).

Trial Registration

Chinese Clinical Trial Registry (ChiCTR), identifier: ChiCTR2600117264. Registered on January 21, 2026 (http://www.chictr.org.cn).

Introduction

Background and rationale

While perioperative mortality has declined substantially over recent decades, postoperative mortality remains a significant global health burden. Approximately 1%–2% of surgical inpatients die within 30 days of surgery, with cardiovascular complications constituting the leading cause of postoperative death [1,2]. Myocardial injury, defined by postoperative troponin elevation attributable to an ischemic etiology, is the most prevalent of these cardiovascular complications. It is now recognized that approximately 8% of surgical inpatients aged >45 years’ experience Myocardial Injury after Noncardiac Surgery (MINS), yet the majority of these events are clinically silent and would remain undetected without systematic troponin surveillance[3,4]. Critically, MINS is not a benign biochemical anomaly; it confers a five-fold increase in 30-day mortality, with risk escalating exponentially as a function of peak troponin concentration [4,5]. Consequently, identification of modifiable perioperative risk factors for MINS is a clinical and research priority.

Inadvertent perioperative hypothermia (IPH), defined as a core body temperature < 36.0 °C, is one such potentially modifiable factor. IPH occurs with a reported incidence ranging from 7% to 90% [6–8], with nationwide surveys in China documenting rates of 39.9% in Beijing [9] and 44.5% across multiple regions [10]. Hypothermia triggers thermoregulatory vasoconstriction, elevating systemic vascular resistance and inducing tachycardia, mechanisms which may precipitate myocardial oxygen supply-demand mismatch [11]. Observational data link IPH to a higher incidence of perioperative cardiovascular events; specifically, normothermic patients exhibit a 1.4% event rate compared to 6.3% in hypothermic cohorts, with corresponding increases in ventricular arrhythmias and myocardial ischemia [12,13]. Beyond cardiovascular sequelae, IPH is associated with coagulopathy, increased transfusion requirements, surgical site infection (19% vs. 6% in normothermic patients), and prolonged hospitalization [14–17].

Despite these associations, the causal link between aggressive normothermia and reduction in major adverse cardiac events remains incompletely defined. A pivotal randomized trial by Frank et al. [12] suggested a cardioprotective signal with active warming; however, the study was of relatively modest sample size and, as conducted before the widespread adoption of high-sensitivity troponin assays, relied on Holter electrocardiography for outcome ascertainment. This diagnostic approach, while appropriate to its era, is substantially less sensitive for detecting perioperative myocardial injury than contemporary biomarker-based definitions [18]. The extent to which these methodological constraints may have influenced the observed treatment effect remains uncertain, underscoring the need for an adequately powered trial using current diagnostic standards.

Elderly patients undergoing major gastrointestinal oncologic surgery are at relatively high risk for inadvertent perioperative hypothermia and associated cardiovascular complications. Age-related physiological decline impairs thermoregulatory vasoconstriction and metabolic heat production, an effect compounded by anesthetic-induced inhibition of the hypothalamic set-point [19,20]. In elderly patients undergoing laparoscopic colorectal procedures, the incidence of intraoperative hypothermia can reach 74% despite routine care [21]. Extensive surgical exposure, prolonged operative times, and the use of cold insufflation or irrigation fluids further accelerate heat dissipation [22,23]. However, clinical evidence reveal that only 20% of at-risk patients receive continuous temperature monitoring, and active warming measures remain markedly underutilized outside of rescue scenarios.

This trial employs an intelligent closed-loop thermoregulation system that could potentially addresses these shortcomings. By utilizing continuous tympanic membrane thermometry integrated with an automated warming device, the system maintains core temperature within a predefined normothermic window (target 36.8 °C) with a precision of 0.1 °C. This approach circumvents the physiological lag of rectal probes and the reactive nature of conventional forced-air warming triggered only at hypothermic thresholds [24,25]. A target set point of 36.8 °C was selected because we reasoned that even with closed-loop feedback control, core temperature would fluctuate around the programmed target owing to the inherent latency of physiological thermoregulatory responses and dynamic adjustment of device power output. This target aims to ensure that intraoperative core temperature remains above the guideline-recommended threshold of 36.0 °C.

We hypothesize that this proactive, precision-based thermoregulatory strategy will confer a clinically meaningful cardioprotective benefit. The present trial is designed to evaluate whether an intelligent closed-loop warming strategy is superior to routine thermal management in reducing the incidence of MINS in elderly patients undergoing gastrointestinal tumor resection [26,27].

Explanation for choice of comparator

The comparator arm employs routine thermal management consistent with current pragmatic clinical practice. This strategy consists of initial passive insulation, with active warming reserved as a rescue intervention triggered only when core temperature declines to ≤ 35.5 °C. This comparator was selected to determine whether a prophylactic, targeted normothermia strategy (maintaining core temperature near 37.0 °C) provides incremental clinical benefit over the reactive, threshold-based approach commonly employed in contemporary operating rooms.

Objectives and hypotheses

Primary objective

The primary objective of this trial is to evaluate the superiority of an intelligent, closed-loop thermoregulation strategy compared to routine thermal management in reducing the incidence of Myocardial Injury after Noncardiac Surgery (MINS) in elderly patients undergoing elective gastrointestinal tumor resection.

Exploratory objective

Furthermore, leveraging the high-resolution continuous temperature data acquired by the closed-loop system, an exploratory objective is to delineate the exposure-response relationship between intraoperative hypothermia duration and MINS. Specifically, we aim to identify a clinically actionable threshold of cumulative hypothermic exposure below which the risk of postoperative myocardial injury is not substantially increased. Such a threshold, if identified, could inform individualized, precision-based perioperative thermal management targets for the geriatric surgical population.

Statistical hypotheses for the primary outcome

The study is designed as a superiority trial. The prespecified superiority margin (Δ) is set at −2 percentage points for the absolute risk difference in MINS incidence. This margin represents the minimum clinically important difference, determined after consultation with clinical experts in perioperative medicine and cardiology. Given the strong prognostic association between MINS and 30-day mortality, a reduction in MINS incidence smaller than 2 percentage points is considered clinically negligible and unlikely to justify a change in perioperative temperature management practices. The formal statistical hypotheses are:

1. Null Hypothesis (H₀): P₁ - P₂ ≥ −2%;

The incidence of MINS in the intelligent thermoregulation group minus the incidence in the routine management group is greater than or equal to −2% (the intelligent strategy does not confer a clinically meaningful reduction).

1. Alternative Hypothesis (H₁): P₁ - P₂ < −2%;

The incidence of MINS in the intelligent thermoregulation group is more than 2 absolute percentage points lower than that in the routine management group.

Consistent with the sample size calculation (one-sided α = 0.05), superiority will be assessed using a one-sided test at the 0.05 significance level. Results will be reported using the two-sided 90% confidence interval for the between-group difference in proportions. Superiority will be declared if the upper bound of the two-sided 90% CI lies entirely below −2 percentage points.

Methods: patient and public involvement, trial design

Patient and public involvement

Patients or the public were not involved in the design of this protocol. The trial design was developed by the investigators and sponsor based on clinical equipoise and logistical feasibility. The results of this trial will be disseminated to participants upon request and through peer-reviewed publication.

Trial design

This is a prospective, multicenter, parallel-group, randomized controlled superiority trial with a 1:1 allocation ratio.

Methods: participants, interventions, and outcomes

Trial setting

The trial will be conducted across seven tertiary academic hospitals in China. The Third Xiangya Hospital of Central South University serves as both the coordinating center and the sponsor-investigator site for this trial. Participating sites include: Cancer Hospital, Chinese Academy of Medical Sciences; Huadong Hospital Affiliated to Fudan University; Henan Cancer Hospital; Hunan Cancer Hospital; Hunan Provincial People’s Hospital; and The First Affiliated Hospital of Nanchang University.

Eligibility criteria for participants

Inclusion Criteria:

1. Age ≥ 65 years.

2. Scheduled for elective gastrointestinal tumor resection under general anesthesia.

3. Expected duration of surgery > 2 hours.

4. Provision of written informed consent.

Exclusion Criteria:

1. Known allergy to device materials.

2. American Society of Anesthesiologists (ASA) physical status ≥ IV.

3. Altered mental status or inability to cooperate with study procedures.

4. Less than 50% body surface area available for active warming.

5. Moderate to severe obesity (Body Mass Index ≥ 30 kg/m²).

6. Preoperative core temperature (tympanic temperature) > 37.5 °C.

7. Pre-existing diagnosis of thyroid dysfunction (hyperthyroidism or hypothyroidism).

8. Severe liver dysfunction (Child-Pugh Class C).

9. Intracranial pathology potentially affecting thermoregulation.

10. Preoperative severe anemia (Hemoglobin < 60 g/L).

11. Any condition deemed unsuitable for participation by the investigator.

Eligibility criteria for sites and providers

All participating sites must be tertiary hospitals with the capacity for major abdominal oncologic surgery and standard operating room equipment for invasive monitoring. Anesthesia providers and surgeons are not required to meet specific credentialing criteria beyond their standard institutional privileges.

Intervention and comparator details

General Environment: Ambient operating room temperature will be maintained between 20 and 23 °C in both groups.

Temperature Monitoring: All participants will have a tympanic membrane temperature probe (WMT104, Beijing Weimai Intelligent Technology (Beijing) Co., Ltd.) placed for continuous core temperature monitoring from operating room entry until PACU discharge.

1. Control Group (Routine Thermal Management): Participants will receive passive insulation (cotton blankets). Active warming (forced-air warming, fluid warmers, etc.) will be initiated only if core temperature decreases to ≤ 35.5 °C. The upper limit for any warming is 37.0 °C.

2. Intervention Group (Intelligent Thermoregulation): Participants will receive proactive, automated warming via a closed-loop system. Based on real-time feedback from the tympanic sensor, the warming device will dynamically adjust output to maintain a target temperature of 36.8 °C, with an upper limit of 37.0 °C.

Criteria for discontinuing or modifying interventions

The allocated thermal management strategy may be modified or discontinued if any of the following occur:

1. Serious Adverse Event (SAE): If continuation of the allocated warming strategy is deemed by the attending anesthesiologist or principal investigator to pose an unacceptable risk to the participant’s safety, the intervention may be modified or discontinued.

2. Prolonged Intraoperative Hypotension: If cumulative intraoperative hypotension exceeds 30 minutes (defined as SBP <90 mmHg, MAP <60 mmHg, or a decrease >20% from baseline SBP or MAP [28]), the investigator should reassess hemodynamic management and may adjust the warming strategy if clinically indicated. Prolonged hypotension will be recorded as a major protocol deviation and will be addressed in a prespecified sensitivity analysis (see Section ‘Additional analyses’). However, participants with prolonged hypotension will not be withdrawn from the trial or excluded from the primary analysis for this reason alone.

3. Major Protocol Violation (other than hypotension): If a major protocol violation is identified, the violation will be recorded and the participant may be excluded from the per-protocol analysis. The participant will remain in the trial and be included in the primary intention-to-treat analysis whenever outcome data are available.

All randomized participants will be included in the primary intention-to-treat analysis according to their assigned group, regardless of whether the allocated intervention is modified or discontinued, provided that primary outcome data can be ascertained. Discontinuation or modification of the intervention does not constitute withdrawal from the trial or exclusion from the primary analysis.

Strategies to improve adherence

Adherence to the allocated thermal management strategy will be monitored through continuous electronic logging of the warming device settings and core temperature data. Anesthesia providers at each site will receive standardized in-service training on the protocol prior to local trial initiation.

Concomitant care

Permitted and recommended practices

Standard general anesthesia practices are permitted. Investigators may use adjunctive regional or local blocks at their discretion. To minimize inter-center variability while maintaining real-world applicability, the following principle-based recommendations are provided:

1. Fluid management: Each center follows its established enhanced recovery after surgery (ERAS) or goal-directed therapy (GDT) protocol. The type and volume of fluids administered (crystalloids, colloids) are recorded in the eCRF.

2. Transfusion: A restrictive transfusion strategy is recommended. Red blood cell transfusion is suggested when hemoglobin falls below 70 g/L, or when hemoglobin is between 70 and 80 g/L with signs of inadequate oxygen delivery. All transfusions are recorded in detail.

3. Vasopressors: Intraoperative blood pressure is maintained within ±20% of the patient’s baseline value. The choice of vasopressor agent is left to the treating anesthesiologist; type, dose, and duration are recorded in the eCRF.

4. Analgesia: Multimodal analgesia is encouraged according to institutional pathways. All analgesic agents and doses are recorded in the eCRF.

Prohibited practices

The use of dexmedetomidine, esketamine, and neuraxial anesthesia is prohibited due to their potential confounding effects on postoperative shivering and thermoregulation.

Hemodynamic management

To minimize confounding from hypotension on MINS, the protocol stipulates maintenance of blood pressure within ±20% of the patient’s baseline value. Vasopressor use is not restricted but must be meticulously documented in the eCRF.

Outcomes

Primary outcome

Incidence of Myocardial Injury after Noncardiac Surgery (MINS) within the first 2 postoperative days. MINS is defined by postoperative troponin elevation adjudicated as resulting from myocardial ischemia without evidence of a non-ischemic etiology [26] (e.g. sepsis, rapid atrial fibrillation, pulmonary embolism et al.).

Measurement Variable: High-sensitivity Cardiac Troponin T (hs-cTnT) or Troponin I (hs-cTnI).

Time Point: Blood samples collected at baseline, Postoperative Day 1 (18 ± 6h), and Postoperative Day 2 (42 ± 6h).

Thresholds for Diagnosis: Based on assay type as specified by the American Heart Association [5]:

1. hs-cTnT: ≥ 65 ng/L, OR 20 to < 65 ng/L with an absolute change of ≥ 5 ng/L from baseline.

2. hs-cTnI (Abbott): ≥ 75 ng/L.

3. hs-cTnI (Siemens): ≥ 60 ng/L.

4. Conventional TnI: ≥ Upper Reference Limit (URL) 99th percentile × 2.

5. Elevated Baseline: For patients with preoperative values > 80% of the above thresholds, a postoperative increase ≥ 20% is required.

Secondary outcomes

Surgical site infection (SSI)

Incidence of superficial incisional, deep incisional, or organ/space SSI from completion of the index operation until hospital discharge, adjudicated using the full CDC/NHSN criterion set (Patient Safety Component Manual, Chapter 9: Surgical Site Infection Event, January 2026).

Transfusion requirement

Administration of any allogeneic blood component (packed red blood cells, fresh frozen plasma, platelets, cryoprecipitate) from the start of surgery through the index hospitalization. Autologous cell-salvaged blood is excluded. Both the proportion of participants transfused and the amount transfused will be recorded.

Hypothermia burden

Cumulative area below the 36.0 °C threshold from operating-room entry to PACU discharge. The total duration of hypothermia will also be reported.

Emergence delirium

Incidence assessed at 30 min after tracheal extubation using the CAM-ICU.

Delayed emergence

Inability to follow a simple verbal command persisting for >60 min after discontinuation of all general anesthetic agents.

Postoperative shivering

Incidence of clinically relevant shivering during PACU stay. Assessment at PACU admission, 15 min, and 30 min, and the highest score recorded will be used for analysis.

Complete operational definitions, including diagnostic criteria, threshold values, and assessment protocols for all the secondary outcomes, are provided in Table S1.

Harms

Adverse Events (AEs) and Serious Adverse Events (SAEs) will be assessed systematically from the initiation of intervention through the 3-year follow-up period. All events will be recorded in the electronic Case Report Form (eCRF). Causality assessment related to the device or study procedures will be performed by the site investigator.

Participant timeline

The schedule of enrollment, interventions, and assessments is summarized in Figure 1. Following a screening period (Days −14 to 0), eligible participants who have provided written informed consent will be randomly allocated on the day of surgery to either the intelligent thermoregulation strategy or routine thermal management. The allocated intervention is applied intraoperatively and is limited to the surgical procedure. Postoperative assessments are conducted at prespecified time points: postoperative day 1 (18 ± 6 h), postoperative day 2 (42 ± 6 h), the day of hospital discharge, and postoperative day 30 (± 2 days). Long-term survival follow-up is conducted by telephone interview every 6 months from Month 6 through Month 36 after surgery (± 5 days). A detailed schedule of all enrollment, intervention, and assessment procedures is shown in Figure 1.

Figure 1.

Time schedule of enrollment interventions, assessments and visits for participants.

Note: ^aConducted every 6 months from postoperative month 6 through month 36, with a visit window of ±5 days. Follow-up may be performed via telephone. ^bIf troponin values remain elevated or trend upward on POD 2, an additional measurement on POD 3 is recommended per clinical judgment. POD: postoperative day; SAE: serious adverse event; ASA: American Society of Anesthesiologists; QoR-15: 15-item Quality of Recovery scale; hs-cTnT/I: high-sensitivity cardiac troponin T/I; CRF: case report form.

Methods: sample size and recruitment

Sample size

The sample size calculation is based on the superiority hypothesis for the primary outcome (MINS). Assuming a control group MINS incidence of 12% in this elderly population [27] and an anticipated reduction to 6% in the intervention group (absolute risk difference of −6 percentage points), with a prespecified superiority margin (Δ) of −2 percentage points, a one-sided significance level of 0.05, and a power of 80%, a total of 1,262 participants (631 per group) is required. This calculation was performed using PASS software (Version: 25.0.6). To account for an anticipated dropout rate of 10%, the target enrollment is set at 1,390 participants (695 per group).

Recruitment

Participants will be identified through pre-admission testing clinics and elective surgical schedules at the seven participating centers. Enrollment will be competitive across sites to ensure timely completion.

Methods: assignment of interventions

Sequence generation

The random allocation sequence will be generated by an independent biostatistician not involved in trial execution using the SAS software (Version 9.4).

Type of randomization

A stratified, permuted-block randomization method will be employed. Randomization will be stratified by study site to ensure balance within each center. Block sizes will be concealed and documented in a separate document unavailable to the enrollment personnel.

Allocation concealment mechanism

Concealment of the allocation sequence will be maintained through a central, secure Interactive Web Response System (IWRS). The sequence is not accessible to site staff prior to participant registration.

Implementation

Eligible participants will be registered in the IWRS by authorized site personnel. The system will assign a unique randomization number and the corresponding group allocation (1:1) only after baseline information is confirmed. Personnel enrolling participants will have no access to the underlying sequence generation algorithm or future assignments.

Blinding

Due to the nature of the intervention (visible use of active warming equipment), blinding of participants and intraoperative anesthesia providers is not feasible. However, the following parties will be blinded:

• Outcome Assessors Investigators adjudicating MINS events and SSI diagnoses will be blinded to group allocation.

• Data Analysts The statistical team will perform primary analyses using masked group identifiers (Group A vs. Group B).

Unblinding procedures

Because the clinical care team is not masked to treatment assignment, formal unblinding procedures do not apply to these personnel. The statisticians performing the primary analyses will remain masked to group allocation until database lock. Emergency unblinding of these statisticians is not anticipated but may be undertaken in exceptional circumstances if deemed necessary by the independent safety monitor or the Ethics Committee to investigate a potential safety signal. Any such unmasking must be approved by the Steering Committee and will be fully documented, including the rationale, date, and individuals involved.

Methods: data collection, management, and analysis

Data collection methods

Data will be collected prospectively using a validated electronic Case Report Form (eCRF) system.

• Physiological Data: The intraoperative temperature and hemodynamic data will be captured directly from the monitors and exported as .CSV files for direct import into the study database to minimize transcription errors.

• Laboratory Data: Local laboratory reference ranges and assay platforms (hs-cTnT/I) will be documented.

• Instrument Validity: The QoR-15 scale and Shivering Scale are validated instruments in perioperative research.

Participant retention and follow-up

 

1. Follow-up at 30 days will be conducted via telephone if the patient is discharged.

2. Long-term survival follow-up (6–36 months) will be conducted via telephone interviews or review of national death registry data every 6 months.

3. For participants who discontinue the intervention early, consent will be sought to continue collection of primary outcome data (MINS).

Data management

Data management will utilize EpiData (version 4.6.0) software. Procedures include:

1. Double Data Verification: Discrepancy queries generated by Clinical Research Associates (CRAs) will be resolved against source documentation.

2. Data Entry: Electronic data capture with built-in range checks and logic validation.

3. Storage: Data will be stored on secure, password-protected servers with regular backups. Only authorized study personnel will have access.

Statistical methods

• Primary Outcome Analysis: The primary analysis will employ the Cochran-Mantel-Haenszel (CMH) test, stratified by study center, to compare the proportion of participants with MINS between groups. Consistent with the sample size calculation, superiority will be tested at a one-sided significance level of 0.05. For clarity of presentation, the between-group difference in MINS incidence will be reported with its two-sided 90% confidence interval, the upper bound of which corresponds to the one-sided 95% upper confidence limit. Superiority will be declared if the upper bound of the two-sided 90% CI lies entirely below the prespecified margin of −2 percentage points.

• Secondary Outcomes Analysis: Continuous variables will be analyzed using independent t-tests or Mann-Whitney U tests as appropriate. Categorical variables will be analyzed using chi-square or Fisher’s exact tests. Repeated measures will be analyzed using mixed-effects models.

Analysis populations

• Intention-to-Treat (ITT)/Full Analysis Set (FAS): All randomized participants, analyzed according to the group to which they were assigned. This is the primary analysis set.

• Per-Protocol Set (PPS): A subset of FAS excluding participants with major protocol violations. This will serve as a sensitivity analysis.

• Safety Analysis Set (SAS): All participants who received any trial intervention, analyzed according to the treatment actually received.

Missing data

For the primary outcome, missing data will be handled using a conservative approach. Sensitivity analyses will be performed using multiple imputation and worst-case/best-case scenario imputation to assess the robustness of the findings. For secondary endpoints, the Last Observation Carried Forward (LOCF) method may be used where appropriate.

Additional analyses

Prespecified subgroup analyses for the primary outcome will explore the consistency of the treatment effect across the following subgroups, including:

1. Surgical Approach (laparoscopic vs. open);

2. History of cardiovascular disease (present vs. absent);

3. Age category (65–74 vs. ≥75 years);

4. Duration of surgery (≤ median vs. > median);

5. Body mass index (< 24.0 vs. ≥ 24.0 kg/m²).

Subgroup analyses are exploratory; the trial is not powered for individual subgroup comparisons. Results will be displayed as forest plots, and interaction tests will be reported.

Sensitivity analyses will include:

1. Per-protocol analysis excluding participants with major protocol violations;

2. As-treated analysis based on the thermal management strategy actually received;

3. Analysis excluding participants with cumulative intraoperative hypotension exceeding 30 minutes;

4. Multiple imputation for missing primary outcome data.

The classification of major versus minor protocol violations will be finalized at the data review meeting prior to database lock.

Methods: monitoring

Data Monitoring Committee

Given that both the investigational closed-loop warming system and the routine thermal management strategy evaluated in this trial employ legally marketed medical devices with established safety profiles, a formal independent Data Monitoring Committee (DMC) with predefined stopping rules is not mandated for this trial. However, a three-tiered safety and data monitoring framework has been established to ensure ongoing trial integrity and participant safety.

1. The principal investigators (W.O. and Q.L.) will convene quarterly internal safety review meetings to audit adverse event rates, protocol compliance, and site performance metrics across all participating centers. Findings will be documented in written reports maintained at the coordinating center.

2. An independent safety monitor (a senior anesthesiologist not otherwise affiliated with the trial) has been appointed to review accumulating safety data at six-month intervals. The independent monitor will have access to unblinded group-level safety data and will report directly to the Ethics Committee of the coordinating center. Should any safety signal be identified, the monitor is empowered to recommend suspension or early termination of the trial to the Steering Committee and the Ethics Committee.

3. Additionally, if emerging data or external evidence indicates that a formal Data Monitoring Committee with predefined stopping boundaries is warranted, the Steering Committee will establish such a committee before further enrollment proceeds.

Interim analyses and stopping guidelines

No formal interim efficacy analyses are planned. The trial may be terminated early for safety concerns or futility based on the recommendation of the sponsor and lead principal investigator.

Trial monitoring

Site monitoring will be conducted by qualified CRAs according to Good Clinical Practice (GCP) guidelines. Monitoring visits will include source data verification, review of regulatory files, and assessment of protocol compliance. The frequency of monitoring visits will be determined by site enrollment rate and data quality metrics.

Protocol adherence will be audited continuously through electronic logging of the allocated thermal management strategy and through regular monitoring visits. The coordinating center will generate quarterly site performance reports, including enrollment rate, protocol deviation rate, completeness of primary outcome data, and timeliness of data entry. These reports will be reviewed by the Steering Committee. Corrective and preventive actions will be implemented for centers with performance concerns, which may include additional training, targeted monitoring, or suspension of enrollment.

Ethics

Research ethics approval

The study will be conducted in accordance with the Declaration of Helsinki. The protocol and informed consent materials will be submitted for review and approval by the Institutional Review Board (IRB) of The Third Xiangya Hospital of Central South University and each participating center prior to enrollment commencement.

Protocol amendments

Any substantial modifications to the protocol will be submitted as formal amendments to the responsible IRBs for approval prior to implementation. All investigators will be notified of approved amendments.

Informed consent

Written informed consent will be obtained by a qualified member of the study team (Principal Investigator or delegated physician) prior to any study-specific procedures. The process will include a detailed discussion of the trial’s purpose, procedures, risks, and the voluntary nature of participation.

Ancillary studies

This protocol does not include the collection of biological specimens for ancillary genetic or biobanking studies.

Confidentiality

Participant data will be identified by a unique randomization number only. All study documents linking personal identifiers to study numbers will be stored separately in locked, access-restricted files at the coordinating center. Electronic data transmission will be encrypted.

Ancillary and post-trial care

The sponsor of this investigator-initiated trial is The Third Xiangya Hospital of Central South University, which assumes overall responsibility for trial conduct and compensation for study-related injuries. Routine post-trial medical care remains the responsibility of the participant and their treating physicians. No protocol-specified ancillary or post-trial care is provided beyond the 36-month follow-up assessment.

Discussion

This multicenter, randomized controlled superiority trial is designed to address a persistent evidence gap in perioperative thermoregulation: whether proactive, precision-guided maintenance of normothermia reduces the incidence of Myocardial Injury after Noncardiac Surgery (MINS) in a high-risk geriatric surgical population.

The association between inadvertent perioperative hypothermia and adverse cardiovascular events has been recognized for decades. Observational studies demonstrate that hypothermic patients exhibit significantly elevated rates of postoperative myocardial ischemia and ventricular arrhythmias [12,13]. Despite this, the causal relationship between aggressive normothermia and cardioprotection remains inadequately established. A randomized trial by Frank et al. [12] suggested a reduction in morbid cardiac events with normothermia; however, the study enrolled only 300 patients and relied predominantly on Holter electrocardiography rather than contemporary high‑sensitivity troponin assays for event ascertainment, potentially missing the majority of true myocardial injury events [18]. The limited number of outcome events (fewer than ten) constrains statistical precision and precludes definitive clinical conclusions. Consequently, current practice guidelines recommend a core temperature target of ≥36.0 °C, a threshold that is arguably below the true physiological norm and may not represent the optimal target for myocardial protection [28].

The present trial overcomes these limitations through three key design features. First, it employs a sample size adequately powered to detect a clinically meaningful reduction in MINS, defined using high-sensitivity troponin thresholds endorsed by the American Heart Association [5]. Second, the continuous tympanic membrane thermometry integrated within the closed-loop system generates unprecedented granularity of intraoperative temperature data, enabling exploratory delineation of the exposure-response relationship between hypothermia duration and myocardial injury. Third, the prespecified superiority margin (Δ = −2 percentage points) defines the minimum between‑group difference considered clinically meaningful, against which the trial results will be assessed. This threshold was selected after consultation with perioperative medicine and cardiology experts: given the strong association between MINS and 30-day mortality, a reduction in MINS incidence smaller than 2 percentage points was judged unlikely to warrant a change in clinical practice. The margin is deliberately conservative relative to the anticipated 6-percentage-point effect, ensuring that any benefit detected by the trial is both clinically relevant and statistically robust.

Achieving the intended between-group temperature separation is critical to trial validity. Under routine care, unwarmed patients undergoing major abdominal surgery typically reach a final intraoperative core temperature near 35.5 °C, whereas the intervention group will be managed using the closed‑loop system with a target set point of 36.8 °C, around which core temperature is expected to fluctuate (upper limit: 37.0 °C). An average between‑group gradient of approximately 1 °C is therefore anticipated. A substantially smaller gradient would reduce the statistical power to detect outcome differences and could compromise the interpretability of the primary analysis. To optimize temperature separation, the intervention is applied proactively from the induction of anesthesia, and the control group receives active warming only as rescue therapy below 35.5 °C, thereby minimizing contamination from concurrent active warming.

Blinding of the intraoperative anesthesia team is not feasible because the active warming equipment used in the intervention arm is physically distinct from standard passive insulation. Nevertheless, the primary outcome (MINS) is a laboratory-based endpoint defined by prespecified troponin thresholds, which are inherently objective. Troponin concentrations will be adjudicated by an independent clinical events committee masked to group allocation. Secondary outcomes, including surgical site infection and transfusion requirements, will also be adjudicated by assessors blinded to treatment assignment. These measures are expected to substantially mitigate the risk of ascertainment bias arising from the open-label design. Additionally, intraoperative hypotension is a recognized competing risk factor for MINS. Although all randomized participants are retained in the primary intention-to-treat analysis, the observed relationship between hypothermia and myocardial injury may be partially confounded by hypotensive episodes. The prespecified sensitivity analysis excluding participants with prolonged hypotension is designed to assess the robustness of the primary findings to this potential confounding.

Elderly patients undergoing gastrointestinal tumor resection face a confluence of predisposing factors that place them at disproportionately high risk for both inadvertent perioperative hypothermia and postoperative cardiovascular complications. Should the intelligent thermoregulation strategy prove superior, the results would provide evidence to support a transition from reactive, threshold‑based warming to proactive, automated normothermia maintenance in this vulnerable cohort. An exploratory analysis of cumulative hypothermic exposure may further define a temperature‑time threshold below which the risk of MINS does not appear substantially elevated, thereby informing individualized, precision‑oriented perioperative thermal management. Conversely, a neutral result would undermine the assumption that mild intraoperative hypothermia is a clinically important causal contributor to myocardial injury and would provide a rationale for redirecting investigative efforts toward alternative modifiable risk factors [27].

Finally, a limitation of this trial should be acknowledged. The SAGE trial is an open-label design, because the active warming equipment is physically distinct from standard passive insulation, blinding of the intraoperative anesthesia team is not feasible. Although the primary outcome (MINS) is objective, the open-label design cannot fully exclude subtle behavioral differences in intraoperative care between groups.

In conclusion, this trial is designed to generate new evidence on the association between intraoperative hypothermia exposure and postoperative cardiovascular complications, with MINS as the primary endpoint, in elderly surgical patients. The trial will evaluate whether a closed‑loop, target‑driven thermoregulation strategy can reduce myocardial injury after gastrointestinal tumor resection by providing more comprehensive perioperative temperature protection. Furthermore, leveraging the large‑scale, high‑resolution continuous temperature data collected across multiple centers, this trial would explore and estimate a hypothermia exposure threshold that is closely associated with postoperative complications in geriatric patients, which could potentially contribute to individualized, evidence‑based perioperative temperature management for this high‑risk population.

Trial status

This trial is registered with the Chinese Clinical Trial Registry (ChiCTR, http://www.chictr.org.cn) under identifier ChiCTR2600117264. The current protocol version is 2.0, dated January 21, 2026. Participant enrollment commenced in April 2026 and is currently ongoing. Recruitment is anticipated to be completed by approximately December 2028.

Protocol version

The current protocol is version 2.0 (December 16, 2025). Amendments in this version include refined diagnostic criteria for myocardial injury after noncardiac surgery (MINS) per the 2021 American Heart Association scientific statement, an updated sample size calculation, and revised secondary outcome definitions.

Funding Statement

This study was supported by Weimai Intelligent Technology (Beijing) Co., Ltd., which provides the investigational closed-loop thermoregulation devices (WMT104) and related technical support. The funder has no role in study design, data collection, analysis, interpretation, or the decision to publish.

Ethics approval and consent to participate

This study was approved by the Ethics Committee of The Third Xiangya Hospital of Central South University on October 22, 2025 (No. R25083). Written informed consent will be obtained from each participant by the study team prior to enrollment.

Consent for publication

Participants will consent to have their data and results published anonymously, just as described in the informed consent form. The results will be published in a journal as a research article when the trial is finished.

Trial governance and committees

The Third Xiangya Hospital of Central South University is the regulatory sponsor and assumes overall responsibility for trial design, conduct, data analysis, and the decision to publish. This is an investigator-initiated trial (IIT). Weimai Intelligent Technology (Beijing) Co., Ltd. provides the investigational closed-loop thermoregulation devices (WMT104) but has no role in study design, data collection, analysis, interpretation, or the decision to publish. The Coordinating Center is based at The Third Xiangya Hospital. A Steering Committee composed of the principal investigators and site leads oversees protocol implementation and publication policy. An independent Endpoint Adjudication Committee, consisting of clinicians masked to treatment allocation, will adjudicate all MINS events and other cardiovascular outcomes. The Data Management Team at the coordinating center ensures data quality and integrity. An independent biostatistician generated the randomization sequence and will perform the blinded primary analyses. Safety monitoring is conducted by the principal investigators and the sponsor’s medical monitor; a formal Data Monitoring Committee was not established given the low-risk nature of the device intervention.

Trial registration

The SAGE trial is registered with the Chinese Clinical Trial Registry (http://www.chictr.org.cn), identifier ChiCTR2600117264. The current protocol version is 2.0, dated January 21, 2026. Participant enrollment commenced in April 2026 and is currently ongoing.

Protocol and statistical analysis plan

The full trial protocol and the detailed statistical analysis plan are available for review on the Chinese Clinical Trial Registry (http://www.chictr.org.cn) under registration identifier ChiCTR2600117264. The current protocol version is 2.0 (January 21, 2026). Requests for additional information may be directed to the corresponding author.

Dissemination policy

Trial results will be submitted for publication in a peer-reviewed journal regardless of outcome. Authorship will follow the recommendations of the International Committee of Medical Journal Editors. A plain-language summary of the results will be made available to trial participants upon request. Aggregate data will be deposited in the Chinese Clinical Trial Registry within 12 months of trial completion. The Steering Committee will determine the appropriate forums for presentation at national and international scientific conferences.

Disclosure statement

All authors declare that they have no other financial or non-financial competing interests.

Availability of data and materials

De-identified individual participant data underlying the primary and secondary outcome results, together with the data dictionary and statistical analysis code, will be available to qualified researchers upon reasonable request beginning 12 months after publication. Proposals should be submitted to the corresponding author and will be reviewed by the Steering Committee. Data sharing requires institutional review board approval and a signed data access agreement.