Accurate monitoring of core temperature is critical in patients under general anesthesia, as it is necessary to detect malignant hyperthermia and prevent perioperative hypothermia. Core temperature can be measured at several anatomical sites, including the pulmonary artery, distal esophagus, nasopharynx, and tympanic membrane (TM) [1,2]. Among these, the pulmonary artery provides the most accurate estimation of the true core temperature [3]; however, it is only routinely accessible in patients undergoing pulmonary artery catheterization. Although esophageal and nasopharyngeal temperatures are commonly measured during general anesthesia [1,4], these probes can only be inserted after endotracheal intubation and must be removed before extubation to reduce patient discomfort, making them unsuitable for continuous use under general anesthesia. Moreover, these probes are impractical for procedures that do not use general anesthesia, such as those performed under regional anesthesia or procedural sedation.

The TM temperature offers a less invasive alternative for core temperature monitoring because it shares a similar vascular supply from the branches of the internal carotid artery with the brain [3,5]. Nevertheless, its susceptibility to ambient temperature influences the anatomical variability of the auditory canal and has several notable limitations [5]. Several studies have used contact thermocouple probes (Mon-a-Therm®) placed in direct contact with the TM and sealed with cotton or gauze to ensure isolation from ambient operating room air [6-8]. However, this technique is inconvenient for routine clinical use. Alternatively, infrared tympanic thermometers are available but are not suitable for continuous intraoperative monitoring during anesthesia.

In the present study, we evaluated the clinical performance and safety of a novel, commercially available contact-type TM probe designed for continuous TM temperature (T-TM) monitoring (Fig. 1). Herein, we compared this probe with both infrared tympanic thermometer temperature (T-IRT) and standard esophageal temperature (T-ESO) using a series of laboratory and clinical investigations.

Overall, this study evaluated the accuracy and clinical utility of a commercially available T-TM probe for core temperature monitoring during general anesthesia. In vitro validation showed excellent agreement between the T-TM and T-ESO probes, with minimal mean bias and tight LoA. Clinical comparisons demonstrated that T-TM closely matched T-IRT, although T-TM consistently yielded slightly higher readings than T-ESO during anesthesia. The detection rate of intraoperative hypothermia (< 36.0°C) was significantly lower with T-TM than with T-ESO, indicating potential differences according to the measurement site.

Continuous monitoring of the TM temperature is typically performed using a contact thermocouple probe placed directly against the TM and sealed with cotton or gauze [6-8]. To overcome the limitations of these previous approaches and enhance both safety and accuracy, we designed a new T-TM probe. Unlike traditional TM thermocouples, which generate voltage based on the temperature gradients between two dissimilar metals, our probe utilizes a thermistor (thermal + resistor). Thermistors are semiconductor materials whose resistance changes predictably in response to temperature, thereby allowing the precise detection of subtle variations and providing cost-effectiveness suitable for single-use applications. The T-TM probe features a soft blunt tip attached to a low-tension microspring to ensure gentle TM contact. The shaft is adjustable in length, whereas sealing is achieved using two or three flexible flanges. To improve the measurement fidelity, the probe was designed to reach an area adjacent to the malleus on the TM, which is considered optimal owing to its high vascularity [12]. It may serve as an alternative to T-ESO monitoring and is suitable for use in various settings such as pre-induction temperature assessment, procedural sedation [13], regional anesthesia [14], oral and maxillofacial surgery, and postoperative monitoring in recovery units or intensive care units.

To ensure consistency and eliminate potential inter-device variability between temperature probes, we first conducted an in vitro experiment comparing the measurement of a water bath by the T-TM and T-ESO probes prior to clinical application. An in vitro comparison between the T-TM probe and T-IRT was not conducted because it is not suitable for accurately measuring water temperature in an open environment. This experiment revealed a minimal mean bias of −0.01°C (95% CI, −0.03 to 0.00°C) and an acceptable LoA (−0.22 to 0.20°C) between the T-TM and T-ESO probes (Fig. 4). A bias and SD within ±0.5°C between the two temperature measurement modalities has generally been regarded as clinically acceptable in previous validation studies of core temperature monitoring devices [11,15,16], and this tolerance limit is consistent with the magnitude of normal circadian variation in healthy individuals [10].

In our clinical comparison between T-TM and T-IRT, we found that the overall mean bias was within the acceptable range [mean bias of 0.04°C (95% CI, 0.00 to 0.07°C) and LoA ranged from −0.33 to 0.41°C] (Fig. 5). However, the bias between the two devices varied considerably across different time points. Overall, the T-IRT measurements tended to be higher than the T-TM pre-anesthesia (T-0), but lower at the end of surgery (T-END). In other words, as body temperature decreased, T-IRT exhibited a greater temperature decline than T-TM. There are two possible explanations for this observation: First, infrared thermometers have a narrow operational range (typically 35 to 42°C); consequently, in the hypothermic range, T-IRT may exhibit a downward measurement bias, producing values lower than the actual core temperature and potentially exaggerating the severity of hypothermia [17-19]. Second, the T-TM probe is enclosed with a thermal mass and insulation material; therefore, it may exhibit a modest time lag in response to dynamic temperature changes compared to infrared thermometers.

Unlike the in vitro validation results, which indicated minimal mean bias between T-TM and T-ESO, regardless of temperature changes, T-TM was consistently higher than T-ESO during anesthesia, with the difference increasing as body temperature decreased. This finding aligns with those of previous studies reporting that the brain temperature is approximately 1°C higher than the body temperature [20,21]. Additionally, a magnetic resonance spectroscopy study in healthy volunteers found that the brain temperature ranged from 36.1 to 40.9°C, with the mean brain temperature (38.5 ± 0.4°C) exceeding the mean oral temperature (36.0 ± 0.5°C) [22]. Other studies have also noted that T-IRT can be slightly higher than the pulmonary artery temperature [23,24]. Overall, the results of the present study show that T-TM was consistently higher than T-ESO, which may be attributed to TM reflecting brain temperature, as well as the effects of redistribution hypothermia and the cooling influence of mechanical ventilation on T-ESO during anesthesia.

When used as the sole method for intraoperative temperature monitoring, T-TM tends to yield slightly higher readings than core temperature probes such as T-ESO, a fact that should be carefully considered. In the present study, at 45 min after induction, when T-ESO was 36.00 ± 0.37°C, T-TM measured 36.65 ± 0.37°C, corresponding to a mean bias of 0.65°C. The temperature difference between T-TM and T-ESO exceeded 0.73°C after one hour of general anesthesia. These findings indicate that T-TM may underestimate the severity and timing of intraoperative hypothermia, potentially delaying appropriate thermal intervention. In clinical practice, a more stringent threshold or adjustment factor may be necessary when using T-TM alone to ensure timely detection and effective prevention of perioperative hypothermia. Additionally, these results indicate that a standard hypothermia threshold of 36°C may not be equally applicable to both measurement sites. To enhance clinical accuracy, it may be beneficial to consider clarifying site-specific definitions of hypothermia. However, further studies are required to establish standardized criteria.

When using a T-TM probe, the primary safety concerns include auditory canal trauma and TM perforation. In the present study, a single case of minor auditory canal trauma occurred during the early phase when the initial prototype was used. The probe design was subsequently improved during the course of the study by incorporating softer biocompatible materials and ergonomic modifications, after which no further improvements were observed. The injury resolved spontaneously without intervention. In rare cases of TM perforation, conservative management is generally sufficient, as most perforations heal spontaneously within 3-6 weeks, and surgical intervention is typically reserved for persistent or complicated cases [25-27]. These findings collectively suggest that with proper techniques and attention to anatomical variability, the T-TM probe can be safely applied in routine anesthetic practice.

This study had some limitations. First, the primary objective of our laboratory-based experiment was to compare the T-TM and T-ESO probes. Although the water-heating model effectively simulated the rapid temperature increase associated with malignant hyperthermia during anesthesia, it did not adequately replicate the gradual temperature decline typically observed during the maintenance phase. This discrepancy may limit the applicability of our findings to clinical scenarios involving hypothermia. Second, the thermistors used in this study were within the accuracy range specified by the manufacturer based on laboratory testing using water heating over a period of approximately one hour. However, larger offsets may occur in real clinical settings, potentially affecting temperature accuracy. Further studies are needed to evaluate this discrepancy in clinical practice. Third, our study focused on patients undergoing ear or nasal surgery, selected because of its relatively short duration and low invasiveness, with support from the otolaryngology department. While this helped minimize procedural variability, it limited the generalizability of our findings. Future research should include a broader range of surgical procedures and larger patient populations. Moreover, studies exploring multimodal temperature monitoring, including T-TM monitoring under general anesthesia, would be valuable for improving our understanding of thermoregulation and preventing thermal complications.

In conclusion, the newly developed T-TM probe showed excellent agreement with standard T-ESO probes in laboratory testing and was comparable to infrared tympanic thermometers in clinical settings. Although T-TM tended to record higher temperatures than T-ESO during anesthesia, further research is required to elucidate the temperature variations among different anatomical measurement sites. Our findings support the use of the T-TM probe as a reliable and practical option for continuous core temperature monitoring in various settings.