Insight into the use of tympanic temperature during target temperature management in emergency and critical care: a scoping review

Background Target temperature management (TTM) is suggested to reduce brain damage in the presence of global or local ischemia. Prompt TTM application may help to improve outcomes, but it is often hindered by technical problems, mainly related to the portability of cooling devices and temperature monitoring systems. Tympanic temperature (TTy) measurement may represent a practical, non-invasive approach for core temperature monitoring in emergency settings, but its accuracy under different TTM protocols is poorly characterized. The present scoping review aimed to collect the available evidence about TTy monitoring in TTM to describe the technique diffusion in various TTM contexts and its accuracy in comparison with other body sites under different cooling protocols and clinical conditions. Methods The scoping review was conducted following the guidelines of the Preferred Reporting Items for Systematic Review and Meta-Analysis extension for scoping reviews (PRISMA-ScR). PubMed, Scopus, and Web of Science electronic databases were systematically searched to identify studies conducted in the last 20 years, where TTy was measured in TTM context with specific focus on pre-hospital or in-hospital emergency settings. Results The systematic search identified 35 studies, 12 performing TTy measurements during TTM in healthy subjects, 17 in patients with acute cardiovascular events, and 6 in patients with acute neurological diseases. The studies showed that TTy was able to track temperature changes induced by either local or whole-body cooling approaches in both pre-hospital and in-hospital settings. Direct comparisons to other core temperature measurements from other body sites were available in 22 studies, which showed a faster and larger change of TTy upon TTM compared to other core temperature measurements. Direct brain temperature measurements were available only in 3 studies and showed a good correlation between TTy and brain temperature, although TTy displayed a tendency to overestimate cooling effects compared to brain temperature. Conclusions TTy was capable to track temperature changes under a variety of TTM protocols and clinical conditions in both pre-hospital and in-hospital settings. Due to the heterogeneity and paucity of comparative temperature data, future studies are needed to fully elucidate the advantages of TTy in emergency settings and its capability to track brain temperature.

Conclusions: T Ty was capable to track temperature changes under a variety of TTM protocols and clinical conditions in both pre-hospital and in-hospital settings. Due to the heterogeneity and paucity of comparative temperature data, future studies are needed to fully elucidate the advantages of T Ty in emergency settings and its capability to track brain temperature.

Background
Targeted temperature management (TTM), former therapeutic hypothermia, is an intentional reduction of core temperature to a selected and strictly controlled [1] range of values, which is aimed to improve outcomes in various clinical conditions, including cardiac arrest (CA), traumatic brain injury, stroke, and myocardial infarction [2][3][4]. By lowering brain temperature, TTM is thought to mitigate brain damage due to global (i.e., CA) or local (i.e., stroke) ischemia, through various mechanisms, including a decrease of cerebral oxygen and glucose consumption, and a reduction of ATP demand [5,6]. Current guidelines and recent trials support the use of TTM (in the range of 32-36°C [2,3]) in all CA patients who remain in a state of coma after return of spontaneous circulation (ROSC) [2,[7][8][9][10][11]. The benefit of systemic and selective TTM in stroke patients is supported by recent trials and meta-analyses [12,13]. Despite a broad consensus on TTM benefits, the application of pre-hospital TTM, for example in out-of-hospital CA, is still controversial [14,15]. Variable outcomes have been reported [16][17][18], which may be partially due to limitations in pre-hospital cooling procedures and/or accuracy of temperature monitoring.
Since discrepancies between brain and systemic temperatures have been described, direct monitoring of brain temperature would be desirable for optimal TTM [19]. However, brain temperature measurement techniques are invasive and impractical in most cases and settings. Among different sites for core temperature measurement (e.g., ear canal, rectum, bladder, esophagus, and pulmonary artery) [20], the ear canal (or tympanic membrane) has been proposed as a surrogate measurement site during TTM procedures, especially in pre-hospital and emergency settings, thanks to its accessibility, minimal invasiveness, and fast response. The vasculature pattern of the tympanic membrane is shared with the brain and mediates a thermal equilibrium between the two sites [21][22][23], which suggests the potential of tympanic temperature (T Ty ) to reflect brain temperature. In addition, the vasculature in the tympanic region is minimally influenced by the thermoregulatory vasomotor response, which guarantees adequate flow conditions [24]. In pre-hospital settings, T Ty has been shown-albeit with mixed results-to be comparable to invasive temperature measurements at hospital admission [20], providing that insulation from the environment is ensured during measurement [25]. On the other hand, T Ty measurement can only be performed if the ear canal is not obstructed (e.g., by blood, cerumen, snow) [20]. T Ty can be biased in situations during which blood flow is absent or inadequate [23,26], and/or it can be affected by anatomical and vascular changes following major ear surgery and large tympanic membrane perforations [27]. T Ty accuracy under different TTM protocols (e.g., local versus whole body), TTM phases (e.g., induction versus maintenance), and different pathophysiological conditions need to be further clarified.
This scoping review aims to identify and to summarize all the available evidence over the last 20 years about T Ty monitoring in the context of TTM from studies performed either in patients with various acute disorders or in healthy subjects. We describe the level of diffusion of the techniques in various TTM contexts with a focus on pre-hospital and in-hospital emergency settings. We provide indications on the accuracy of tympanic measurements in comparison to other body sites under different TTM phases, cooling protocols, and clinical conditions.

Methods
The scoping review was conducted following the guidelines of the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) extension for scoping reviews (PRISMA-ScR) [28].

Eligibility criteria
The literature search was performed by two authors (AM and MM) to identify studies, conducted in the last 20 years, that used T Ty during TTM approaches. The search strategy is schematized by the inclusion criteria in Table 1, categorized according to the broad Population-Concept-Context (PCC) mnemonic, recommended for scoping reviews [29,30]. The scoping review was focused on pre-hospital and in-hospital emergency settings. We considered both studies testing TTM approaches in healthy subjects and studies where TTM was performed in patients experiencing different emergency conditions. Studies about accidental hypothermia, drug-induced hypothermia, and perioperative and postoperative hypothermia were excluded. The range of TTM temperatures was set to 32-36°C according to TTM definition in [2,3], while studies on normothermia maintenance in patients with fever were not considered. The search was restricted to articles published in English in peer-reviewed journals. No restriction on study design was posed. Abstracts presentations, conference proceedings, and reviews were excluded.
Information sources, search strategy, and study selection A systematic search was performed in PubMed, Scopus, and ISI Web of Science electronic databases to identify primary references from January 2000 to April 2020. The following search strings were used: ("aural" OR "tympanic" OR "epitympanic" OR "ear" OR "ear canal" OR "in-ear" OR "ear-in" OR "earbud" OR "earpiece" OR "earable") AND ("temperature" OR "temperature monitoring" OR "core temperature" OR "core body temperature" OR "body temperature") AND ("hypothermia" OR "hypothermic" OR "therapeutic hypothermia" OR "hypothermic treatment" OR "target temperature management" OR "TTM" OR "body cooling" OR "low temperature" OR "low body temperature"). The database search was followed by a review of the citations from eligible studies. Studies were selected based on title and abstract using the online platform Rayyan [31]. Selected studies were read thoroughly to identify those suitable for inclusion in the scoping review.

Data extraction
Two reviewers (MM and AM) independently extracted the demographic and experimental data from the selected studies. When disagreement occurred, they reviewed the papers together to reach consensus. For each study, the following relevant information was extracted and summarized: the characteristics of the investigated study population; TTM protocols (body cooling modality, target temperature); the experimental and/or clinical settings of application; the available temperature measurements (presence and location of comparative/ reference temperature measurements in addition to the tympanic one); and the main results of the studies in terms of feasibility of the tympanic measurements and comparability of T Ty with core temperature measurements from other body sites.

Selected studies
The database search identified a total of 725 relevant references once duplicates were removed (Fig. 1). A total of 681 references were excluded after reading title and abstract and 44 were retrieved for further evaluation. Of these, 9 studies were excluded, because they did not fulfill the inclusion criteria. Following the selection process, 35 studies were included in the scoping review. Of these studies, 12 measured T Ty during tests of TTM protocols in healthy subjects, 17 during TTM in patients with acute cardiovascular events, and 6 during TTM in patients with acute neurological disorders. The studies are described in the next paragraphs and summarized in Tables 2, 3, and 4.

Tympanic temperature measurement during testing of TTM approaches in healthy subjects
The literature search identified 12 studies that monitored T Ty to test the effects of TTM protocols in healthy subjects. These studies are summarized in Table 2. In 10 studies [32][33][34][35][36][37][38][39][40]42], TTM was achieved using surface cooling garments, such as head and neck or chest and thighs cooling devices. In the remaining two studies [41,43], endovascular cold solutions were used. Comparative core-temperature measurements were present in eight studies [32][33][34][35][36][37][38][39], where rectal/intestinal sites were monitored. Consistently among studies, T Ty showed more pronounced changes than rectal [33][34][35] or intestinal temperature [32]. During chest and thighs surface cooling, the difference between tympanic and rectal temperature was maximal during induction of hypothermia and decreased during its maintenance [36]. Compared to other measurement sites, during head cooling T Ty temperature showed lower variations than skin temperature [35] and more reliable data than sublingual temperature measurements [42]. Overall the studies showed that T Ty was useful in the validation of novel cooling strategies in healthy subjects, where T Ty was able to track temperature variations Table 1 Inclusion criteria for the scoping review summarized according to the Population-Concept-Context (PCC) mnemonic, recommended for scoping reviews [29,30] Population • Healthy adults (testing of target temperature management approaches).
• Patients undergoing target temperature approaches under emergency conditions, including cardiovascular and neurologic emergencies. • Any gender.

Concept
• Tympanic temperature measurement in the context of target temperature management.

Context
• Testing of target temperature management approaches in healthy subjects; target temperature management in patients in pre-hospital and in-hospital emergency settings. • Original peer-reviewed research articles (any study design), published in English in the last 20 years.
Poli et al. [  The T Ty -toe T gradient decreased more during the intervention when the arms and legs were covered (9.1 ± 5.7°C) than uncovered (11.7 ± 4.2°C, p < 0.001).
Data are numbers (N), percentages (%), mean ± standard deviation, median (interquartile range) or [range], as available. *Hypothermia was performed for neuroprotection only in 6 patients CH, cerebral haematoma; IRTT, infrared tympanic thermometer; ICH, intracerebral haemorrage; ICP, intracranial pressure; IS, ischemic stroke; NICU, neurointensive care unit; NIHSS, National Institutes of Health Stroke Scale; NR, not reported; rtPA, recombinant tissue-type plasminogen activator; SAH, subarachnoid hemorrhage; SSS, Scandinavian Stroke Scale (SSS) score; SU, stroke unit; T, temperature; TH, therapeutic hypothermia; TM, temperature measurement; TTM, target temperature management; T Ty , tympanic temperature; TBI, traumatic brain injury; y, years (overall bias of − 0.6°C [50] and − 1.03°C [51]) and in a shorter cooling-time duration [51]. In pre-hospital settings, T Ty was capable of tracking the effects of prompt post-ROSC application of TTM by cold saline infusion [57] or by a hypothermia cap [58] in CA patients, as well as the effects of cold saline and surface pads in patients with acute myocardial infarction [60]. However, tympanic measurements showed to be biased by external factors, such as variations in the environmental temperature [52] or the presence of snow/ice in the ear canal [45]. In the in-hospital setting, tympanic measurements were able to track temperature changes associated with nasal/ pharyngeal or head/neck cooling [44,48,49,53], cold saline infusion [47,57], or a combination of local and whole body cooling [50,51,[53][54][55][56]. In patients with acute myocardial infarction [59,60], the tympanic site was used to complement catheter tip measurements, when the latter were not available.

Tympanic temperature measurement during TTM in acute neurological disorders
Six studies tracked T Ty during TTM in patients with acute neurological disorders, which included ischemic or hemorrhagic stroke, subarachnoid hemorrhage, and cerebral hematoma after traumatic brain injury. The studies are summarized in Table 4. TTM protocols were applied in-hospital in all the retrieved studies. In three studies [61][62][63], patients were intubated and deeply sedated. TTM protocols differed among the studies in terms of cooling devices, target temperature measurement sites, starting temperature, and target temperature (mostly mild hypothermia). The applied cooling techniques included (i) whole body cooling by intravenous injection of cold saline solutions [62,65] and (ii) local body cooling by nasopharyngeal [61,62] or head/neck cooling devices [63], body surface wraps/sponges [66], and/or the "forced air" method [64]. In most of the studies, comparative core measurements were available for the bladder, rectal, and esophageal sites [61][62][63]67], and in one study also for the pulmonary artery [61]. T Ty showed a larger drop compared to other core-temperature measurement sites during pharyngeal cooling [61], while it showed strong correlation with rectal temperature during surface cooling by "forced air" method [64]. In three studies [61][62][63], brain temperature measurements from a probe inserted below the cortical surface were available. T Ty correlated well with brain temperature during whole body cooling induced by intravenous cold saline solution in stroke patients, although it displayed lower values already at baseline with a bias of − 0.4°C [62]. During nasopharyngeal cooling [61,62] or head and neck cooling [63], T Ty overestimated brain cooling, showing a more marked decrease (drop in the first hour of cooling ranging from −1.69 to −2.2°C at the tympanum vs. −0.32 to −1.4°C at the brain), while other core-temperature measurement sites underestimated brain cooling, displaying a lower decrease (temperature drop ranging from − 0.18°C to − 1.1°C) [61,63].
T Ty displayed capability to track temperature changes induced by either local or whole cooling. In addition, when used in combination with skin temperature, it depicted the risk of shivering during surface cooling [66].

Discussion
The main findings of the present scoping review, aimed at assessing the diffusion, feasibility, and accuracy of T Ty monitoring during TTM, are: (i) T Ty was capable to track temperature changes induced by a variety of TTM approaches, including local or whole body cooling, in both pre-hospital and in-hospital settings and under different clinical conditions; (ii) T Ty may have selective advantages for TTM in prehospital settings, where it is often the sole temperature measurement available; and (iii) limited evidence is available about T Ty accuracy in relation to reliable core body and brain sites.

Feasibility and performance of T Ty monitoring in emergency and critical care
The evidence provided by the 35 identified studies generally supported the capability of T Ty to follow temperature changes induced by either local or wholebody cooling strategies. The most common application fields for T Ty were the testing of novel cooling strategies in healthy subjects and the monitoring of TTM in patients with acute cardiovascular events, while applications in patients with acute neurological disorders were sparser. In patients with acute cardiac disease, T Ty monitoring was applied in both pre-hospital and in-hospital emergency settings. In the former setting, tympanic monitoring was mostly used as the sole temperature measurement, which may indicate a selective advantage of T Ty in this condition. Thanks to its reduced invasiveness and easy application, ear probe measurements may allow prompt TTM initiation [68]. In comparison, esophageal temperature probes usually require an intubated patient and rectal temperature measurements may be not easily accessible [68]. However, factors limiting T Ty reliability should be properly considered for an appropriate use of the technique. T Ty measurements may be influenced by alterations in the blood flow to the brain, as demonstrated for instance by tilting maneuvers [23]. Therefore, T Ty measurements in CA patients should be considered reliable only after the patient has regained a stable spontaneous circulation. Moreover, pre-hospital studies showed T Ty measurements to be affected by external factors, such as variations in the environmental temperature or the presence of snow/ice in the ear canal [45]. Consistently, previous studies pointed out the necessity of performing T Ty measurements in a clean and dry ear canal and the importance of properly insulating the tympanic probe, especially when operating in settings exposed to environmental factors (e.g., cold, wind) [25,69].
Comparison of the tympanic site versus other coretemperature measurement sites In emergency and critical care settings, alternative coretemperature measurement sites are available to track temperature; thus, the performance and eventual advantages of T Ty in comparison to other measures need to be evaluated. Tympanic measurements were combined with other measurements in 30 studies [32-39, 41-55, 59-64, 66], of which 22 studies [32-39, 44, 46-54, 61-64] provided a direct comparison with temperatures measured at different core or brain sites. These studies presented heterogeneity in terms of studied population, cooling protocols and devices, tympanic thermometer type (IRTTs or thermistor/thermocouple thermometers), and comparative/reference sites. All these variability factors hindered the calculation of an overall figure of merit for tympanic measurement site. As an additional limitation, rectal temperature was mostly used as comparator among studies. The rectal site has known limitations and a slow response in dynamic conditions [20,[70][71][72], which is mainly attributed to the buffering influence and heat-sink effect of rectal tissue and stool in the rectum [73,74].
The comparison of T ty with the most reliable core sites (i.e., esophageal, jugular, and pulmonary artery sites), available from studies in emergency and critical care settings, showed that T Ty performance depended on the cooling protocols and reference site considered. Compared to esophageal temperature, T Ty showed comparable cooling times during intranasal cooling [49], and low bias and high correlation during cold saline infusion [47]. A larger bias with esophageal temperature was observed during head and neck cooling [54]. Compared to the pulmonary artery, the tympanic site showed the highest bias during cold saline infusion [50,51]. Although the pulmonary artery temperature is usually considered as the gold standard core-temperature to guide clinical mild hypothermia [75,76], previous studies have shown that the temperature of the pulmonary artery blood may reflect less well brain temperature when hypothermia is induced or reversed, during either selective head cooling or rapid intravascular cooling [77,78]. Instead, esophageal temperature responds rapidly to changes in the temperature of the blood perfusing the heart and great vessels [77,79] and it showed a better relationship with brain temperature when inducing hypothermia and at early TTM, in either selective head or whole-body cooling [77]. The better agreement of T Ty with the esophageal than with the pulmonary artery temperature may thus suggest the reliability of T Ty to track brain temperature. The capability of T Ty to reflect brain temperature is supported by the assumption that the tympanic membrane is supplied by vasculature from the same sources that supply the brain (i.e., branches of the basilar and internal carotid arteries, which join anastomoses with branches of the external carotid artery in the region around and within the tympanic membrane [21,22,80]), which guarantees thermal equilibrium between the two sites [23]. However, although T Ty is currently the most commonly used non-invasive method for brain temperature estimation [81], being the sole anatomical structure close to the brain that is accessible without surgery [82,83], concerns remain on its accuracy, mainly due to measurement errors, measuring devices, and/or real temperature differences between the ears [80]. In the specific setting of TTM, the present review revealed a gap of evidence in the literature about the capability of T Ty to reflect brain temperature. We identified only three studies in acute neurological patients [61][62][63], which provided comparative direct brain temperature measurements at a single subcortex site. The studies displayed heterogeneity in terms of patient characteristics and underlying disorders, cooling procedures, and target temperature. The results showed a high correlation between T Ty and brain temperature during whole body cooling. Nonetheless, T Ty generally overestimated brain cooling in either whole body [61,62] or local body cooling [62], with more severe overestimation during head and neck cooling [63]. Of note, other sites for core-temperature measurements generally underestimated cooling effects [61]. Although these results may suggest the potential of T Ty to track brain temperature with similar performance to other more invasive distal measurement sites, the larger response of T Ty may result in an overestimation of cooling effects through different TTM phases and thus in a shorter cooling-time duration [51], with the risk for patients to stay outside the ideal temperature range during TTM induction and steady state.

Future perspectives
Simultaneous measurements at different brain and core temperature sites according to well-defined protocols should be performed during both local and whole-body cooling procedures. The characterization of the spatiotemporal temperature patterns under various TTM approaches by a continuous temperature acquisition through the different TTM phases is desirable. In experimental studies, brain temperature should be monitored at multiple sites, since a single site may not reflect temperature across the brain, especially in the presence of head cooling and marked temperature gradients [84][85][86]. The systematic assessment of bias and correlation between T Ty and brain or other core-temperature measurement sites and the comparison with therapeutic outcome may allow to define sharp recommendation and safe target ranges for T Ty under different TTM applications. T Ty performance may be improved with proper recalibration of target temperature values, as T Ty often led to an underestimation of core temperature even at baseline but showed a moderate to high correlation with esophageal temperature. Finally, clinical, experimental, and industrial research should synergistically concur to develop wearable temperature trackers [80], able to overcome the limitations of current tympanic thermometers [25,66,69,80,87] and to grant fix probe positioning and protection from external environmental conditions [80]. These developments may improve temperature monitoring and allow early TTM extension under logistically challenging critical conditions.

Conclusions
The results of the present scoping review provided evidence about the capability of T Ty to track temperature changes induced by either local or whole-body cooling in both prehospital and in-hospital TTM applications. However, there is a paucity of studies performing a systematic comparison of T Ty performance with reliable core and brain temperature measurement sites, which hinders a thorough evaluation of T Ty advantages in emergency settings and of the capability of T Ty to track brain temperature. Future experimental and clinical studies should bridge this gap of evidence by providing reliable devices and dedicated temperature ranges for safe application of T Ty in TTM and by clarifying the relationship between T Ty and brain temperature. Thanks to its easy use and reduced invasiveness, T Ty may have selective advantage in pre-hospital settings, when practical limitations may hinder temperature acquisition from more invasive sites.