The protocol of the Heatstroke Study was approved by the Teikyo University Ethical Review Board for Medical and Health Research Involving Human Subjects (No. 17-021-5). Informed consent was provided in a manner specified by the ethics committee of each institution for this study between 2020 and 2021. From 2017 to 2019, the requirement for informed consent was waived by the Ethics Review Board of Teikyo University Hospital because of the observational nature of this study. The study was performed in accordance with the ethical standards in the Declaration of Helsinki and ethical guidelines for medical research in Japan.

This study was a secondary analysis of the Heatstroke Study database. The Heatstroke Study is a nationwide prospective observational multicenter study conducted by the JAAM Heatstroke and Hypothermia Surveillance Committee. It is conducted every summer (between July and September) and includes emergency departments (EDs) in both tertiary care hospitals and local secondary hospitals (115 hospitals participated in 2017 and 2018, 109 hospitals in 2019, and 165 hospitals in 2020 and 2021) [11–13]. The Heatstroke Study registered patients clinically diagnosed with heat-related illness by physicians based on symptoms (pyrexia, dehydration, dizziness, myalgia, headache, nausea, disturbance of consciousness, and convulsions) and a history of heat exposure according to the JAAM Heatstroke Guidelines 2015 [15,16]. The other details of the Heatstroke Study have been reported previously [11–13]. We merged data of common variables from 2017 to 2021 for analysis.

This study included pediatric patients (<18 years) with heat-related illness who were transferred to EDs and registered in the Heatstroke Study in the summer season (from July to September) between 2017 and 2021. Patients whose ages were unknown were excluded.

We extracted clinical data as follows: age, sex, pre-event disability status (defined as modified Rankin scale [mRS] 2 or higher), past medical history, date and time of ED registration, occurrence situation (outdoor/indoor, sports), ambulance usage, occurrence of seizure, vital signs at ED (Glasgow Coma Scale [GCS], systolic blood pressure, heart rate, body temperature), and measurement site of body temperature. Regarding body temperature, if core body temperature was available, it was used; otherwise, the surface body temperature was used for analysis. We identified cases with the JAAM-Heatstroke scale grade 3, which met any one of the following criteria: acute kidney injury (AKI), liver injury, coagulopathy, or impaired consciousness (Fig. 1). Any patient meeting grade 3 criteria is recommended for admission to a hospital in the JAAM guidelines [16–18]. Referring to previous research, AKI was categorized based on the KDIGO (Kidney Disease: Improving Global Outcomes) guideline for AKI stage 1 (increase in serum creatinine by 0.3 mg/dL or to 1.5 times the 50th percentile of serum creatinine adjusting for age) [17–20]. Liver injury was identified by serum total bilirubin value 1.2 mg/dL or greater, coagulopathy was a JAAM disseminated intravascular coagulation (DIC) score 4 or higher, and impaired consciousness was defined as GCS 14 or less or observed seizure [17–19]. The JAAM-DIC criteria are used to identify DIC, particularly in critically ill patients, including platelet count, fibrin/fibrinogen degradation products, prothrombin time, and systemic inflammatory response syndrome scores (Supplementary Table 1) [21]. We also obtained the following in-hospital information: cooling methods (intravenous [IV] cold fluid, evaporative cooling, gastric or bladder lavage, gel pad or blanket, continuous renal replacement therapy), disposition (admission to general ward, admission to the intensive care unit [ICU], discharge from the ED, and died in the ED), mechanical ventilation, length of stay, in-hospital mortality, and functional outcome (mRS) on discharge. We also obtained available laboratory data at ED admission (hemoglobin, hematocrit, white blood cell count, platelet count, blood urea nitrogen [BUN], serum creatinine, serum lactate, total bilirubin, aspartate aminotransferase, alanine aminotransferase, sodium, potassium, chloride, glucose, and C-reactive protein [CRP]). The detailed explanation of the variables is shown in Supplementary Table 2.

The outcomes were in-hospital mortality and unfavorable functional outcome (moderate disability to death), defined as mRS of 3 or higher at hospital discharge.

We described the clinical characteristics of all patients and of age categories. Missing data were indicated as “missing” or “unknown,” and imputation was not applied because of the nature of the descriptive study. To investigate the clinical features of minor and severe cases, we showed the trends of vital signs and laboratory data of patients who were discharged from the ED, admitted to the general ward, or needed ICU admission or died in the ED [22]. Also, we analyzed the characteristics after excluding the cases that died in the ED. Variables were indicated using median and interquartile range (IQR) for continuous variables, number and proportion for categorical variables, or boxplots. A sample size estimation was not performed since this study did not intend to test a specific hypothesis. All statistical analyses were performed with R ver. 4.1.2 (R Foundation for Statistical Computing).

Of 3,154 patients registered in the Heatstroke Study database, 146 pediatric patients were analyzed (Fig. 2), of which 87 (59.6%) were male. The median age was 15 years (IQR, 13–16 years), and the most common age group was high school (age 15–17 years; 60 patients, 41.1%), followed by middle school (age 12–14 years; 56 patients, 38.4%). The smallest age group was infants and preschoolers (age 0–5 years; 7 patients, 4.8%). Patients with any medical history comprised 17.8% of the study population (26 of 146 patients). Patient characteristics are shown in Table 1 (missing data are described in the Supplementary Table 3).

Heat-related illness occurred most frequently in August (69 patients, 47.3%), followed by July (62 patients, 42.5%). The afternoon hours (12:00–18:00) saw the highest incidence (76 patients, 52.1%), followed by evenings (18:00–24:00; 32 patients, 21.9%) (Fig. 3). Most cases occurred in outdoor settings (102 patients, 69.9%) and were related to sports activities (117 patients, 80.1%). Most cases (117 patients, 80.1%) were primarily in students aged 6–17 years who were engaged in sports activities.

Regarding presentation to the ED, 90 patients (61.6%) arrived at the ED by ambulance (Table 1). Median body temperature was 37.2 ℃ (IQR, 36.7–38.0 °C), predominantly determined via axillary measurement (106 patients, 72.6%). Patients with JAAM-Heatstroke scale grade 3 comprised 91 patients (62.3%), including cases with AKI (52 patients, 35.6%), liver injury (37 patients, 25.3%), and GCS score of ≤14 or seizure (57 patients, 39.0%). Most school-aged patients were grade 3 on the JAAM-Heatstroke scale, with 39 of 56 patients (69.6%) in middle school and 42 of 60 patients (70.0%) in high school. Only five patients (3.4%) had a body temperature of 40 ℃ or higher, which is one of the classical criteria for heatstroke suggested by Bouchama and Knochel [23] Only one case met the JAAM-DIC criteria for coagulopathy.

Regarding cooling methods, IV cold fluid was most frequently performed (48 patients, 32.9%), while invasive cooling procedures were rarely performed (Table 2). Following evaluation and treatment in the ED, the majority of cases (110 patients, 75.3%) was admitted to general wards, and 23 patients (15.8%) were admitted to the ICU. Two patients (1.4%) who were in cardiac arrest upon arrival at the ED died there; no additional in-hospital mortality was observed. The median length of stay was 2 days (IQR, 2–3 days), and no patients with unfavorable functional outcomes defined as mRS 3 or higher at discharge were observed, other than the ED deaths.

Among patients admitted to the ICU or mortality cases at the ED, almost all (22 of 25 patients, 88.0%) met the criteria for JAAM-Heatstroke scale grade 3, including patients with altered levels of consciousness or seizures (16 patients, 64.0%), AKI (15 patients, 60.0%), and liver injury (9 patients, 36.0%) (Table 3). The median body temperature of patients admitted to the ICU was 38.0 °C (IQR, 37.1–39.4 °C), while that of patients admitted to general wards or discharged from the ED was 37.2 °C (IQR, 36.7–37.7 °) or 36.9 °C (IQR, 36.7–37.4 °C), respectively. Laboratory results for patients admitted to the hospital indicated dehydration with high hemoglobin, hematocrit, BUN, creatinine, serum sodium, chloride, and lactate (Fig. 4, Supplementary Fig. 1). Additionally, these patients had high values of creatine kinase indicating mild rhabdomyolysis. The characteristics excluding patients who died in the ED are shown in Supplementary Table 4 and Supplementary Fig. 2.

This observational study showed that most pediatric cases of heat-related illness who were transferred to EDs in Japan involved middle or high school students and were related to outdoor sports activities in August. Most hospitalized patients met the criteria for JAAM-Heatstroke scale grade 3, presenting with altered consciousness, AKI, and dehydration. Two cases of out-of-hospital cardiac arrest resulted in death at the ED; there were no other fatalities or cases of functional disability.

First, this study showed that most pediatric cases of heat-related illness admitted to EDs involved middle or high school students engaged in sports activities. This finding suggests that demographic as an ideal target population for prevention of heat-related illnesses among the pediatric population in Japan. Previously, a database of school accidents managed by the Japan Sports Council reported 26 deaths due to heat-related illnesses in school life from 2006 to 2021, with 23 (88.5%) involving middle or high school students and related to sports activities [10]. Additionally, in 2012, 4,495 of 4,971 cases (90.9%) with heat-related illnesses in schools involved middle or high school students, and approximately 84% were related to sports activities [10]. These trends can be explained by the common participation of middle or high school students in sports [24]. In addition, the results of the present study support the conclusion that middle and high school students participating in sports activities should be a target population for prevention of heat-related illnesses in Japan. To implement effective preventive measures, it is essential to raise public awareness and share these findings with key stakeholders, such as government sectors responsible for health, education, and sports. For example, regulations could be proposed to restrict outdoor sports during the afternoon or evening in July and August in the absence of appropriate precautions. Moreover, it is crucial to highlight these findings on a global scale, particularly in East Asian countries with similar climates, to promote broader preventive efforts.

This study notably revealed that three of seven infant or preschool patients experienced severe outcomes (two fatalities and one admission to the ICU). Although detailed information about the events or situations leading to these outcomes was unavailable, we speculate that infants and preschoolers may not communicate effectively their distress to parents or guardians, potentially delaying recognition and worsening their conditions. While this hypothesis remains speculative, further investigation into the circumstances leading to these incidents is essential for developing targeted preventive measures.

Second, we suggest that the description of clinical features among pediatric patients in this study can help us understand heat-related pathogenesis and treatment strategies. This study indicated that most patients admitted to a hospital met the criteria for JAAM-Heatstroke scale grade 3 (88% among patients admitted to the ICU and 60% among patients admitted to a ward), while a smaller proportion of the patients discharged from the ED (3 of 11, 27.3%) met the criteria. The Japanese clinical guideline for heat-related illness from 2015 recommends the JAAM-Heatstroke scale grade 3 as the indication for hospital admission, and this approach has been used in some public awareness initiatives [16,25]. Accordingly, our study results can be interpreted as showing good concordance between the criteria and clinical practice among pediatric patients. However, due to the limited number of patients discharged from the ED, caution may be needed in interpreting the results.

In contrast, cases with a body temperature of ≥40 ℃ at the ED, which is part of the classic criteria for the most severe subset of heat-related illness (heatstroke), were rare (3.6%). This prevalence is markedly lower than that of adult cases registered in the Heatstroke Study, where 25% to 35% of cases had a body temperature of ≥40 °C [13]. This rarity is potentially because most cases were suspected in the early phase and received cooling measures and first-aid before ED arrival. Accordingly, even in severe cases, the body temperature may have decreased below 40 ℃ by the time they reached the ED. Additionally, this could be due to children’s smaller body sizes, making them more responsive to cooling methods compared to adults.

Among patients admitted to the hospital, high hemoglobin, hematocrit, lactate, BUN, serum creatinine, sodium, and chloride were observed in blood test results at the ED, which suggests that they suffered from dehydration. Mild rhabdomyolysis was also observed. These findings support hydration and IV fluid as essential treatments for pediatric heat-related illness. Interestingly, the CRP values were within the normal range in most cases in this study. Considering that CRP has been reported as a biomarker for sepsis among pediatric patients [26], this result suggests that CRP can differentiate between heat-related illness and sepsis among pediatric patients with fever. Fortunately, in this study, unfavorable outcomes such as mortality or discharge with functional disability were limited, and these results can be helpful for clinicians when discussing prognosis with families.

This study had several limitations. First, it was not based on population-wide data and only included patients who were transferred to or visited the participating hospitals, which were predominantly tertiary care centers and not children’s hospitals. Thus, the study did not account for those who visited primary care clinics, those who did not visit any medical facility, or patients confirmed dead at the scene. Specifically, pediatric mortality cases involving heatstroke in children locked in cars [27] might have been missed, possibly introducing selection bias. Second, while the Heatstroke Study was conducted during the summer season (July to September), cases of heat-related illness occur in other months, such as May or June. According to occurrence reports from the Japan Fire and Disaster Management Agency, 5% to 10% of total cases of heat-related illness during the period from May to September were observed in May or June. However, the present study did not investigate the incidence of cases during this period. Third, there might have been measurement biases in some clinical data. For example, although rectal monitoring is preferred for measuring core body temperature due to its accuracy, its invasiveness limited its use. Axillary measurements, which are more common, may not accurately reflect core body temperature. Furthermore, since the Heatstroke Study does not exclusively focus on pediatric patients, some scales (e.g., GCS, JAAM-Heatstroke scale) and measurements (e.g., JAAM-DIC score or criteria for liver injury) might not accurately reflect the clinical conditions of pediatric patients. Finally, as this was a descriptive study, it could not determine any associations or causal relationships between factors or the incidence of heat-related illnesses or clinical outcomes. Further research is necessary to identify factors associated with clinical outcomes and to develop effective treatment strategies.

This descriptive study indicated that most pediatric cases with heat-related illness transferred to the EDs were of middle or high school age, occurred in the afternoon during the month of August, and were related to outdoor sports activity. Patients admitted to hospitals suffered from altered consciousness, dehydration, and AKI. These results could help us to understand the clinical features and discuss preventive measures.