Heat stroke management: clinical phenotypes and implications for cooling strategies.

backgroundHeat stroke is a life-threatening condition demanding rapid core temperature reduction. However, optimizing cooling strategies is challenging due to pronounced patient heterogeneity. This study aimed to identify distinct clinical phenotypes of heat stroke to enable personalized cooling management.

methodsThis two-stage study analyzed heat stroke patients from seven tertiary hospitals in China. The prospective cohort was derived from an ongoing, prospectively registered clinical trial (NCT05923931). Latent class analysis was applied to this prospective cohort to derive a phenotype model, which was validated in an independent retrospective cohort and then applied to the total cohort (N = 298) for analysis. Phenotype-specific cooling responses were evaluated by testing interactions between phenotype and cooling modalities and by comparing the effectiveness of practical cooling bundles within each phenotype. Associations between early infusion volume and 48-hour biomarker trajectories were explored using generalized additive mixed models.

resultsThe analysis identified two distinct clinical phenotypes: a non-severe phenotype (47%) and a severe phenotype (53%). The severe phenotype was characterized by more profound organ injury, a slower cooling rate, and higher mortality. Crucially, the efficacy of core cooling modalities was phenotype-specific (P for interaction = 0.038): evaporation cooling was highly effective in the non-severe phenotype but had attenuated effect in the severe phenotype. For the severe phenotype, a bundled strategy of high-volume cold-fluid infusion combined with evaporation cooling was necessary to achieve the fastest cooling (β = 0.03, p < 0.001), whereas either modality alone was insufficient. Exploratory analysis further revealed that early infusion volume was differentially associated with subsequent changes in organ stress biomarkers between the two phenotypes. The direction of all key associations was consistent across cohorts, and the primary phenotype-specific response patterns were reproduced, confirming the robustness of the findings.

conclusionsThis study identifies a clinically actionable phenotype classification for heat stroke. Our findings suggest a phenotype-guided management protocol: for non-severe cases, prioritize evaporation cooling with conservative fluids; for severe cases, a bundled strategy of high-volume cold-fluid infusion plus evaporation is necessary. This hypothesis-generating framework offers a potential pathway toward early, tailored intervention at the point of care, pending confirmation in future studies.