BackgroundEarly diagnosis of hospitalized elderly patients with infectious stress hyperglycaemia (ISH) is clinically important, especially under the global coronavirus disease 2019 (COVID‐19) pandemic, as without timely prevention and effective treatment, it is likely to deteriorate into septic shock, thus worsening patient survival and complications. Moreover, cumulative studies have showed that patients with COVID‐19 are reported to have a greater prevalence of hyperglycaemia. However, the underlying mechanism remained unknown.Aim and methodSystematic screening of specific biomarkers of serum exosome‐derived microRNAs (sE‐miRNAs) from ISH patient has not yet been reported. In this study, sE‐miRNAs were derived from 10 elderly patients with ISH and 5 control patients with disease‐match without hyperglycaemia (non‐ISH). RNA sequencing identified that a total number of 49 sE‐miRNAs with differential expression between ISH and control group. Of which, top 22 miRNAs ranked by sensitivity × specificity were chosen for further research. Moreover, 7 out of 22 miRNAs that related to glucose metabolism or immune disorder were picked up for further validation in an independent cohort consisting of 52 participants (31 ISH and 21 non‐ISH).ResultA validation analysis revealed that three miRNAs (hsa‐miR‐21‐5p, hsa‐miR‐335‐5p and hsa‐miR‐28‐3p) were statistically up‐regulated in exosomes from ISH patients. In the validation cohort and discovery cohort, the AUC of three individual miRNAs ranged from 0.73 to 0.88. A logistic model combining three miRNAs achieved an AUC of 0.96. Besides, sE‐miRNAs‐based signatures effectively characterized patients' poor clinical outcome. Survival curve analysis showed that hsa‐miR‐335‐5p, hsa‐miR‐28‐3p but not hsa‐miR‐21‐5p, were significantly closely related to mortality, and the combination of these three miRNAs could also predict patients outcome (p < .05).ConclusionThis study depicted the circulating exosomal miRNAs change in ISH patient, which could be used as a promising biomarker to detect ISH at an early stage and predict patients clinical outcome.