Prehospital lactate-glucose interaction in acute life-threatening illnesses: metabolic response and short-term mortality.

Background and importance: Lactate is an already recognized biomarker for short-term mortality. However, how glycemia and diabetes affect the predictive ability of lactate needs to be revealed.

Objective: To determine how hypoglycemia, normoglycemia, and hyperglycemia modify the predictive ability of lactate for short-term mortality (3 days). The secondary objective was to evaluate the predictive ability of lactate in diabetic patients.

Design, settings and participants: Prospective, observational study performed between 26 October 2018 and 31 December 2022. Multicenter, EMS-delivery, ambulance-based study, considering 38 basic life support units and 5 advanced life support units referring to four tertiary care hospitals (Spain). Eligible patients were adults recruited from among all phone requests for emergency assistance who were later evacuated to emergency departments.

Outcomes measure and analysis: The primary outcome was in-hospital mortality from any cause within the third day following EMS attendance. The main predictors considered were lactate, blood glucose levels and previous diabetes.

Main results: A total of 6341 participants fulfilled the inclusion criteria. 68 years (IQR: 51-80); 41.4% were female. The 3-day in-hospital mortality rate was 3.5%. The predictive capacity of lactate for 3-day mortality was only significantly different between normo-glycemia and hyperglycemia. The best predictive result was for normo-glycemia - AUC = 0.897 (95% CI: 0.881-0.913) - then hyperglycemia - AUC = 0.819 (95% CI: 0.770-0.868) and finally, hypoglycemia - AUC = 0.703 (95% CI: 0.422-0.983). The stratification according to diabetes presented no statistically significant difference, and the predictive results were AUC = 0.924 (95% CI: 0.892-0.956), AUC = 0.906 (95% CI: 0.884-0.928), and AUC = 0.872 (95% CI: 0.817-0.927) for nondiabetes, uncomplicated cases, and end-organ damage diabetes, respectively.

Conclusion: Our results demonstrated that glycemia, but not diabetes, alters the predictive ability of lactate. Therefore, hyperglycemia should be considered when interpreting lactate, since this could improve screening to detect cryptic shock conditions.