BackgroundPatients with shock treated by emergency medical services (EMS) have high morbidity and mortality. Knowledge of prehospital factors predicting outcomes in patients with shock remains limited. We aimed to describe the prehospital predictors of mortality in patients with non-traumatic shock transported to hospital by EMS.MethodThis is a retrospective cohort study of consecutive ambulance attendances for non-traumatic shock in Victoria, Australia (January 2015–June 2019) linked with government-held administrative data (emergency, admissions and mortality records). Predictors of 30-day mortality were assessed using Cox proportional regressions. The primary outcome was 30-day all-cause mortality.ResultsOverall, 21 334 patients with non-traumatic shock (median age 69 years, 54.8% female) were successfully linked with state administrative records. Among this cohort, 9 149 (43%) patients died within 30-days. Compared with survivors, non-survivors had a longer median on-scene time: 60 (35–98) versus 30 (19–50), p <0.001. Non-survivors were more likely to be older (median age in years: 74 (61–84) vs 65 (47–78), p<0.001), had prehospital cardiac arrest requiring cardiopulmonary resuscitation (adjusted HR (aHR)=6.26, 95% CI 5.87, 6.69) and had prehospital intubation (aHR=1.07, CI 1.00, 1.14). Reduced 30-day mortality was associated with administration of epinephrine (aHR=0.66, CI 0.62, 0.71) and systolic blood pressures above 80 mm Hg in the prehospital setting.ConclusionThe 30-day mortality from non-traumatic shock is high at 43%. Independent predictors of mortality included age, prehospital cardiac arrest and endotracheal intubation. Interventions that target reversible causes of short-term mortality in patients with non-traumatic shock are a high priority.