Detection of subclinical brain dysfunction by sensory evoked potentials in patients with severe diabetic ketoacidosis

Objective: Subclinical brain dysfunction is a potentially deleterious complication of diabetic ketoacidosis but is rarely recognized. Thus, we investigated the diagnostic value of sensory evoked potentials for detecting subclinical brain dysfunction in patients with diabetic ketoacidosis. Design: Prospective trial. Setting: Intensive care unit in a university hospital. Patients: 5 neurologically asymptomatic patients (Glasgow Coma Scale score 15, slight drowsiness; aged 20 to 66 years) with an established diagnosis of severe diabetic ketoacidosis were studied. Measurements and results: Short- and long-latency sensory evoked potentials were recorded within 2 h of initiation of therapy for ketoacidosis and 7 days after normalization of ketoacidosis, respectively. Two hours after starting therapy, sensory evoked potential peak latencies were prolonged in all five patients compared to age-matched healthy subjects [cervical N13 to cortical N20 interpeak latency of short-latency evoked potentials (mean) 5.8 vs 5.3 ms, p < 0.05; N35 peak latency 40 vs 34 ms, p < 0.05; N70 peak latency of long-latency evoked potentials 102 vs 76 ms, p < 0.01]. In all five patients, cervical N13 to cortical N20 interpeak latency and N35 and N70 peak latency reverted to normal 7 days after recovery from diabetic ketoacidosis. Conclusions: Our study indicates that the recording of sensory evoked potentials is a sensitive method of detecting subclinical brain dysfunction in patients with severe diabetic ketoacidosis. Since sensory evoked potentials were significantly prolonged in all five patients, this strongly suggests that subclinical brain dysfunction occurs more frequently than is generally recognized.