IMMEDIATE ELECTRONYSTAGMOGRAPHY IN THE DIAGNOSIS OF THE DIZZY PATIENT

Study objectives: To determine whether the results of electronystagmog raphy (ENG) testing improve an emergency physician's diagnosis of dizz iness. Design: Prospective, one-year. Setting: University and three co mmunity hospital emergency departments. Type of participants: Ninety-t hree consecutive patients presenting with dizziness. Interventions: ED impression was recorded after complete ED evaluation. An ENG was perf ormed within one hour by an audiologist, who gave a reading of ''centr al,'' ''peripheral,'' or ''normal.'' The result was given to the emerg ency physician, who was invited to revise his or her impression (the ' 'ED impression after ENG result''). Final diagnosis was based on the E D impression and by contact with the patient's physician(s) as well as the patient by telephone after one and four weeks. Accuracy of ENG wa s assessed by comparing ENG reading with the final diagnosis using the chi2 test. In addition, the contribution of ENG to ED diagnosis was a ssessed by comparing the accuracy of the ED impression after ENG readi ng with the ED impression alone using McNemar's test (hit versus no-hi t). Measurements and main results: Both ED impression and the ENG sign ificantly correlated with the final diagnostic category (chi2 = 104.9, P < .001; chi2 = 70.79, P < .001, respectively). ENG correctly diagno sed nine of 11 patients with central dizziness. Of 23 patients with un determined cause after ED evaluation, ENG correctly identified seven p atients with peripheral dizziness and three with central dizziness. ED impression after ENG reading was more accurate than ED impression alo ne (chi2 = 6.13, P < .05). Conclusion: Emergency physicians correctly categorized most dizzy patients, but audiologist performance and inter pretation of an ENG significantly improved this categorization. ENG ma y have the potential to identify clinically unsuspected central dizzin ess and to categorize dizziness of ''unknown'' cause. Further study is needed to determine whether ENG could be performed by modifying certa in types of heart monitors available in the ED.