Prediction of outcome in resuscitated patients by clinical course and early somatosensory-evoked potentials

Abstract

Objective and setting: The aim of this prospective study was to compare the clinical predictive values of the Glasgow Coma Scale (GCS) and somatosensory-evoked potentials (SEP) in the assessment of neurological outcome in pati-ents with anoxic brain damage following cardiopulmonary resuscitation. Subjects and interventions: We repeatedly obtained GCS values and recorded cervicomedullary and cortical SEPs during the clinical course of 20 consecutive patients (median age: 61 years). SEPs were performed according to standard techniques and recordings were done using the interna-tional 10/20 system of electrode placement. In the majority of patients early brainstem auditory-evoked potentials (BAEP; n=14) and cranial computed tomography (CCT; n=12) were initially recorded. Endpoint: The neurological outcome score determined both at the time after completing intensive care treatment and at clinical follow-up (1-5 months) served as a reference. The patients were subdivided into three groups according to neurological outcome (I=excellent, II=moderate to severe disabilities and III=bad, including death and persistent vegetative state). Measurements and main results: Ten of 12 patients in Group III died and two remained in a persistent vegetative state; neither of these regained consciousness. Only two patients in both Groups I and II died, the others had no (n=4; Group I) or moderate (n=2; Group II) neurological deficits. The GCS values which were initially determined (GCS I) could not sufficiently predict outcome. In contrast, the GCS obtained at the time of discharge from ICU (GCS III) showed a distinct difference between Groups I and III. The positive predictive value (PPV) of GCS I for outcome in Groups I and II was 67% but the PPV of GCS III in this setting was 89%. Eight of 12 patients in Group III had bilaterally, and two unilaterally absent cortical N20 waves in their initial SEP recordings, whereas in all eight patients of both Groups I and II the cortical N20 wave was preserved. The PPV of the absent N20 wave for outcome Group III was 100%. In only one patient CCT demonstrated a specific sign of acute brain injury (brain oedema). The initial BAEP recordings did not significantly differ in terms of altered latencies or amplitudes among the different outcome groups and no missing peaks were observed. Conclusions: In patients with hypoxic brain damage following cardiopulmonary resuscitation the bilateral absence of the N20 wave in cortical SEP recordings predicts a fatal clinical outcome in virtually all cases at an early stage. The initial outcome prediction by the GCS is hampered by individual variations and the influence of medications but repeated testing during the clinical course correlates well with the final outcome.