Visual evoked responses as a monitor of intracranial pressure during hyperosmolar blood-brain barrier disruption.

Intracarotid mannitol (HBBBD = hyperosmotic BBB disruption) as a method of transiently increasing solute/drug delivery to brain parenchyma has been associated in animals with a 1-2% increase in brain water and an increase in cisternal ICP. To determine whether these changes are clinically significant, we investigated ICP changes associated with HBBBD in 33 patients with malignant brain tumor utilizing flash VERs in which the N2 latency correlates well with ICP (N2 > 80 ms linearly corresponds to ICP > 20 cm H2O). VERs were obtained prior to, 4 and 24 h after 114 HBBBD/chemotherapy procedures. Additionally, in 10 patients (37 HBBBDs), VER monitoring was performed during the procedure. In 112/150 (75%) HBBBDs, good barrier opening was obtained (radionuclide brain scan). Postoperative mean N2 latencies did not differ significantly from pre-HBBBD values (N2 = 86 +/- 3.3 pre, 90 +/- 3.0/4 h, 87 +/- 3.6/24 h); there was no significant difference in N2 latencies in those patients with good vs poor BBBD. In the 10 patients monitored during HBBBD, peak N2 latency = 94.9 +/- 1.6, however, was significantly above pre- and post-HBBBD values (p < 0.04). We conclude that flash VERs are a useful noninvasive measure of ICP, and that HBBBD is associated with mild transient increase in ICP which is not clinically detrimental.