Our infant rat model of traumatic subarchnoid hemorrhage combines violent shaking and hypoxia to produce subdural hemorrhaging and progressive cortical degeneration similar to that seen in victims of the shaken baby syndrome. Anesthetized, 6-day-old male rats were subjected to one episode of shaking under hypoxic conditions. Brain histologies revealed moderate-to-severe cortical hemorrhaging at 48 h postinjury and progressive cortical degeneration, as indicated by a 15.3% and 20.2% reduction in cortical wet weight, at 7 and 14 days postinjury, respectively. The purpose of the present study was to assess the effects of two antioxidant lipid peroxidation inhibitors (tirilazad mesylate and PNU-101033E), and the glutamate release inhibitor (riluzole), upon the brain pathology seen in this model. A significant, 54.3-75.3%, reduction in cortical hemorrhaging was observed in rats that were treated with a total of three doses of tirilazad (10 mg/kg, i.p.): 10 min before or 5-30 min after injury, and again at 2 and 24 h postinjury (p < 0.01 vs. vehicle). However, treatment with tirilazad or the more potent, brain-penetrating pyrrolopyrimidine, PNU-101033E (10 min before plus 2, 24, 48, and 72 h after), did not attenuate the progressive cortical degeneration typically seen at 14 days postinjury. These results suggest that free radicals play an important role in the pathophysiology of secondary brain hemorrhaging due to shaking + hypoxia, but may not be critical in the mediation of the subsequent neurodegeneration. Rather, glutamate neurotoxicity may be a key factor here. This is suggested by our observation that the glutamate release inhibitor, riluzole, significantly reduced cortical degeneration when it was administered up to 1 h postinjury in the present model. Specifically, the cortical wet weights of rats treated with 8 mg/kg riluzole (i.p.) 10 min before or 1 h after shaking + hypoxia (and again at 24 h postinjury) were 95.3% and 97.4% of noninjured controls, respectively, at 14 days postinjury (p < 0.02 vs. vehicle). Riluzole treatment beyond 1 h (e.g., 2 or 4 h postinjury) did not reduce the neurodegeneration. Lastly, we attempted to demonstrate that the therapeutic window for riluzole-induced attenuation of cortical degeneration could be extended beyond 1 h through the use of combination therapy. In this experiment, rat pups were treated with 10 mg/kg tirilazad (i.p.) at 30 min postinjury followed by 8 mg/kg riluzole (i.p.) at 4 and 24 h postinjury. At 14 days postinjury, the cortical wet weights of these rats were 94.5% of noninjured controls, thus demonstrating significant neuroprotection (p < 0.05 vs. vehicle) and a widening of the therapeutic window from 1 to 4 h in length. These results suggest that early attenuation of free radical-induced lipid peroxidation may slow down the biochemical cascade of events related to glutamate-induced excitotoxicity and, in doing so, prolong the time during which a glutamate release inhibitor, such as riluzole, is effective.