In patients who experience sudden death from spontaneous subarachnoid hemorrhage, more than 90% present with acute pulmonary edema. The underlying pathogenesis of this complication is poorly understood. In addition, the specific role of the extravasated blood products and the associated elevation in intracranial pressure leading to the systemic and pulmonary effects during subarachnoid hemorrhage are not well established. The authors tested a new model of acute and severe subarachnoid hemorrhage comparing fresh whole autologous blood (n = 20) with 5% albumin (n = 19) injected at two different rates (35 seconds versus 24 minutes) into the cisterna magna of anesthetized, mechanically ventilated rats. Cerebral and systemic hemodynamics and the corresponding pulmonary function were evaluated. The type of fluid injected had no influence on survival or hemodynamic and respiratory parameters. Rapid infusion of either blood or albumin (n = 14) produced an acute and transient rise in intracranial pressure (37.9 +/- 3.5 mm Hg) associated with systemic hypertension and increased cerebral perfusion pressure that was sustained in survivors but not in nonsurvivors. Slow infusion (n = 23) produced a more progressive increase in intracranial pressure to 31.2 +/- 7.1 mm Hg with a parallel and sustained increase of systemic blood pressure and preserved cerebral perfusion pressure in survivors, but produced a pattern of more severe hypertension followed by hypotension in nonsurvivors. Sixty-four percent of animals (rapid infusion) and 48% of animals (slow infusion) survived the challenge and presented no pulmonary alterations. In contrast, nonsurviving rats developed reduced lung compliance and gas exchange, an increased alveolar-arterial protein concentration ratio (0.36 +/- 0.02 versus 0.17 +/- 0.03 in survivors; P <.0001), and increased lung weight (5.7 +/- 0.3 g versus 2.0 +/- 0.1 g; P <.0001), demonstrating a fulminant increased permeability pulmonary edema, leading to death within one hour. These results indicate that the chosen rapid- and slow-injection rates resulted in a similar death rate of 50%. Mortality was similar for blood and albumin administration, pulmonary edema occurred in nonsurvivors in both the rapid- and slow-injection groups, and pulmonary edema is associated with more severe hypertension in the slow-injection group. Furthermore, these results suggest that the development of neurogenic pulmonary edema that is characterized by an acutely increased capillary permeability to proteins is independent of the degree of intracranial pressure increase or the type of fluid administrated.