To the Editor.
—Heffner and Sahn, in a recent review article published in the April Archives (1983;143:765-769), suggested controlled, prolonged hyperventilation (HV) as a therapeutic modality in several neurological conditions associated with increased intracranial pressure (ICP). This recommendation is based on the fact that cerebral vasoconstriction is caused by reduced PaCO2. The authors' recommendation should, however, be tempered by published data showing that any reduction in ICP is likely to be brief and that risks of subsequent brain expansion and accentuated intracranial hypertension are considerable.More than ten years ago, Raichle et al1 published results of animal and human experiments in which cerebral blood flow (CBF) was monitored during and after prolonged HV; in this context, blood flow alterations reflect changes in brain volume and ICP. In dogs with PaCO2 reduced to 20 mm Hg, CBF initially fell by almost 60% and then appeared to fall another