Background and Objectives The purpose of this study was to test the hypothesis that loss of cerebrospinal fluid (CSF) causes cerebral vasodilatation, which is reversible with peridural injection of autologous blood.

Methods Ten pigs were anesthetized with an infusion of propofol and remifentanil and mechanically ventilated to normocapnia with air and oxygen (60%). Cisternal puncture was performed and increments of 1 mL of cerebrospinal fluid were aspirated. After each milliliter was removed, hemodynamic and respiratory variables and cerebral blood flow (CBF) were measured, the latter with a transdural laser Doppler flowmeter (BLF 21; Transonic Systems Inc, Ithaca, NY) through a cranial burr hole. After 9 mL of CSF had been removed, 10 mL autologous blood was injected into the lumbar epidural space, and the CBF and other variables were measured immediately and 5 minutes thereafter. Ten milliliters of autologous blood was then injected subdurally and the measurements repeated. Data were analyzed for significant differences from the baseline and previous values by repeated analysis of variance.

Results CBF increased from 44.7 ± 7.97 tissue perfusion units (TPU) (mean ± SEM) at baseline to 75.3 ± 13.53 TPU after removal of the first 7 mL of CSF (P < .0001). Following injection of 10 mL of blood into the epidural space, CBF immediately decreased to 47.6 ± 9.18 TPU. After subdural injection of blood, the CBF decreased further to 20 ± 3.77 TPU.

Conclusions The increase in CBF probably represents cerebral vasodilatation. The immediate return of CBF to baseline values after epidural injections of blood, and to lower values after subdural injections of blood, was probably due to vasoconstriction. The data suggest that postdural puncture headache, and its successful treatment with epidural blood patch, can probably be ascribed to cerebrovascular dynamics.