Blood sludging and clotting during stasis in cardiac arrest might hamper reperfusion with standard external CPR (SECPR}, particularly after long cardiac arrest (CA) times without CPR. Rabbits were lightly anesthetized for preparation. After tracheal intubation and insertion of monitoring sensors, a parietal burr hole was made, and a platinum H2-sensitive electrode was inserted 2 mm into the cerebral cortex by micromanipulator. Cortical cerebral blood flow (CBF) by cortical H 2 desaturation was recorded, and semi-log plots of clearance curves were obtained. CO2 sensitivity under normotension was established {n5). Motion artifacts were controlled. In 21 CA experiments, under IPPV, 2 or 3 control CBF values were obtained. Anesthesia was discontinued. After saturation for 15 min with 10% H2/50% N20/40% 02, the heart was stopped with KC1. CA was permitted to persist for 1, 3, 5, 7, or 9 rain in randomized sequence. At the end of CA, SECPR was started with IPPV/FiO2 = 1, and sternal compressions with 2 fingers (80-100/min}, trying to optimize arterial pressure. H2 desaturation curves were obtained during SECPR. Prearrest cortical CBF remained consistent with _+ 10% (23-40 mL/100g/ min}. After arrest, during SECPK, H 2 washout curves showed cortical CBF values above viability limit (above 20% of normal} only after 1 min arrest time (AT), with CBF 12% to 31% of control. When AT was increased to 3 min, CBF during SECPR was only 9% of control; after 5 min AT, 6%; and after 7 and 9 rain AT, CBF was 0. MAP during SECPR was 42 + 17 after 1 rain AT, and 15 + 2 after 7 min AT. After 15 min of SECPR, open-chest CPR (OC-CPR) raised MAP but not CBF. The longer the CA time without CPR, the worse the CBF generated by subsequent SECPR. SECPR should be started as rapidly as possible, and must be improved for use after prolonged (unwitnessed) cardiac arrest.