We retrospectively analyzed our 2-year experience with venous access for 363 therapeutic plasma exchanges in 46 patients with neurological disease, including acute Guillain-Barré syndrome (N = 20), myasthenia gravis (N = 17), and chronic inflammatory demyelinating polyneuropathy (N = 9). Twenty-three patients (50%) completed the planned course of therapy using only peripheral venous access, and 28 central venous catheters were placed in the remaining 23 patients. Patients utilizing central venous access did not undergo a greater number of procedures, but they were more likely to have acute Guillain-Barré syndrome (P < 0.02) or to be hospitalized in a medical intensive care unit (P < 0.01). Three types of central catheters were used, and although our experience was predominantly with 1 type, differences were noted. Only 3% of procedures (3 of 96) done with a Quinton-Mahurkar catheter were associated with a catheter failure, compared to 27% (4 of 15, P < 0.01) with a Hickman catheter and 67% (2 of 3) with a triple-lumen catheter. Life-threatening complications occurred with 3 of 28 (11%) central catheters. To optimize the success of therapeutic plasma exchange using central access, it is critical that hemapheresis personnel advise each patient's primary physician regarding the type of central venous catheter required. Currently, we recommend use of a Quinton-Mahurkar or other dual-lumen hemodialysis catheter.