BACKGROUND: When performing spinal anesthesia, glucose is frequently added to control the extent of the anesthesia by increasing the specific gravity of the local anesthetic solution. It is not clearly known whether the added glucose directly affects the nerve blockade or not. The purpose of this study is to examine the effect of glucose solution on isolated nerve fibers in vitro. METHODS: Compound action potentials (CAPs) of A-fiber range were recorded from isolated nerves of adult Sprague-Dawley rats (300 400 gm). Tonic (0.5 Hz) and phasic (30 Hz) supramaximal stimuli were repeatedly applied to one end of the nerves and the recordings were made on the other end. Nerves were perfused with modified Krebs solution for 45 minutes initially to get baseline data and then perfused with test solutions containing different concentrations of glucose (2.5%, 5%, 7.5%) for 30 minutes. The same experiments were repeated with mannitol at the same osmolality as the glucose. RESULTS: Glucose produced a decrease in the amplitude of CAPs in a dose-dependent manner (79.2 +/- 3.4, 50.3 +/- 3.7, 34.6 +/- 4.0 for 2.5%, 5%, 7.5% glucose solutions, respectively). At each of the same concentration levels, the degree of nerve conduction blockade did not have any significant difference within the glucose groups and mannitol groups. CONCLUSIONS: Glucose, in clinically employed concentration range, directly depressed peripheral nerve conduction in vitro, probably via osmotic effect.