Serum phenytoin levels of two groups of brain-injured patients on gastrostomy tube (GT) feeding were retrospectively examined. In the first group (group C = clamped, n = 13), GT was clamped for one hour after the phenytoin dose (2-3 doses per day), and in the second group (group NC = not clamped, n = 9) GT feedings were continued without interruption. Mean +/- SD of the serum phenytoin levels was 14.4 +/- 4.7 micrograms/ml for group C, and 9.2 +/- 6.8 micrograms/ml for group NC. When serum phenytoin levels were adjusted for decreased serum albumin levels, the results were 19.8 +/- 6.4 micrograms/ml for group C and 11.7 +/- 7.9 micrograms/ml for group NC. Statistical analysis showed a significant difference between the means in both the phenytoin and adjusted phenytoin levels between the two groups (p < 0.05). No significant difference was found in phenytoin dosage per kg of body weight, protein intakes, calcium, magnesium and sodium per 24 hours between the means for the two groups. Results of this study suggest that clamping the GT for one hour after the phenytoin dose results in higher serum levels of the drug.