To ascertain the relationship among 2,5-hexanedione (2,5-HD) concentrations in nerve, serum and urine, rats were injected subcutaneously with 2.6 mmol/kg 2,5-HD alone, or together with 2.6 or 13.0 mmol/kg of methyl ethyl ketone, acetone and toluene. 2,5-HD concentrations in sciatic nerve (NC), serum (SC) and urine (UC) were determined, and the linear regression between each two of NC, SC, and UC were calculated. There was good correlation between NC and SC, SC and UC in the 2,5-HD alone group, and good correlation between NC and SC in the co-treated groups. Go-treatment solvent had little effect on the relationship between SC and NC. 13.0 mmol/kg cotreated solvent tended to decrease the regression coefficients compared with 2.6 mmol/kg co-heated solvent. These results show that SC can be used in estimating NC in the 2,5-HD alone or cotreated groups, and UC can be used in estimating SC in the 2,5-HD alone group.