A method for the simultaneous determination of sulfadiazine and trimethoprim in plasma from Beagle dogs was developed and validated. Samples were deproteinized with acetonitrile and extracted with ethyl acetate. Sulfachloropyridazine and ormethoprim were used as internal standards for the sulfadiazine and trimethoprim analysis, respectively. The chromatography was carried out both on an LC-UV (liquid chromatography-ultraviolet detection) and ion-trap LC-MS(n) (liquid chromatography-mass spectrometric detection) instrument, operating in the positive APCI mode (atmospheric pressure chemical ionization). The purpose of this work was to compare the quantification results of both methods. Both the LC-UV and LC-MS-MS methods were validated for their linearity, accuracy, precision, limit of detection and limit of quantification, according to the requirements defined by the European Community. Calibration curves using plasma fortified between 0.1 and 1 microg/ml of sulfadiazine, 0.1 and 2 microg/ml of trimethoprim, 1 and 20 microg/ml of sulfadiazine showed a good linear correlation (r> or =0.9990, goodness-of-fit< or =8.4%). The results for the accuracy and precision at 1 microg/ml of sulfadiazine and trimethoprim and at 20 microg/ml of sulfadiazine fell within the ranges specified. The limits of quantification of both methods were 0.1 microg/ml. The limits of detection were 0.019 microg/ml of sulfadiazine and 0.024 microg/ml of trimethoprim for the LC-UV method, and 0.020 microg/ml of sulfadiazine and 0.062 microg/ml of trimethoprim for the LC-MS-MS method. The methods have been successfully applied in a pharmacokinetic study to determine the drug concentrations in plasma samples from dogs. A good correlation between the results of both methods was observed (R=0.9724, slope=1.0239, intercept=-0.2080 microg/ml for sulfadiazine and R=0.9357, slope=1.0433, intercept=0.0325 microg/ml for trimethoprim). The precision of both methods was also tested on the results of the same samples using an F-test (alpha=0.05), indicating that both methods did not differ in precision.