The design and execution of the Cooperative Crohn's Disease Study in Sweden are described in this paper. A double-blind, double-dummy, crossover (2 X 4 mo) technique was used to compare the suppressive efficacy of metronidazole (0.4 g b.i.d.) and sulfasalazine (1.5 g b.i.d.). The number of randomized patients (78) presented approximately one-third of the available population. The Crohn's Disease Activity Index and the plasma level of orosomucoid were the main variables for clinical evaluation. Results were analyzed primarily in the first treatment period by ranking the clinical outcome of every patient according to a uniform and detailed scheme and applying Wilcoxon nonparametric statistics. The cross-over data only served as additional information. Thirty-six patients had had earlier and mostly positive experience with sulfasalazine. Repeated plasma drug analysis indicated good compliance. The blindness of the trial was tested and appeared satisfactory. The coordination of the trial proceeded as planned. A lack of full conformity in the electroimmunoassay of orosomucoid was taken care of satisfactorily.