Twenty-one cystic fibrosis patients under 3 years of age were enrolled in an open multicenter study to assess the feasibility of the study design and to compare selected pharmacologic features of cephalexin or dicloxacillin administered orally for 2 months. Patient tolerance and compliance were significantly less for dicloxacillin (p less than .01 and p less than .001, respectively). Superficial Candida infections were more common in the cephalexin group (p = 0.02), however increased stool frequency and nonspecific diaper rashes were more prevalent in patients receiving dicloxacillin (p less than .05). Staphylococcus aureus was isolated from respiratory secretions after 2 months from two dicloxacillin and no cephalexin patients. Areas under the curve and peak serum concentrations were higher for cephalexin (p less than .05 and p = .02), but antistaphylococcal activity in serum was higher for dicloxacillin (p less than .05) due to a lower mean MIC compared to cephalexin. Deep pharyngeal plus routine throat culture yielded more pathogens than either method alone. Express mail and central processing of respiratory specimens was efficient for most organisms, however there was some loss of Streptococcus pneumoniae and Haemophilus influenzae. Cephalexin was associated with better patient acceptance and compliance despite higher rates of superficial fungal infections as compared to dicloxacillin. Cephalexin, routine bacteriologic throat swabs processed locally or centrally, mail-in urine compliance assessment and a multicenter design are feasible components for a long-term prospective evaluation of antibiotic prophylaxis in patients with cystic fibrosis.