The 30-,ug cefprozil disk and the cephalosporin class (30-Fg cephalothin) disk were compared for their abilities to predict cefprozil susceptibility by agar disk diffusion testing. High error (5.02% major and 14.11% minor) rates were encountered with the cephalothin disk and were most frequently observed for Escherichia coli and Enterococcusfaecalis. The cefprozil disks resulted in rates of 0.05% very major, 0% major, and 3.72% minor errors. Thus, the 30-,ug cefprozil disk is more desirable for predicting cefprozil susceptibility. Cefprozil (BMY 28100) is a new orally absorbed cephalosporin with activity against many pathogens that cause urinary tract infections, skin and soft tissue infections, upper and lower respiratory tract infections, and acute otitis media. Compared with other oral cephalosporins such as cefaclor, cephalexin, and cefixime, cefprozil is more active against streptococci and methicillin-susceptible staphylococci (3-5, 7-9, 15). Against members of the family Enterobacteriaceae, cefprozil was comparable to cefaclor and amoxicillin-clavulanate but superior to cephalexin and ampicillin (3, 4, 7, 9, 15). Cefprozil and cefaclor have similar activities against Haemophilus influenza and Moraxella (Branhamella) catarrhalis (1,. 3, 8, 15). Cefprozil is more stable to a broader range of ,3-lactamases than cefaclor (3, 5, 15, 16). Susceptibility to cefaclor and several other cephalosporins has been accurately determined with the 30-,ug cephalothin class disk (11). In this study, we compare the performance of the 30-,ug cefprozil disk and the 30-,ig cephalothin disk in predicting susceptibility to cefprozil. Cefprozil was made by Bristol-Myers Squibb Company, Syracuse, N.Y. Cefprozil and cephalothin disks were manufactured by BBL Microbiology Systems, Cockeysville, Md. The study was performed with approximately 2,300 bacterial isolates received from the cefprozil clinical trials. These strains represent the first isolate of each bacterial species obtained from each patient; thus, each strain is represented only once in this datum set. The agar-dilution MICs and disk diffusion zone diameters were determined by the National Committee for Clinical Laboratory Standards' standardized procedures (13, 14). H. influenza was tested by agar-dilution with Haemophilus test medium (13, 14). Least-squares analysis was used to calculate the regression lines and correlation coefficients (10). A scattergram comparing cefprozil MICs and zones of inhibition around the 30-,ug cefprozil disk for 2,317 isolates is shown in Fig. 1. The regression line computed was y = 25.116-1.652x (where y = zone diameter and x = log2 MIC) with a correlation coefficient of 0.732. In calculating the regression line, off-scale MICs (s0.008 and >128 ,ug/ml) and zone diameters (c6 mm) were excluded (12). The regression line analysis confirms the cefprozil zone diameter and MIC breakpoints recommended previously (6), i.e., the interpretative zone for susceptibility is .18 mm (MIC correlate, c8