Antibiotic restriction policies are being used to control antibiotic costs. We performed a retrospective review to determine the effect of changes in antibiotic restrictions on the outcome of community acquired pneumonia (CAP) in patients admitted to an urban county teaching hospital (Ben Taub General Hospital). There were two time periods of study: 10/94 to 3/95 (T1) and 10/95 to 3/96 (T2). Beginning in 4/95, the use of ampicillin-sulbactam required approval of an infectious diseases attending physician due to increasing resistance of certain bacterial pathogens to ampicillin-sulbactam and due to the lower daily cost (including that of administration) of low ($27.90) and high ($51.80) dose ceftriaxone vs ampicillin-sulbactam ($36.32 and $54.36, low & high dose, respectively) in our institution. Charts from 115 and 97 patients with CAP in T1 and T2, respectively, were reviewed. The demographics of the study population in T1 and T2 were similar: 46.1 vs 47.5 yrs, 68 M/47 F vs 53 M/44 F, 76% vs 76% non-white, severity of illness (clinical prediction scores 86 and 87), and ICU admission (18% vs 29%, P=.10), respectively. The restriction policy significantly reduced the percentage of patients receiving ampicillin-sulbactam (33% vs 12%, respectively, P<.001) and increased the utilization of ceftriaxone (44% vs 74% respectively P<.001). Outcome measures were similar during the two study periods: time to 1st antibiotic dose (12 vs 8 hr), mortality (13% vs 13%), mean (8.6 vs 7.4 d) or median (6 vs 5 d) length of hospital stay, and mean number of ICU days per ICU patient (10.6 vs 5.4 d, P=.09). Thus, the antibiotic restriction policy had the intended effect of reducing the use of ampicillin-sulbactam for the treatment of CAP without adversely affecting the outcome variables examined.