Objective: To determine if a change in the pulse oximeter goal range and high alarm limit for oxygen saturation (SpO 2 ) alters the distribution of SpO 2 for premature infants in oxygen.Study design: This was a prospective, observational analysis. For group 1 (February 2002 to April 2002, n ¼ 23), pulse oximeter alarms were set at 80% (low) and 96% (high), and the goal range was 90-95%. For group 2 (May 2002 to August 2003, n ¼ 49), the high alarm was lowered to 94%, and the goal range was 88 to 94%. The SpO 2 values for 24 h were downloaded from Nellcor pulse oximeters during the two periods and the percent time within, above and below the goal range was derived and compared.Results: Groups were similar except for use of post-natal steroids (group 2>1). The percent time within (57.7±9.8 vs 59.4±12.4%), above (15.4±10.6 vs 14±9.4%) and below (26.9±9.7 vs 26.6±10.2%) the goal range was similar for groups 1 and 2, respectively. However, the percent time with SpO 2 <80% increased significantly for group 2 (4.0±2.7 vs 1.9±1.4%).Conclusions: Changes in pulse oximeter policy and alarms in labile, sick premature infants need evaluation for their effects on the distribution of SpO 2 values before routine use.