Objective: Extremely low birth weight (ELBW) infants frequently undergo transfusion because they are critically ill, often need artificial ventilation, and have the highest blood sampling loss in relation to their weight. During the last decade our transfusion guidelines were changed 3 times to become more restrictive. We hypothesized that these modifications substantially decreased the number of transfusions in our ELBW infants. Methods: We performed a single-center analysis of 256 infants with birth weights from 500 to 999 g who were admitted from 1989 to 1997 and included 3 study periods, each starting with newly modified transfusion guidelines in April 1989, September 1991, and January 1995. We evaluated prospectively recorded clinical data and retrospective chart analysis for transfusion-related information. Results: The median number of transfusions per infant decreased from 7 in the first period to 2 in the third period, whereas donor exposure decreased from 5 to 1 and blood volume transfused decreased from 131 to 37 mL/kg birth weight ( P < .01). The median venous hematocrit measured before transfusion decreased from 43% to 35% in infants who underwent ventilation and from 41% to 31% in spontaneously breathing infants. The median birth weight decreased from 870 to 740 g and the median gestational age from 27 to 25 completed weeks ( P < .01). The overall survival rate was 75% and did not change. The incidences of retinopathy, intraventricular hemorrhage, and patent ductus arteriosus remained unchanged. Conclusion: Over this 9-year period with increasingly restrictive transfusion guidelines, the transfusion number decreased by 71% and the donor exposure by 80% in ELBW infants without adverse clinical effects. (J Pediatr 2000;136:220-4)