To determine, retrospectively, the age of packed red blood cell (PRBC) units transfused to patients admitted to the ICU with the diagnosis of severe sepsis and to correlate this vanable with outcome.

All patients admitted to the ICU during 1992 with a diagnosis of severe sepsis were selected retrospectively. The criteria for the diagnosis of severe sepsis and septic shock were based on established guidelines. For each patient the total number of PRBC units transfused, the number of units transfused before, during and after the septic episode, and the age of each PRBC unit transfused were recorded.

Of the 31 patients admitted to the ICU with severe sepsis, 19 died and 12 survived. No statistical differences between survivors and nonsurvivors were found with respect to age, sex, number of days in ICU, duration of sepsis, incidence of septic shock, admission Apache Il score or total number of PRBC units transfused. During sepsis the median age of PRBC units transfused to survivors was 17 days (range 5–35)vs 25 days (range 9–36) for nonsurvivors (P < 0.0001). A negative correlation (r = −0.73) was found between the proportion of PRBC units of a given age transfused to survivors and increasing age of PRBC.

This is the first study to report a correlation of mortality with the age of PRBC transfused. The cause of this association is unclear. If this association is confirmed by a prospective randomised tnal it would have major implications for the use of PRBC in severe sepsis.