Objective: To evaluate the safety of human albumin administered for therapeutic purposes.
Design: Retrospective compilation of spontaneously reported serious adverse events.
Setting: Records of serious adverse event reports received from 1990 through 1997 by nine major suppliers of therapeutic human albumin worldwide.
Patients: Primarily hospitalized patients.
Interventions: Administration of human albumin.
Measurements and main results: The number of 40-g doses distributed by the nine suppliers during the study period was 95.4 x 10(6), corresponding to 3.82 x 10(6) kg albumin, and reported serious adverse events totaled 123. The incidence of all serious adverse events was 1.29 per 10(6) doses (95% confidence interval, 1.07 per 10(6) to 1.54 per 10(6) doses). No patient death was judged to be probably attributable to albumin administration. The incidence of fatal serious adverse events possibly related to albumin was 5.24 per 10(8) doses (95% confidence interval, 1.70 per 10(8) to 12.24 per 10(8) doses).
Conclusions: Although underreporting must be recognized as a limitation of spontaneous adverse event reports, this study encompassing approximately 100 million albumin doses provides evidence that both nonfatal and fatal serious adverse events in albumin recipients are very rare. These results provide further support for the excellent long-term safety record of human albumin.