BACKGROUND: Studies from the Spanish influenza era reported that transfusion of influenza-convalescent human blood products reduced mortality in patients with influenza complicated by pneumonia. Treatments for H5N1 influenza are unsatisfactory, and convalescent human plasma containing H5N1 antibodies could be an effective therapy during outbreaks and pandemics. PURPOSE: To see whether transfusion with influenza-convalescent human blood products reduced the risk for death in patients with Spanish influenza pneumonia. DATA SOURCES: Manual search of English-language journals from 1918 to 1925. Citations from retrieved studies were also searched. STUDY SELECTION: Published English-language studies that had at least 10 patients in the treatment group, used convalescent blood products to treat Spanish influenza pneumonia in a hospital setting, and reported on a control or comparison group. DATA EXTRACTION: Two investigators independently extracted data on study characteristics, outcomes, adverse events, and quality. DATA SYNTHESIS: Eight relevant studies involving 1703 patients were found. Treated patients, who were often selected because of more severe illness, were compared with untreated controls with influenza pneumonia in the same hospital or ward. The overall crude case-fatality rate was 16% (54 of 336) among treated patients and 37% (452 of 1219) among controls. The range of absolute risk differences in mortality between the treatment and control groups was 8% to 26% (pooled risk difference, 21% [95% CI, 15% to 27%]). The overall crude case-fatality rate was 19% (28 of 148) among patients who received early treatment (after <4 days of pneumonia complications) and 59% (49 of 83) among patients who received late treatment (after >/=4 days of pneumonia complications). The range of absolute risk differences in mortality between the early treatment group and the late treatment group was 26% to 50% (pooled risk difference, 41% [CI, 29% to 54%]). Adverse effects included chill reactions and possible exacerbations of symptoms in a few patients. Limitations: Studies were few and had many methodologic limitations. No study was a blinded, randomized, or placebo-controlled trial. Some pertinent studies may have been missed. CONCLUSIONS: Patients with Spanish influenza pneumonia who received influenza-convalescent human blood products may have experienced a clinically important reduction in the risk for death. Convalescent human H5N1 plasma could be an effective, timely, and widely available treatment that should be studied in clinical trials.